The Bridge and Structural Engineer by IABSE

The Bridge & Structural Engineer Indian National Group of the International Association for Bridge and Structural Engineering ING - IABSE

Contents :

Volume 47, Number 4 : December, 2017

Editorial ●

Editorial column from Chairman, Editorial Board : Mr. Alok Bhowmick

●

From the Desk of Guest Editor : Dr. Prem Krishna

CFD as a Tool for Assessing Wind Loading R. Panneer Selvam

Assessment of Damping from Field Studies Arunachalam Srinivasan

Tuned Mass Dampers to Control Oscillations of Wind-Sensitive Flexible Structures O.R. Jaiswal, Vikas Thakur, Aditya Ghushe

Effective Static Wind Load Distributions for Large Roofs John Holmes

Special Wind Engineering Considerations for Very Tall and Slender Buildings K. Suresh Kumar

Tall Building Façade Designs -Impact of Wind and Associated Loadings Rajan Govind

Wind Related Studies of an Unusual Building : A Case Study S. Selvi Rajan, P. Harikrishna, G. Ramesh Babu, N. Lakshmanan

Wind Engineering for a 107 M Tall Statue : A Case Study Nicholas Truong, Antonios W Rofail

State of the Art of Long Span Bridge Aerodynamics Toshio Miyata, Hitoshi Yamada, Hiroshi Katsuchi

10.

Wind Tunnel Testing of Long Span Cable-Stayed Bridge Models M. Keerthana, S. Selvi Rajan, P. Harikrishna, G. Ramesh Babu, A. Abraham, S. Chitra Ganapathi

Contents

Special Topic : Wind Sensitive Structures

Paper carried over from September, 2017 Issue: 1.

Sustainable Engineering for Indian Metros with U Shape Viaducts Pankaj Kumar Jain, Anand Pandey, Serge Montens

Panorma ●

Obituary

●

Office Bearers and Managing Committee - 2017

●

Highlights of the Ing-Iabse Workshop on “Application of Irc:112 - Code of Practice for Concrete Road Bridges” held at Raipur on 13th and 14th October, 2017

●

Highlights of the Ing-Iabse & IRC International Seminar on “Repair, Rehabilitation and Retrofitting of Bridges and Structures” held at Jaipur (Rajasthan) on 15th and 16th December 2017

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

The Bridge & Structural Engineer ING - IABSE

Journal of the Indian National Group of the International Association For Bridge & Structural Engineering

March 2018 Issue of the Journal will be a Special Issue with focus on Extradosed Bridges Salient Topics to be covered are : 1. Conceptual Engineering 2. Preliminary Design & Member Sizing 3. Detailed Analysis & Structural Design 4. Stay Cable Technology 5. Construction Issues 6. Case Studies Those interested to contribute Technical Papers on above themes shall submit the abstract by 15th January 2018 and full paper by 31st January 2018 in a prescribed format, at e-mail id : ingiabse@bol.net.in .

The Bridge & Structural Engineer ING - IABSE

Journal of the Indian National Group of the International Association For Bridge & Structural Engineering

June 2018 Issue of the Journal will be a Special Issue with focus on Geotechnics of Transportation Infrastructure Salient Topics to be covered are : 1. 2. 3. 4. 5. 6. 7. 8. 9.

Challenges in design and construction of Pavements and Embankments. Design and construction of substructures for Highways, High speed Railway and Metro Projects. Advances in Bridge Foundations, Waterways, Airfields and Pipeline transport geotechnic. Slope stability, Landslides, Debris flows and Avalanches on hilly roads and remedial measures. Sub-surface sensing, Investigation and Monitoring in transport geotechnics Use of Geosynthetics and Non traditional materials in transport geotechnics. Transport geotechnics in complex Underground Construction Ground Improvement techniques for transport geoetchnics Emerging trends in transport geotechnics– Unsaturated Soil Mechanics, Macro and Nano Technology, Climate Change and Sustainability

Those interested to contribute Technical Papers on above themes shall submit the full paper by 30th April, 2018 in a prescribed format, at e-mail id : ingiabse@bol.net.in .

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

December, 2017

B&SE: The Bridge and Structural Engineer, is a Quarterly journal published by ING-IABSE. It is one of the oldest and the foremost structural engineering Journal of its kind and repute in India. It was founded way back in 1957 and since then the journal is relentlessly disseminating latest technological progress in the spheres of structural engineering and bridging the gap between professionals and academics. Articles in this journal are written by practicing engineers as well as academia from around the world.

All material published in this B&SE journal undergoes peer review to ensure fair balance, objectivity, independence and relevance. The Contents of this journal are however contributions of individual authors and reflect their independent opinions. Neither the members of the editorial board, nor its publishers will be liable for any direct, indirect, consequential, special, exemplary, or other damages arising from any misrepresentation in the papers. The advertisers & the advertisement in this Journal have no influence on editorial content or presentation. The posting of particular advertisement in this Journal does not imply endorsement of the product or the company selling them by ING-IABSE, the B&SE Journal or its Editors.

Front Cover : Top Left : Rare photograph of the original Tacoma Narrows Bridge (which collapsed in 1940 within 4 months of its opening) where the roadway twisted and vibrated violently under 40-mile-per-hour (64 km/h) winds on the day of the collapse (on 7th November 1940). Bottom Left : Scale pressure model of Shanghai Tower in NRCC wind tunnel. Right : Façade cladding pressures on 107 m (305 ft) tall statue in the form of Lord Shiva, to be constructed in Nathdwara, India. The pressure contours are obtained by combining the wind tunnel results from the 1:100 and 1:200 scale model test. using the multi-sector methods.

Editorial Board Chair: Alok Bhowmick, Managing Director, B&S Engineering Consultants Pvt. Ltd., Noida

Members: D.O. Tawade, Chairman, ING-IABSE & Member (Technical), NHAI Mahesh Tandon, Managing Director, Tandon Consultants Pvt. Ltd., New Delhi A.K. Banerjee, Former Member (Tech) NHAI, New Delhi Harshavardhan Subbarao, Chairman & MD, Construma Consultancy Pvt. Ltd., Mumbai Nirmalya Bandyopadhyay, Director, STUP Consultants Pvt. Ltd., New Delhi Jose Kurian, Former Chief Engineer, DTTDC Ltd., New Delhi S.C. Mehrotra, Chief Executive, Mehro Consultants, New Delhi

Advisors: A.D. Narain, Former DG (RD) & Additional Secretary to the GOI N.K. Sinha, Former DG (RD) & Special Secretary to the GOI G. Sharan, Former DG (RD) & Special Secretary to the GOI A.V. Sinha, Former DG (RD) & Special Secretary to the GOI S.K. Puri, Former DG (RD) & Special Secretary to the GOI R.P. Indoria, Former DG (RD) & Special Secretary to the GOI S.S. Chakraborty, Former Chairman, CES (I) Pvt. Ltd., New Delhi B.C. Roy, Advisor Transportation, AECOM & Chief Executive RUPL, New Delhi Published: Quarterly: March, June, September and December Publisher: ING-IABSE C/o Secretary, Indian National Group of the IABSE IDA Building, Ground Floor (Room Nos. 11 and 12) Jamnagar House, Shahjahan Road New Delhi-110011, India Phone: 91+011+23388132 and 91+011+23386724 E-mail: ingiabse@bol.net.in, ingiabse@hotmail.com, secy.ingiabse@bol.net.in

The Bridge & Structural Engineer, December 2017

Disclaimer :

Submission of Papers: All editorial communications should be addressed to Chairman, Editorial Board of Indian National Group of the IABSE, IDA Building, Ground Floor, Jamnagar House, Shahjahan Road, New Delhi-110011. Advertising: All enquiries and correspondence in connection with advertising and the Equipments/Materials and Industry News Sections, should be addressed to Shri I.K. Pandey, Secretary, Indian National Group of the IABSE, IDA Building, Ground Floor, Jamnagar House, Shahjahan Road, New Delhi-110011.

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

Journal of the Indian National Group of the International Association for Bridge & Structural Engineering

The Bridge and Structural Engineer

iii

Editorial column from Chairman, Editorial Board

Though this issue is dated December 2017, due to delay in its publication, I am writing this editorial column when I have already said good bye to the old and welcome to the new year 2018. I take this opportunity to wish Happy New Year to our readership and contributors, to members of the editorial and advisory board, to many independent expert reviewers, guest editors, to all the executive members and managing committee members of ING-IABSE and staff of INGIABSE secretariat. In this concluding issue of the past year (2017), I would like to take the opportunity to express my sincere gratitude to all of them for their valuable contributions to the quality and the prosperity of this journal, which is steadily but surely gaining popularity amongst the readers and becoming a high level reference journal in structural engineering. The New Year (2018) will see a new editorial team with new ideas for further improvement in the journal. This special issue of the Journal is dedicated to the theme of “WIND SENSITIVE STRUCTURES”. The objective of dedicating this issue of the journal to the captioned theme is to advance professional knowledge and improve the practice of wind engineering in civil engineering design, construction and for advancement of scientific knowledge and practice in wind engineering-related problems. Wind has unique characteristics. While it is a source of non-polluting and renewable energy, it is also a undesirable loading on structures, which can play havoc. All structures, including the main wind force-resisting system of the buildings and all components and cladding thereof, shall be designed and constructed to resist wind loads. Indian sub-continent is one of the worst affected cyclonic regions of the world, having a coast line of 7516 kms, which is exposed to nearly 10% of the world’s Tropical Cyclones. The Indian coastal areas are thus vulnerable to cyclone disasters. Recurring cyclones account for large number of deaths, loss of

Volume 47 │ Number 4 │ December, 2017

livelihood opportunities, loss of public and private property and severe damage to infrastructure, thus seriously reversing the developmental gains at regular intervals. For this special issue, our Guest Editor is Prof (Dr.) Prem Krishna, who is no stranger to our readers. He is an International expert on the subject of ‘Wind Engineering’. He has made significant contributions towards wind disaster mitigation efforts of the country. He has been a member of the expert group which produced the Vulnerability Atlas of India, published in 1997 and revised in 2007. He has held several important professional memberships/ positions in the industry, which includes Fellow (1987 onwards) and Vice – President , Indian National Academy of Engineering (2008-2013) ;Founder President, Indian Society of Wind Engineering (19932000); President, International Association for Wind Engineering (1991-95) ; Chairman, Research Council, CBRI, Roorkee (2010 – 2017). A Commemorative Volume was released in his honour, at the National Conference on Wind Engineering, Nagpur, February 2004. There could not have been a better choice of Guest Editor for us for this issue & we are indeed thankful to Prof. Prem Krishna for the painstaking effort made by him in bringing this issue of the journal to this shape with authors from all round the globe. I hope readers will find this issue thought provoking and interesting. Happy reading !

(Alok Bhowmick)

The Bridge and Structural Engineer

From the Desk of Guest Editor

Some Thoughts on Wind Engineering Natural Hazards have been the scourge of living beings from times immemorial. These come in many forms – earthquakes, floods, landslides, tsunamis, volcanic eruptions, wind storms. The size of the disasters caused by these hazards has gone on increasing at a regular pace despite huge investments made the world over for mitigating them. The last century has been worse as compared to the previous ones. Due to the mitigation efforts, the only relief noticed, over the last few decades, is in the losses of life, but damage to property, industrial production, insured losses go on increasing unabated. One may believe that this trend persists because the world has more to lose now than ever before, besides climatic changes also leading to changes for the worst as far as the hazards themselves are concerned. Understandably, the third world (including Asia – Pacific) losses of life are greater than other parts of the world. Likewise, the more prosperous and better prepared countries suffer more in economic terms. Amongst the various hazards listed above, the three most lethal are earthquakes, floods and wind storms – the last named is related to the subject of this volume. Strong wind storms are like unwelcome guests which are unavoidable. The only recourse left to engineers & scientists is to accept the occurrence of strong winds; try to understand the character of wind as best as possible; asses through experimental or numerical means the wind loading on structures, and, minimize the same through appropriate choice of geometric form; design the structures to resist wind effects by keeping the response within limits; suppress the same where necessary and possible by adopting appropriate design measures, and, to

The Bridge and Structural Engineer

thus protect and keep the humanity as comfortable as possible, and, with as little expense as possible. This is the wind scientists’ mandate, if one does not include agricultural and aerospace aspects. Wind flow generation takes place on account of differential pressure differentials in the atmosphere, and, manifests itself, in the forms of, Gales and Monsonic Winds, Cyclones/Typhoons/Hurricanes, Thunderstorms, Tornados, and, other localized storms. Fig. 1 shows typical photographs of some of these.

Fig. 1 : Typical Views of Wind Storms

In basic terms, an obstruction to the flow of wind, howsoever mild in velocity, causes the obstruction to experience a pressure on its surface. Such obstructions, in our context are the various kinds of structural forms – buildings, bridges, towers, industrial units such as chimneys and cooling towers, transmission systems, and, so on. Wind being a randomly varying dynamic phenomenon (see Fig. 1), the pressure experienced by a structure is necessarily dynamic, and, so therefore is its response. In a sense, therefore, all structures are Volume 47 │ Number 4 │ December, 2017

‘wind-sensitive’ - there is only a variation in degree. Short, stocky or stiff structures, such as a majority of buildings, many of the magnificent monuments and stone bridges of the old era, evoke a nearly static or ‘quasi static’ response. The slender and more flexible structures, such as, the skyscrapers of today, tall towers, long span bridges and roofs – mostly of the cable-supported type, nevertheless experience a predominantly aerodynamic response. When the predominant frequencies of wind velocity fluctuations (a measure of its turbulence) tend to be close to those in the structure, there is the possibility of ‘resonance’. Fortunately, though, because of the inherent nature of randomness in the fluctuating wind, we mostly get away with quasi-resonant conditions and reduction in consequent damage that could occur due to resonance. The response of the structures in the above cases though, increases with increasing wind speeds. A different kind of aerodynamic oscillation problem can be caused by the shedding of von Karman vortices which can occur when wind strikes such cylindrical structures as chimneys, tall buildings, cables, transmission lines, or, bridge decks. Particularly in smoother flows, the eddies can become better ‘organized’ and can cause resonant conditions, and, lead sometimes to a catastrophic condition. This may happen even at low wind speeds.

is not uncommon that research and development in an area of engineering or technology is dictated by the status of applications in that field, or, events which open up hitherto unknown vistas. Thus, the early efforts in wind related studies were related to understanding the wind characteristics and, evaluating its effects on small or stocky buildings and structures with hardly any dynamic sensitivity. Developments, from about the 1950s onwards, in materials of higher strength-weight ratio, and, increasing slenderness of buildings/bridges changed that picture. The dramatic collapse of the Tacoma Narrows Suspension bridge in 1940, the boom of skyscrapers in the 1960s, particularly in USA, and the collapse of Ferry bridge cooling towers in the UK in 1965, provided the clarion call, if one were needed.

Fig. 2 : Typical Trace of Wind Velocity, V, with Time, T

Cermak’s establishment of a boundary layer wind tunnel in 1950, Jensen’s experiment to show the importance of modelling the boundary layer flow with appropriate turbulence content, Davenport’s pioneering work in making wind loading better understandable for engineering application, and Scruton’s contributions in aeroelasticity (particularly for bridges), and so on, set the tone for further developments in wind engineering, and there has been no looking back. In this context the first conference on Wind Effects in Buildings and Structures held in UK in 1963 provided great impetus. Off course this field of research has also since thrown up many prominent experts who have made significant impact on further developments. Helped by their efforts and due to the tremendous growth of the electronic technology, affecting great sophistication in instrumentation and computing power, there have been very substantial advances, in respect of wind resistance and safety of buildings and structures.

The frontiers in Science and Technology R&D are ever changing. The advancements, particularly now, are faster than one can keep pace with. That is why a person or a group of persons are always chasing the frontiers, and also are able to grasp or comprehend only a part of the vast repertoire of knowledge on any specific area of their enquiry. This is applicable in equal measure to the field of wind engineering. The history of formal research pertaining to wind effects on buildings & structures is no more than about 100 years. In fact, the term “Wind Engineering” is only about 50 years old. It

Those regions of the world, or societies, which have been able to absorb the advances better, enjoy a reasonable degree of safety. In regions where the economic resources do not permit the above status, risks are greater. Irrespective of the region nevertheless, continued efforts are yet needed. As mentioned already in the first paragraph, losses due to strong winds, particularly to property have been continuing to mount over the decades. Furthermore, the trends in using taller and wider span structures continue to place increasing demands upon the wind engineer, and, continue to push the frontiers. Added

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

to this, is the challenge mounted by developments in structure architecture, encouraged by modern day capabilities in computing and drafting, besides technologies for fabrication and construction – straight lines have given way to the most complex and curvilinear geometries. The problem this poses for the structural designer is that wind loading for most of these new forms is not covered by wind loading codes and literature. Obviously therefore the battle is far from won and there is a continuing need to strive on. Indian Scenario of related development will indicate the same trends as those occurring internationally, except that, in terms of wind-sensitive structures, India is way behind the frontline situation. In bridges, our longest spans are in the region of 500 m, as compared to the near 2 km mark

(the Akashi Kaikyo bridge in Japan) achieved internationally, and longer spans being a real possibility. Likewise, in buildings, our highest ones are yet to cross the 300 m mark, whereas the world is already making the 1 km tall building. Perhaps, it is for this reason that our status in wind engineering is only moderate, in terms of related R&D, facilities for experimental work, training of the manpower resource, codal development, design expertise, and, so on. There is however no need to doubt that the wind engineering of this country will move into the higher orbit sooner than later, given the huge challenge facing the country to add to its greatly deficient infrastructure. The collage of photographs in Fig. 3, gives a glimpse of developments of structures falling into the windsensitive category.

Fig. 3 : A Collage A. Some Recent Buildings – Curvilinear, Medium Rise, Fabric Roofing

Views of Wind Tunnel Experimentation

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

vii

It is commendable that the ING-IABSE has thought it fit to bring out a special issue of their Journal, “THE BRIDGE & STRUCTURES Engineer”, to focus on this aspect of engineering, and bring it to the attention of its readership. The volume has 10 papers from authors, within the country and outside, who can be acclaimed to be leading researchers, designers, or, practitioners, in their related specialisations in wind engineering. I place on record my appreciation for the ready co-operation of all these authors, who acceded to my request and found time from their busy schedules to prepare these excellent contributions. The composition of the volume covers state-of-art of some leading issues/problems that the wind/structural engineer faces today, or, is bound to be confronted with as the

viii

Volume 47 │ Number 4 │ December, 2017

profession moves towards the construction of larger spans or taller structures, and, case studies related to a few very challenging wind engineering studies. I am sure the volume will provide for interesting reading, and, help advance this important discipline of engineering. I also place on record my thankfulness to the Editorin-Chief, Er. Alok Bhowmick, for inviting me for my association in bringing out this volume, and, for his cooperation and advice.

(PREM KRISHNA)

The Bridge and Structural Engineer

Brief Profile of Dr. Prem Krishna In an engineering career commencing in 1959 and spanning approximately 58 years to date, Dr. Prem Krishna has had a unique opportunity of studying and teaching at Civil Engineering Departments of the highest quality and reputation, namely, the University of Roorkee (Now Indian Institute of Technology Roorkee), the University of Illinois, Urbana, USA and the Imperial College of Science & Technology London, UK. In addition to being a committed teacher, he has been keenly active in R&D and maintained close interaction with industry in the areas of structural engineering, wind engineering & disaster mitigation. Born March, 1938, he obtained his early degrees from the University of Roorkee - BE(Civil) in 1959, ME(Structures) in 1961. He obtained his Doctorate from Imperial College, London in 1964 and joined the faculty of Civil Engineering at the University of Roorkee in 1965. He retired from there in 1998 but maintained contact with the institution in adjunct positions till 2007. Professor Prem Krishna set after in earnest to develop the field of Wind Engineering in India in 1978. He has worked assiduously since to foster research in this important field, providing leadership in building capacity in terms of human resources, facilities for testing, developing design support material. Thus, the country has moved towards self - sufficiency to a great extent in tackling the wind load related problems concerning the large numbers of wind sensitive structures – towers, chimneys, buildings, bridges –required for infrastructure development. The efforts have been recognized internationally and the country is squarely on the World Map of Wind Engineering. Dr. Prem Krishna took the initiative to organize the first Asia-Pacific Symposium on Wind Engineering at the University of Roorke in 1985

The Bridge and Structural Engineer

– this has since become a four-yearly conference, revolving amongst the AP economies. He along with a few other colleagues set up the Indian Society of Wind Engineering in 1993 to create a platform for interaction between engineers and scientists interested in the area of wind. He served as its first President for seven years. The ISWE organized successfully the 9th International Conference on Wind Engineering at Delhi in 1995. Prem Krishna has made significant contributions towards wind disaster mitigation efforts of the country. He has been a member of the expert group which produced the Vulnerability Atlas of India, published in 1997 and revised in 2007. This is a single most useful document to provide the basic data for planning the natural disaster mitigation programme of the country. As Chairman of Bureau of Indian Standards (BIS) Sectional Committee on Cyclone Resistant Structures, he has helped develop relevant standards. He has also been actively involved in the development of other wind-related design codes. Important Professional Memberships/Positions he has held, include, Fellow (1987 onwards) and Vice – President, Indian National Academy of Engineering (2008-2013); Founder President, Indian Society of Wind Engineering (1993-2000); President, International Association for Wind Engineering (1991-95); Chairman, Research Council, CBRI, Roorkee (2010 – 2017), and, the awards and honours include, Distinguished Alumnus Award, 2012, IIT Roorkee; Life-time Achievement Award by the ISWE, celebrating 20 years of its establishment, Gourav Award 2012, by the Indian Association of Consulting Civil Engineers, Bangalore, apart from several others. A Commemorative Volume was released in his honour, at the National Conference on Wind Engineering, Nagpur, February 2004.

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

CFD as a Tool for Assessing Wind Loading R. Panneer Selvam Womble & University Professor , University of Arkansas Fayetteville, USA E-mail: rps@uark.edu

Summary CFD application to wind engineering is reviewed. Current status and future challenges are enumerated. First the review focuses on modelling of vortexshedding, galloping and flutter for wind sensitive structures. Then the review of building aerodynamics and extreme wind effects on structures is reported. In the building aerodynamic study, the effects of turbulence on structures are also reviewed. Keywords: Computational fluid dynamics, computational wind engineering, fluid-structure interaction, galloping, flutter, tornado, bridge aerodynamics, building aerodynamics.

Introduction

CFD is becoming an alternative tool in the area of wind engineering. In recent years extensive developments have happened in Computational Wind Engineering (CWE) due to availability of high performance computers and storage systems. Extensive papers on CWE are available in most of the wind engineering journals. This review will cover developments beyond that reported in Selvam [1-4]. For some historic review one can refer to [4].

Computational Tools

2.1 Numerical methods Generally Finite Difference Method (FDM) or Control Volume Method (CVM) is used in CWE. Most of the commercial softwares available are based on CVM. Only very few Finite Element Methods (FEM) based software are utilized for bridge aerodynamics and tornado flow around building and hills [1-7]. Selvam and his group used it mainly because of its accuracy and capability to transport complex flows. The

The Bridge and Structural Engineer

R.P. Selvam, born 1955, received his civil engineering degree from the University of Madras, India. He got his PhD from Texas Tech University before becoming Professor at the University of Arkansas. His main area of research is related to computational wind engineering.

superiority of the FEM for convective modelling is reported in Selvam [1]. To obtain much higher accuracy in CWE p-adaptive FEM or spectral method was used for bridge aerodynamics as reported in [8-10]. In the p-adaptive FEM, the order of the polynomial used is up to p=6. Even this high level of accuracy is not enough for transporting inflow turbulence. Lim et al. in [11] used CVM to compute the extreme pressures on a cube due to inflow turbulence. Even with their 8.2 million grid points, the authors could only transport high frequency spectrum up to fh/U=3, and, with reasonable accuracy up to fh/U=2 only. Here f is the frequency of the wind, h is the reference height of the building and U is the reference velocity. This particular grid is equivalent to 10 points per cycle for the high frequency spectrum of fh/U=3. For this the authors used 32 processors for their parallel computing study. This limitation is mainly because CVM needs extensive computer points. It is estimated that about 6750 million points are needed to transport fh/U=10 level of high frequency spectrum with reasonable accuracy. Currently, due to storage and computer limitations it is very difficult to use that level of grid points. Other alternatives have been investigated in our laboratory in the recent years to face this challenge. For this, Fourier spectral and p-FEM are investigated. Some of our recent investigations in this regard are reported briefly in section 6. 2.2 Turbulence models Nowadays, it has become very common to use Large eddy simulation (LES) as turbulence model, Reynolds stress based k-ε turbulence models are used very rarely. For details of turbulence models one can refer to [12-13]. Volume 47 │ Number 4 │ December, 2017

2.3 Navier-Stokes solvers and methods for solving Ax=b Different ways of solving the velocities and pressure in a sequential way for the incompressible NavierStokes equations are well documented in [12-14]. Coupled solution techniques are used very rarely. Most of the commercial software uses some variant of SIMPLE procedure. In the CWE applications millions of equations need to be solved. Selection of proper solver saves enormous computer time. Solvers like preconditioned conjugate gradient (PCG) and multigrid are commonly used for both serial and parallel computing. The details of the method are reported in [12 & 14]. It will be efficient to use multigrid with PCG.

Wind Sensitive Structures and Bridge Aerodynamics

The dynamic force created by turbulence, vortex shedding and motion of the structure causes aeroelastic responses as reported by Davenport [15] and Dyrbye and Hansen [16]. The vortex induced vibration (VIV) caused by vortex shedding and galloping and flutter instability condition produced by motion of the structure or negative aerodynamic damping can cause severe damage to chimneys, light poles, traffic signal poles, tall buildings and bridges as reported in [16-20]. These deformations can vary with respect to thickness d of the structure of the order of 1 to 30. The design procedure suggested by Griffin and Ramberg [16] for VIV problem is very useful. In the above three phenomena, galloping and flutter produce large deformation due to negative damping and also the oscillation frequencies are closer to the structural frequencies. This means the effect is more due to structural motion. On the other hand, motion due to vortex shedding is more around the Strouhal frequency and the amplitude is higher when the resonance occurs. During galloping the y/D ratio is much higher than one, whereas during vortex shedding the ratio is less than one. Also there is no galloping for circular cylinder. In the vibration due to vortex shedding and galloping only displacements in the cross wind direction are considered, whereas in the flutter calculations, both the movement in the cross wind direction and rotation perpendicular to the plane of the section are considered. Fluid-structure interaction (FSI) study can be conducted by a combination of wind tunnel or CFD and analytical approach or using only CFD. To use

Volume 47 │ Number 4 │ December, 2017

the analytical approach, the drag and lift coefficients Cd and Cl have to be measured or computedfor various angles of attack in the case of galloping and aerodynamic derivatives have to be measured or computed for critical velocity for flutter calculations. This method is called forced motion of the structure. In the second approach complete CFD is used and Selvam et al. [21] introduced it to wind engineering. This method is called free motion of the structure. The details of the approaches used by different researchers for bridge aerodynamics are reported in [21]. The forced motion study is a linear assumption and hence it cannot consider the nonlinearities of the flow. Whereas in the free motion study, all the nonlinearities developed in the flow are captured in CFD and also the computation required may be less compared to forced motion method. Now many researchers have started to use the free motion method for bridge aerodynamics. 3.1 Current status of galloping study The motion induced vibration was first modeled by Den Hartog [22] using simple one-degree of freedom (DOF) spring oscillator model. He developed a quasistatic instability criterion for galloping, which is still applicable for many practical problems. The Den Hartog instability criterion is: ∂Cl/∂α+Cd<0

(1)

where Cl and Cd are mean lift and drag coefficients. These values are measured at various angles of attack α for a fixed cylinder in a wind tunnel as reported by Parkinson and Smith [23] for a square cylinder, and, Deniz and Staubli [24] for rectangular and octagonal section. Now this can be accomplished by CFD also as reported in Gomez et al. [25]. This equation is extensively used for prediction of the critical velocity for galloping. Later two and three DOF models were developed and for recent review one can refer to Zuo et al. in [20]. The galloping study is conducted using wind tunnel, field measurement, computational fluid dynamics and by analytical method. The analytical methods will be useful for design if verified by any of the above methods. Using CFD, galloping response can be calculated by finding the Cl vs angle relation using CFD and then using analytical approach or by complete fluid-structure interaction (FSI) study using CFD. 3.1.1 Analytical method for galloping calculation: First, drag force coefficient Cd and lift coefficient Cl

The Bridge and Structural Engineer

are calculated for various angles of attack by wind tunnel tests or CFD. Gomez et al. [25] used CFD for a Z cross section to compute Cl for various angles. The analytical model used can be a one dimensional spring model used by Parkinson and Smith [23]. The galloping fluid force created in the direction normal to the flow of wind is given by F*=0.363CFyρaV2dL/2. This force only includes the negative damping part but not the lift force due to vortex shedding. The lift force due to vortex shedding is considered to be very small compared to VIV forces. The CFy is provided in a polynomial form as reported in [23]. The CFy is calculated from the drag and lift coefficients Cd and Cl for various angles of incidence for a fixed cross section in a wind tunnel or using CFD. The equation of motionin a non-dimensional form is given in (2). The reference values considered for the derivation are: d-reference length and 1/ω as reference time. The non-dimensional equation is as follows: ∂2ttY+Y=nA1(U-U0)∂tY

(2)

where, Y

Amplitude y/d

mass ratio=ρad2L/2M

reduced air velocity=V/(ωd). Here d is the side perpendicular to the flow

A1 first order coefficient of CFy U0 minimum reduced air velocity for damping= 2ξ/ (nA1) where ξ is the damping coefficient In equation (2) only one term of the CFy is considered. The Y in Equation (2) is a non-dimensional number. The above equation can be integrated in time by 4th order Runge-Kutta method (RKM) or Newmark-β method. Both methods worked well. The RKM algorithm is given in Thompson [26]. Since the equation is similar to Duffing’s equation several methods are developed to solve it analytically. A review of recent approaches is provided in Vio et al. [27] and Liu et al [28]. It is found that Newmark-β method can use longer time step and hence implicit solution like Newmark-β method is preferred to RKM method. 3.1.2 FSI study using CFD for galloping: Robertson et al [29] and Dettmer and Peric [30]used CFD for FSI applications. In this study, for each time step the flow and the pressure around the structure is computed using CFD and then using the pressures and forces on the structure, the dynamic equation for the structure is solved to find the new position. Hence complete CFD is used for FSI. The major issue one needs to

The Bridge and Structural Engineer

take care of is to satisfy the geometric conservation law. Different methods to take care of this issue are reported in Selvam et al. [21] for bridge aerodynamics and this is applicable here also. 3.2 Bridge aerodynamics using CFD As discussed in the beginning of this section for bridge aerodynamics, forced and free motion methods are used to calculate the critical velocity for bridge aerodynamics. Most of the initial work is based on the forced motion method. In the recent years, free motion method is used more often due to the computational advantages. Overall the critical velocity for flutter can be calculated with reasonable accuracy with the existing computer models. 3.2.1 Forced motion method for bridge aerodynamics: Simiu and Scanlan [31] details the procedure using wind tunnel measurements. This approach is used by Walther [32] and Larsen and Walther [33] using CFD. Walther [32] performed 13 computer runs to calculate the flutter derivatives and then used the analytical equation to calculate the critical velocity for flutter. Hence it takes more computer time than the free motion method which needs about 5 to 7 computer runs. The method is based on linear assumptions and hence it cannot consider the nonlinear effect of the fluid. 3.2.2 Free motion method for bridge aerodynamics: Selvam et al. [21] introduced the free motion method to calculate the critical flutter velocity by transferring the knowledge from aeronautical area as reported in Selvam et al. [34]. Later Robertson et al., [35], Frandsen [36] and Braun and Awruch [37]used the method. In the recent years most of the work is based on free motion method. Recently Selvam and Bosch in [38] used the model to investigate the effect of initial conditions and grid refinements effect on critical velocity for flutter.

Building Aerodynamics

4.1 Low rise building Paterson and Aplet [39] first used k-ε turbulence model to compute the mean pressure around building and then Murakami and Mochida [40] used LES. Later Selvam [41-42] used k-ε model and LES with inflow turbulence to compute the pressure around the Texas Tech University (TTU) building. From then on extensive works were reported using CFD. Most of the available work can compute the mean pressures with reasonable accuracy. Still there are extensive

Volume 47 │ Number 4 │ December, 2017

challenges in computing the peak pressures using CFD. Recent work by Lim et al. [11] used inflow turbulence to compute the mean and peak pressures on a cube. The authors used 8.2 million points and the computation is conducted with 32 processors. As seen, the mean values are comparable with measurements but not the peak value. One of the reasons is that the required high frequency spectrum of the turbulence is not captured with the grid that could be used. Further work is needed to improve the accuracy. Either more grid points need to be used with CVM or more accurate numerical methods for the same number of grid points have to be deployed. 4.2 High rise building Recently several researchers [43-47] used inflow turbulence to compute the peak pressures around the tall building. Inflow turbulence models are improved by different researchers as reviewed by Aboshosha et al. [43]. These authors have also provided a MATLAB program to generate the inflow turbulence. One of the drawbacks in all the reported work is the grid resolution to carry the necessary high turbulence spectrum. For reporting the spectrum of the inflow turbulence, points far away from the building on the wind ward side much larger than the wavelength of the high frequency turbulence reported in the inflow, have been used. None clearly reported the turbulent spectra at the location of the building. Because of different grid spacing and the accuracy of the method used to transport the turbulence, the spectra will vary. More investigation is needed in this regard. Also the reported peak or rms pressures are not completely in agreement with field or measured values.

Extreme Wind Effects on Buildings and Structures

Since hurricane wind effects are considered to be more like straight wind, the methods discussed in the previous section are applicable. Here we will consider the extreme wind effect like tornado on buildings. Only very limited work has been done on downdraft wind effect on structures and hence it is not reviewed.

In the analytical tornado vortex model, the tornado at every instant of time is described by a mathematical equation as presented in Selvam and Millett [48-49]. This model considers the tangential velocities and translational velocity of a tornado but not any radial or vertical velocity. The tangential wind profiles prescribed by the Rankine combined vortex and Vatistas model reasonably compare with field and laboratory measurements and the details are reported in Strasser and Selvam [50]. This model has been used extensively by Selvam and his research group for more than 25 years. They applied the model to compute time variation of forces and flow features on 2D cylinders [51], 3D buildings [48-49, 52] and for tornado-terrain interaction studies [53-54]. Selvam and Millett in [49] reported that the vortex type interaction on a cubic building produced more than twice the force on the roof compared to straight wind for the same reference wind speed. In the tornado vortex generation chamber models, a tornado like vortex is developed by providing tangential velocity Vθ and radial velocity Vr around a circular cylinder as inlet velocity and vertical velocity at the top of the circular hole as outlet velocity as shown in Figure 1 (a) and (b). In a stationary vortex generation chamber, the building is moved and the chamber is kept stationary as shown in Figure 1 (a); and in the moving vortex chamber, the chamber is moved and building is kept stationary as shown in Figure 1 (b). By controlling the Vθ, Vr, top radius Ro and the spacing between the floor and vortex chamber Ho, several type of tornado vortex can be developed in the vortex chamber. The above factors are related by the term swirl ratio and are defined as: S= Vθ/(2AVr)=tanθ/2A

(3)

where aspect ratio A=Ho/Ro and tanθ= Vθ /Vr.

5.1 Tornado Models for Engineering Applications Tornado models used for engineering applications can be classified into three major categories: (1) Analytical vortex model (2) Stationary vortex generation chamber and (3) Moving vortex generation chamber. 5.2 Analytical vortex model

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

building using analytical tornado model and vortex chamber model needs to be made.

Conclusions and Future Challenges

From the review we can conclude that for computing mean pressure around building, and, critical flutter velocity for bridges, one can use CFD with some level of confidence. However, computing extreme pressures on building due to turbulence is still a challenge. Fig. 1 : (a) Stationary vortex chamber (b) Moving vortex chamber

5.3 Stationary vortex chamber Computer modelling work is conducted by Lewellen et al. [55], Hangan and Kim [56] and Natarajan and Hangan [57] to study the wind field due to the effect of different parameters. Recently Lewellen [58] investigated the tornadic flow over hill using immersed boundary method. Here they moved the hill with respect to the simulator. The major drawback with this approach is that there is no proper grid resolution on the surface of the hill and hence the investigation of the interaction of tornado-hill may not be that accurate. Zu et al. [59] study the flow and pressure around a dome due to tornado by moving the dome. Here they moved the dome by dynamic mesh method. They did not use enough grid resolution close to the building and also no comparison was made with any tunnel study. Nasir and Bitsuamlak [60] studied the flow around a building by moving the building using an overset mesh. They didn’t report any detailed comparison of the force and pressure coefficients with straight wind flow. 5.4 Moving vortex chamber Kuai et al. [61], Sengupta et al. [62] and Phuc et al. [63] moved the vortex chamber utilizing sliding mesh algorithm. Kuai et al. [61] reported only wind field study and Sengupta et al. [62] reported CFD study for a microburst interacting on a cube. Only Phuc et al. in [63] reported computed maximum and minimum pressure coefficients on a cube for a swirl ratio of 0.65. They didn’t have any comparison with wind tunnel study. From the tornado force on building review, we can conclude that there is more room for improvement in the existing validation for pressure around buildings due to tornado. Also comparison of pressures around

The Bridge and Structural Engineer

One of the major drawbacks with the current methods like control volume method (CVM), finite difference method (FDM) and finite element method (FEM) is their numerical accuracy to transport high frequency turbulence. Much more accurate spectral method (SM) [64-66] may have to be used to transport with less number of grid points. In transporting a sine wave using CVM or FDM with 10 points per wave, there may be considerable loss in its amplitude due to numerical diffusion, depending upon the distance it is transported from the inlet. To illustrate this issue, a sine wave is transported for a distance of 5 times the wavelength and the final amplitude using different methods are shown at the right side of Figure 2 (a). Here the values are specified at the left side and outflow boundary condition is specified on the right side. The wave is transported using convection equation without any diffusion. So any loss in amplitude is due to numerical diffusion. The methods considered are FDM, FEM and SM. For the SM method, Chebyshev polynomial is used. One can see that the SM method preserves the shape and amplitude of the sine wave much better than that by FDM or FEM. This error is further magnified if the boundary condition is periodic on both sides. In the periodic case, when the wave is transported for more time, there is more error using FDM or FEM. To show the performance of Fourier spectral method (FSM) with FDM and FEM; a sample run is made for the same 10 points/wave and the wave is transported for about one full circle or 60δ where δ is the grid spacing as shown in Figure 2 (b). If it would have been transported couple of more times, FDM or FEM would have lost almost all the amplitude and FSM would have preserved the same amplitude. But the spectral methods are full domain methods and need extensive computer memory for SM using Chebyshev polynomial. Currently research is underway to use these methods for wind engineering applications.

Volume 47 │ Number 4 │ December, 2017

10.

11. Fig. 2 : Sine wave transport using 10 points/wavelength (a) nonperiodic boundary conditions (b) periodic boundary conditions

12.

References

Selvam, R.P. (1998), Computational procedures in grid based computational bridge aerodynamics, in Bridge Aerodynamics, Larsen, A. and Esdahl, S. (eds), Balkema, Rotterdam, 1998, pp. 327-336. Selvam, R.P. (2002), Computer modeling for bridge aerodynamics, in Wind Engineering, by K. Kumar (Ed), Phoenix Publishing House, New Delhi, India, 2002, pp. 11-25 Selvam, R.P. (2008), Developments in computational wind engineering, Journal of Wind and Engineering, 5, 47-54 Selvam, R.P., (2010), Building and bridge aerodynamics using computational wind engineering, Proceedings: International workshop on wind engineering research and practice, May 28-29, Chapel Hill, NC, USA. Ahmed, N. and R. P. Selvam (2015) Ridge effects on tornado path deviation, International Journal of Civil and Structural Engineering Research, 3, 273-294 Strasser, M.N., M.A.A.Yousef and R. P. Selvam (2016), Defining the vortex loading period and application to assess dynamic amplification of tornado-like wind loading, Journal of Fluids and Structures, 63, 188-209

Volume 47 │ Number 4 │ December, 2017

13.

14.

15.

16. 17.

18.

19.

Yousef, M.A. and R.P. Selvam (2017), The influence of tangential to translational velocity ratio on tornado force coefficients on building using CFD, Proceedings: The 13th Americas Conference on Wind Engineering (13ACWE), Gainesville, Florida USA, May 21-24 Selvam, R.P. and Z. Qu (2001), Further work on P-adaptive FEM for flow around circular cylinders, Conference Proceedings, Americas Conference on Wind Engineering-2001, Clemson University, June 4-6th, 2001 Selvam, R.P. and Z. Qu (2002), Adaptive p-finite element method for wind engineering, Wind & Structures, 5, 301-316 Selvam, R.P. (2008), Adaptive stabilized hp-FEM and large eddy simulation for wind engineering, The Proceedings of the 4th International conference on Advances in Wind and Structures (AWAS’08), (Eds.) C.K. Choi, J.D. Holems, Y.D. Kim and H.G. Kwak, May 29-31, Jeju, Korea, pp. 15-24 Lim, H.C, T.G. Thomas, I.P. Castro (2009), Flow around a cube in a turbulent boundary layer: LES and experiment, Journal of Wind Engineering and Industrial Aerodynamics, 97(2),96–109. Ferziger, J.H. and M. Peric (2002), Computational Methods for Fluid Dynamics, Springer, 3rd edition. Versteeg, H.K. and W. Malalasekera (2007), An introduction to Computational Fluid Dynamics, Prentice Hall, 2ndedition Selvam, R.P. (2017), Introduction to Computational Fluid Dynamics and Heat Transfer, class notes, University of Arkansas. Davenport, A.G. (1995), How can we simplify and generalize wind loads? Journal of Wind Engineering and Industrial Aerodynamics, 54/55, 657-669 Dyrbye, C. and S. O. Hansen, (1997), Wind Loads on Structures, John Wiley & Sons Griffin, O.M. and S.E. Ramberg (1982), Some recent studies of vortex shedding with application marine tubulars and risers, Journal of Energy Resources and Technology, 104, 2-13 Caracoglia L, N.P. Jones (2007), Numerical and experimental study of vibration mitigation for highway light poles. Engineering Structures, 29, 821–831. Zuo, D. (2009) Wind-induced vibration of slender structures with tapered circular cylinders, The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

The Bridge and Structural Engineer

20. Zuo, D. L.Wu, D.A. Smith, S. M. Morse (2017), Experimental and analytical study of galloping of a slender tower, Engineering Structures 132, 44– 60 21. Selvam, R.P., S. Govindaswamy, H. Bosch (2002). Aeroelastic analysis of bridges using FEM and moving grids. Wind and Structures, 5, 257-266 22. Den Hartog, J.P., (1932), Transmission line vibration due to sleet, Transactions of AIEE, 51, 1074–1086. 23. Parkinson, G. V. &J.D. Smith (1964), The square prism as an aeroelastic non-linear oscillator, Quarterly Journal of Mechanics and Applied Mathematics 17, 225-239 24. Deniz, S. and T. Staubli (1998), Oscillating rectangular and octagonal profiles: modeling of fluid forces, Journal of Fluids and Structures (1998) 12, 859-882 25. Gomez, I. et al. (2014), Numerical Investigation of Galloping Instabilities in Z-Shaped Profiles, The Scientific World Journal, Volume 2014, Article ID 363274, 14 pages 26. Thompson, W.T. (1993), Theory of vibration with applications, Prentice Hall, 4th Edition. 27. Vio, G.A., G. Dimitriadis, and J. E. Cooper, (2007), Bifurcation analysis and limit cycle oscillation amplitude prediction methods applied to the aeroelastic galloping problem, Journal of Fluids and Structures, vol. 23, no. 7, pp. 983–1011 28. Liu et al. (2014) A contrast on conductor galloping amplitude calculated by three mathematical models with different DOFs, Shock and Vibration, Volume 2014, Article ID 781304, 10 pages 29. Robertson, I., L. Li, S. J. Sherwin, and P.W. Bearman (2003), A numerical study of rotational and transverse galloping rectangular bodies, Journal of Fluids and Structures, vol. 17, no. 5, pp. 681–699, 2003. 30. Dettmer, W.G., D. Perić (2006), A computational frame work for fluid rigid body interaction: finite element formulation and applications, Comput. Methods Appl. Mech.Eng., 195, 1633–1666 31. Simu, E. and R.H. Scanlan (1978), Wind effects on structures: An introduction to wind engineering, John Wiley & Sons. 32. Walther, J.H. (1994), Discrete vortex method for two-dimensional flow past bodies of arbitrary shape undergoing prescribed rotary and translational motion, PhD thesis, Technical University of Denmark.

The Bridge and Structural Engineer

33. Larsen, A. and J.H. Walther (1998), Discrete vortex simulation of flow around five generic bridge deck sections. J. Wind Eng. Ind. Aerod 7778, 591–602. 34. Selvam, R.P., M.R.Visbal and S.A. Morton (1998), Computation of nonlinear viscous panel flutter using a fully-impilicit aeroelastic solver, AIAA98-1844, also in the Proceedings of the 39th AIAA structures, structural dynamics and material conference, Long Beach, CA, April 20-23, 1998, Part 2: pp. 1263-1272 35. Robertson, I., S.J. Sherwin, P.W. Bearman (2003), Flutter instability prediction techniques for bridge deck sections, International Journal of Numerical Methods in Fluids, 43, 1239-1256. 36. Frandsen, J.B., 2004. Numerical bridge deck studies using finite elements. Part I: flutter. J. Fluids Struct 19, 171–191. 37. Braun, A. L., and A.M. Awruch, (2008), Finite element simulation of the wind action over bridge sectional models: Application to the Guamà River Bridge (Parà State, Brazil). Finite elem. Anal. Des. 44, 105-122 38. Selvam, R.P. and H. Bosch (2016), Effect of initial conditions and grid refinements on bridge flutter calculations using CFD and HPC, (BBAA VIII) 8th International Colloquium on Bluff Body aerodynamics and Applications, Northwestern University, Boston, June 7-11 39. Paterson, D. and C.J. Aplet (1986),Computation of wind flow over three-dimensional buildings, Journal of Wind Engineering and Industrial Aerodynamics, 24, 193-213 40. Murakami, S. and A. Mochida (1987),Threedimensional numerical simulation of air flow around a cubic model by means of large eddy simulation, Journal of Wind Engineering and Industrial Aerodynamics, 25, 291-305 41. Selvam, R.P. (1992), Computation of Pressures on Texas Tech Building, Journal of Wind Engineering and Industrial Aerodynamics, 41-44, 1619-1627. 42. Selvam, R.P. (1997), Computation of Pressures on Texas Tech Building Using Large Eddy Simulation, Journal of Wind Engineering and Industrial Aerodynamics, 67 & 68, 647-657 43. Aboshosha H, Elshaer A, Bitsuamlak G, El Damatty A (2015), Consistent inflow turbulence generator for LES evaluation of wind-induced responses for tall buildings. J Wind Eng Ind Aerodyn, 142:198– 216.

Volume 47 │ Number 4 │ December, 2017

44. Elshaer, A., H. Aboshosha, G. Bitsuamlak, A. Damatty, A. Dagnew (2016), LES evaluation of wind-induced responses for an isolated and a surrounded tall building, Engineering Structures, 115, 179-195. 45. Zheng, D., A. Shang and M. Gu (2012), Improvement of inflow boundary condition in large eddy simulation of flow around tall building, Engineering Applications of Computational Fluid Mechanics, 6, 633–647 46. Gousseau, P., B. Blocken, G.J.F. van Heijst (2013), Quality assessment of Large-Eddy Simulation of wind flow around a high-rise building: Validation and solution verification, Computers & Fluids, 79, 120-133. 47. Daniels, SJ, Castro, IP and Xie, ZT (2013) Peak loading and surface pressure fluctuations of a tall model building, Journal of Wind Engineering and Industrial Aerodynamics, 120, 19-28. 48. Selvam, R.P., and P.C.Millett (2003), Computer modeling of tornado forces on buildings, Wind & Structures, Vol. 6, pp. 209-220 49. Selvam, R.P. and P.C. Millett, (2005), Large eddy simulation of the tornado-structure interaction to determine structural loadings, Wind & Structures, Vol. 8, pp. 49-60. 50. Strasser, M.N. and R. P. Selvam (2015), A Review of Viscous Vortex Tangential Velocity Profiles for Application in CFD, Journal of the Arkansas Academy of Sciences, 69, 88-97 51. Strasser, M.N., M.A.A.Yousef and R. P. Selvam (2016), Defining the vortex loading period and application to assess dynamic amplification of tornado-like wind loading, Journal of Fluids and Structures, 63, 188-209 52. Alrasheedi, N.H. and Selvam, R.P. (2011) Tornado forces on different building sizes using computer modeling. In ASME Early Career Technical Conference, Fayetteville, AR, USA, 31 March–2 April 2011. 53. Gorecki, P.M. and R. P. Selvam (2015) Rankin combined vortex interaction with rectangular prism, International Journal of Computational Fluid Dynamics, 29, 120-132. 54. Gorecki, P. M., & Selvam, R. P. (2015). Visualization of tornado-like vortex interacting with wide tornado-break wall. Journal of Visualization, 18, 393-406 55. Lewellen, W.S. and D.C. Lewellen (1997), LargeEddy simulation of a tornado’s interaction with the surface, Journal of the Atmospheric Sciences, 54, 581-605 8

Volume 47 │ Number 4 │ December, 2017

56. Hangan, H. and J.D. Kim (2008), Swirl ratio effects on tornado vortices in relation to the Fujita scale, Wind and Structures, 11, 291-302. 57. Natarajan, D. and H. Hangan (2012), Large eddy simulations of translation and surface roughness effects on tornado-like vortices, Journal of Wind Engineering and Industrial Aerodyanmics, 104106, 577-584 58. Lewellen, D. C., (2010), Effects of topography on tornado dynamics: A simulation study. Preprints, 26th Conference on Severe Local Storms, AMS, Nashville, TN, paper 4B.1. 59. Zu, J., G. Yan and C. Li (2016), Investigation of wind pressure of translating tornado on spherical dome structures, Proceeding: 8th International Colloquium on Bluff Body Aerodynamics and Applications, Northeastern University, Boston, Massachusetts, USA June 7 – 11 60. Nasir, Z., G.T., Bitsuamlak, and H.Hangan, (2014), Computational modeling of tornadic load on a building, 6th International Symposium on Computational Wind Engineering, Hamburg, Germany, June 8-12, 2014. 61. Kuai, L.F. L. Haan, W. A. Gallus and P. P. Sarkar, (2008), CFD simulations of the flow field of a laboratory simulated tornado for parameter sensitivity studies and comparison with field measurements, Wind and Structures, 11, 75-96. 62. Sengupta, A., F.L. Haan, P. P. Sarkar and V. Balaramudu (2008), Transient loads on buildings in microbusts and tornado winds, J. Wind Engineering and Industrial Aerodynamics, 96, 2173-2187. 63. Phuc, P.V., T. Nozu, K. Nozawa, and H. Kikuchi, (2012), A Numerical Study of the Effects of Moving Tornado-Like Vortex on a Cube, The Seventh International Colloquium on Bluff Body Aerodynamics and Applications (BBAA7), Shanghai, China; September 2-6 64. Canuto, C. et al. (1988). Spectral methods in fluid dynamics. New York: Springer-Verlag. Canuto, C., M.Y. Hussaini, A. Quarteroni and T.A. Zang (2006), Spectral Methods: Fundamentals in Single Domains, Springer, New York 65. Canuto, C., M.Y. Hussaini, A. Quarteroni and T.A. Zang (2007), Spectral Methods: Evolution to Complex Geometries and Applications to Fluid Dynamics, Springer, New York 66. Karniadakis, G.E.& S. Sherwin (2005), Spectral/ hp Element Methods for Computational Fluid Dynamics, Oxford University Press, Oxford, Second edition

The Bridge and Structural Engineer

ASSESSMENT OF DAMPING FROM FIELD STUDIES Arunachalam SRINIVASAN Director JP-WINCENTRE Jaypee University of Engg. & Technology, Raghogarh, Guna, M.P. E-mail: sarunacha@yahoo.co.in

Summary Structural damping is a very important dynamic property of a tall building. Despite several fullscale measurements reportedly conducted on tall buildings to estimate damping, its behaviour is found to be complex and a complete understanding of damping mechanism still remains elusive. This paper critically reviews various issues related to assessment of damping from full-scale experiments. Design recommendations for estimating damping, based on the work of Tamura’s research group, are currently regarded as more reliable and are discussed. More full-scale experiments are also required to be undertaken in India to improve the data base. Keywords: Structural damping; tall buildings; estimation techniques; field experiments; wind actions.

Introduction

The dynamic response of a given structure such as a tall building, when exposed to random wind loads primarily depends on its dynamic properties, which include its mass, stiffness and structural damping. Damping in the structure is the most important parameter but the most complicated in its treatment as compared to that of mass and stiffness. In most cases, accurate estimation of structural damping prior to completion of construction is very difficult, since damping significantly depends on characteristics of construction material, the process of its manufacturing, and, the types of structural connections adopted for different members of the structural system. This introduces high level of uncertainty in the estimation of the damping value and affects its reliability. Besides,

The Bridge and Structural Engineer

After serving CSIR-SERC, Chennai as Chief Scientist for 30 years, presently working as Director, JP-WINCENTRE at Jaypee University of Engineering and Technology, Guna in Madhya Pradesh. His areas of interest and specialisation include wind engineering, wind effects on structures, boundary layer wind tunnel testing, and cyclone disaster mitigation.

because of non-linear characteristics of materials and scaling problems, prediction of structural damping values of buildings based on model experiments is not considered effective and reliable. Hence the practice of conducting full scale field measurements on tall buildings to estimate reliable damping ratios started as early as in 1970s. Despite such extensive studies, only limited quality information on damping is available in the literature [1]. Typically structural damping ratio (DR)ζ can have a value of 1% for a concrete chimney, whereas its value for an unlined steel chimney can be as low as 0.2%. Similarly for tall buildings several international codes of practice on wind loads recommend structural damping ratio, ζ, ranging from as low as 0.5% to as high as 6% [2] implying a large scatter. Research all over the world and in particular by the teams led by Yukio Tamura in Japan and by Ahsan Kareem in USA, during the last two decades, has brought about significant improvement in the estimation of structural damping by conducting quality full-scale experiments on buildings, and improved measurement techniques and estimation procedures. This paper critically reviews, conventional estimation methods for assessment of structural damping ratios, for tall buildings, and recent advances in the understanding of structural damping mechanism, its amplitude dependency, decreasing trend in high damping value regions, studies on improved semi-analytical techniques for estimation of ζ , points to be observed when conducting quality experiments for measurements of reliable data. It is observed that despite such efforts, more full scale experiments are required to improve the accuracy of data base on structural damping ratio, ζ, and, the reliability in structural design.

Volume 47 │ Number 4 │ December, 2017

Complexities damping

assessing

structural

Damping in real structures often does not follow the simple ideal viscous model. Other types of damping such as friction damping, radiation damping, and aerodynamic damping and soil-structure interactions further complicate the damping behaviour. In practice, the concept of equivalent damping coefficient is adopted and considered appropriate [3]. This implies that the energy dissipated by damping is equivalent to the energy dissipated in all the damping mechanisms combined and present in the actual structure. Generally, the damping ratios are by themselves very small in magnitude (for example, typically ζ is about 0.002 for a steel building). Thus it is not uncommon that a large scatter is reported on the measured values of ζ based on full-scale experiment. For example, the coefficient of variation on ζ is reported to be as high as 70% [4]. This implies that if we assume a typical estimate of ζ =2%, the actual values can vary from 0.6% to 3.4%. This would typically correspond to acceleration response estimates of about 1.8 units to 0.8 units, which has a reduction ratio of about 2.25 between these response estimates. This would evidently affect the reliability of the system. Hence use of proper testing methods, measurement and evaluation techniques, such as, frequency decay decomposition (FDD) and multimode random decrement (MRD) techniques, is essential for reliable estimation of damping ratio [3].

Full scale measurements on structural damping

Over the years, several databases have been created based on full-scale measurements on concrete and steel buildings of various heights, to be able to provide empirical relationships for acceptable level of estimates for design practice. Generally two types of field measurement techniques are reported to measure dynamic properties of tall buildings. These include excitations due to (i) mechanical shakers and (ii) ambient wind conditions. Kwok et al [5] report synchronised human excitation and crane excitation as additional testing methods for full-scale investigations. All these tests use accelerometers for response measurements which enable determination of natural frequencies of vibration, damping values and mode shapes in both translation and torsion modes. The mechanical shaker method has a distinct advantage over ambient wind excitation method. It permits a better control on the

Volume 47 │ Number 4 │ December, 2017

time and frequency of input force. The shaker can be used to generate an input force at resonance and to sustain it for convenient length of period till vibrations are built-up to a sizable amplitude (to get steady-state motion) and then can be stopped suddenly to yield a free vibration decay curve. This can be then analysed to obtain better estimates of damping values. Further, by using shakers with different masses, one can generate different amplitudes of building vibration and investigate the amplitude effect on damping and natural frequency. Such techniques have been used in full scale studies by Jeary [6], Littler and Ellis [7], Butterworth [8], and Kwok [5] among others. The mechanical shaker method, however, is relatively expensive and calls for use of heavy exciters for testing. Lifting, positioning and handling of such shakers at a higher elevation with a large mass in a building, particularly when the building is already occupied, poses problems of portability and safety. Further, in order to avoid influence on the measured structural response due to unwanted or external wind excitations, such tests are required to be performed under light-wind or no-wind conditions. In view of the above, in subsequent times, conducting full scale field studies under ambient conditions has become popular and viable. [1,5,9,10]. Such studies are relatively less expensive and do not require use of heavy exciters. Also all or part of measurements can be used as references and Multi-Input Multi-Output (MIMO) techniques can be used for modal analysis, thus enabling easy handling of closely spaced and even repeated modes for evaluation of damping values [1]. A few case- studies reported in the literature are briefly described and discussed in the following Section 3.1. 3.1 Studies by Davenport and Hill-Carroll Davenport and Hill-Carroll [11] compiled full scale measurement data on structural damping from 165 tall buildings. They found that the ratio of rms tip displacement (σy) to the height of the building, H, was the prime variable influencing the structural damping ratio, ζ and proposed the following empirical relationship.

...(1)

where A is a constant and ‘’n’’ is power law exponent. Both these variables depend on the building type (whether concrete or steel) and the number of storeys as given in Table.1.

The Bridge and Structural Engineer

Table: 1 Empirical constants ‘A’ and ‘n’ for estimation of “ζ” [11] Building Type Steel Concrete

No. of storey 5 to 20 >20 5 to 20 >20

0.03 0.02 0.03 0.025

0.075 0.11 0.11 0.11

It can be seen from Table 1 that damping ratio, ζ, increases with vibration amplitude raised to the 1/10 power approximately. 3.2 Studies on damping mechanism and damping predictor Full- scale experiments were conducted using induced vibration techniques on nine buildings by BRE, UK. Typical building dimensions varied between 12 m to 50 m indicating that low and mid-rise buildings were considered for the tests. Using measured acceleration response data, Jeary [6] observed that the logarithm of rate of increase of damping was proportional to the square root of a linear dimension of the building, in the direction of the mode of vibration. Following the Stick-Slip concept (STICTION) on damping originally proposed by Wyatt [12], and by quoting the explanation given by Griffith[13] that, “when amplitude increases many surface cracks participate and the increase in crack length represents an energy sink which is measured as damping in practical terms”, Jeary[6] proposed the mechanism of damping as consisting of (a) low amplitude plateau, (b) midamplitude region where damping was increasing with amplitude, and, (c) a high-amplitude plateau region. The following empirical expression was proposed for damping predictions. where j = mode number; ζ0 =

...(2) , where T0 is the

natural period (in sec). and

, where D is

representative dimension, in m, aligned in the direction of motion and xH is tip amplitude of vibration in m and H is the height of the building. 3.3 Studies on dynamic properties on prototype buildings-Emphasis on quality procedure Ellis from BRE, UK has proposed the simple formula, f1=46/H, (where H is the height of building in m) for prediction of natural frequency of a tall building and this has been widely recommended in many of

The Bridge and Structural Engineer

the codes of practice [3, 6]. In one of their studies, Littler and Ellis reported measurements on dynamic properties on two buildings (i) Herbershon Court and (ii) Hume Point [14] . Forced vibration technique was adopted using mechanical exciters. Through analysis of (i) response spectrum and (ii) exponential envelope fitting of part of the decay curve, for the Habershon Court, damping values were deduced as 1.75% and 1.88% respectively and the agreement was reported to be excellent. The corresponding values of natural frequency were reported as 1.931 Hz and 1.92 Hz. Besides collection of quality data, Littler and Ellis have laid emphasis on use of proper procedure for analysing the field data. Stationarity, Variance error and bias error were considered highly important in estimation of damping values. Stationary data basically implies that statistical properties do not change with time. Particularly for data collection under ambient wind conditions, stationarity condition is to be satisfied; otherwise the damping ratios are likely to be overestimated. Variance errors imply that a large number of sample records need to be collected and averaged. Variance error is equal to (1/√N) where N is the number of records considered. For example, if N=100, variance error is equal to ± 10%, which reduces to ± 4.4% for N=500. Thus general practice of performing analysis on limited number of records should be avoided. Similarly, bias error occurs whenever there is a sharp change in the amplitude of response with respect to frequency, and there are only few spectral lines present in the spectrum, for example, near the region of natural frequency. This implies that by resorting to large number of frequency resolution and increased total time of recording, bias error can be minimised. However, this leads to a conflict condition that the larger the length of recording, the lesser will be the stationarity of the data. Under such situations, Littler and Ellis suggest the use of selective ensemble averaging technique. Here the long continuous data collected will be divided into several individual records of suitable lengths, and the average wind speed and direction of the individual records will be stored. Subsequently by considering corresponding records obtained under similar wind conditions, the stationarity can be improved. This technique was applied by the investigators in the full scale study on Hume Pont structure where large duration of data of about 2250 hours was considered for analysis. 3.4 Damping studies using full scale data on buildings in Japan A systematic and well planned field program on

Volume 47 │ Number 4 │ December, 2017

collection of full scale data on dynamic properties of buildings in Japan was undertaken by Damping Evaluation Committee of Architectural Institute of Japan (AIJ) led by Yukio Tamura. Due to space constraints, only salient details of the results published by Satake et. al. [14] and Tamura [1, 3, and 15] are discussed in this section. The full scale data were pertaining to a total of 284 tall buildings and other tower-like structures, which included 137 steel farmed buildings (SB), 25 reinforced concrete (RC) buildings, 43 steel framed reinforced concrete (SRC) buildings and 79 tower-like structures. The collected data primarily corresponded to small amplitude region i.e., normalised ratio of tip amplitude to height of the building, (xH/H) ≤ 2 x 10-5 and were regarded as highly reliable in the literature [2, 9]. The following empirical relations were recommended by Tamura et al.[1] for the natural frequency, f1 in Hz. Table 2 : Suggested empirical relation for natural frequency by Tamura et al [1] Sl. No. 1 2

Type of building

Amplitude level Low High

RC buildings Steel buildings

f1=67/H f1=50/H

f1=56/H f1=42/H

Ellis’s equation for f1 f1 =46/H, H is height in m.

It can be seen that as compared to Ellis’s equation of f1 =46/H, expressions given in Table.2, predict higher values of natural frequency for RC buildings, both in lower and higher amplitude regions, and also for steel buildings in region of lower amplitudes.

viz., (xH/H) <2 x 10-5. It was specifically reported by Tamura [1] that although expressions for ζ1 in Eqns. (4) and (5) contain a term depending on frequency, no clear reason could be attributed to frequency. Hence, in a subsequent paper Tamura [3] has reported an improved version of the first terms in Eqns. (4) and (5), as a function of height, H, rather than the function of frequency, f1. The updated damping predictors for RC and steel buildings are respectively given as: (e) ζ1=0.93/H+470*(xH/H)-0.0018...(6) (f) ζ1=0.65/H+400*(xH/H)+0.0029..(7) It is repeatedly stressed in the literature that use of proper damping estimating technique is essential for deducing reliable values of damping ratio. Damping is found to increase only in the small amplitude regions and in case of higher levels of amplitude, say exceeding around (xH/H) >2 x 10-5, there is no evidence of an increase in damping value. It was further reported that once all the friction between structural and non-structural elements is mobilised, a critical tip drift ratio is achieved and with further increase in amplitude, there will be a decreasing trend in the value of damping. The result is based on full scale experimental data on a 100 m tall steel building analysed by Jeary [6] and also based on a 3-storey low-rise building analysed by Fukuwa et al. [16], with results reproduced in Fig. 1. For example, the reported deduced values of critical tip drift ratio, are respectively 10-4 and 3x10-5 for the two cases discussed in Fig. 1 (a) and (b).

Frequency dependency and amplitude dependency of structural damping ratio, ζ, were also studied and following predictors were suggested based on regression analysis. (i) Frequency dependency of ζ: (a) For steel frame buildings: ζ1=0.013f1; (b) for RC buildings: ζ1=0.014f1 ..(.3) (ii) Amplitude dependency of ζ: ..(4)

Eq.(4) is valid for buildings with heights, 10 m< H <300 m and in small amplitude conditions, viz., (xH/H) <2 x 10-5. (d) Similarly,

for

Steel

farmed

buildings, ..(5)

Eq.(5) is applicable for buildings with heights, 30 m < H < 200 m and in small amplitude conditions, 12

Volume 47 │ Number 4 │ December, 2017

Fig.1 (a) and (b) Amplitude dependency of damping (i) ( steel building (H=100m); (ii) low rise building (H=10m) [1,3].

The Bridge and Structural Engineer

Based on their detailed investigations, the following are the design damping ratios for RC and steel buildings in Japan[3]. Table 3 : Recommended values of design damping ratio [1,3] Building Building ζ1 Values Type Height Habitability Safety level level (Elastic range) RC

10-100m 2.5% to 0.8% 3% to 1.2%

Steel

30-200m 1.8% to 0.7% 2% to 1%

3.6 Studies by Kwok et al. Kwok et al presented results of full scale

measurements of dynamic properties of four tall buildings in Hong Kong [5]. The buildings have heights varying between 130m and 256m. Details of instrumentation, test methodology, test results, merits and limitations of four different testing methods employed were discussed. The four testing methods included ambient wind excitations, human synchronisation method, crane excitation and mechanical shaker excitations. Based on field acceleration measurements, values of natural frequencies, damping and mode shapes of both translational and torsional modes were presented. Spectral analysis and free vibration decay curves were primarily employed as estimation techniques. The salient results reported are given in Tables.4 and 5 below:

Table 4 : Comparison of measured and predicted values of fundamental natural frequencies [5] S. No.

Building

1 2 3 4

A B C D

Building height (m) 130 255 256 236

Measured lowest frequency(Hz) 0.703 0.254 0.244 0.273

Ellis’s formula f1=(46/H)Hz 0.354 0.180 0.180 0.195

Formula suggested by Tamura group f1=(67/H)Hz 0.515 0.263 0.262 0.284

Table 5 : Comparison of measured and predicted average values of structural damping [5] S. No. Building

1 2 3 4 5

Building Mode height (m)

A C

130 256

236

1 1(x) 2(y) 1(x) 2(y)

Measured As per Jeary As per As per average formula ζ (%) formula by Davenport damping ratio, [6] Tamura group and Carroll ζ ζ (%) ζ (%) [1] (%) [11] 0.61 0.73 0.84 1.09 0.55 2.02 0.34 1.3 0.67 1.36 0.36 1.23 0.55 1.42 0.36 1.29 0.62 1.32 0.36 1.25

From the Table 4, it is seen that measured values of f1 show significant difference with respect to predicted values, by Ellis’s formula whereas they show better agreement with f1=67/H, suggested by Tamura group. This was attributed by Kwok et al. [5] to the possible similarity between buildings in Hong Kong and Japan in terms of building type, height, construction practice leading to load resisting capacity etc. Similarly from the Table 5 it is seen that the measured average value of ζ is showing relatively better agreement with predicted values by Tamura group, whereas predictions by Jeary’s and Davenport’s predictors are much higher [1, 11]. Even in case of predictions by Tamura formula, one can observe that the measured

The Bridge and Structural Engineer

values are in excess of nearly 52% to 86% with corresponding values. All these bring about the hard fact that assessment of structural damping on buildings is very difficult owing to its complex behaviour and its small magnitude. It is also noted that in this study recent advanced techniques such as FDD and MOD proposed by Tamura have not been used. 3.7 Studies at CSIR- SERC, Chennai In India, only limited full-scale studies have been carried out to investigate wind characteristics and wind induced response studies on structures, although the awareness is increasing for conducting more such studies. Shanmugasundaram et al. [17] carried out Volume 47 │ Number 4 │ December, 2017

full scale experiments on a 52 m tall lattice tower located at CSIR-SERC campus, to investigate wind characteristics and wind induced structural response. Data were collected during normal winds and under cyclonic wind conditions, which occurred during June 1996 and in December 1996. Based on autospectral analysis of acceleration responses, at top of the tower, the lowest natural frequency was reported as 1.17 Hz. This agreed well with the 1.8Hz obtained based on analytical methods. Spectral analysis of measurements during cyclone wind conditions had clearly showed the presence of significant wind energy extending beyond 1Hz, even up to 10 Hz. The damping values were evaluated using half power bandwidth method. The average value of ζ during normal winds was estimated as 1.6%, which was found to increase to 1.72% during June 1996 cyclone and to 2.55% during December 1996 cyclone. It was further reported that the increase in mean wind speed increases overall damping ratio due to an increase in aerodynamic damping and modal damping. Hari Krishna et al. [10] investigated wind, terrain and structural damping characteristics on a 50 m tall guyed lattice mast located on east coast of India under ambient wind conditions... The mast was instrumented with three cup anemometers with direction vanes at 10m, 17m, 29m and 50m levels to measure wind speed and directions. Tri-axial accelerometers were positioned at 29m and at 50m levels, to measure acceleration response at each level both in X and Y directions. Based on auto-spectra of measured responses, it was reported that peak frequencies were observed at 3.05 Hz and 2.03Hz at 29m and 50m respectively. The spectra also showed closely spaced modes of guyed mast. Using half power bandwidth method, values of structural damping ratio, ζ were reportedly scattered between 1% and 3% with an average value of 1.6%. This reasonably compared with 2% value normally suggested for bolted steel structure. 3.8 Investigations on damping studies by Spence and Kareem Significant contributions have been made by the research group at the University of Notre Dame, USA over the years under the leadership of Ahsan Kareem, (i) on the understanding of mechanism of structural damping on tall buildings, (ii) improved analytical methods for estimation of ζ, (iii) programs on fullscale experiments for monitoring of tall buildings under wind actions [18, 19]. As described in previous sections based on studies by different researchers, it is clear that the structural damping, ζ of a building is 14

Volume 47 │ Number 4 │ December, 2017

non-linearly varying with amplitude, x0, and further the natural frequency, f0, also shows a softening tendency with an increase in amplitude. In a recent study by Spence and Kareem [9], it was pointed out that the sources of frictional forces are generally generated from structural connections, and from rubbing actions between structural and non-structural members, which are not only unavoidable in buildings but also become less realistic while attempting to describe frictional forces through deterministic models. Consequence of such an assumption implies that the number of slip surfaces will continue to grow indefinitely with increase in amplitude, whereas as reported by Spence et al., it would be expected that there would be a tendency for saturation of slip surfaces as amplitude increases. Further it was reported that while the concept of randomly distributed slip surfaces was considered by Davenport and Carroll [11], modelling of amplitudes at which these slip surfaces are initiated as random variables was not considered therein. Thus Spence et al., have proposed a new concept-based probabilistic model for amplitude dependent damping ratio. The frictional forces, f0, and the random vibration amplitudes, x0, were considered through their joint probability function, pfoxo. In other words, in the conventional dynamic equilibrium equation for the response of the building, the experimentally observed amplitude-dependent damping and natural frequency have been additionally included. Expressions were derived for (i) amplitude-dependent damping ratio and (ii) amplitude dependent natural frequency. The above model has been calibrated by Spence et al., by applying it to a high-quality, larger data-base on damping on tall buildings from the literature [20]. The performance of the damping model proposed showed a good agreement with the experimental data. The study also pointed out that (i) while the height of the building is generally believed to be a better predictor for damping, the structural system seemed to be a more important parameter in deciding the damping capacity of the building and (ii) damping estimates based on frequency domain analysis tend to yield higher values than those based on time domain methods, supporting similar observations earlier reported in literature.

Discussions

Structural damping is a very important dynamic property of a tall building. However, its behaviour is found to be complex and a complete understanding of damping mechanism still remains elusive. The various full-scale experiments conducted in different parts of the world as seen in previous sections, tend

The Bridge and Structural Engineer

to reflect the character of the buildings tested, the type of connections, type of construction material and construction practice used, structural systems adopted etc., Since above parameters are expected to differ at different locations, generalisation of structural damping estimates for design is rather difficult. Presently in India, such full-scale experiments conducted on tall buildings are highly limited in number and it is required that more such experiments are conducted in the near future to improve the quality of available data base. Currently design recommendations on values of structural damping suggested by the Japanese group are regarded as more reliable. Aerodynamic damping which is more important for design of tall circular chimneys particularly for across-wind responses due to vortex shedding is generally not considered severe relatively for tall buildings.

TAMURA, Y., “Amplitude Dependency of Damping in Buildings and Critical Tip Drift Ratio”, International Journal of high rise buildings, 2012, Vol.1, pp: 1-13.

HAVILAND R., “A study of the uncertainties in the fundamental translational periods and damping values for real buildings”, Massachusetts Institute of Technology, 1976, PB-253, pp. 188.

KWOK, K.C.S., TSE, K.T., and Campbell, S., “Field Measurements of Dynamic Properties of High-rise Buildings”, Advances in Structural Engineering, 2012, Vol.14, No.6, and pp: 11071128.

Jeary, A.P., Damping in Tall Buildings: A mechanism and a predictor: Earthquake Engineering and Structural Dynamics, 1986, Vol.14, pp: 733-750.

LITTLER, J.D., and ELLIS, B.R., “Measuring the Dynamic Characteristics of Prototype Structures, A State of The Art in Wind Engineering, Proceedings of 9th International Conference on Wind Engineering, 1995, pp: 133-154.

BUTTERWORTH, J., LEE, J.H., and Davidson, B., “Experimental Determination of Modal Damping From Full-Scale Testing”, 13th International Conference on Erathquake Enginering”, 2004, Paper No. 310.

SPENCE, S.M.J., and KAREEM, A., “Tall Buildings and Damping: A Concept Based DataDriven Model”, 2014, Journal of Structural Engineering ASCE, 2014, 140(5): pp 1-15.

Summary and concluding remarks

A critical survey of literature on importance and assessment of structural damping in tall buildings based on field studies is presented in this paper. Despite several research efforts, it is observed that there is a considerable scatter in the literature in the full-scale testing methods, estimation techniques, and consequently in values of structural damping ratio, ζ recommended in some of the codes and standards. The STICTION model developed by Wyatt is considered to be the most realistic model for explaining the mechanism of damping in buildings. Experiments have shown that both damping and natural frequency vary non-linearly with vibration amplitudes. The database, compiled by Architectural Institute of Japan (AIJ) is regarded in the literature as one of the most highly reliable and comprehensive data base on dynamic properties of tall buildings. Advanced estimation techniques such as frequency domain decomposition, and multi-mode random decrement technique proposed by Tamura group have been shown to have the potential of accurate and efficient determination of dynamic properties including damping and natural frequencies. Since damping behaviour is complex, more quality, full-scale experiments on tall buildings are required to be undertaken.

References

TAMURA, Y., “Damping in Buildings and Estimation Techniques”, Proceedings of 7th International Advanced School on Wind Engineering, 2010, pp: 315-330.

of 3rd International Workshop on Regional Harmonisation of Wind Loading and Wind Environmental Specifications in Asia-Pacific Economies, (APEC-WW-2006), 2006, pp: 9-14.

KWOK, K.C.S., “Codification Dynamic Structural Properties of Buildings”, Proceedings

The Bridge and Structural Engineer

10. Hari Krishna. P., Annadurai, A., Shanmugasundaram, J., and Lakshmanan, N., “Full-scale Measurements of the Structural Response of a 50 m Guyed Mast under Wind Loading”, Engineering Structures, 2003, Vol.25, pp: 859-867. 11. Davenport, A.G., and Hill - Carroll, P., “Damping in tall buildings: its variability and treatment in design”, 1986, Session: Building Motion in Wind, Proceedings of ASCE Convention, Seattle, pp: 42-57. Volume 47 │ Number 4 │ December, 2017

12. WYATT, T. A., “Mechanisms of damping.” Proceedings of Symposium on Dynamic Behaviour of Bridges, Transport and Road Research Laboratory, Crowthorne, U.K., 1977, pp: 10–21. 13. Griffith, A.A., “Theory of Rupture of Brittle Materials”, Trans., Royal Society Ser., A , 1921 Vol. 21, pp: 163-198. 14. Satake, N., Suda, K., Arakawa, T., Sasaki, A., and Tamura, Y., “Damping evaluation using full-scale data of buildings in Japan”, Journal of Structural Engineering, ASCE, 2003, April, pp. 470~477. 15. Tamura, Y., Suda, K., and Sasaki, A., “Damping in Buildings for Wind Resistant Design”., Proceedings o International Symposium on Wind and Structures for the 21st Century, Korea, 2000, pp: 115-130. 16. Fukuwa, N., Nishizawa, R., Yagi, S., Tanaka, K., and Tamura, Y. “Field measurement of damping and natural frequency of an actual steel-framed building over a wide range of amplitude”,

Volume 47 │ Number 4 │ December, 2017

Journal of Wind Engineering and Industrial Aerodynamics, 1996, 59, pp. 325~347. 17. Shanmugasundaram, J., Harikrishna, P., Gomathinayagam, S., and Lakshmanan, N., “Wind, Terrain and Structural Damping Characteristics under Tropical Cyclonic Conditions, Engineering Structures. 1999, Vol. 21, pp: 1006-1014. 18. Kareem, A., and Gurley, K... “Damping in structures: Its evaluation and treatment of uncertainty.” J. Wind Eng. Ind. Aerodyn., 1996, 59(2–3), 131–157. 19. Kijewski-Correa, T., et al. . “Validating windinduced response of tall buildings: Synopsis of the Chicago full scale monitoring program.” J. Struct. Eng., 2006, 10.1061/ (ASCE) 07339445(2006)132:10 (1509), 1509–1523. 20. Aquino, R. E. R., and Tamura, Y... “Damping based on EPP spring models of stick-slip surfaces.” (CD–ROM), 2012, 13th Int. Conf. on Wind Eng., International Association for Wind Engineering (IAWE), Atsugi, Kanagawa, Japan.

The Bridge and Structural Engineer

Tuned Mass Dampers to control oscillations of Wind-Sensitive Flexible Structures

O.R. Jaiswal

Vikas Thakur

Aditya Ghushe

Professor Visvesvaraya National Institute of Technology, Nagpur, India ojaiswalvnit@gmail.com

Research Scholar Visvesvaraya National Institute of Technology, Nagpur, India vikasmsthakur@gmail.com

PG Student Visvesvaraya National Institute of Technology, Nagpur, India apghushe47@gmail.com

Dr. O.R. Jaiswal is presently working as Professor at the Dept. of Applied Mechanics, Visvesvaraya National Institute of Technology, Nagpur. He received his Bachelor degree in Civil Engineering from VRCE Nagpur, completed M Sc (by research) and Ph D from IISc Bangalore. He was a research associate (Post Doctoral) at University of Liverpool, UK. He is a member of Bureau of Indian Standards (BIS) committee on seismic design of structures (CED 39) and wind analysis of structures (CED 57). He was also Executive Committee member of the Indian Society for Wind Engineering (ISWE), 2009-2011.

V.M. Thakur is presently working as Assistant Professor at the Dept. of Civil Engineering, J D College of Engineering and management, Nagpur. He received his Bachelor degree in Civil Engineering and Master’s degree from RCOEM Nagpur. He is currently pursuing Ph D at VNIT, Nagpur.

Aditya Ghushe is presently pursuing his Master’s degree in Structural dynamics and Earthquake Engineering at the Dept. of Applied Mechanics, Visvesvaraya National Institute of Technology, Nagpur. He received his Bachelor degree in Civil Engineering from Govt. College of Engineering.

Abstract

In this paper, issues associated with the use of Tuned Mass Damper (TMD) in controlling the structural vibrations are described. Since its first use in 1909 by Frahm, TMD has now become a robust and reliable device to control vibrations due to wind as well as earthquake loads. Analytical, numerical and experimental work, have led to optimum parameters of TMD. In many tall buildings, towers and in bridges, TMDs are being commonly used to control wind oscillations. TMD is usually installed as a large mass supported on required stiffness, and provisions for suitable damping are made. In some of the recent studies and applications, TMD has been derived from the existing components of structure. This approach eliminates the deployment of additional mass and also no additional maintenance is required. In such a case, diligently provisions of stiffness and damping are to be made. Use of existing components as TMD offers an attractive alternative to structural designers.

Wind analysis of structures is based on stochastic modeling of wind loads, wherein, temporal fluctuations of wind velocity are modeled as stochastic process which is quantified in terms of Power Spectral Density (PSD). The work by Davenport1 laid the basis for this procedure of wind analysis of structures. Subsequently many models for PSD of wind loads are given by Refs. [2, 3, 4, 5]. The standard text books on Wind Engineering include these methods of wind analysis6-8. This is referred as along-wind response. Similarly, the across-wind response, which occurs due to vortex shedding, is also analyzed using PSD of across wind loads9, 10. These analyses help in properly estimating the along- wind and across- wind response of wind-sensitive structures and accordingly structural engineers make provisions for stiffness and strength to ensure the safety of structures. Structural designers have also used vibration control techniques for civil engineering structures under wind and earthquake

The Bridge and Structural Engineer

Introduction

Volume 47 │ Number 4 │ December, 2017

loads. Vibration control techniques are broadly classified in four types namely: Passive control, Active control, Semi-active control and Hybrid control. Fig. 1 gives details of various tools used in each of these control techniques. Active control devices require continuous power, are costly, require regular maintenance and hence have been used in very selected structures. In order to overcome these limitations, semi-active and hybrid control techniques are proposed and they have also been used in structures. Review articles give detail information on Active, Semi-active and Hybrid control of structures11-14. On the other side, the passive control devices are simple, do not require costly maintenance, and in spite of their limited effectiveness, have been used in many civil engineering structures. Amongst the passive control devices (Fig. 1), metallic Yield Damper, Friction Damper and Viscous Dampers enhance the damping to facilitate energy dissipation. Tuned Mass Damper (TMD) and Tuned Liquid Damper (TLD) act as dynamic vibration absorber and they suitably modify the stiffness, and damping of the structure which helps in energy dissipation. Since the beginning of this century, many research studies have been reported on passive control devices.

For each of these passive control devices, analytical and experimental results are available, and their effectiveness, advantages and limitations has been well discussed15. Present paper focuses on Tuned Mass Damper, which was used for the first time in 1909 by Frahm to reduce oscillations of ship haul. In this paper, first, the issues of optimum parameters of TMD are described, then, limitations due to tuning to one frequency and its remedy, is discussed. Various structures, in which TMD is deployed are mentioned. These TMDs are used for wind and earthquake loads. Further, those structures, wherein, TMD has been derived from the existing components of the structure are discussed. If the TMD is derived from the existing component, then, there is no need of any additional mass and no additional maintenance is required. However, in such case, damping of TMD may not be same as that required for optimum condition. The effectiveness of TMD with non-optimum damping is also discussed. It is emphasized that use of existing part of structure as TMD offers an attractive alternative to structural designers.

Fig. 1 : Various Types of Control Systems

Tuned Mass Dampers

TMD is an auxiliary or supplementary device mounted on main structure, whose dynamic response is to be controlled. In studies demonstrating effectiveness of TMD, the main structure is considered as single degree of freedom (SDOF) 18

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

system. Fig. 2 schematically shows a SDOF with TMD. Here M, K and C are, respectively, the mass, stiffness and damping of the main mass and mt, kt and ct are the mass, stiffness and damping values of the TMD, respectively. The equations of motion for free vibration are

The Bridge and Structural Engineer

... (1)

Rearranging these equations and using the non-dimensional parameters one gets

Where, f is ratio of natural frequency of TMD and natural frequency of main system , is the mass ratio, is the damping ratio of main system and is the damping ratio of TMD. For TMD to be effective, two parameters, frequency ratio (f) and TMD damping ratio (ξ) are to be suitably tuned. This tuning is to be done for given mass ratio, μ and damping of main system β. The optimum parameters refers to optimum frequency ratio, fopt and optimum TMD damping ratio, ξopt for a given value of mass ratio, μ and damping ratio of main system, β.

Fig. 2 : Single Degree of Freedom (SDOF) System with TMD

Optimum TMD Parameters

A TMD gives best results i.e. maximum reduction in dynamic response of main system, if its parameters are optimum. Since the work of Ref. [16], many research studies have been reported on optimum parameters of TMD. There are two approaches towards optimum

... (2)

TMD parameters one is loading-based optimization and second is structure-based optimization. 3.1 Loading Based Optimization Approach In loading based optimization approach, optimum parameters are based on the type of loading. In the past optimum parameters are derived for harmonic loading, white noise loading, band limited noise or as loading with specific type of spectra like Kanai-Tajimi spectra. Ref. [17] developed the optimum parameters for harmonic loading. Optimum parameters were obtained by minimizing the steady state response of support excited damped SDOF system. Ref. [18] presented the optimum parameters for TMD subjected to harmonic and white noise random loading. Ref. [19], using numerical searching technique, obtained mathematical expressions of optimum parameters of TMD for white noise excitation. Ref. [20] derived the closed form expressions for optimum TMD parameters installed on linear damped systems for white noise excitation. The significance of response quantity to be optimized (displacement, velocity and acceleration) was also discussed. Ref. [21] proposed TMD parameters by considering seismic input modeled by a non-stationery modulated Kanai-Tajimi filtered stochastic process. Ref. [22] have obtained optimum parameters for four types of loading i.e., harmonic and white noise applied as force on primary structure and as base acceleration. Further, they have proposed closed form expression for optimum parameters through curve fitting. Table 1 gives expressions for optimum frequency ratio and damping ratio for some of the cases.

Table 1 : Expressions for optimum parameters based on loading dependent approach from various studies (f = frequency ratio ξ = damping ratio of TMD, µ = mass ratio, β = damping ratio of main system, λ (ground frequency ratio) =ωg/ ωs)

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

3.2 Structure Based Optimization Approach In this approach, the optimum parameters are obtained based on characteristics of structure alone and there is no dependency on loading. The optimum parameters are obtained from free vibration characteristics of main system. One of the widely mentioned disadvantage of TMD is that its efficient only near tuning frequency. In the past studies, it is demonstrated that this limitation is removed either by multiple TMD with different time period or by using large damping in TMD. Such large damping is needed in optimum TMD based on free

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

vibration characteristics. These TMDs are quite robust and are not sensitive to variation in loading frequency content and design parameters. One of the first studies on this approach is by Ref. [23]. He suggested that for a system, the optimum TMD parameters are those, at which damping ratio of two modes become equal to the average of damping ratio of main structure and TMD. Ref. [24] extended work done by Ref. [23], and obtained the optimum parameters by equating the modal damping ratio and the natural frequencies of two modes. They also pointed out that for higher

The Bridge and Structural Engineer

mass ratio, it is not possible to achieve the condition specified by Ref. [23]. They further showed that the damping ratio in two modes become equal but they are higher than the average of the damping ratios of main structure and TMD. Ref. [24] obtained optimum parameters for a wide range of system through trial and error approach. They fitted-in a curve and gave closed form expression for the optimum parameters, of undamped (Eq. (3)) and damped system (Eq. (4)).

... (3)

... (4)

Ref. [25] obtained optimum parameters by equating modal damping ratio. He got the same optimum parameters as those given by Ref. [24]. Ref. [26] noted that in the vicinity of the optimum parameters suggested by Ref. [24], the modal mass in two modes tend to become equal. They obtained optimum frequency ratio for undamped system with undamped TMD, as:

... (5)

Refs. [26, 27, 28], also pointed out that if modal damping ratios are made equal, then multiple solutions exist for optimum parameters. The optimum TMD criteria discussed here have been used in various applications and TMDs have been deployed in various structures.

Fig. 3 : Tuned Mass Damper in Taipei 101

Chiba Port Tower is a 125m high steel frame building, with first mode time period of 2.25 s in X direction and 2.7 s in Y direction. It has been fitted with TMD at top as shown in Fig. 4. The mass is a concrete block and stiffness is provided by lateral spring and time period of TMD is 2.24 s and 2.72 s respectively in X and Y direction. TMD has damping of 15% and this TMD is shown to give 30-40% reduction in dynamic response32. For tall chimeys, TMD in the form of outer ring are used in a 183 m tall RC Chimney33. This ring is supported on five supports from RC shell, and the stiffness and damping is provided by bearings (Fig. 5).

Structures Fitted with TMD

TMD is the most commonly used control device to suppress oscillations due to wind loads29. The analytical, numerical and experimental research has decisively established the effectiveness of TMD, and by now many structures are fitted in with TMDs. The information about structures fitted with TMD is given in Ref. [30]. These structures include multistory buildings, towers and bridges. Typical form of TMDs are pendulum type and mass resting on spring with dampers. Geometrical details of TMD used in some structures are described here. Taipei 101, which is a 101 storey building in Taiwan is fitted-in with TMD in the form of pendulum (Fig. 3) 730 ton steel ball suspended with ropes from the 88th floor and its length covers the five floors. The steel sphere, which constitutes the mass of TMD is made of steel plates. The damping is provided with the help of telescopic dampers (Fig. 3)31.

The Bridge and Structural Engineer

Fig. 4 : Tuned Mass Damper of Chiba Port Tower, Japan32

Fig. 5 : Tuned Mass Damper in Chimney33

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

For a 90 m tall steel chimney in Rayong, Thailand, a TMD in form of outer ring suspended with wire ropes is used (Fig. 6). The damping is provided with the help of viscous dampers. The ring is suspended with the help of wire ropes from brackets provided in stack, and the bottom portion of ring is resting on viscous damper arrangement (Fig. 6). The total mass of chimney is 195 ton and mass of TMD is 3.6 ton. The optimal damping ratio required is around 11-12%, but since the maximum relative movement between the ring and stack was limited, the damping ratio of TMD is raised to 40%. The effective damping of stack was 0.5%, which after installation of TMD was increased to about 3.6-4.5%34. Fig. 7 : Tuned Mass Damper - Huis Ten Bosch Tower, Nagasaki31

Fig. 6 : Pendulum TMD and its Component: (a) Steel Ring, (b) Viscous Damper, (c) Suspended Wire Rope, (d) Drawings Showing the Assembly33

The Huis Ten Bosch Tower in Nagasaki, Japan is having height of 105 m and is fitted with one TMD of 7.8 ton comprising of material made of asphalt between steel plates of laminated rubber bearings (Fig. 7). The time period of TMD is 1.54 seconds. Response reduction of 33 to 50% was obtained by use of TMD. Citycorp Tower in New York is equipped with the TMD located in 63rd floor having mass ratio of around 0.7%. The stiffness in TMD is provided by pneumatic springs which is tuned to the frequency of main system and adjustable linear damping in the dampers is from 8-14%. The TMD is having a 400 ton concrete block bearing on a thin film of oil (Fig. 8). 22

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Fig. 8 : Citycorp Tower, New York center TMD

TMD from Existing Components of Structure

Recent studies on TMD have demonstrated that the limitation of tuning with respect to a particular frequency can be overcome by providing large damping, as given by the optimum parameters based on free vibration criteria or structure based optimization. This means that, if TMD has large damping then, its negative effect due to mistuning does not exist, and this makes TMD quite robust. Further, recent studies have also demonstrated that TMD can have any given damping and frequency can be tuned to that particular damping35,36. However, such a TMD may not be as effective as that of TMD with optimum damping. This is a significant and

The Bridge and Structural Engineer

useful finding in view of the fact that, it may not be always possible to provide the optimum TMD damping. Another major concern with TMD is that, it requires reasonably large mass to be deployed. In order to avoid deployment of additional large mass, in some studies, existing component of structure is used as TMD. This approach is quite attractive, since it eliminates the requirement of any additional mass and also, since the existing part is used, there are no additional maintenance requirements. While using the existing components, one has to suitably provide the required stiffness and explore the possibility of providing additional damping. Further, depending on the damping available for this existing component, one can suitably tune the TMD frequency. It may be noted here that if TMD does not have optimum damping, and has a given value of damping, then also TMD helps in reducing the vibrations, but the extent of reduction is less. A typical TMD with optimum damping gives 30-40% reduction, where as a TMD with damping same as that of main system, gives 15-20% reduction. Some of the studies wherein, existing components are used as TMD are discussed here. Ref. [37] have deployed TMD in the form of

Fig. 9a : Weak Storey as TMD

Fig. 10 : Comparison of Seismic Response of Water Tank with and without TMD (Jaiswal 2004)

The Bridge and Structural Engineer

soft storey at the top of the building (Fig. 9a). This storey comprised of RC slab, beams and column, and is tuned to the first natural frequency of building. Mass ratio of TMD was 5.8%. Though this TMD does not have optimum damping, it reduced seismic response by 15 to 20%. This TMD, which covers the entire floor of the building, was shown to be effective in X and Y-direction, and for irregular buildings it was effective in reducing the twisting oscillations also. Ref. [38] has used TMD in the form of limber roof top moment frame. This frame is quite flexible and is tuned suitably to act as TMD. Roof TMD floor mass was approximately 11.6% of the original mass of building. For elevated water tanks, Refs. [39, 40] proposed TMD derived from the existing components, i.e. roof slab and column inside the container, which support roof slab (Fig. 9b). Here, roof slab is detached from the container wall, and hence, roof slab supported by inner columns constitute a TMD. This TMD which does not require any additional arrangement is very lucrative and gives 15-20% reduction under earthquake loading. Fig. 10 shows comparison of seismic response of a particular tank with and without TMD.

Fig. 9b : Existing Components as TMD in Water Tank

Ref. [41] proposed a roof isolation technique to reduce the building vibrations to seismic forces. This system is realized by the insertion of flexible laminated rubber bearings between roof and the columns which support this roof, and the addition of high damping viscous dampers connected between isolated roof and rest of the building (Fig. 11). Isolated roof, flexible bearings, and viscous dampers respectively constitute the mass, spring, and dashpot of such a TMD.

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Fig. 13a : Geometric Details of Maintenance Platform Fig. 11 : TMD in the form of Roof Isolation42

Ref. [42] designed TMD for Minato bridge in Osaka, Japan. Existing bearing of the bridge were replaced by a new floor deck-isolation system. Floor decks and isolation systems together can be viewed as a large mass TMD (Fig. 12), making mass ratio 77%. KanaiTazimi spectrum was used to model earthquake force.

Fig. 13b : Details of Maintenance Platform and Elastomeric Bearing as TMD

Fig. 12 : Tuned Mass Damper in Minato Bridge, Osaka, Japan

Refs. [43, 44] used maintenance platform of RC Chimney as TMD. In the tall RC chimney, maintenance platforms are provided at regular intervals along the height. One such platform at the top level (Fig. 13a) was used as TMD. For this purpose it was proposed that a bearing will be inserted at the support of the platform (Fig. 13b). This mass of platform acts as TMD mass, and the stiffness and damping is provided by the bearings. The effect of such a TMD was studied for along- wind and across- wind response. A 273 m tall RC chimney was considered and the top most platform was converted into TMD with mass ratio of about 1% and TMD damping is taken as 5%. A typical result showing effect on along-wind response (bending moment) is shown in Fig. 14a and effect of TMD on peak deflection in across wind response is shown in Fig. 14b. There is substantial reduction in both the responses. Thus, for tall RC chimneys, use of maintenance platform as TMD offers a good option to suppress along-wind and across- wind response. 24

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

(a) Bending Moment in Along-Wind Response

(b) Tip deflection in across wind analysis Fig. 14 : Effect of TMD on Wind Response

Ref. [28] conducted study on a 5-storey model building and sequentially isolating top two storey, one

The Bridge and Structural Engineer

storey and floor of top storey. The required stiffness and damping is provided by flexible pads (Fig. 15). In such cases, mass ratio, even exceeds 100%. They showed that such TMDs derived from existing part lead to 30-50% reduction in seismic response.

Fig. 15 : Structural Model with and without Inter-Storey Isolation System

Ref. [36] conducted an experimental study on a steel frame model on shake table. The rigid mass at the top of the frame was used as TMD by inserting HDRB (High-Damping Rubber Bearing) (Fig. 16). They fixed the damping of TMD at 12% and obtained optimum frequency ratio by minimizing root mean square displacement of main system. They showed that such TMD gives reduction from 30 to 60%.

Fig. 16 : Details of TMD Device (Angelis et al. 2012)

Conclusions

Amongst various control devices, TMD has been widely used to control the dynamic response of structures due to wind and earthquake loads. Since its first application in 1909 by Frahm, a great deal of research and experimental work has established the efficiency of TMD in reducing the vibration due to wind and earthquake loads. Research work on TMD has been directed at obtaining the optimum parameters, i.e. stiffness and damping of TMD. There have been two approaches to get optimum parameters. In the first approach, optimum parameters are obtained for a particular loading like harmonic force, harmonic

The Bridge and Structural Engineer

base excitation of displacement type, and acceleration type, white noise and band limited loading etc. These optimum parameters are also specific to response, i.e. whether displacement is to be minimized or whether acceleration response is to be minimized. A variety of such optimum parameters are available (Table 1). The second approach is to obtain optimum parameters based on free vibration characteristics. (i.e. structure based approach). In this approach, modal damping ratio and modal frequencies are made equal. In this approach one gets higher value of optimum damping ratio and for certain range of values of mass ratio, μ and damping ratio of main system, β, there are multiple solutions45. Presence of higher damping and higher mass ratio of TMD is shown to eliminate the limitation due to tuning to one single frequency. It has been demonstrated that TMD with higher damping ratio are quite robust and do not give any negative effect as is the case with TMDs with low damping values. Another interesting research finding is that, even if TMD does not have optimum damping, then also, for any given value of TMD damping, optimum frequency can be obtained. Such TMD are also shown to be effective under wind and earthquake loads. Based on analytical, numerical and experimental studies the efficiency of TMD is well established and in last fifty years TMDs have been developed in many buildings, towers and chimneys to control wind response and also for earthquake response. The simple configuration, low maintenance and economy have encouraged the use of TMD throughout the world. In India, the only application of TMD is reported in the concept paper by Ref. [46] for the Air Traffic Tower at Delhi Airport. For 102 m tall ATC tower, a 50t TMD is deployed at 92 m height. Though details of structural configuration of TMD are not mentioned, the TMD enhanced the damping from 1% to around 7% in one direction and 4% in other direction. In recent studies, some more simplification is achieved by diligently exploring the possibility of using existing part of structure as TMD. Such an approach eliminates the requirement of deploying additional mass and there is no additional maintenance. However in this approach, suitable provision is to be made for stiffness and damping of TMD. Once existing part is used as TMD, then, displacement of TMD is also of concern, since it shall not affect the functionality of that part. Ref. [41] used roof top as TMD and provided bearings to get required stiffness and damping. The seismic response was reduced by about 20%. For the elevated water tanks Ref. [39] Volume 47 │ Number 4 │ December, 2017

has used roof slab of the container as TMD. This was achieved by detaching the lateral connection between the container roof and wall. Reduction of 15 to 20% was achieved under seismic loading. It is imperative that if additional damping is provided to TMD then more reduction can be achieved. Another interesting study on use of existing part is by Ref. [43], wherein for a tall RC chimney of 273 m height, maintenance platform at the top level is shown to be effective TMD to reduce along- as well as across- wind response. TMD from existing part offers an attractive alternative to structural designers and depending on the geometry and functional requirements, one can use some existing part as TMD and make suitable provisions to deploy dampers for better response reduction.

References 1.

Davenport AG. The Application of Statistical Concepts to the Wind Loading of Structures. Proceedings of the Institute of Civil Engineers, 1961; 449-472. 2. Simiu E. Wind Spectra and Dynamic Along Wind Response. Journal of the Structural Division, 1974; 100(9):1897-1910. 3. Kareem A. Wind-Excited Response of Buildings in Higher Modes. Journal of the Structural Division, 1981; 107(4):701-706. 4. Solari G. Along Wind Response Estimation: Structural Classification. Journal of Structural Engineering. 1983;109(2):575-580. 5. Solari G. Analytical Estimation of the Along Wind Response of Structures. Journal of Wind Engineering and Industrial Aerodynamics. 1983;14(1):467-477. 6. Simiu E, Scanlan RH. Wind Effects on Structures-Fundamentals and Applications to Design. John Willey & John, INC, 3rd Edition, New York, 1996. 7. Liu H. Wind Engineering. A Handbook for Structural Engineers. Prentice Hall, Englewood Cliffs, N.J, 1991. 8. Homes JD. Wind Loading of Structures. Spon Press, London, 2001. 9. Kareem A. Model for Predicting the Across Wind Response of Buildings. Engineering Structures. 1984; 6:136-141. 10. Liang S, Liu S, Li QS, Zhang L, Gu M. Mathematical Model of Acrosswind Dynamic Loads on Rectangular Tall Buildings. Journal of Wind Engineering and Industrial Aerodynamics. 2002; 90: 1757-1770.

Volume 47 │ Number 4 │ December, 2017

11. Fisco NR, Adeli H. Smart structures: Part I Active and Semi-Active Control. Scientia Iranica, Transactions A: Civil Engineering. 2011; 18: 275–284. 12. Fisco NR, Adeli H. Smart Structures: Part II Hybrid Control Systems and Control Strategies. Scientia Iranica, Transactions A: Civil Engineering. 2011; 18: 285–295. 13. Housner GW, Bergman LA, Caughey TK, Chassiakos AG, Claus RO, Masri SF, Skelton RE, Soong TT, Spencer BF, Yao JTP. Structural Control: Past, Present, and Future. J. Eng. Mech. 1997; 123(9): 897-971. 14. Spencer BF, Nagarajaiah S. State of the Art of Structural Control. Journal of Structural Engineering. 2003; 129 (7): 15. Soong TT, Dargush GF. Passive Energy Dissipation Systems in Structural Engineering. John Wiley & Sons, New York, 1997. 16. Den Hartog JP. Mechanical Vibrations. McGraw-Hill, New York, 1956. 17. Tsai H, Lin G. Optimum Tuned Mass Dampers for Minimizing Steady State Response of Support-Excited And Damped Systems. Earthquake Engineering and Structural Dynamics 1993; 22:957-973. 18. Warburton GB. Optimum Absorber Parameters for Various Combinations of Response and Excitation Parameters. Earthquake Engineering and Structural Dynamics. 1982;10(3):381–401. 19. Bakre S, Jangid R. Optimum Parameters of Tuned Mass Dampers for Damped Main System. Structural Control and Health Monitoring. 2006; 14:448–470. 20. Tigli O. Optimum Vibration Absorber (Tuned Mass Dampers) Design for Linear Damped Systems Subjected to Random Loads. Journal of Sound and Vibration. 2012; 331:3035–3049. 21. Marano GC, Trentadue F, Greco R. Stochastic Optimum Design Criterion of Added Viscous Dampers for Buildings Seismic Protection. Structural Engineering and Mechanics.2007; 25(1):21-37. 22. Salvi J, Rizzi E. Closed-form Optimum Tuning Formulas for Passive Tuned Mass Dampers under Benchmark Excitations. Smart structures and Systems.2016; 17 (2):231-256. 23. Villaverde R. Reduction in Seismic Response with Heavily-Damped Vibration Absorbers. Earthquake Engineering and Structural Dynamics. 1985; 13:33–42.

The Bridge and Structural Engineer

24. Sadek F, Mohraz B, Taylor A, Chung R. A Method of Estimating the Parameters of Tuned Mass Dampers for Seismic Applications. Earthquake Engineering and Structural Dynamics. 1997; 26, 617-635. 25. Miranda J. On Tuned Mass Dampers for Reducing the Seismic Response of Structures. Earthquake Engineering and Structural Dynamics. 2005; 34:847–865. 26. Gajre A, Dhage S, Jaiswal O. A New Approach to Obtain Optimum TMD Parameters using Equal Modal Mass Criteria. 14th Symposium on Earthquake Engineering, Roorkee, India, 2010. 27. Moutinho C. An Alternative Methodology for Designing Tuned Mass Dampers to Reduce Seismic Vibrations in Building Structures. Earthquake Engineering and Structural Dynamics. 2012; 41:2059–2073. 28. Reggio A, De Angelis M. Optimal EnergyBased Seismic Design of Non-Conventional Tuned Mass Damper (TMD) Implemented Via Inter-Story Isolation. Earthquake Engineering and Structural Dynamics. 2015; 44:1623–1642. 29. Aly AM. Proposed Robust Tuned Mass Damper for Response Mitigation in Buildings Exposed to Multidirectional Wind. Struct. Design Tall Spec. Build. 2014; 23(9): 664–691. 30. Soto MG, Adeli H. Tuned Mass Damper. Arch Comput Methods Eng. 2013; 20:419–431. 31. Li QS, Zhi L, Tuan AY, Kao C, Su S, Wu C. Dynamic Behavior of Taipei 101 Tower: Field Measurement and Numerical Analysis. Journal of Structural. Engineering. 2011, 137(1): 143-155. 32. Connor J, Laflamme S. Structural Motion Engineering. Springer International Publishing, Switzerland, 2014. 33. Brownjohn J, Carden E, Goddard C, Oudin G. Real Time Performance Monitoring of Tuned Mass Damper System for a 183 m Reinforced Concrete Chimney. Journal of Wind Engineering and Industrial Aerodynamics. 2010; 98:169-179. 34. Areemit N, Warnitchai P. Vibration Suppression of a 90 m Tall Steel Stack by using a HighDamping Tuned Mass Damper. Proceedings of the 8th East Asia-Pacific Conference on Structural Engineering and Construction, Singapore, December, 2001.

The Bridge and Structural Engineer

35. Miranda J. Discussion of System Intrinsic Parameters of Tuned Mass Dampers used for Seismic Response Reduction. Structural Control and Health Monitoring. 2016; 23:349–368. 36. De Angelis M, Perno S, Reggio A. Dynamic Response and Optimal Design of Structures with Large Mass Ratio TMD. Earthquake Engineering and Structural Dynamics. 2012; 41: 41–60. 37. Jaiswal OR, Bakre SV. Use of Weak Storey at Top Floor as Tuned Mass Damper for Controlling Seismic Response of Multistoried Buildings. 12th Symposium on Earthquake Engineering. Roorkee, India, 2010. 38. Johnson J, Reaveley L, Pantelides C. A Rooftop Tuned Mass Damper Frame. Earthquake Engineering and Structural Dynamics. 2003; 32:965–984. 39. Jaiswal OR. Simple Tuned Mass Damper to Control Seismic Response of Elevated Tank. 13th World Conference on Earthquake Engineering. Vancouver, B.C., Canada, 2004. 40. Dhage S. Design of Tuned Mass Dampers for Elevated Water Tanks. MTech Thesis, VNIT, Nagpur, India, 2012. 41. Villaverde R. Roof Isolation System to Reduce the Seismic Response of Buildings: A Preliminary Assessment. Earthquake Spectra. 1998; 14(3):521–532. 42. Hoang N, Fujino Y, Warnitchai P. Optimal Tuned Mass Damper for Seismic Application and Practical Design Formulas. Engineering Structure. 2008; 30:707 -715. 43. K R C Reddy, O R Jaiswal, P N Godbole, 2011, “Wind Response Control of Tall RC Chimneys”, Journal of Wind & Structures, Vol. 8, No.1, pp 1-9. 44. Reddy K. Control of Wind Response of Tall RC Chimneys Using Tuned Mass Dampers. PhD Thesis, VNIT, Nagpur, India, 2012. 45. Thakur V, Jaiswal OR. A New Method to Obtain TMD Parameters for Seismic Applications. Manuscript under Review. 2017. 46. Ho S, Mccormick F, Sheppard J, Caetana E. The Delhi Air Traffic Control Tower: Engineering, Architecture and Design with TMD for the Tallest ATCT in India.18th Congress of IABSE. Seoul, 2012.

Volume 47 │ Number 4 │ December, 2017

Effective static wind load distributions for large roofs John Holmes, has a Bachelor degree from University of Southampton UK, and a Ph.D from Monash University, Australia. He worked at James Cook University and CSIRO Australia, before founding JDH Consulting in 1996. His main area of interest is wind loading of structures, and has received a number of awards, including the A.G. Davenport Senior Medal from the International Association for Wind Engineering in 2011.

John Holmes Director, JDH Consulting Mentone, Victoria, Australia John.Holmes@jdhconsult.com

Summary A general approach to the development of wind load distributions for large roofs should incorporate mean, background, and for some very large structures, resonant components. Based on the author’s experience with development of wind loading for large stadium structures over twenty years, the paper discusses the options available for acquisition, and post-processing of wind-tunnel data, to provide rational wind loading distributions for structural designers of large roofs.

Introduction

Wind is the critical loading in the design of large roofs of partially-enclosed outdoor sports stadiums, and enclosed indoor facilities. However, the response of these structures can be complex, as wind can produce a variety of load distributions and members in the roof structure may respond differently to them. For very large roofs, resonant dynamic response may be significant, although it is unlikely to be dominant for any roof (unlike long-span bridges or very tall structures). Generally simplified approaches in codes and standards are not appropriate for large roofs. Code loads although generally conservative, may, in some cases, result in under design, because the typical quasi-steady load distribution is not necessarily the worst load case for individual members, which may have complex influence lines, and hence will respond differently to the varying spatial distributions that occur in turbulent winds. Although fluctuating wind pressures on large roofs are complex in nature, varying both temporally 28

Volume 47 │ Number 4 │ December, 2017

and spatially, the structural designers of these large structures generally prefer Effective Static Wind Load (ESWL) distributions in a similar format to those provided in the codes and standards. However, it is a challenge for wind engineers to provide rational distributions for structures which are large in dimensions in relation to the length scales in the fluctuating external pressure field acting, have complex structural influence lines, and may have some low-frequency resonant modes with complex mode shapes. Over the last twenty years, new approaches have been developed for processing wind-tunnel model measurements to handle the complexities of wind loading, and to generate appropriate and realistic load distributions. These are reviewed in the present paper.

Wind tunnel pressure measurements

2.1 Two Approaches to Effective Static Load Distributions Because of the large fluctuating component in the wind loading on large roofs, and the nature of the separating-re-attaching flow that produces it, the statistical correlation between pressures at points separated by large distances is small. This can be of economic advantage to the cost of the structure when the reduced correlations are incorporated into the ESWL distributions, either explicitly or implicitly. One of two methods can be used when processing wind-tunnel measurements of fluctuating pressures on large roofs: ●

In the first ‘correlation’ approach, the correlations between pressure fluctuations at different parts

The Bridge and Structural Engineer

of the roof are measured directly, and expected pressure distributions corresponding to peak load effects are obtained. ●

A direct, or ‘time history’, approach in which simultaneous time histories of fluctuating pressures from the whole roof are recorded and stored. These are subsequently weighted with structural influence coefficients to obtain time histories of load effects. The instantaneous pressure distributions coinciding with peak load effects are then identified and averaged. This approach also incorporates the correlation effects, which are implicit in the fluctuating components of the pressures at various points, or panels on the roof.

The first (correlation) approach is based on the LRC formula, developed by Kasperski and Niemann[1], which provides the ‘expected’ (in a statistical sense) distribution of instantaneous pressure, that will most likely coincide with a maximum value of a load effect, such as the tension force or bending moment in a structural member of a roof. This equation can be written in its simplest form as:

LRC

at a point on the roof = a peak factor

reason, the calculation of ρLRC requires knowledge of correlation coefficients for every pair of points, or panels, all over the roof, and are fairly complex calculations for a large roof. However, the matrix functions in EXCEL or MATLAB can be used to advantage for these calculations, as discussed in a following section. The direct approach is conceptually simpler, but also requires structural influence coefficients, and a considerable amount of wind-tunnel time and subsequent computation, because averaging over multiple samples is necessary to achieve ‘stable’ load distributions. 2.2 Relative Advantages and Disadvantages The relative advantages of the two methods are summarized below: Correlation method Advantages: a)

less computation time – this is because of the averaging carried out once in the calculation of correlations between the fluctuating pressures,

much less storage of data from the wind-tunnel testing is required.

× correlation coefficient between the pressure and the load effect

Direct method

× standard deviation of pressure at that point

Advantages:

... (1)

as defined above should be added to the mean, LRC or static, component of loading, and to any resonant components if they are significant. Of the terms on the right-hand side, the first and third are relatively straightforward to understand and determine. The standard deviation of fluctuating pressure is simply the root-mean-square fluctuating point pressure that is routinely measured in wind-tunnel tests, in the form of a non-dimensional coefficient. The peak factor is typically a number between +/-3 and +/-5 (often +/-4 is used) that depends on the probability distribution of the fluctuating pressure and the sample time used (e.g. 10 minutes to 1 hour in full-scale time). The second term is the Load-Response Correlation (LRC) coefficient (ρLRC), and is less easy to determine. It is the correlation coefficient between the fluctuating pressure at the point of interest, and the load effect (tensile force, bending moment etc.). The latter may depend on pressures from all parts of the roof, with the relationship determined by a set of influence coefficients obtained by structural analysis. For this The Bridge and Structural Engineer

conceptually simpler to understand for the nonwind engineer,

it is somewhat easier to calculate dynamic (resonant) components of loading with this method.

The two methods have been compared by Holmes and Wood[2] for the same structure, and shown to give very similar equivalent static wind load distributions, within the statistical variability inherent in the two approaches.

Influence Coefficients and the LRC Coefficient

3.1 Influence Coefficient The term influence ‘coefficient’ is somewhat of a misnomer, as ‘coefficient’ is normally used for nondimensional quantities, as in ‘pressure coefficient’. An influence coefficient, in the structural context, is normally dimensional. For example, for a deflection influence by a panel pressure, the unit might be ‘metres per kPa’. In that case, the influence coefficient would Volume 47 │ Number 4 │ December, 2017

be the deflection at a point (say ‘j’), produced by a unit pressure applied to a panel somewhere else on the roof (say ‘i’). 3.2 Matrix Calculation of the LRC Coefficient For a roof that has been divided into N discrete panels, the LRC coefficient can be written (e.g. [3], Eq. 8):

... (2)

This is an equation for the correlation coefficients between the time-varying pressure at a panel j, pj(t), and the structural response, r(t), which, for a linear structure, is given by:

... (3)

where βi is the influence coefficient - that is the value of the response, or load effect, r, when a unit pressure is applied to Panel i. Hence, Eq. (3) gives the total response as a summation of contributions from all N panels on the roof (including underside or interior panels). Equation (2) can be written:

Equation (5) can be further expanded to calculate ρLRC for a number of load effects or responses (say r = 1 to M), in a single set of matrix calculations.

... (4)

where ρpi,pj is the correlation coefficient between the fluctuating pressure on Panel i and that on Panel j, σpi and σpj are the standard deviations of the pressure on Panels i and j respectively, and σr is the standard deviation of the load effect. For all N panels (j = 1, N):

The importance of resonant contributions to the wind loading of large roofs (as opposed to long-span bridges, or very tall buildings) is often overstated. The lack of correlation of the applied pressures over a large area means that the generalized forces required to drive the complex resonant mode shapes of a large roof, are usually small in magnitude. However, for some large roof spans (such as the roof of the new Wembley Stadium – see later discussion) with low natural frequencies, the resonant contributions can be significant. 4.1 Resonant Response and Resonant Contribution to Equivalent Static Load Distributions

For resonant mode k, the peak inertial load on a roof panel, i, is:

or in shorthand form:

... (5)

where [ρpi,pj] is a symmetric matrix (N by N) of correlation coefficients for every pair of panels 1 to N. [σpi] is a diagonal matrix of the standard deviations of the panel pressures for i = 1, N. The standard deviation of the load effect, σr, that appears on the left-hand side of Eq. (5) can also be written in a matrix form [4]: 30

Volume 47 │ Number 4 │ December, 2017

... (6)

where, mi is the mass of the panel, nk is the natural frequency of the kth mode in Hertz, μik is the mode shape coordinate for panel i and mode k, and gk is a peak factor that can be taken as (assuming Gaussian response, [5]):

Incorporation of Resonant Components

For low damping, the effective load distribution due to the resonant dynamic response is given by the following expression for inertial force (i.e. mass times acceleration).

... (7)

The total contribution, summed over all panels, from the inertial forces of mode k to the maximum, or minimum, value of a load effect, such as the internal forces in a structural member, is given by:

... (8)

Both μik and βi can take positive and negative values, and it can be seen from Equation (8) that the resonant contribution of a particular mode to a member response may be significant, if the mode shape μik has a similar spatial shape to that of the influence coefficient, βi. Conversely, a dynamic mode with a mode shape that is anti symmetric with respect to the centreline of a

The Bridge and Structural Engineer

roof will make no contribution to a member force that has a symmetric influence line. σak is the square root of the resonant component of the mean square generalized deflection in the kth mode of vibration. For narrow-band response, it can be calculated by the well-known approximation [6]:

... (9)

where Gk is the generalized mass in mode k; ηk is the critical damping ratio, and SFK (ηk) is the spectral density of the generalized force, Fk, in the kth mode, evaluated at the natural frequency nk. SFK (ηk) is most easily calculated from the time histories of the fluctuating pressures, weighted by the mode shape.

of the influence lines to pressure correlation length, and to the effective span-wise loading distributions for the load effects. A practical example of a multispan beam is a roof purlin. Fig. 1 shows the beam with four simple supports, and three equal spans. The influence line for the bending moment at the centre of the beam, P, is also shown. This plot shows the bending moment at the point P, for a unit point load applied any position along the span; the load position is plotted on the abscissa of the influence line plot. It can be noted that the influence line is positive in the central span, but negative in the side spans. An influence line of varying sign is also a common characteristic of structural members in large roofs.

4.2 Combination of Mean, Background and Resonant Components to Equivalent Static Load Distributions The correct method of combining the mean, background and resonant components of the ESWL, associated with a load effect or response, is by the following equation:

... (10)

where the absolute values of the weighting factors, Wback, Wres,1 are given by: ... (11) ... (12) Note that since both maximum and minimum responses can be important, both positive and negative values of the weighting factors Wb and Wres* may be required. The above equations are based on the assumption that fluctuating background, and resonant components of loading and structural response are uncorrelated with each other. It can also be shown that the peak structural response obtained from Eq. (10) is correctly given by: ... (13)

Examples

5.1 Three-Span Beam Case A simple three-span beam,with relatively simple influence lines for reactions and bending moments, can usefully be used to demonstrate the relationship The Bridge and Structural Engineer

Fig. 1 : Influence Line for the Bending Moment at the Centre of a 3-Span Beam

Now consider the beam to be subjected to a turbulent fluctuating wind load of zero mean, but with uniform standard deviation, and an exponential correlation function of the following form:

... (14)

where y is the separation distance of any pair of points along the beam, L is the overall beam length, and C is the ratio of the integral length scale to L CL). (i.e. As discussed previously, the effective static wind load distribution to maximize a load effect such as the central bending moment, depends on the correlation properties of the loading – as defined by Eq. (14), and on the influence line – in this case given by the function shown in Fig. 1. Based on the LRC formula, the normalized ESWL distributions to maximize the central bending moment have been calculated for three values of C: 0.2 (low correlation), 2, and 20 (high correlation), and are shown in Fig. 2. Volume 47 │ Number 4 │ December, 2017

The upper graph in Fig. 2 shows that for low correlation lengths with respect to the beam length, the effective static load distribution takes a form similar to the influence line of Fig. 1; that is the structural influence line strongly affects the ESWL appropriate to the central bending moment. This case is similar to that of a large roof on a stadium with low pressure correlations resulting from the separating-re-attaching flow.

Wind-tunnel tests at a scale of 1/400 were carried by MEL Consultants in 1997, using the Monash University wind tunnel. 16 pressure measurement panels, with internally machined manifolding connecting the pressure taps, were provided on the fixed West roof, and 12 on the North roof, with two panels on the underside surfaces in each case. One of the moving roofs was provided with 14 manifolded panels on the upper surface, and 2 underneath.

However, for high correlation/beam lengths (the lower graph in Fig. 2), the influence line has little effect, and the ESWL is close to the distribution of maximum pressures along the beam. This is similar to that of a small structure for which the quasi-steady model of wind loading works quite well, and is the basis of most wind codes and standards.

The correlation method, discussed earlier in this paper, was used to determine structural loads and ESWL distributions for a variety of representative structural members, four different roof open/closed conditions, and for sixteen wind directions. Due to the inherent stiffness of the structures, resonant response was neglected and only the mean and background components of wind loading were considered.

Fig. 3 : Docklands Stadium, Melbourne

Fig. 2 : Effective Static Wind Load Distributions for a 3-Span Beam

5.2 Docklands Stadium, Melbourne The Docklands Stadium in Melbourne (Figs. 3, 4) is the second largest stadium in Melbourne (after the Melbourne Cricket Ground), and the largest in Australia with moving roofs. The stadium has an oval plan form with fixed roofs on four sides supported by large trusses. There are two moving roofs, and in the fully-closed position, the arena is totally enclosed. In the open position, the moving roofs retract over the fixed roofs on the east and west sides. The structure is thus comprised of six independent roofs. 32

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Fig. 4 : Wind-Tunnel Model of Docklands Stadium, Melbourne

5.3 Olympic Stadium, Sydney The Olympic Stadium at Homebush in Sydney was completed in 1999 in time for the Olympic games of 2000. The Stadium is on an oval plan form, and the east and west roofs form part of a hyperbolic paraboloid, but open over the playing area. The two roofs are supported along their leading edges by two large space frame trusses. The wind-tunnel testing was

The Bridge and Structural Engineer

carried out at the University of Sydney in 1996, at a scaling ratio of 1:500. The tests were carried with two configurations: Olympics and Post Olympics, with the former case containing extra seating on the north and south sides. Relying on the symmetry of the stadium, only the west roof of the model was instrumented with pressure tappings. Twenty-five pressure measurement panels were used, including one on the underside, with up to four individual tappings being manifolded together within each panel. The pressure correlation approach was also used for this roof; however, in addition time histories of the fluctuating panel pressures were recorded and stored for the critical wind directions. The latter were used to determine the resonant response contributions for three critical damping ratios, using the approach discussed earlier in this paper. However, the resonant contributions were found to be quite small, and were included in the final design wind load distributions as a simple 10% increase. 5.4 Wembley Stadium, London The ‘new’ Wembley Stadium was completed in 2007, although the design work for the roof was carried out much earlier, with the wind-tunnel testing undertaken by BMT Fluid Mechanics in 2000, and additional consulting on wind loads provided by JDH Consulting. The signature feature of the stadium is the huge east-west running arch, rising to 133 metres, with a span of 315 metres (Fig. 5). The arch has a circular lattice cross section of 7 metres diameter with circular members, and supports the gravity and wind loads from the entire north roof. There are ‘doors’ on the east and west of the roof which can be open or shut, and both configurations were studied by BMT.

made the results unreliable), and the direct wind loads on the arch itself were calculated separately in a ‘desktop’ study by JDH Consulting; however, wind loads from the roof transmitted to the arch were calculated in the post-processing of the wind-tunnel data. Extensive processing of the recorded wind-tunnel pressures was undertaken by both BMT and JDH Consulting, with BMT adopting the ‘direct’ approach based on time histories, and JDH the ‘correlation’ method. Resonant dynamic response was included in the calculations for selected members, with as many as thirteen modes included. An interesting aspect was the variability of the dynamic response factor (d.r.f.), depending on the degree of similarity between the influence line and the mode shape(s) (see earlier discussion of this topic). For example, the d.r.f. varied from 1.05 to 1.69 over the range of calculated peak member forces and moments.

Conclusions

Modern approaches to determination of wind loading distributions on large roofs such as those on large sports facilities have been described. Both methods described make use of pressure measurements on wind-tunnel models, and the post processing incorporates structural influence coefficients to determine effective static wind load distributions and can be used to directly determine individual member internal forces and bending moments. For very large roofs, resonant response can be significant, and a method is given for incorporating inertial load distributions with the mean and background components to give combined effective loads. Finally, it should be noted that attempting to use conventional modal analysis to construct total structural responses (i.e. with background and resonant contributions from each mode), although apparently mathematically possible, is a poor approach for two reasons: a)

Contributions from numerous higher dynamic modes are required to accurately construct the total background response, even though their resonant contributions are negligible. This was demonstrated in 1995 by Vickery [7].

The method cannot be used to construct equivalent static load distributions, as the distributions for the background and resonant components are quite different from each other.

Fig. 5 : Wembley Stadium, London

The wind-tunnel model included 118 measurement panels, including those on the ‘doors’, and testing was carried out at ten degree intervals. No wind load measurements were made on the arch for scaling reasons (Reynolds Number differences would have

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

References 1.

Kasperski, M. and Niemann, H-J. (1992). “The LRC (Load-Response-Correlation) Method: A General Method of Estimating Unfavourable Wind Load Distributions for Linear and Non-Linear Structural Behaviour”, Journal of Wind Engineering and Industrial Aerodynamics, 1992, Vol. 43, pp. 1753-1763. Holmes, J.D. and Wood, G.S., “The Determination of Structural Wind Loads for the Roofs of Several Venues for the 2000 Olympics”, ASCE Structures Congress, Washington, D.C., May 23-25, 2001. Holmes, J.D., “Effective Static Load Distributions in Wind Engineering”, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 90, 2002, pp. 91-109.

Volume 47 │ Number 4 │ December, 2017

Holmes, J.D. and Best, R.J. “An Approach to the Determination of Wind Load Effects on LowRise Buildings”, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 7, 1981, pp. 273-287.

Davenport, A.G., “Note on the Distribution of the Largest Value of a Random Function with Application to Gust Loading”, Proceedings, Institution of Civil Engineers, Vol. 19, 1964, pp. 449-471.

Crandall S.H. and Mark W.D., Random Vibration in Mechanical Systems, Academic Press, New York, USA, 1963.

Vickery B.J., “The Response of Chimneys and Tower-Like Structures to Wind Loading”, in A State of the Art in Wind Engineering, Wiley Eastern Limited, 1995.

The Bridge and Structural Engineer

Special Wind Engineering Considerations for Very Tall and Slender Buildings Suresh Kumar spent many years as a wind engineering researcher/ consultant in India, Canada and Netherlands. He had the privilege to work as a wind consultant on many iconic structures around the world.

K. Suresh Kumar PhD, PEng Principal / Regional Manager Wind Engineering Consultant RWDI India Trivandrum, India suresh.kumar@rwdi.com

Summary There seem no limits to heights or slenderness when it comes to design of tall buildings. Currently, a very tall building of height 1 km and very slender building of height/width ratio 22 are under construction. There are many upcoming buildings with heights much higher than 500m and slenderness greater than 15. Special wind engineering considerations are required to make these buildings appropriately wind resistant and buildable. Considering the ever-increasing height/slenderness of these buildings beyond normal building height/slenderness, attention must be paid to the wind climate variations along the height of the building. Further, proper aerodynamic shapes must be adopted to make the building wind resistant in a natural way. Finally, advanced aeroelastic tests as well as large scale model tests are often required to confirm the traditional low Reynolds number small scale test findings from rigid models. These topics are elaborated in this paper with the help of case studies.

Introduction

So far the building heights have been mentioned in meters. This is going to change soon with the introduction of 1 km high Kingdom Tower in Jeddah which is under construction now. Further, there are other studies done, or in the process, for evaluating wind-induced response of tall buildings with heights in the range of 1 to 1.2 km. There are also discussions going on in the community about a mile-high tower. The other important parameter is slenderness. Considering the non-availability of large land areas, the foot prints of the buildings are getting smaller

The Bridge and Structural Engineer

and thus the slenderness (height/width) is on the rise. There are many buildings coming up with slenderness in the range between 10 and 20. Building with a slenderness of 22 is already under construction in New York, which is the 430-m high Steinway Tower at 111 West 57th Street. Tall, extremely flexible towers are piercing into the unexplored portion of the atmosphere. These will attract large wind loads. But a judicious treatment of wind climate variations along the height of the building, and proper geometric shaping, can potentially reduce the wind loads on such super tall/slender buildings. Advanced wind tunnel tests such as aeroelastic tests and large scale model tests can also help in optimizing the wind-induced response of these super tall/slender buildings.

Wind Climate at High Altitudes

Traditionally, data from wind tunnel testing is combined with statistical wind climate model of the site to arrive at the wind-induced loads on tall buildings. The current climate model is based on surface measurements taken at airports / meteorological stations extrapolated over the height of the tower, using established wind engineering models. Quite often established power law method is used to convert the surface speeds to the speed at the top of the tower or at some reference heights in the range of 500m. However, applicability of these models is uncertain beyond 500m especially in the case of peculiar wind storms such as Shamal winds and thunderstorms. Figure 1 shows typical mean wind profiles of different storm types. Historical wind data is recorded at surface level height of about 10 m. One can see from Fig. 1 that the vertical scaling of

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

the surface data depends entirely on the storm types. The popular power law method for scaling predicts higher speed as the height increases; this model can represent the wind profiles corresponding to synoptic winds and extreme winds such as cyclones/hurricanes. However, this model will not work for shamal and thunderstorm wind events, where upper level speeds are much lower than the speed at the lower level.

Fig. 1 : Typical Mean Wind Profiles of Different Storm Types

Two of the projects considered here are (1) the current world’s tallest tower, Burj Khalifa in Dubai, and (2) the upcoming world’s tallest tower, Kingdom tower in Jeddah. While Burj Khalifa rose to 828 m above ground, Kingdom tower is soaring to one kilometer above ground. Heights of both towers are far beyond the 500 m range, hence requiring careful attention. 2.1 Burj Khalifa, Dubai, UAE Strong northerly and northwesterly winds, known as the “Shamal” (meaning ’north’), produce the most widespread hazardous weather known in the Arabian (Persian) Gulf region and are caused by the presence of a large pressure gradient that develops behind a cold front passage. Other extreme winds in the region are typically associated with mid-latitude cyclones and severe thunderstorm systems. A common phenomenon with Shamal and thunderstorm wind events is the high speed jet flow near the lower boundary layer and the reducing wind speeds above this layer. In these cases, the vertical profile of wind speed between the surface and the level of maximum wind speed can be unusually steep. A general lack of upper-level wind observation data is an obstacle to understand the characteristics of vertical wind profiles in Shamal and other high wind events in Dubai. For determining design wind speeds for the Burj Khalifa project, RWDI previously used long-term hourly wind data from the surface station in Dubai

Volume 47 │ Number 4 │ December, 2017

to predict the 50-year return period wind speed. The resultant speed was scaled to the top of the boundary layer using a traditional power-law extrapolation. In addition, upper air data from Abu Dhabi (which is about 150 km away) were analyzed to provide a direct prediction of the 50-year wind speed at the top of the boundary layer. The Abu Dhabi upper air data, however, were not hourly measurements (useable readings available at an average rate of approximately two per day). As such, these data were considered to provide a less reliable estimate of 50-year wind speed than the surface data at Dubai. Given the unusual characteristics of vertical wind profiles during Shamal and thunderstorm events, the question remains as to how the surface data from Dubai should be scaled to the top of the boundary layer. Additional information on upper-level winds can be obtained from the National Center for Atmospheric Research/National Center for Environmental Prediction (NCAR/NCEP) global reanalysis data set. These data are based on world-wide meteorological observations interpolated to the 3-dimensional grid by means of meteorological modeling. However, the horizontal and vertical grid resolutions of the output are too coarse for a thorough study of local wind profiles at the study site. Due to the data resolution issues mentioned above, neither the upper air analysis at Abu Dhabi, nor the NCAR/NCEP Reanalysis alone can provide a thorough assessment of vertical profiles during strong wind events in Dubai. To address this issue, the upper air observation data and the NCAR/NCEP reanalysis data with a highresolution numerical mesoscale meteorological model were combined to reproduce high-resolution 3-D wind fields for a selection of historical extreme wind events in the study area. The model, known as the Fifth-Generation Penn State/NCAR Mesoscale Model (MM5), is a limited-area, nonhydrostatic, terrain-following model designed to predict mesoscale atmospheric circulations. MM5 is a widely used meteorological model that is based on solving the fundamental equation of atmospheric motion on a 3-dimensional grid. The model incorporates parameterizations for the various grid scale and sub-grid scale physical processes that influence atmospheric conditions such as convection, cloud formation, precipitation, radiation, surface heat transfer and moisture flux, etc.

The Bridge and Structural Engineer

The Shamal wind events are large in scale and, to model them effectively, it is necessary to model a large geographic area. This was done efficiently by using a series of nested model domains as shown in Fig. 2, with each nest being at 3 times higher spatial resolution than the domain in which it is embedded. The simulation had 31 vertical layers and, to ensure that vertical profiles in the boundary layer were well characterized, 15 of those layers were placed in the lowest 1500 m above the surface.

highest wind core (U > 24 m/s) is to the northwest of Dubai over the Arabian Gulf. A low-level jet (yellowgreen band at the bottom of the vertical slice) exists above the entire Arabian Gulf, degrading over the southern part of the Arabian Peninsula. Wind speeds are reduced significantly above 1500 m (blue band), and then increase again above 5 km (orange band). The low-level wind feature is a typical Shamal event.

Fig. 3 : 3D view of Example Output from Mesoscale Modelling

Fig. 2 : Mesoscale Modelling – Nested Grids

Fig. 3 shows an example of output from the meteorological model. This 3-D view illustrates a horizontal slice of the wind field, at 700m above the surface, as well as a vertical slice passing through Dubai, at 0400 GMT on May 4, 1999. At this moment, northwesterly wind dominates the region and the wind speed is about 20 m/s at 700 m above Dubai. The

Fig. 4 shows examples of vertical profiles that were extracted from the model. This Shamal event shows strong wind shears for both wind speeds and wind directions. Wind speed is peaking to 18 m/s, and wind direction is shifting by about 100 degrees between surface and 1400 m. In Fig. 4, the balloon data obtained from Abu Dhabi International Airport at 04:00 1992/5/27 (dotted line) are alike to the MM5 modeling data at the same location and time. The advantages of such modeling include the level of detail as well as the freedom to obtain modeling data for any location and time.

Fig. 4 : Extracted Vertical Profiles from MM5 Data

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

After extensive numerical simulations and analysis of the wind climate of Dubai, it is found that overall the wind climate model based on surface data from Dubai does not require any significant adjustment for wind direction to be representative of the top of the boundary layer. Based on the analysis, it is found that factor of 1.7 would be appropriate to scale the surface wind speed result to 600 m, which is consistent with a typical power law assumption for open terrain. This would yield wind speed value of about 38 m/s, which is about 15% lower than suggested wind speed by Dubai regulatory board. Thus, the statistical analyses of the Dubai surface data, scaled to a height of 600 m using a typical power law, gives values that are somewhat higher but reasonably consistent with the value obtained from the corrected upper air data at Abu Dhabi. The details of this study can be found in Vincent et al[1]. 2.2 Kingdom Tower, Jeddah, Saudi Arabia The long term historical measured surface data from the King Abdulaziz International Airport provides the most complete record for predicting the frequency of wind events near the ground surface in the Jeddah area. For most buildings, this information can adequately be extrapolated to the full height of the building using standard boundary layer profile assumptions. Due to the extreme height of the Kingdom Tower, there is considerable uncertainty in relying on the surface data exclusively. The boundary layer effects of the wind such as roughness of the surface, marine intrusions, and drainage flows off the mountains are strongest near the surface, decreasing with height. It is likely in the case of the Kingdom Tower that the top of the Tower will often be beyond the impact of the surface boundary layer, where the wind conditions are more dominated by the larger scale synoptic air masses. The historical balloon sounding data from the King Abdulaziz International Airport provides information on the true wind vertical profile. However, since the measurements are typically only twice daily, there is a high likelihood that the peak wind speed from a wind event, or even the entire wind event may be missed. As well, there are issues in releasing weather balloons under high wind conditions, so the data set may be biased against higher wind events, which became evident in our analysis.

Volume 47 │ Number 4 │ December, 2017

To augment the historical measured data available for the area and to obtain a better understanding of the upper level winds that will affect the tower, mesoscale modeling of historic weather conditions was leveraged to obtain a long-term record of hourly vertical wind profiles at the Kingdom Tower site. The Weather Research and Forecast Model (WRF) is a limited-area, non-hydrostatic, terrain-following sigma-coordinate model designed to simulate or predict atmospheric circulation on scales ranging from 100’s of meters through to 1000’s of kilometers. The latest National Oceanic and Atmospheric Administration (NOAA) Global Forecast System (GFS) re-analysis data (at 1.0-degree resolution globally) was used to set the WRF initial and boundary conditions. This dataset is based on the historical archives of a worldwide meteorological observation network, including surface and upper air balloon measurements, satellite and radar measurements, etc. The data are available as a 3- dimensional grid developed using mesoscale meteorological modeling run in a “hind-cast” mode, meaning the results are continuously calibrated using measured values. RWDI executed the Weather Research and Forecast (WRF) model, over a nested model domain centered over the Arabian Peninsula. The model was nested from the GFS 1 degree reanalysis data in progressively higher resolution domains, starting with a domain with horizontal grid spacing of 36 km, then 12 km, and down to 4 km in the Jeddah area. This model was run for the period from 2001 to 2010 inclusive, resulting in a total of more than 87,000 hourly records per grid cell. From this model, vertical profiles for the entire period were extracted from the 4km domain at the Kingdom Tower site and used for the subsequent analysis. The details of this study can be found in Valerie et al[2]. Fig. 5 shows a vertical cross section of the winds over Jeddah generated from vertical layers of WRF data. This figure depicts a wind event with maximum speeds at approximately 300 m level, and considerably lower speeds at about 1000 m level. Since high wind events are the primarily concern, the top wind events based on speed were extracted from the WRF and airport data for review, and an example time series were plotted for comparison in Fig. 6. For the majority of the events reviewed, those with winds generally from

The Bridge and Structural Engineer

the South showed the best agreement between the measured data and the WRF data (as seen in Fig. 6),

Fig. 5 : WRF Model Cross Section of Wind Speeds Over Jeddah (Date-Time: 2006-03-22 - 1:00 GMT)

Individual vertical profiles of winds were extracted from the WRF data and reviewed, with the focus being the higher wind events at each height. Fig. 7 shows some of the vertical profiles from one of the highest wind events in the 10-year period modelled. The profiles indicate a very deep boundary layer for much of the storm, indicated by the wind speed consistently increasing with height. More frequently though, there is evidence that the winds near the

followed by those from the North. This comparison gives confidence on our simulation.

Fig. 6 : Comparison of Wind Speed Variations

surface are mechanically decoupled from the winds at the top of the tower also shows an example of a thunderstorm wind event where this is the case. The winds near the surface are quite high and from East-Northeast up to a height of approximately 160 m, at which point the wind speed starts to decrease. Near the height of the tower at 1000 m, however, the wind speeds are much lower than at the surface, and are coming more from West-Northwest.

Fig. 7 : Selected Wind Profiles from an Aziab Wind event & Thunderstorm event

The Bridge and Structural Engineer

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Based on the detailed study, the design wind speeds at the surface were extrapolated to upper levels using the standard open profile and compared to the results from the upper level extreme value analyses of the balloon and WRF data in Fig. 8. The left graph shows the relationship between the 10-year wind speeds used for serviceability and the predicted wind speeds from the balloon and WRF upper air data,

where the graph on the right shows the same for the 50-year building code wind speed (corresponding to a 42 m/s 3-second gust). This indicated that the open profile assumed for the boundary layer applied to 10 m wind speeds is sufficiently conservative in accounting for the wind speeds aloft. These results helped in improving the serviceability performance of the tower[3].

Fig. 8 : Comparison of Predicted Wind Speeds to Standard Open Profile

Aerodynamic Treatments

Aerodynamic shaping is well known to reduce drag loads on aircraft structures, cars, ships etc. Further, in sports aerodynamics, body postures during skating and bi-cycling competitions are extremely important to reduce the drag to achieve higher speeds. In addition, the surface roughness on cricket/golf balls once again helps to reduce drag and cover maximum distance and/or swing in different directions. So in many parts of our life, aerodynamics plays a crucial role. Likewise, in wind loading on tall buildings, aerodynamics plays a major role as well. In tall buildings, one of the key wind loading issue is caused by the occurrence of alternate vortex shedding at the sides leading to cross-wind oscillations perpendicular to the wind. By shaping the towers, especially the aerodynamic treatments at the corners, one could disrupt the intense vortex shedding and reduce the impact of such phenomenon[4, 5, 6]. This technique has been investigated in the past on many projects, one of

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

them being the famous 505m tall Taipei 101 tower in Taiwan, where double step corners were introduced to disrupt the intense vortex shedding. Now that the towers are becoming super tall and super slender, the importance of aerodynamic shaping is getting even more crucial than before. 3.1 Shanghai Tower, Shanghai, PRC The Shanghai tower is a 632-metre, 128-story mega tall skyscraper in Lujiazui, Pudong, Shanghai. The preliminary wind tunnel studies on the original building configuration indicated, that due to its relatively streamlined shape, the Shanghai Tower has much lower drag coefficient (i.e., shape factor) than the standard code value of 1.3. Therefore, the wind loads in the along-wind direction were only about 75% of the code prediction. However, due to the across-wind dynamic response, the resultant structural wind loads were much higher than the along-wind loads. The observed across-wind dynamic response is associated

The Bridge and Structural Engineer

with vortex shedding at critical velocity near the 100-year return period winds. To better understand the building’s aerodynamic behavior, a benchmark test was conducted for a traditional square box

building that has the same height and tapered width as the Shanghai Tower, shown in Fig. 9. The details of testing and results are reported by Calin et al.[7].

Tapered Box

100º Configuration

110º Configuration

120º Configuration

180º Configuration

Final Configuration

Fig. 9 : Wind Tunnel Photos Showing Various Twisting Configurations of Shanghai Tower

The overall wind-induced overturning moments, shear forces and torsional moments acting at the grade level have been predicted for the design return period for all test configurations. Based on the preliminary results,

The Bridge and Structural Engineer

initially the 110º twist is chosen by the project team. Fig. 10 indicates that the 110º twist configuration can be designed with about 20% less wind loads compared with the traditional tapered box building, Volume 47 │ Number 4 │ December, 2017

assuming both having the same dynamic properties. However, given the height of the Shanghai Tower and the importance of wind loads in structural design, an aerodynamic optimization of the building shape to further reduce wind loads was desired and considered feasible.

Fig. 10 : Comparison of Tapered Box with 110º Twist – Shanghai Tower

The aerodynamic optimization study was focusing on reducing across-wind dynamic response. With the basic requirement of not significantly changing the architectural intent, the aerodynamic optimization was mainly to investigate the potential benefits by further twisting the building shape. More twisting of the building shape will lead to less correlation of vortex shedding along the height, so that the acrosswind response can be reduced.

tunnels in the presence of all surroundings within a full-scale radius of 600 m and all major structures within a full-scale radius of 1200 m. The studies were first conducted for the 100º twist and 180º twist, a feasible range of twisting in practice. The test results verified the effectiveness of twisting in terms of reducing across-wind response. By examining the results and considering other factors including aesthetical aspects and the building envelop design, further tests were conducted for 110º twist and 120º twist. Few of the tested configurations are shown in Fig. 9 along with the final configuration. In addition to the building shape optimization, the building orientation was also investigated. The concept consisted of orienting the entire building in such a way that the worst aerodynamic direction of the building stayed away from the prevailing strong wind direction at the site, so that the wind directionality reduction effects could be maximized. In connection with this, the 110º shape of the Shanghai Tower was tested in different footprint locations by rotating the tower on the study site from the original footprint location. It was found that the footprint position that generates the lowest wind loads acting on the tower is obtained by rotating the 110º Shape by 30º counterclockwise. With this information confirmed, the 120º shape of the tower was also tested in the same position (footprint rotated by 30 degrees counterclockwise). Another wind tunnel test was performed by rotating the 120º Shape by 40º counterclockwise. It was found that this new footprint position (40º footprint rotation counterclockwise) generates wind loads lower that the 30º footprint rotation. This conclusion is evident from the results summarized in Table 1.

The wind tunnel studies for aerodynamic optimization were conducted using the High-Frequency Force Balance (HFFB) technique. The details of HFFB technique can be found elsewhere[8]. 1:500 scale models were tested in one of RWDI’s boundary wind Table 1 : Summary of Predicted Overall Structural Wind Loads at Grade – Shanghai Tower

The test results did show significant benefits from optimal orientation. However, due to other design considerations, it was decided that the final building design should remain at its original orientation. The final configuration, that accounts for all design aspects, is shown in Fig. 9, which includes roof top openings for wind turbines and has a twist angle of 120º. Overall, twisting the envelope and reorienting the footprint helped in reducing the wind -induced 42

Volume 47 │ Number 4 │ December, 2017

response of the Shanghai Tower significantly. Once again the importance of aerodynamic design of the envelope of very tall and slender buildings is reinforced through this example. 3.2 Burj Khalifa, Dubai, UAE Initially, several High Frequency Force Balance (HFFB) tests at 1:500 scale model were carried out to determine preliminary design loads on Burj Khalifa.

The Bridge and Structural Engineer

The wind tunnel data were then combined with the dynamic properties of the tower to compute the overall wind-induced response of the tower. The HFFB results were used as early inputs for the structural design and several parametric runs tweaking the tower stiffness and mass distributions were undertaken. Further, several rounds of force balance tests were undertaken to refine the tower geometry architecturally and aerodynamically. Fig. 11 shows a pictorial view of the tower in final form along with the roof plan showing the axis system for loading. In general, the building has Y-shape geometry with three noses A, B and C and three wings between the noses. This plan form is torsionally very stiff. Initially, the setbacks at each nose region were all at the same height, but then through series of HFFB wind tunnel tests, these setbacks were altered between the noses at different heights. After several wind tunnel tests, based on the results, the tower was reshaped to minimize the wind effects. In general, the number and spacing of the setbacks were altered and optimized in such a way as to confuse the wind and thus, substantial reduction in wind loading on the tower was resulted.

Fig. 11 : Burj Khalifa in Dubai and Plan View of the Tower

Further, based on the results of HFFB results, it was noticed that wind impact into the wing is far more critical and higher than wind impacting the nose. This information was used while reorienting the

Burj Khalifa in such a way as to deviate the wings away from pointing the most frequent strong wind directions of Dubai. Significant improvements in wind-induced response of Burj Khalifa were observed because of reshaping and reorienting the tower. Initial 5-year return period accelerations were at about 37 milli-g, and later this value was reduced to about 19 milli-g. About half of this improvement belongs to the improved knowledge of wind statistics and the remaining through re-orientation, reshaping and structural improvements[9]. 3.3 The Steinway Tower, New York, USA The under-construction Steinway tower, an upcoming residential property in New York, will be an exclusive project as far as slenderness is concerned. This will become the most slender building in the world with height/width ratio of 22. Considering the super slenderness and inherent flexibility, it is anticipated that the geometry is very crucial to reduce the windinduced response of this tower. Tremendous efforts have been made to shape the structure with setbacks and top porous pinnacle to make the structure more aerodynamic instead of just blocking the wind flow. Fig. 12 shows a pictorial view of the Steinway tower with the 1:500 scale HFFB model in RWDIâ&#x20AC;&#x2122;s wind tunnel. The raw overall sway moments on the tower at the grade level are shown in Fig. 13 for two damping levels. As one can notice, the cross-wind moments are not significant compared to the along-wind moments. Cross-wind moments are expected at the wind angles when the mean moment is close to zero (circled region in Fig. 13). Essentially, along-wind moments dominate the design. The details of testing and results were reported by Cicci et al.[10]. This is made possible by the aerodynamic shaping of the tower. Even after all the aerodynamic treatment of the envelope, the building will be housed by an 800-ton tuned mass damper designed by RWDI to arrest excessive motions during extreme wind and seismic events.

Fig. 12 : Image of The Steinway Tower in New York and Wind Tunnel Model

The Bridge and Structural Engineer

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Fig. 13 : Raw Overall Base Moments Versus Wind Direction - Steinway Tower

3.4 Kingdom Tower, Jeddah, Saudi Arabia Similar to other towers, the latest kilometer high Kingdom tower in Jeddah also went through careful aerodynamic shaping to reduce wind-induced responses as much as possible. The tower has a tapered façade in contrary to the setback facades in Burj Khalifa. However, the overall base plan of Y-shape is like Burj Khalifa, which is torsionally stiff. Further, the noses of Y plan have been chamfered to mitigate intense vortex shedding. At RWDI, initially the studies were started with HFFB models of scale 1:800 and later, High Frequency Pressure Integration (HFPI) model at 1:600 scale was used to fine tune the responses. The details of modelling, testing and results on Kingdom Tower can be found in Ref[11]. Fig. 14 shows both HFFB and HFPI models of Kingdom tower. The HFPI model helped in accommodating the higher order modal effects as well as arriving at better wind load distribution along the height of the tower. HFPI technique details can be found in Ref[12]. Table 2 shows the comparison of the loads derived from HFFB and HFPI models for one of the configurations. The HFPI tests resulted in somewhat lower loads primarily due to the increase in stiffness of the structural system.

Fig. 14 : 1:800 Scale HFFB Model and 1:600 Scale HFPI Model of Kingdom Tower, Jeddah

Volume 47 │ Number 4 │ December, 2017

Table 2 : Summary of Overall Base Loads on Kingdom Tower, Jeddah

Advanced Wind Tunnel Tests

Traditionally High Frequency Force Balance (HFFB) tests have been widely used for determining windinduced loads on tall buildings. Occasionally, this is supplemented with High Frequency Pressure Integration (HFPI) studies with a view to account higher modes and obtain better wind load distributions. In case of super tall/slender buildings, other important effects such as aerodynamic damping, better estimation of peak factors, sensitivity to Reynolds number are to be evaluated for the betterment of wind-induced response predictions. In this connection, advanced wind tunnel tests such as aeroelastic tests as well as large scale wind tunnel tests are recommended. 4.1 Shanghai Tower, Shanghai, PRC Reynolds number is an index of the ratio of flow inertial to flow viscous forces. Wind tunnel tests are commonly conducted at very low Reynolds number, while a full-scale building in strong winds is in a very high Reynolds number. The effects of the mismatch of Reynolds number between the wind tunnel tests and full scale response tend to be more significant for a building with rounded surfaces, such as the Shanghai tower, than for buildings with sharp corners. For a cylinder, the Reynolds number effects include potential overestimate of static drag force and dynamic lift force from wind tunnel prediction. The Shanghai Tower has rounded smooth surfaces, but its shape is so different from a circular cylinder that

The Bridge and Structural Engineer

the existing knowledge on circular cylinders cannot be simply applied. High Reynolds number testing is therefore needed to fully understand the potential Reynolds number effects. The high Reynolds number tests for the Shanghai Tower were conducted using a 1:85 scale pressure model, as shown in Fig. 15. The model was tested in the 9 m × 9 m wind tunnel of the National Research Council of Canada (NRCC).

heights with a view to fine tune the loading, which is normal for a tower of this height. Fig. 16 shows one of the aeroelastic models of Burj in one of RWDI’s wind tunnels in Canada. An aeroelastic model is flexible in the same manner as the real building, with properly scaled stiffness, mass and damping. The details of modelling, testing and results on Burj Khalifa can be found in Ref[13]. Note that the aeroelastic model could model the first six sway modes, which includes the third order modes. During the aeroelastic tests, bending moments were measured at base, as well as at several higher levels. Accelerations were also measured at various levels using accelerometers. The aeroelastic model tests produced significantly lower wind-induced response compared to the HFFB tests. During the aeroelastic tests at multiple speeds, the dropping of moment coefficients with test speed gave an indication that Reynolds number effects were present.

Fig. 15 : 1:85 Scale Pressure Model of Shanghai Tower in NRCC wind tunnel

The test results indicate that the mean drag force of the Shanghai Tower could be lower at full scale than the wind tunnel prediction from the previous 1:500 scale model, especially when wind blows into the broad face. However, the dynamic lift force was not found to be very sensitive to the Reynolds number. Since the design wind loads of the Shanghai Tower are governed by the across-wind dynamic response, which is only affected by dynamic lift coefficients, the potential reduction on drag force tends to have a limited impact on the final structural design loads. The critical velocity for vortex shedding was found to be consistent between the previous 1:500 scale model and the high Reynolds model. Therefore, no corrections needed to be applied to the acrosswind response obtained from the 1:500 scale model studies. 4.2 Burj Khalifa, Dubai, UAE 4.2.1 Aeroelastic Tests of the Entire Tower Followed by several HFFB tests, numerous aeroelastic tests were carried out on Burj Khalifa for different

The Bridge and Structural Engineer

Fig. 16 : 1:500 Scale Aeroelastic Model of Burj Khalifa, Dubai

Fig. 17 illustrates the relative change in mean base moment coefficient on the aeroelastic model as a function of wind speed for two wind directions. The 10-year accelerations at habitable level also reduced to about 12 milli-g range in the aeroelastic tests from the previously predicted value of about 19 milli-g from the final HFFB test. The wind-induced loading at the base also reduced significantly in the aeroelastic tests and one such example loading on a particular scheme of Burj is shown in Table 3. The reasons for such a significant difference in wind-induced responses between aeroelastic and HFFB tests are (1) lower Reynolds number of the force balance tests, (2) presence of aerodynamic damping in aeroelastic tests, (3) influence of higher order modes, and (4) lower peak factors resulted out of aeroelastic tests. In summary, aeroelastic tests are vital especially when we are anticipating influence of Reynolds number, aerodynamic damping, higher order modes and peak factors. Volume 47 │ Number 4 │ December, 2017

Fig. 17 : Normalized Mean Drag Moment Coefficient Variation with Wind Speed

Table 3 : Summary of Predicted Overall Structural Wind Loads at Grade - Burj Khalifa, Dubai

4.2.2 Aeroelastic tests of spire In addition to the several aeroelastic tests on entire Burj tower, special studies were undertaken to accurately determine the wind-induced responses of the steel Spire portion of the Burj Khalifa above elevation 574.5 m during construction stages by directly measuring the responses on a flexible dynamic model which included aeroelastic effects. Two configurations were considered for construction stages: (1) with a construction crane in place; and (2) without crane in place. The primary objectives of this study were to: (i) provide data for the structural design of the Spire portion of the tower during construction stages; and, (ii) to determine the wind forces acting on the ties of the crane. The model study was carried out using a 1:200 scale aeroelastic model of the Spire in one of RWDI’s boundary-layer wind tunnels in Canada. The details of the testing, analysis and results can be found in Suresh Kumar et al.[14]. Photographs of the wind tunnel model of the Spire with and without the crane are shown in Fig. 18. Based on the provided structural properties, an aeroelastic model of the Spire as well as the crane were designed and built using the appropriate scaling principles to simulate the 1st, 2nd and 3rd order modes of the Spire in both the X and Y directions and the 1st order modes of the crane. The torsional motions were not simulated on this model since the torsional responses were found to be small from the previous HFFB studies on the entire tower.

Volume 47 │ Number 4 │ December, 2017

Fig. 18 : 1:200 Scale Aeroelastic Model Burj Spire Above Elevation 574.5 m with and without Tower Crane

4.2.3 Larger scale wind tunnel tests Most of the tests in Burj Khalifa were carried out at 1:500 scale model. Considering the curved facades of the Burj Khalifa closer to the noses and also the variation of the mean moment coefficients with wind speed seen in aeroelastic tests, there were some concerns regarding the influence of lower Reynolds number simulation at 1:500 scale tests, especially with HFFB tests at low velocities. In order to understand the Reynolds number sensitivity to façade pressures and aerodynamic coefficients, top portion of Burj was built at 1:50 scale model and tested at NRCC wind tunnel of 9 mx 9m cross section. The large-scale model and its instrumentation is shown in Fig. 19.

Fig. 19 : 1:50 Scale Rigid Pressure Tapped Model of Top Portion of Burj Khalifa

The Bridge and Structural Engineer

Tests were carried out at wind speeds up to 55 m/s. Mean and instantaneous pressure distributions around six cross sections of the tower were measured and compared with similar measurements made at 1:500 scale in RWDI’s 2.4 m x 1.9 m wind tunnel. The comparison of the high Reynolds number results with those at normal test Reynolds number revealed the fact that aerodynamic coefficients as well as pressure coefficients resulted out of 1:500 scale aeroelastic model and pressure model tests reached closely to asymptotic state. This was made possible by simulating high enough Reynolds numbers during 1:500 scale aeroelastic and pressure tests, and therefore, based on this conclusion there were no special Reynolds number corrections made on the 1:500 scale model test results. 4.3 The Steinway Tower, New York, USA Advanced aeroelastic tests were carried out on this tower to take advantage of the potential positive impact of aerodynamic damping and reduced peak factors. As expected and learned from previous cases, once again the aeroelastic test results were lower than those of HFFB tests. In addition to the reasons stated above, the HFFB method includes some assumptions concerning the distribution of the generalized modal forces with height. The effect of these assumptions may be exaggerated for this tower due to: i) its varying shape with height, which causes a variation of the dynamic drag and dynamic lift force coefficients with height; and, to a lesser extent, ii) the nonlinear fundamental mode shapes. The aeroelastic model technique directly simulates the generalized modal forces by simulating the variations in both the geometry and mode shapes over height. Table 4 shows the comparison of aeroelastic and HFFB tests. Table 4 : Summary Overall Base Loads on The Stainway Tower, New York

4.4 Kingdom Tower, Jeddah, Saudi Arabia Similar to other towers, RWDI has also carried out aeroelastic tests on 1:600 scale model of the Kingdom tower which is shown in Fig. 20. As expected from other similar tests, aeroelastic tests arrived at lower loads than HFFB and HFPI test results, which is shown in Table 2. The difference was very evident in comparing the HFFB results with the aeroelastic results. Part of the difference is due to the stiffer

The Bridge and Structural Engineer

structure during aeroelastic stage as well as slightly lesser speed. The remaining difference is caused by positive aerodynamic damping and lower peak factors resulted out of aeroelastic tests like the case of other towers.

Fig. 20 : 1:600 Scale Aeroelastic Model of Kingdom Tower in RWDI’s Wind Tunnel

Concluding Remarks

Super tall/slender towers are on the rise. The challenges that these towers pose to builders/engineers are multifold. Since wind being the predominant lateral force acting on these towers, careful attention must be exercised while determining wind-induced response of these towers. Typically, codes/standards are used to evaluate wind loads on tall buildings at least in the preliminary design stage, if not for final design in case of intermediate height buildings. However, the most prominent international wind codes of countries namely Australia, America and Canada have made wind tunnel tests mandatory when buildings are beyond certain heights, slenderness and flexibility[15]. In fact, Australian wind code states clearly that the provision covers only buildings less than 200m high and frequency greater than 0.2 Hz. Further, the issues addressed in this paper such as wind climate at high altitudes, aerodynamic treatments to improve responses and advance wind tunnel tests are beyond the reach of any wind codes as they are cumbersome to codify with a view to maintain simplicity of the codes. In these circumstances, wind engineers must think beyond these traditional procedures to arrive at optimum solutions for such towers. This paper discusses some of the special wind engineering considerations required for the very tall structures. Wind climate at high altitudes is one of the much sought out area for consideration in case of super tall buildings, since the traditional scaling of surface Volume 47 │ Number 4 │ December, 2017

winds to heights beyond 500 m seems uncertain/ conservative. Detailed study of higher altitude wind climate using numerical simulations is a promising option to accurately capture the vertical variation of wind speeds and directions at any site. Aerodynamic shaping of the towers is also equally important to make the buildings more aerodynamic and highly resistant to extreme winds. Simply, similar to the way we shape our air crafts, posture in bike racing competitions, roughen the surface of the golf/ cricket balls, shape the cars/ships, we also need to pay attention to shape the buildings to make them more aerodynamic resulting in lesser wind-induced responses. Further, instead of traditional rigid model High Frequency Force Balance tests, detailed aeroelastic tests are very useful in taking care of aeroelastic effects, lower peak factors and higher order modal effects. Most of the time, aeroelastic tests give lower wind-induced loads than the traditional HFFB tests, which is beneficial for optimizing the structure. In addition, large scale model tests are also conducted at times to check the Reynolds number sensitivity of flow.

References 1.

Vincent Tan et al., “Numerical Modelling of Extreme Wind Events for the Burj Dubai Project, Dubai, UAE”, RWDI Report#03-1381, October 4, 2004.

Valerie, S. et al., “Upper Level Wind Climate Assessment Kingdom Tower, Jeddah, Saudi Arabia”, RWDI Report# 1011301, January 18, 2012.

Jon Galsworthy, “Rising to the Clouds with Confidence: The upper Level Wind Climate Assessment for the Jeddah tower”, Structure Magazine, June 2016, pp. 37-40.

Dutton, R., & Isyumov, N., “Reduction of Tall Building Motion by Aerodynamic Treatments”, Journal of Wind Engineering and Industrial Aerodynamics, 36 (2), 1990, 739-747. Kawai, H., “Effect of Corner Modifications on Aeroelastic Instabilities of Tall Buildings”,

Volume 47 │ Number 4 │ December, 2017

Journal of Wind Engineering and Industrial Aerodynamics, 74-76, 1998, 719-729. 6.

Maruta, E., Kandaa, M. & Satob, J., “Effects on Surface Roughness for Wind Pressure on Glass and Cladding of Buildings”, Journal of Wind Engineering and Industrial Aerodynamics, 74-76, 1998, 651-663.

Calin Dragoiescu et al. “Wind-Induced Structural Response Studies - Shanghai Center Tower, Shanghai, P.R.C”, RWDI Report # 08-1440, May 19, 2009.

Calin Dragoiescu, Garber, J. and Suresh Kumar, K., “A Comparison of Force Balance and Pressure Integration Techniques for Predicting Wind-Induced Responses of Tall Buildings”, ASCE Structures Congress, St.Louis, Missouri, 2006.

Peter A. Irwin, and William F. Baker, “The Burj Dubai Tower: Wind Engineering”, STRUCTURE magazine, June 2006, pp. 28-31.

10. Cicci, M.D. et al. “Aeroelastic Model Wind Tunnel study, 105- 111 West 57th Street, New York, USA”, RWDI Report#1400320, June 16, 2015. 11. Gary Stone et al., “Wind-Induced Structural Responses – Kingdom Tower, Jeddah”, RWDI Report# 1011301, January 18, 2012. 12. Calin Dragoiescu, Garber J., Suresh Kumar, K., “Tap Resolution Related to the Accuracy of Pressure Integrated Wind Loads.”, Proceedings of the 5th European African Conference on Wind Engineering (5EACWE), Florence, Italy, 2009. 13. Gary Stone et al., “Aeroelastic Model Wind Tunnel Study – Burj Khalifa, Dubai, UAE”, RWDI Report# 03-1381, April 14, 2004. 14. Suresh Kumar, K. et al. “Aeroelastic Model Wind Tunnel Study Burj Dubai, Aero 2 -13 Scheme Spire, Construction Stage, Dubai, UAE”, RWDI Report# 03-1381, December 8, 2005. 15. Suresh Kumar, K. “Wind Tunnel Testing of Tall Buildings: When to Carry Out?”, ISSE Journal Mumbai, Vol. 15-1, p. 3-4, Jan 2013.

The Bridge and Structural Engineer

Tall Building Façade Designs -Impact of Wind and Associated Loadings Rajan has 22 years of international experience within the facade industry with Structural engineering background. With his wide International and regional experience with on-going leading role on number of projects, he has attained extensive knowledge on façade and construction industry in the region. Known face to regional facade industry, he has attained good reputation with high quality works completed.

Rajan Govind Façade Specialist – Director at BES Consultants Pvt. Ltd., rajan@besconsultants.net www.besconsultants.net

Abstract This paper illustrates and shares author experiences on High rise building façade design challenges specific to wind associated and structural design aspects.

Introduction

Building facades are increasingly becoming sophisticated and interesting in order to adapt modern building architecture and challenging requirements. Facades are receiving special attention in the construction industry as it is a specialized construction field which has well adopted latest technologies. Façade designs, constructions,

implementations are not set by any limitations or boundaries and thus leave opportunities for designers or owners to make their design ideas a reality. This paper focusses on the façade systems structural engineering aspects of high rise buildings, presented with experiences and actual project details.

High Rise Design Challenges

Facades for high rise buildings have specific design. Few major aspects are given in Table 1. These challenges are not just limited by engineering and science but many other practical aspects related to fabrication and implementation.

Table 1 : Major Aspectsof Façade Design ●

Design Structural

●

Weather Performances

●

Logistics

Visual architectural

●

Durability

●

Access

●

Performances

●

Finishes

●

Material handling

●

Framing sizes and shapes ●

Corrosive environment

●

Site testing

●

Brackets - adoptability to ● building structure

Local conditions

●

Limitation due to building form

Design Challenges are illustrated in the graphical image (Fig. 1). Following are some key design principles which shall be focused by façade engineers/designers from engineering point of view. ●

Direct load transfer

●

Structurally efficient

●

Allow movement/rotations

●

Ease of fixing

The Bridge and Structural Engineer

Site

●

Simplicity in detailing

●

Ability to interface cleanly with main building

2.1 Structural Design Aspects Wind pressures being a prominent design load, which varies from normal to high and extreme pressures, and non-uniform pressures among various elevations, lead to challenges on achieving optimum design balance between practical and technical aspects. Volume 47 │ Number 4 │ December, 2017

Fig. 3 : Typical Floor Bracket and Curtain Wall Profiles

3. Fig. 1 : Design Challenges

2.2 Building Movements Following are primary aspects for the façade movements on tall buildings. Refer to Fig. 2.

Glazing Systems

There are several ways the systems shall be classified. However, from the construction point of view, following are the widely used glazing systems. 1. 2. 3.

Stick Glazing system Semi Unitized or Cassette system Unitized system or panel system

3.1 Stick Glazing System

Fig. 2 : Panel Mode of Movement Due to Sway

Sway or Inter-Story Drift: Horizontal or lateral movements at floor level due to wind or seismic movements. This will induce racking and in - planar forces on façade panels.

Fig. 4 Illustrates the basis of the system in which Glass has been supported by Aluminium grid work framing system. Vertical frames (called as “Mullion”) span between floors supported by brackets at each floor. Horizontal framing (called as “Transom”) connected to verticals is used to glaze the glass units. All materials are delivered as components to the project site and installations are carried sequentially. These types of systems are site fabricated and generally adapted for low rise buildings, small size applications, using standard available designs. This system is adapted for the basic requirements which do not require customized design solutions.

Fig. 2 shows typical panel deformation mode due to building sway. As a good design guidance, the façade systems are recommended to accommodate H/500 (H-floor height) lateral movement. Floor movements: Building floors may have different movement behaviours due to imposed load and other aspects. Hence it is recommended that the façade fixings and floor interfaces are adequately designed to address these movements. Ignoring this key design aspect may have adverse impact such as panel cracking or other failures. Fig. 3 illustrates a typical floor fixing which has vertical movement allowances.

Fig. 4 : Stick Glazing System

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

3.2 Semi Unitized System Also known as Cassette system, this is similar to Stick system as defined in sec. 3.1. However, the glass units are glazing with Aluminium sash frames at factory and then delivered to the site. These glazed units are hooked on to the Aluminium grid works erected at the site. Hence this system is partially pre-fabricated and has less site work. All materials are delivered as components to the project site and installations are carried out sequentially. Refer Fig. 5. This system has slightly lesser site work as compared to Stick glazing system. However, the system weather performances rely on site workmanship as it involves the primary sealing between glazing cassettes to be maintained with high integrity, so that no water or air enters the system. Fig. 6 : Unitized/Panel System

High Rise Façade Systems

Facades of tall buildings use Unitised panelised glazing, which follows a pre-fabricated factory finished façade system. Fig. 7 shows a factory finished glazed panel.

Fig. 5 : Semi Unitized System

3.3 Unitised or Panel Systems This employs most modern fabrication and construction techniques of pre-fabrication approach. The glazed panels are fully finished at the factory and delivered to the site and ready to be erected in place. This results in very little work at the site and achieves high quality and faster completion. In recent times, most modern buildings globally adapt this approach. These types of systems were implemented internationally and proven to be time tested design solutions without any compromise on quality and time frame. The following sections give a detailed description of this particular façade systems widely adopted for modern high rise/tall building facades. Refer Fig. 6.

The Bridge and Structural Engineer

Fig. 7 : High Rise Façade System

●

Irrespective of any façade system, following are basic design principles of tall building façade systems.

●

Uses Light weight aluminium framing system structurally glazed with glass panels with silicone structural sealant/bonding technique. Volume 47 │ Number 4 │ December, 2017

●

Panels are supported on to building using bracketing and fixing at each floor.

●

Systems allow for large movement and building sway. Each panel is supported and hung from top, hence allowing large movements and building

●

sway without affecting or breaking glass units. Pre-fabricated and pre-finished panels with very less site works.

Key Structural Aspects

Critical structural aspects of tall building facades are elaborated in Table 2.

Table 2: Key Structural Aspects Aspect

Critical Impact

How to Address

Wind

Predominant load, overall design controlling

Appropriate Structural design of framing, in fill panels and support systems.

Seismic

Fixing designs and interface with building.

Fixings and connections to allow large movement. Appropriate bracket designs

Movement

Design to address thermal and building movements.

Panel joints and gaskets are designed for large movement and equally without compromise on weather performances

Wind load being a predominant design aspect in the overall designs, following section elaborates this aspect.

NBC 2016, calculated as per recommended methods. 2.

Wind tunnel (laboratory based) – Physical scaled building model with surrounding terrain were tested in a laboratory setup, and results measured using sensors and instrumentation. The test results are expected to give correct prediction of wind pressures based on actual site conditions considering surroundings.

CFD - (Computer simulations based)Computational Fluid Dynamics techniques will help to predict reasonably accurately, the building façade pressures for complex and more detailed localized study otherwise impossible by above methods.

Wind Loads

Wind load being a predominant design aspect in the overall designs, has impact on the buildings in many ways as elaborated below.

6.2 Wind Patterns – Indian Cities Images in Figs. 9 and 10 show the Wind Directional Distribution (for Mumbai region) as per IMD information, wind speed calculated for 50 Years return Period and 10 Years Return Period. Fig. 8 : Wind Load Design Aspects

Engineer requires good understanding of wind loads and how the building behaves under wind pressures. Appropriate wind pressure derivations and prediction of its effects are fundamental to high rise façade design. These change from building to building, hence engineer’s understanding of these aspects is critical. 6.1 Wind Load Calculations Wind loads are derived in following ways 1.

Code based: Indian Standard IS 875 Part 3 and

Volume 47 │ Number 4 │ December, 2017

It is noted that the North-West Direction is the most dominant for Mumbai region. Whereas South-West is also critical as highest wind speed is experienced through this direction. It is also important for engineers to consider not just peak wind pressures or 50 years return period. Often for design efficiency several non-critical façade elements are designed for 10 years return wind speed. Hence engineers’ understanding of local effects and wind patterns are highly important.

The Bridge and Structural Engineer

Fig. 9 : 50 Year Return Period

Case Study – Mumbai High-Rise

This section illustrates wind load derived from three different approaches as explained above. This has been extracted from an actual high-rise building project completed in recent times. There is no single approach that may yield the best outcome - quite often comparison of different methods will justify the Engineers decisions. Due to complex nature of wind distributions on tall building facades, engineers need

Fig. 11 : Wind Pressure Thro IS Code

Fig. 10 : 10 Year Return Period

to carry out detailed analysis and study of particular locations on micro level. This will avoid any structural failures due to possible underestimation of high local pressures. Hence engineers can’t just limit to code based approach for high-rise building facades - a detailed approach through a wind tunnel or CFD based approach will help accurate predictions of wind pressures.

Fig. 12 : Wind Pressure Thro Wind Tunnel

Discussions

Facades for high rise buildings often face several design and implementation challenges; each building may pose specific sets of challenges. Hence engineers and

The Bridge and Structural Engineer

Fig. 13 : Wind Pressure Thro CFD

designers should approach these requirements with an innovative and most appropriate approach best suited for that particular building requirements. Following are Author’s own assessment and experiences.

Volume 47 │ Number 4 │ December, 2017

Engineers Limitations

A typical Building structural engineer may not be aware of several key aspects which impact the designs, such as joints and connection detailing on facades which may have great impact on structural aspects. In the Façade engineering of high-rise building, there are situations wherein framing and fixing sizes are

just not controlled by wind loads or structural aspects, but are rather limited by non-engineering challenges like functional, visuals etc,. Hence the overall quality of structural design and project outcomes are relying on experience of the engineer. Table 3 elaborates the critical aspects of these points.

Table 3 : Engineers Skills and Quality of Designs Design Aspects

Design Check

Skills Required

Impact

Structural

Basic Design Check

Structural and Façade Engineering

Check adequate for short term performances only



Joints

Secondary Check

Façade Expertise

Check long term performances.



Movements

Secondary check & Building Interface Checks

Façade Expertise & Building Engineering

Check long term performances and potential risk of Large movements



10. Knowledge and Capability Availability of local expertise on this domain is limited due to several factors but few key factors are noted below. ● ● ● ●

Limited high-rise building designs were taught in the institutions Knowledge gap on practical Vs theoretical i.e., Academic learning not as per real world. Very less participation or collaboration of private/industry experts with academics/ institutions Limited built references and knowledge were not harvested or shared to wider engineer’s community

Volume 47 │ Number 4 │ December, 2017

Design Quality & Relevance

Hence due to several factors, high rise building façade engineering design knowledge and techniques are largely infused from international experts, collaborations etc., Experiences and lessons learned from wider international experts with leading local institutions adopting and learning appropriate and modern engineering designs will help this domain. The capability is improving dramatically with the recent developments of high rise buildings in many regions of the country, with improved availability of quality and experienced engineers and designers for high rise building facades.

The Bridge and Structural Engineer

Wind Related Studies of an Unusual Building : A Case Study

S. Selvi Rajan

P. Harikrishna

G. Ramesh Babu

N. Lakshmanan

Chief Scientist CSIR-SERC, Chennai TN, India sselvi@serc.res.in

Sr. Principal Scientist CSIR-SERC, Chennai TN, India hari@serc.res.in

Principal Scientist CSIR-SERC, Chennai TN, India gramesh@serc.res.in

Formerly Director and Project Advisor CSIR-SERC, Chennai TN, India

Selvi Rajan received her Bachelor of Technology in Civil Engineering from IIT Madras and Masters and Ph.D. from Anna University, Chennai. She is currently heading Wind Engineering Lab. of CSIRSERC, Chennai

Harikrishna received his Bachelor of Engineering in Civil Engineering from Andhra University and Masters from Anna University, Chennai and Doctorate from IIT Madras. He is currently a Senior Pr. Scientist in Wind Engineering Lab. of CSIR-SERC, Chennai

Ramesh Babu received his Bachelor of Engineering in Civil Engineering from SV Univeristy Tirupathi and Masters from JNTU-College of Engg., Anantapur and Prsuing Doctorate at IIT Madras. He is currently a Pr. Scientist in Wind Engineering Lab. of CSIRSERC, Chennai

Lakshmanan received his Bachelor of Engineering in Civil Engineering from A.C. Tecxh Karaikudi, Anna University and Masters and Ph.D. from IIT Madras. He was former Director of CSIRSERC, Chennai and an INAE Fellow

Summary Tall buildings with novel architecture having unusual size and shape, have become usual in this age, for purposes such as advertising, or to copy other famous buildings, or to serve as landmarks. One such large tubular structure to be located in terrain category 3 (sub-urban) was considered for aeroelastic and aerodynamic tests using Boundary Layer Wind Tunnel Facility of CSIR-SERC, Chennai, to determine the dynamic response of the structure. By considering the height, breadth and width of the prototype complex building, a geometric scale of 1:250 had been chosen to ensure negligible blockage effects during wind tunnel experiments. The mean force coefficients obtained from pressure measurement studies for chosen critical angles of wind incidence were used to evaluate the mean wind load distribution at design wind speed. The mean wind loads are multiplied with the design gust response factor to obtain the dynamic wind load distributions.

Introduction

Tall buildings with newer and innovative plan shapes are being built by architects and engineers, particularly in the central urban areas of cities. Curved structures

The Bridge and Structural Engineer

with circular/elliptical shape in plan/cross-section are generally preferred for functional/architectural purposes. The design of these unusual buildings is mostly governed by wind/earthquake loading. Since wind loads are randomly varying in nature, rational assessment of wind loads, and of structural behaviour are important requirements for ensuring safety and economy in the design of the structure. For asymmetric and unusual tall buildings with different plans at different levels in different orientations, appropriate wind loading coefficients are not readily available in the wind loading standards of practice. Further for a large three-dimensional structure like tall buildings, cooling towers, etc., pressures may not be fully correlated over the entire height of the structure. The Boundary Layer Wind Tunnel (BLWT) study continues to be the reliable method of proper assessment of wind loads on such unusual buildings[1]. The unusual building with complex structural form selected for the present case study is a large threedimensional tubular structure having three structural frequencies lower than 1.0 Hz. For a structure with such flexibility, it is mandatory to carry out aeroelastic model testing in the wind tunnel, in addition to the studies on pressures on rigid model to assess the Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

dynamic action owing to the spectral loading due to turbulence. In addition, the elliptic structural shape of vertical and transverse structural components of this unusual building may create vortex shedding leading to additional dynamic loading along with interference. Further, flow separations and reattachments would modify the spectral characteristics as well. Hence, dynamic behaviour of this type of building becomes very important, and is required to be studied using boundary layer wind tunnel.

Dimensions of the Unusual Tall Building

The overall height of the prototype building is 101 m. The complex building consists of five ellipsoids. Two horizontal ellipsoids which are

primarily under flexure, having large sectional dimensions (Tubes B &D), have elliptic cross section in one of the side elevations and are oriented in two orthogonal directions. These horizontal ellipsoids are supported at two different levels on three vertical ellipsoids (Tubes A, C & E), which have elliptic crosssections in plan. The unusual tall building also has a building (Tube F) at the ground level with a parabolic shape in one of the side elevations. The schematic view of the complex building with the tubular structural portions, viz., Tubes A, B, C, D, E and F, as mentioned above are shown in Fig. 1. The dimensions of various components of the tall unusual building are given in Table 1.

(a)

(b)

Fig. 1 : Schematic View of Tall Unusual Building and its FEM Model

Table 1 : Full-Scale Dimensions of the Tall Unusual Building (Fig. 1) Tubular Structural Portions A B C D E F

Major dia. (m) x Minor dia. (m) 15.0 x 13.0 33.0 x 27.0 21.0 x 18.0 33.0 x 27.0 15.0 x 13.0 44.0 (breadth)

Details of Wind Tunnel Experiment

Pressure measurement study and an aeroelastic study, respectively, on the scaled model of the tall unusual building has been carried out using the state-of-art Boundary Layer Wind Tunnel (BLWT) available at CSIR-SERC, Chennai. The complex building is to be located in the terrain category 3, as per IS:875 (Part 3) 1987[2]. Accordingly, mean velocity profile, turbulence intensity profile and spectrum of horizontal

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Length (m) 113.5 89.5 93.3

Height from Base to Top (m) 99.0 93.6 101.0 45.0 56.7 39.0

wind speed corresponding to sub-urban terrain with a length scale of 1:250 have been simulated in the wind tunnel by using a trip board followed by a set of boards with wooden cuboidal roughness elements as vortex generators. The results of the simulation in terms of profiles of mean velocity and turbulence intensity, and spectrum of longitudinal velocity fluctuations are shown in Fig. 2. The power law coefficient of the mean velocity profile was experimentally found to be

The Bridge and Structural Engineer

equal to 0.21. The simulated profiles of mean velocity and turbulence intensity, and spectrum satisfactorily

a) Mean Velocity Profile

represent the wind characteristics of a sub-urban terrain to a geometrical scale ratio of 1:250.

b) Turbulence Intensity Profile

c) Spectrum of Fluctuating Wind Velocity

Fig. 2 : Simulation of Atmospheric Wind Characteristics in Wind Tunnel

3.1. Fabrication of Pressure Model A rigid model approach has been selected for the measurement of pressures acting circumferentially on all tubes of the complex building. With a geometric scale of 1:250, the maximum height of the model of the building corresponds to 404 mm, which can be accommodated well within the boundary layer developed in the wind tunnel. All linear dimensions of the structural component details in Table 1 are suitably scaled down. Acrylic material has been used for fabricating the rigid model. Use of this material permitted reasonable transparency and workmanship, and helps to avoid any kinks while connecting the pressure tubes to the pressure sensors. The coordinates of the ellipsoids are geometrically matched to a scale of 1:250. Individual tubes (Tubes A, B, C, D & E) are fabricated separately and the building model is integrated together along with the parabolic building (Tube F), as shown in Fig. 3. Suitable arrangements have been made to house 5 high-speed pressure sensor modules each comprising of 32/64 channels inside different tubes of the model. To provide a rigid base connectivity, a thicker acrylic sheet is used as base plate between the model and the turn table of wind tunnel. Typical view of the fabricated/instrumented model is shown in Fig. 3.

Fig. 1. By considering the height, breadth and width of the prototype complex building, a geometric scale of 1:250 has been chosen to ensure negligible blockage effects during wind tunnel experiments. To account for dynamic scaling, the mass and the stiffness are to be scaled to attain the scaled frequencies and mode shapes. The dead and live loads on the structure of all the six portions (Tubes A, B, C, D, E & F) are given in Table 2. Based on these loads, the weight of each portion of the model is computed. Unlike earthquake loading, wind loading occurs over longer duration, and also has much higher probability of occurrence. Hence, total live load is considered, as this would lead to lower natural frequencies. As the frequencies become lower, the energy available for excitation due to turbulence in atmospheric wind becomes larger, leading to increased dynamic response.

3.2. Fabrication of Aeroelastic Model The unusual tall building has two ellipsoids of large sectional dimensions, termed as beams (elements B & D) supported at two different levels on three vertical ellipsoidal columns (elements A, C & E), as shown in

Fig. 3 : View of the Unusual Tall Building Model for Pressure Measurement

The Bridge and Structural Engineer

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

Table 2 : Weight Distribution in Various Structural Portions Tubular Structural Portions A B C D E F Total

Weight, t Dead Live Total 20750 16085 31125 2440 2590 230 75290

6875 4600 8480 280 375 300 20910

27625 20685 39605 2720 2965 2600 96200

Weight of Model, kg 1.745 1.324 2.535 0.174 0.190 0.166 6.134

The natural frequencies and mode shapes have been obtained using FEM modeling. The first mode roughly corresponds to rotation of the top level ellipsoidal beam about one of the columns, the second mode corresponds to lateral translation of the ellipsoidal beam at top, the third mode corresponds to the translation of the lower ellipsoidal beam, and the fourth mode corresponds to the rotation of the lower level ellipsoidal beam. The first three modes involve dominant motion in the horizontal plane. The weight distributions of all the tubular structural portions were scaled down to represent the corresponding masses on the prototype structure to a linear scale ratio of 1:250 (Table 2). If metal plates are used in model, they can be of thin cross-section, leading to local defects and also local resonances. Hence, all the tubular structural portions are made as composites with a rigid pipe at the centre surrounded by light weight core and a thin metal sheet bonded to the core (Fig. 4). Additional M.S. steel plates are added at the ends of the two horizontal beams and three vertical portions to simulate the total mass. The aluminium pipes located at the centre of the vertical columnsare projected at the base to introduce flexibility at the base of the columns. All the tubular structural components are rigid. Thus, elastic elements have been introduced at the base of each of the tubular portions (columns/Tubes A, C and E) to obtain the required frequencies and mode shapes for the first four fundamental modes. The height of the flexible elements is adjusted to obtain the required natural frequencies and mode shapes. The proposed scheme in modeling would ensure linear mode shape in the vertical direction. With the simulated masses, this would lead to slight lowering of natural frequencies. The three vertical columns (Tubes A, C and E) have been instrumented with strain-gauges to 58

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

measure moment vectors Mx and My along the body axes of the structure (Fig. 4). The strain gauges in the direction of X-axis are connected to 3 channels, and the strain gauges in the direction of Y-axis are connected to other 3 channels of the data acquisition system. The aeroelastic model is then rigidly coupled to the turn table of wind tunnel for conducting wind tunnel experiments. Typical view of the fabricated and instrumented model is shown in Fig. 4.

Fig. 4 : View of the Instrumented Aeroelastic Model inside Wind Tunnel

Forced vibration testing of the model is conducted using an electro-dynamic shaker, and the frequency response of the model is determined after mounting it in the final position on the turn table. The instrumentation essentially consisted of accelerometers to measure acceleration response, an electrodynamic shaker to excite the building to respective modes, a signal generator to generate sinusoidal function for various range of frequencies, power and charge amplifiers to amplify the signals. Table 3 gives the expected natural frequencies based on model scale ratio for the first three modes and the actual frequencies realised on the model using forced vibration testing and through the spectral plots of measured My and Mx responses. Table 3 : Natural Frequencies of Model Mode Expected Forced No. Frequency, Vibration Hz Tests 1 66 53 2 83 68 3 117 91

Wind Tunnel Tests 53 69 97

Ratio

0.80 0.83 0.83

The Bridge and Structural Engineer

All the three frequencies have a near constant ratio with the expected frequencies, and the mode shapes are well captured. The percentage of damping is experimentally found to be approximately 5.5. The correction for damping is suitably carried out by enhancing the variance obtained in experiments in the resonant range and the enhanced values of standard deviation are termed as modified standard deviation.

been computed. With reference to a fixed set of structure axes (X, Y and Z), orientations for forces Fx, Fy and Fz along with drag and lift directions in horizontal plane corresponding to angle of wind incidence ‘θ’ are defined in Fig. 5.

3.3 Wind Tunnel Test Cases A total of 196 pressure taps have been provided covering the entire envelope of the building model. The respective tributary areas contributed by each pressure measurement zone are manifested, covering the entire surface of the building to obtain integrated load effect. The pressure tubing systems with suitable restrictor having flat frequency response up to 300 Hz have been used to satisfactorily capture the fluctuating pressures upto 250 Hz and consequent attenuation of peak pressures. The experiments are conducted for two different wind speeds of about 6.6 m/s and 12.0 m/s, measured at the height corresponding to top of the model. As the tall unusual building model is complex and asymmetric in appearance, it is tested for 24 angles of wind incidence (θ) from 0º to 345º, at 15º intervals, for both the tests. The pressure data are acquired for a sampling duration of 18 seconds and with a sampling rate of 500 samples per second per channel. The aeroelastic model has been instrumented with six accelerometers and six strain gauges to measure the individual forces in X and Y directions respectively at the top and bottom of vertical structural portions of Tubes A, C and E, projecting above the horizontal structural portions. All the test cases are carried out at two mean wind speeds of 17.5 m/s and 23.0 m/s, measured at the height of the model and the results are further extrapolated to design wind speed.

Evaluation of Aerodynamic Coefficients

Mean pressure and various mean force (Fx, Fy, Fz, Fr, FD, FL) coefficients are evaluated from the measured pressure data. All the mean pressure coefficients are deduced with respect to the reference pressure at respective heights. The coefficients for maximum, minimum and standard deviation of pressures have

The Bridge and Structural Engineer

Fig. 5 : Schematic Diagram in Plan Showing Body Fixed Axes X and Y, θ, α and β

The forces per unit height/width Fx, Fy and Fz along the X-axis, Y-axis and Z-axis (body fixed axes), respectively, are computed by integrating the measured pressures along with the respective tributary arc lengths/widths, orientation of the pressure tap location in horizontal plane (α) and inclination of pressure tap location in vertical plane (β) as illustrated in Fig. 5. By resolving Fx and Fy in the direction of wind and perpendicular to the direction of wind, the drag force (FD) and the lift force (FL) are evaluated using the angle of wind incidence (θ). A MATLAB program has been developed for the analysis of all the strain gauge and accelerometer based data, after applying appropriate sensitivity factors and correction factors. Spectral analysis of measured time history records of all the strain gauges and accelerometers are carried out using FFT functions of MATLAB program. The typical spectra obtained from the acceleration trace is shown in Fig. 6. Spectral analysis of time history records of all the accelerometers clearly revealed that responses of accelerometers has dominant response under mode 1 corresponding to about 52 Hz. Similarly, the dominant response under mode 2 corresponded to about 70 Hz (Fig. 6). Volume 47 │ Number 4 │ December, 2017

Fig. 6 : Spectra of Accelerations in X & Y Direction at 00 Angle of Wind Incidence (23 m/s)

Results and Conclusion

From the variations of mean force coefficients with angle of wind incidence, critical angles of wind incidence have been identified for different tubes. Based on the identified critical angles of wind incidence for all the tubular structural elements, the critical angles of wind incidence have been recognized for the evaluation of wind loads on the building. The recommended force coefficients derived from the rigid model tests in the wind tunnel are used for the evaluation of mean wind loads. Further, the results based on pressure measurement studies are taken into account to compute the sum of the entire mean base bending moments from all six tubular elements. Comparison of the results are made by normalising the mean base bending moments with the respective maximum values of base bending moments from the variations with respect to angle of wind incidence. Normalising mean My, Mx moments evaluated from pressure measurements with the corresponding maximum values, the normalised variations of mean Mx, and mean My, are evaluated and are shown in Fig. 7. The normalised profiles of My and Mx of base bending moments obtained from aeroelastic tests at mean wind speed of 23.0 m/s, are also shown in Fig. 7. These show similar trend, and compare well with the profiles predicted from pressure measurement studies. The contribution of fluctuating component including resonant component at the natural frequencies of the building are considered using the gust response factor, G. The gust response factor, G has been evaluated based on dynamic response measurements from the aeroelastic studies. A design gust response factor of 60

Volume 47 │ Number 4 │ December, 2017

2.8 has been suggested based on these experimental studies. Using the measured acceleration values, the peak displacements of the prototype structure at design wind speed of 43.2 m/s has been evaluated and are shown in Fig. 8. The spectral plots of accelerations and dynamic strains did not indicate vortex shedding over any specific frequency band. Based on the experimental studies, 0º and 180º angles of attack are to be considered as critical for My and 90º and 270º angles of attack for Mx, apart from 45º, 135º, 225º and 315º for combined My and Mx.

Fig. 7 : Variation of Normalised Mean My and Mx based on Pressure and Aeroelastic Studies

Fig. 8 : Predicted Peak Total Displacement for the Prototype Structure

The Bridge and Structural Engineer

Acknowledgements Authors thank the financial support provided by M/s. ONGC Kolkata. The support rendered by all the staff of the Wind Engineering Laboratory, towards fabricating the building model and during the course of the experiment, is gratefully acknowledged.

Selvi Rajan S., Arunachalam S., Ramesh Babu G., Harikrishna P., Abraham A., Chitra Ganapathi S., Sankar S., and Nagesh R. Iyer., â&#x20AC;&#x153;Wind Tunnel Pressure Measurement Studies on the Model of a Tubular Buildingâ&#x20AC;?, SSP 08841, CSIR-SERC Report 1, November 2011.

Selvi Rajan S., Arunachalam S., Ramesh Babu G., Harikrishna P., Abraham A., Chitra Ganapathi S., Sankar S., and Nagesh R. Iyer., â&#x20AC;&#x153;Wind Tunnel Studies on the Aeroelastic Model of a Tubular Buildingâ&#x20AC;?, SSP 08841, CSIR-SERC Report 2, December 2011.

Lakshmanan N., Selvi Rajan S., Harikrishna P., Ramesh Babu G., and Arunachalam S., â&#x20AC;&#x153;Aeroelastic Studies on a Complex Tubular Building Modelâ&#x20AC;?, CSIR- SERC Research Report No. MLP 14041-17, December 2011.

References 1.

Simiu E., and Scanlan R.H., â&#x20AC;&#x153;Wind Effects on Structuresâ&#x20AC;?, Third edition, John Wiley & Sons, Inc., New York, 1996.

IS:875 (Part 3) â&#x20AC;&#x201C; (1987) â&#x20AC;&#x201C; â&#x20AC;&#x153;Indian Standard Code of Practice for Design Loads (other than Earthquake) for Buildings and Structures, Part 3, Wind Loadsâ&#x20AC;?, Bureau of Indian Standards, New Delhi, 1989.

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

6HFWRUV &RYHUHG z +LJKZD\V %ULGJHV z 7UDIILF 7UDQVSRUWDWLRQ z /DQG 'HYHORSPHQW z %XLOGLQJV z $LUSRUWV z :DWHU :DVWHZDWHU z 3RZHU z 3RUWV z 5DLOZD\V

6HUYLFHV 3URYLGHG z )HDVLELOLW\ 6WXGLHV z (QJLQHHULQJ 'HVLJQ z 3URRI &KHFNLQJ z 3URMHFW 0DQDJHPHQW z ,QGHSHQGHQW (QJLQHHU 6HUYLFHV z $XWKRULW\ (QJLQHHUV 6HUYLFHV z /HQGHUÂˇV (QJLQHHU 6HUYLFHV z 'XH 'LOLJHQFH 6WXGLHV

2YHU <HDUV RI &RQVXOWLQJ ([FHOOHQFH

) U L V F K P D Q Q 3 U D E K X

R X U F R P P L W P H Q W W R G H O L Y H U L Q J H [ F H O O H Q F H

%DOJRYLQG :DGL 1HZ 3UDEKDGHYL 5RDG 3UDEKDGHYL 0XPEDL ,1',$ 7HO z )D[ z H PDLO SIPXPEDL#ISLQGLD FRP z ZHEVLWH ZZZ ISLQGLD FRP

WKURXJK LQQRYDWLRQ

H [ F H O O H Q F H W K U R X J K L Q Q R Y D W L R Q

The Bridge and Structural Engineer

$1$% $&&5(',7('

$%06 %6

,62 )6

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

+DOI 3DJH $GYHUWLVHPHQW 6L]H + [ : FPV

Wind Engineering for a 107 m Tall Statue : A Case Study

Nicholas Truong Associate Director, Windtech Consultants Pty Ltd., Sydney, Australia nicholas@windtechglobal.com

Antonios W Rofail Director, Windtech Consultants Pty Ltd Sydney, Australia trofail@windtechglobal.com

Nicholas Truong received his Bachelor degree in Mechanical Engineering from the University of Queensland and Doctorate from the University of Oxford. He has over 15 years of consulting, research and development experience in the field of fluid dynamics and wind engineering. Since joining Windtech Consultants in 2008 Nicholas’ particular interest has been using wind tunnel testing to design unusual structures under wind loading.

Antonios Rofail received his Bachelor and Masters degrees in Civil Engineering from the University of Sydney. He is a member of the Standards Australia committee for Wind Actions on Structures and AWES Committee for the Quality Assurance Manual for Wind Engineering. He co-founded Windtech Consultants in 1991 and has over 30 year’s experience in wind engineering, working on over 2500 major projects worldwide.

Abstract Windtech Consultants was engaged to model the extreme loads and pressures acting on a 107 m (305 ft) tall statue in the form of Lord Shiva, to be constructed in Nathdwara, India. To add to the complexity, the statue includes a trident held from the base and having the form of a 58 m high mast. The loads on the main structure were determined from wind tunnel measurements using the high frequency force balance and high frequency pressure integration methods. Good agreement was found between the mean wind tunnel coefficients for the two methods. The final loads were calculated by combining the wind tunnel results with the local directional wind climate using the multi-sector method. Wind loads on the trident were found using the equivalent static wind loads techniques.

Introduction

The site for the statue is Nathdwara in the state of Rajasthan, India. The site is approximately 500 km south-west of New Delhi and 600km north of Mumbai. It is proposed that the statue will be constructed within a newly landscaped parkland and that the statue will be positioned on the top of a ridge line. The statue will be constructed from a concrete skin over steel 62

framework. A sprayed metal finish will be then applied to the concrete. The trident will be constructed from a concrete staff and the head of the trident is a metal skin over a steel frame (Fig. 1).

Volume 47 │ Number 4 │ December, 2017

Fig. 1 : Statue in Elevation

To determine the extreme wind loads on the statute a detailed wind tunnel study was undertaken and the following areas were considered: analysis of the local wind climate, determination of the wind loads on

The Bridge and Structural Engineer

the main statue, determination of the wind loads on the trident and determination of the surface pressure distribution on the statue and the trident.

Wind Tunnel Model

A small model of the statue was initially sculpted by a renowned artist and then three-dimensionally scanned. This three-dimensional CAD file was used to construct the wind tunnel model using three-

(a)

dimensional printing. The full statue was constructed at a scale of 1:200 and additionally a detailed model of the head of the trident and top of the statue was constructed at 1:100 scale (Fig. 2). These models were used in the subsequent wind tunnel testing. The 1:200 scale model was used to study the wind loads on the whole statue, whereas the 1:100 scale model was used for the detailed study of the wind loads on the trident.

(b)

Fig. 2 : (a) 1:200 Scale Model Setup in the Wind Tunnel, (b) 1:100 Scale Model Setup in the Wind Tunnel

Subsequent to the force balance testing the 1:200 scale model was instrumented with over 700 pressure sensors (Fig. 3).

Fig. 3 : Pressure Sensor Locations (a) Front Elevation (b) Plan

Local Wind Climate

A detailed wind climate analysis was conducted for the Nathdwara region using data from the nearby Jaipur and Kota Airports. Other nearby airports, such as Jodhpur were not used due to topographical and

The Bridge and Structural Engineer

geographical differences between the development site and the airports. The recorded wind speed measurements were terrain corrected using the ESDU terrain transition method (ESDU, 2002). The annual and monthly extremes of the two sites were analyzed together and separately using Gringorton’s extreme value analysis technique (Holmes, 2001). Due to the projected long lifetime of the structure, a wind speed based on a mean recurrence interval of 250 years was specified by the client. Fig. 4 presents a comparison of the results of the wind climate analysis with the Indian wind loading code (Bureau of Indian Standards, 2007) and the AsiaPacific wind speed handbook (Holmes & Weller, 2002). Although the magnitude of the 250 year wind speeds varies significantly between the three sources, the wind-speed to return-period gradients are comparable. Furthermore the 47 m/s (50 year) basic wind speed region from the Indian Standard covers approximately twenty percent of mainly the north of India and typically standards reflect the upper limit of wind speeds within a region plus a small safety factor.

Volume 47 │ Number 4 │ December, 2017

measurements with a patch area and moment arm. The patch areas and moment arms were determined from the three dimensional CAD model. For example the times series of the base moments about the y-axis can be calculated as: The time series can then be used in any time or frequency domain based analysis method. Fig. 4 : Comparison of Basic 3 Second Gust Wind Speeds for the Nathdwara Region (Referenced to 10 m Height in Open Terrain)

For subsequent analysis the 250 year basic gust wind speed of 49.2 m/s as determined by Windtech was used. Directional wind speed probability distribution was also determined for this site. The directional probability densities were modified to account for uncertainties in the wind climate data.

Wind Loads on the Main Statue

4.1 Method The wind loads on the main statue structure were determined using the high frequency pressure integration method. The high frequency pressure integration method determines the wind loads by integrating simultaneously recorded surface pressure

(a)

The high frequency pressure integration method was preferred over the high frequency force balance method due to the practical issue that there was the potential for the vibrational natural frequency of the wind tunnel model to conflict with the first mode natural frequency of the completed statue, during the calculation of the resonant response. This conflict occurs when the reduced frequency of the wind tunnel model is close in value to the reduce frequency of the dominant frequencies of the full scale structure. Due to the complex form of the statue, the reliability of the integration patch areas and moment arms was confirmed by testing the statue using the high frequency force balance method. Due to the issue outlined above it was only relevant to compare the mean base moment response. The mean overturning base moment from the two methods were compared and good agreement was found (Fig. 5).

(b)

Fig. 5 : Comparison of Mean Base Moments High Frequency Pressure Integration (HFPI) vs High Frequency Force Balance (HFFB), (a) Moments about the Y-Axis, (b) Moments about the X-Axis

There is a small overestimate in the base moments for moments about the Y-axis for winds from the west. This may be attributed to the wind flow between the arm and the statue and through the gap in the statue mound. Some refinements to the integration methods were tested. However, as will be shown in subsequent analysis, these moments do not dominate the design of the structure and it was decided to proceed with the analysis using the HFPI results.

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

The wind tunnel results from the high frequency pressure integration test were combined with the directional wind climate model to determine the base moments. The base moments were calculated using the multi-sector (directional probability integration) method (Holmes, 1990) which accounts for the probability of winds occurring from various directions.

The Bridge and Structural Engineer

Briefly, the multi-sector method used the following procedure: 1.

The directional wind speed probability distribution is known from the wind climate analysis

The directional response of the structure as a function of wind speed is known from the wind tunnel testing.

The inverse of the functions from points 1 and 2 are combined such that the directional probability can be calculated for a given response level

The response level is calculated from the functions from point 3 by applying the constraint that the total of the directional probabilities needs to equal the design probability (in this case 1/250).

(a)

4.2 Results Figs. 6 and 7 present the axis diagram and the directional contribution to the moments about the X-axis, Y-axis and Z-axis. Figs. 6b and 7 show that the response of the statue to wind loading is not overly directional. The probability integration method estimates the loads to be 9% lower for moments about the X-axis, 19% lower for loads about the Y-axis and 16% lower for loads about the Z-axis, compared with the traditional sector by sector calculation method. These modest reductions are typical for structures which have a broad directional response and are consistent with similar studies (Truong & Rofail, 2016). The broad along wind response of the structure can also be seen in Fig. 8 which presents the base moment coefficients without the resonant contributions.

(b)

Fig. 6 : Main Statue Results (a) Axis Diagram (b) Peak Base Moment about X-Axis

(a)

(b)

Fig. 7 : Main Statue Results (a) Peak Base Moment about Y-Axis (b) Peak Base Moment about Z-Axis

The Bridge and Structural Engineer

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

(a)

(b)

Fig. 8 : Main Statue Results Base Moment Coefficients Excluding the Resonant Response (a) about X-Axis (b) about Y-Axis

Façade Cladding Pressures

Façade cladding pressures were determined by combining the wind tunnel results from the 1:100 and 1:200 scale model test. Reliable estimations of the facade cladding pressures for the various convex and concaved surfaces of the statue would have been difficult without the use of wind tunnel testing. The final façade pressures were calculated using the multisector method described above (Fig. 9).

Fig. 9 : Main Statue Pressure Results (a) Peak Positive Pressures (b) Peak Negative Pressures

The probability integration method estimates the range of overall peak pressures to be 30% lower than the traditional sector by sector calculation. For the 90th percentile range of peak pressure the decrease is 23%. The magnitude of these decreases are consistent with the application of the probability integration method, where the most impact occurs for the largest values.

Wind Loads on the Trident

6.1 Method The statue design team were considering two

Volume 47 │ Number 4 │ December, 2017

structural linkage design options for the trident. In the first option, the trident is linked to the statue only at the base of the trident on Level 12 and in the second option the trident is also structurally linked to the statue by the tail of the snake. In the case of the first option the snake is only ornamental. Wind loads were estimated for the first case with the results of this study to inform the decision on whether a structural link, using the snake, should be incorporated into the final design. The wind loads on the trident were determined using the effective static wind loads method ((Holmes, 1996), (Holmes, 2002)). In this method the mean, background and resonant response are calculated separately and then combined to provide an effective load distribution. The effective static wind loads method calculates the correct equivalent pressure distribution by utilizing the 'load-response correlation' method (Kasperski, 1992) . The use of the 'loadresponse correlation' method allows the calculation of an effective static wind load for each load effect. For example, a different pressure distribution is calculated for peak base shear compared to the shear force at the half height of the structure. The effective static wind loads method was chosen for this application as it would allow for a detailed consideration of the wind loads due to the dynamic response of the trident in comparison with an envelope pressure distribution approach. It also includes the positive aerodynamic damping which is associated with the movement of the large trident head. To enable the application of the effective static wind loads method the drag coefficient of the head of the trident was determined using a 1:100 scale wind tunnel model. As the head and shoulders of

The Bridge and Structural Engineer

the statue is located with 15 m of the center of the trident, the statue may funnel winds on to the trident head and increase the wind loads. The impact of the statue in increasing the wind speed around the trident

(a)

head for selected wind angles was measured using a wind tunnel test including the top of the statue (Fig. 10). These increases were included in the final load calculations.

(b)

Fig. 10 : (a) Trident Axis Diagram; (b) Comparison of Drag Coefficients for Cases with and without the Effect of the Statue

6.2 Results The mean, background and resonant response of the trident were calculated using the effective static loads method and the estimated pressure distributions are determined. For this case the closed form approximations for the various load effects have been used (Holmes, 1996) and the load distributions are only a function of the relative height of the load effect. Two load distributions are presented below, the

(a)

first is for base reactions and the second is for junction between the head of the trident and the concrete staff (Figs. 11 and 12). The mean and background pressure distributions all follow the expected trends. However, the effective resonant pressure distribution has a large peak at Level 24. This is associated with the transition from a concrete staff to a steel framed trident head as well as the mass of the cross bar of the trident.

(b)

Fig. 11 : Trident Results at Base (a) Pressure Distribution (b) Force Distribution

(a)

(b)

Fig. 12 : Trident Results at Junction Between Trident and Concrete Staff (a) Pressure Distribution (b) Force Distribution

The Bridge and Structural Engineer

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

A feature of the effective static loads method is the inclusion of the effect of aerodynamic damping. For this structure the aerodynamic damping was estimated to be 1.4% of critical which is approximately twothirds of the total damping of the structure. The final force distribution is calculated by using the combined pressure distribution with the drag coefficients from the 1:100 scale model tests. The drag coefficient for the shaft was based on the surface roughness and was taken from the Australian and New Zealand Standard on wind actions (Standards Australia, 2016). Figs. 11 and 12 show the final shear force distribution as a percentage of the total shear force.

Conclusions

A comprehensive wind tunnel study has been conducted on the proposed Shiva statue located near Nathdwara in the state of Rajasthan, India. The overall bases moments, façade cladding pressures and point loads on the trident structure have been used to verify and inform the design of the statue.

Acknowledgments

The authors wish to acknowledge the input provided by Dr. John D. Holmes of JDH Consulting as well as the contributions from Dr. Prem Krishna, formerly Professor at the University of Roorkee, and Dr. Abhay Gupta of Skeleton Engineers.

ESDU, 2002. Strong Winds in the Atmospheric Boundary Layer, Part 1 : Hourly Mean Wind Speeds (Amdts A to E) (ESDU 82026), London: Engineering Science Data Unit.

Holmes, J., 1990. Directional Effects on Extreme Wind Loads. Aus Civil Eng Trans, CE32(1), pp. 45-50.

Holmes, J., 2001. Wind Loading Of Structures. London: Spon Press.

Holmes, J. D., 1996. Along-wind Response of Lattice Towers: Part III – Effective Load Distributions. Engng. Struct., 18(7), pp. 489-494.

Holmes, J. D., 2002. Effective Static Load Distributions in Wind Engineering. JWEIA, Volume 90, pp. 91-109.

Holmes, J. D. & Weller, R., 2002. Design Wind Speeds for the Asia-Pacific Region, Sydney: Standards Australia International.

Kasperski, M., 1992. Extreme Wind Load Distributions for Linear and Nonlinear Design. Engineering Structures, Volume 13, pp. 27-34.

Standards Australia, 2016. Australian and New Zealand Standards Structural Design Actions: Part 2 Wind Actions (AS/NZS 1170.2-2011 incl Amdt-1 to 4), Sydney: Standards Australia.

Truong, N. & Rofail, A., 2016. Multi Sector Directional Probability Integration of Wind Loads: Comparison against the Sector Method. McLaren Vale, South Australia, 18th AWESW.

References 1.

Bureau of Indian Standards, 2007. Code of Practice for Design Loads (other than Earthquakes) for Buildings and Structures. Part 3 Wind Loads. IS 875, New Delhi: Manak Bhavan.

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

State of the Art of Long Span Bridge Aerodynamics

Toshio MIYATA

Hitoshi YAMADA

Hiroshi KATSUCHI

Professor Emeritus Yokohama National Univ. Yokohama, Japan miyata@ace.ocn.ne.jp

Professor, Yokohama National Univ. Yokohama, Japan yamada-hitoshi-cj@ynu.ac.jp

Professor, Yokohama National Univ. Yokohama, Japan katsuchi@ynu.ac.jp

Toshio Miyata received his Bachelor, Master and Doctor degrees in Civil Engineering from University of Tokyo. He was appointed Professor at Yokohama National University and currently Professor Emeritus of Yokohama National University.

Hitoshi Yamada received his Bachelor, Master and Doctor degrees in Civil Engineering from University of Tokyo. He is currently Professor at Yokohama National University

Hiroshi Katsuchi received his Bachelor degree in Civil Engineering from Tokyo Institute of Technology, Master degree from Johns Hopkins University and Doctor degree from Yokohama National University. He is currently Professor at Yokohama National University

Summary Suspension bridge span length has become longer and longer since the start of construction of modern suspension bridges in 19th century. Due to technological development, in particular, wind-resistant design methodology, the span length has reached almost 2,000 m in the Akashi Kaikyo Bridge completed in 1998 in Japan. This paper revisits the wind-resistant design of the Akashi Kaikyo Bridge and presents some new recent trends in long-span bridge aerodynamics as well.

come to govern the structural design of suspension bridges as the main span length increases. In fact, there were many difficulties to be solved in the wind-resistant design of the Akashi Kaikyo Bridge. In addition, new findings and innovations of bridge aerodynamics were brought during the design work of the Akashi Kaikyo Bridge. This paper revisits the wind-resistant design of the Akashi Kaikyo Bridge and also presents some new recent trends in long-span bridge aerodynamics.

Introduction

Since the analytical theory of a suspension bridge was developed in the 19th century, construction of modern suspension bridges started. Furthermore, the deflection theory was developed by Moisseiff in 1901, by which the main span length of a suspension bridge has increased dramatically. In the end of 20th century, the Akashi Kaikyo Bridge with the world’s current longest span was completed in Japan in 1998, as shown in Fig. 1. The span length is 1,991 m which was increased from 1,410 m in the Humber Bridge in UK at that time. Since then, many long-span suspension bridges have been built. However the Akashi Kaikyo Bridge still holds the world’s longest span. As known widely, wind effects

The Bridge and Structural Engineer

Fig. 1 : Akashi Kaikyo Bridge

Difficulties in Ensuring Flutter Stability

In the wind-resistant design of long-span bridges, ensuring the stability against flutter is the most challenging issue. As long as a deck is suspended from Volume 47 │ Number 4 │ December, 2017

cables hanging on two pylons, natural frequencies of the bridge decrease as the center-span length increases, as shown in Fig. 2[1, 2].

fundamental torsional frequency and deck width) makes the situation more clear, as shown in Fig. 4[3]. As already described, the flutter speed required is around 80 m/s and deck width is determined by the number of traffic lanes. The reduced wind speed in Fig. 3 is mostly determined by the fundamental torsional frequency. Besides, it is known that torsional flutter occurs in a relatively low reduced wind speed, i.e., 5, and that coupled flutter occurs in a higher range, i.e., more than 10, as shown in Fig. 4[1-3]. This means that very long-span bridges more than the AkashiKaikyo Bridge must tackle both similar flutters, but different mechanisms.

Fig. 2 : Relationship between Natural Frequency and Center-Span Length[1]

It is well known that the onset wind speed of torsional and coupled flutter is proportional to the fundamental torsional frequency and the frequency ratio between the fundamental torsional and bending frequencies for coupled flutter. Fig. 3 shows the onset flutter wind speeds calculated with the Selberg’s formula for various suspension bridges[3]. Since the design requirement for flutter speed of long-span bridges in Japan is around 80 m/s, the reduction of the flutter onset wind speed must be compensated by aerodynamic methods, for instance, modification of a deck cross section. The extent of this effort will be intensified more as the center-span length increases.

Fig. 3 : Flutter Onset Wind Speed with Selberg’s Formula Against Span Length[3]

Looking at this difficulty ensuring flutter stability for long-span bridges from another aspect, plotting the design requirement for flutter speed in a reduced wind speed (flutter speed required divided by the 70

Volume 47 │ Number 4 │ December, 2017

Fig. 4 : Flutter Requirement in Reduced Wind Speed vs. Span Length[3]

In order to ensure the flutter stability against those difficulties, the basic concept of large torsional rigidity and deck shape optimization has been adopted so far in the history of the wind-resistant design of long-span bridges in Japan. Typical solutions are a truss-stiffened girder or a closed-box deck with aerodynamic stabilizing devices such as fairings, spoilers, splitter plates and so on. However, there may be a limit of realizing large rigidity under constraints of traffic demand and self weight. Besides, the aerodynamic optimization of a deck cross section is very sensitive and nonlinear to the change of geometry. Since flutter is one of the aeroelastic vibration modes, an aerodynamic aspect or aero-elastic aspect in a narrow sense must be replaced by an aero-elastic aspect in a broad sense. That is, flutter should be dealt with a vibration of a whole structure. Therefore, there must be a better solution against flutter of longspan bridges from the viewpoint of an aero-elastic relationship of the whole bridge. The importance of 3D wind-resistant design instead of the 2D design should be emphasized.

The Bridge and Structural Engineer

2D Deck-Shape Flutter Control

As already described, the history of flutter suppression or control is to optimize a deck cross section so as to be aerodynamically stable. However, it is a very critical process because the aerodynamic effects are very sensitive and nonlinear to the change of the cross section. Besides, this process becomes much more difficult as the span length increases. Examples of this process in the case of the AkashiKaikyo Bridge is shown in Fig. 5[4]. Single closed box decks failed to meet the code requirement for flutter instability, provided that the conventional design was made by the minimum thickness rule under prescribed loading. To raise their critical wind speeds, it should be essential to increase the use of steel in spite of economical disadvantage or to produce possible aerodynamic devices such as open slots

and a median barrier. Some double box girders with a wide open slot and median barrier meet the code requirement. However, they are not advantageous because the estimated dead load of the deck was heavier than that of a truss structure. In addition, wind tunnel tests reveal that they suffer from sever vortex excitation. “Compound stiffness girder” is an attempt to ensure aerodynamic stability by using conventional girder configuration. It consists of two single box sections of same width but different depth. They are arranged such that shallow sections are used in the middle portions of both main and side spans to attain sufficient aerodynamic stability, and deep sections are used at the end portions of each span to provide appropriate rigidity. Finally, it was decided that a truss stiffened structure should be adopted by considering the achievement in other long-span bridges in Japan, erection method, maintenance cost, etc.

Fig. 5 : Flutter Onset Wind Speed for Various Deck Configurations[4]

Significance of 3D Coupled Flutter Characteristics

During the wind-resistant design and aerodynamic investigation of the Akashi Kaikyo Bridge, a serious concern arose whether the 2D section model test investigation in a smooth flow is sufficient or not. In order to respond to this question, a large aeroelastic model test program was planned. The main purpose of this program is to examine three-dimensional aerodynamic responses and to confirm the safety of the Akashi Kaikyo Bridge. For this, the test was conducted in both a smooth flow and a boundary layer turbulent flow. The safety against flutter instability, gust response and static wind load were fully investigated. The second purpose is the establishment of a rational wind tunnel testing method for longspan bridges. A section model supported by springs

The Bridge and Structural Engineer

has been traditionally used to study the aerodynamic stability of a bridge deck section. However, the truss girder of the Akashi Kaikyo Bridge exhibit coupled flutter characteristics and moreover strong aerodynamic interference between the girder and main cables at the center portion of the main span. For such bridges, it is difficult to estimate the threedimensional behavior of the whole structure from only the sectional model responses. Therefore, the applicability of the analytical method by using the strip theory was studied simultaneously. In this method, the aerostatic and unsteady aerodynamic forces of the girder are measured by using the sectional models, and these forces were applied to calculate the threedimensional responses. The analytical predictions of flutter onset wind speed, gust responses and windinduced deformation were compared with those of the full model test. Volume 47 │ Number 4 │ December, 2017

Since the purpose of the aeroelastic model test was to confirm the applicability and the limitation of a sectional model test, the scale of the aeroelastic model test was chosen to be 1/100 which was the same as that of the sectional model test to be compared. In addition, the aeroelastic model was specially designed and fabricated to reproduce full-scale bridge behaviour as accurate as possible. Fig. 6 shows the response records during coupled vibration of three degrees’ under large lateral

(a) Deformation by Wind Loading

deflection and windward-down rotation at the center by wind loading[5, 6]. It was believed that torsional flutter dominated bridge deck flutter for ordinary span length. Even for longer span length like the Akashi Kaikyo Bridge, coupled flutter was believed to occur with the coupling of torsion and heaving motions in the similar manner to an airplane wing in early days. However, Fig. 6 shows that this coupled flutter is coupled among torsion, heaving and lateral motions with certain phase lags.

(b) Time History of Flutter at Midpoint of Center Span

(c) Complex Flutter Mode Shape at Every 1/8 Cycle Fig. 6 : 3D Flutter Characteristics in 1/100-Scaled Aeroelastic Model of Akshi Kaikyo Bridge[5]

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

Multi-mode Coupled Flutter Analysis

Flutter instability of a long-span bridge as a sort of wind-structure interaction, or an aeroelastic problem has been simply expressed by a homogeneous equation of motion as in Eq. (1). The motion-dependent forces in the right hand side are usually defined in the linearized form with coefficients written in reduced wind speed.

... (1)

where h, α and p are vertical heaving, torsional and sway motions, L, M and D are lift, pitching moment and drag forces respectively, K = Bω/U is reduced frequency defined by width of deck (B) and wind speed (U). The coefficients H1* - H6*, A1* - A6* and P1* - P6* are motion-dependent force coefficients or flutter derivatives which are functions of the reduced frequency. Eq. (2) gives a complete 3D formulation. In early days when Scanlan developed an original one, only lift and pitching moment components associated with heaving and torsional motions (H1* - H3*, A1* - A3*). After the complicated 3D coupled behavior in the Akashi Kaikyo Bridge, it was expanded to Eq. (2) by incorporating drag and lateral motion components. Once identifying those motion-dependent force coefficients experimentally in a reasonable fashion, coupled flutter instability can be analyzed in the form of complex eigen-value analysis to get critical wind speed, response aerodynamic damping, flutter mode shape and response frequency. Assuming a three-dimensional structural frame model (Fig. 7) of an entire suspension bridge structure deformed under specified wind loads, the governing equation of motion can be described as follows;

The Bridge and Structural Engineer

When writing the linear equation in such a way, the instability as a result of complex eigen-value analysis is just to represent the boundary of negative or positive damping of the aeroelastic system concerned over or below the critical wind speed. There are several proposed formulations for the motion-dependent force coefficient of a bridge deck since Bleich et al.[7] after the Tacoma Narrows Bridge collapse. The most popular one may be the formulation developed by Scanlan[8, 9] as shown in Eq. (2).

... (3)

... (2)

where [u] is a deformed displacement vector. Now, it may be likely deduced that calculated mode shapes and variables of flutter instability delicately depend on structural properties as well as aerodynamic forces. Therefore, the direct FEM analysis method to analyze whole equation of motion should have a remarkable advantage in the sense of avoiding uncertainty in initial inputs. Of course, if an accurate image of flutter behavior is given in advance of the analysis, which mode shape results in the lowest critical wind speed, the modal analysis method of smaller calculation size must be a very good tool to derive the results by more saving time and cost. Above-mentioned 3D FEM analysis of flutter instability was as expected a very good tool to describe those complicated behaviors observed in a variety of investigations of wind effects on the Akashi Kaikyo Br., in particular flutter instability in the 1/100-scaled aeroelastic wind tunnel model study. It has the longest span length in the world of a three-hinged and threespan truss stiffening girder; 960 + 1, 991 + 960 m long. The scaled model had about 40 m long, situated in a large wind tunnel of 41 m wide, 4 m high and 30m long test section. In the case of originally designed truss deck, complicatedly coupled behavior of flutter instability occurred at a wind speed slightly higher than the requirement in the design code. The coupled Volume 47 │ Number 4 │ December, 2017

three-dimensional motion of dominant torsion and heaving as well as lateral was accompanied by random

(a) FEM Model

gust responses under large lateral deflection and windward-down twist at the center by wind loads.

(b) Wind Force Loading

Fig. 7 : 3D Frame Model of a Suspension Bridge and Wind Force Loading

Fig. 8 presents a comparison of the change in response aerodynamic damping and frequency with wind speed increased, measured in the wind tunnel test, and those of calculated ones by using analytical method interpreted above. Looking at measured change in response aerodynamic damping in Fig. 8(a), positive damping over lower wind speed range suddenly turns down to negative to cross the lateral axis of wind speed at around 8 m/s, to be defined the critical wind speed. Calculated results by the former idea with H1*-H4* and A1*-A4* of coupling only vertical heaving and torsional motions to excite flutter instability are in a good agreement with measured ones at wind speed

(a) Response Aerodynamic Damping in Logarithmic Decrement

of 7 m/s or lower while a large gap is apparent at higher wind speed. In this connection, coupling with the lateral sway motion was believed to play some role in flutter onset. Inclusion of the drag components as well as other force components (H1* - H4*, A1* A4*, P1* - P3* and P5*) yielded the result in a pretty good agreement between measured and calculated damping. Particularly, it succeeded in producing turn down from positive to negative damping around wind speed of 8 m/s. As for the change of response frequency, latter idea to include the effects of three degrees’ motions also demonstrates a pretty good agreement with the measured one[5].

(b) Response Frequency in Torsional Branch

Fig. 8 : Change of Aerodynamic Response of 1/100 Scaled Aeroelastic Model of Akashi Kaikyo Bridge with Wind Tunnel Speed[5]

Contribution of drag force components to the behaviour of flutter onset can be interpreted by means of investigating kinematics energy done by aerodynamic forces. Fig. 9 (a) shows calculated

Volume 47 │ Number 4 │ December, 2017

energy distribution to share among each element of the bridge at around wind speeds of flutter onset. It can be seen that the total energy changes from damping range to exciting range with wind speed

The Bridge and Structural Engineer

increased, in which the contributions of lift as well as drag forces to the change are pretty high. Analyzing particular effects of drag force further, as illustrated in Fig. 9 (b), the energy component caused by torsional motion more significantly contributes to the sub total by the drag force than the component by vertical heaving motion[5].

(a) Components (%) of Energy by All Forces on Each Element of Bridge (lower), (b) Components (%) of Energy by Drag Force on Deck caused by Each Motion (upper) Fig. 9 : Kinematics Energy Balance Distribution around Wind Speeds Close to Flutter Onset[5]

3D Structural Flutter Control

Improvement of the stability against the flutter in long-span suspension bridges has been made by 1) modification of deck configuration as an aerodynamic improvement to get more stable section and 2) adjustment of structural dimensions and setup to increase torsional frequency of the whole structure. Those procedures have traditionally been the primary wind resistant design process, where a lot of experimental investigations have been carried out in wind tunnels and a variety of structural design works have been repeated on the basis of past experiences. However, furthermore, from a viewpoint to suppress coupled flutter instability concerned, the control approach of flutter mode shape can be another possible solution[10]. Onset of coupled flutter instability has been understood as a solution of a two-dimensional modal flutter equation based on the fundamental torsional mode and the fundamental vertical heaving mode as an approximation of the flutter mode shape, usually both the symmetric 1st in case of three-hinged and

The Bridge and Structural Engineer

three-span suspension bridge. In coupling terms of this flutter equation, there is a combination of the mode shapes. When resemblance of those mode shapes assumed firstly is not perfect, it should affect the solution of critical wind speed of flutter onset. As presented above in the 3D flutter onset of the full-scale wind tunnel model of the Akashi Kaikyo Bridge, the flutter mode shape is originally so complex. As a result, the presence of coupling between three degrees’ motions as well as phase lag among those motions and along the bridge axis is to represent unique distribution of kinematics energy balance to result in flutter occurrence. It will be quite easy to imagine that this approach can be the third method to increase the stability, if the flutter mode shape and/or phase lag can be optimized artificially by any structural modification to increase the critical wind speed. It comes to be actualized after introduction of the direct flutter FEM analysis, which can analyze every detail of flutter occurrence including the distribution of phase lag in the flutter mode shape. Configurations of entire structure of a very long-span suspension bridge are apt to be selected by consideration of both static and dynamic effects. When this idea of mode shape control is involved in early decision of the structural configuration, additional improvement for the flutter stability will be possibly able to be achieved after/ before the conventional brushing-up in wind tunnel experiments. Recently, some improved bridge decks such as a twin-box deck, a multiple-box deck, a spindle deck[11], etc. have been adopted or proposed. In addition, a new structural configuration has been adopted as seen in the third Bosporus Bridge. Using a FE structural model for such innovative structures, structural and dynamic properties including mass, natural frequencies and mode shapes cab be easily obtained. Once flutter derivatives for such structures are obtained, flutter characteristics including the critical wind speed can be predicted. It is noted that flutter derivatives are obtained by not only an experimental method but also a CFD technique very recently. The CFD technique still needs improvement, however it will save the cost and time. It will be a new wind-resistant design procedure in future.

Conclusions

This paper revisits the wind-resistant design of the Akashi Kaikyo Bridge and presents some new recent trends in long-span bridge aerodynamics. The windresistant design work of the Akashi Kaikyo Bridge brought many significant findings of 3D coupled Volume 47 │ Number 4 │ December, 2017

flutter and established a 3D flutter analysis method. In addition, it has changed the wind-resistant design method from exclusive wind-tunnel dependent to hybrid of a 3D structural analysis and a wind-tunnel experiment. It has also enhanced the future possibility of inclusion of a CFD technique.

References 1.

Miyata, T., Yokoyama, K., Yasuda, M. and Hikami, Y., Akashi Kaikyo Bridge: Wind Effects and Full Model Wind Tunnel Tests, Aerodynamics of Large Bridges, A. Larsen (ed.), Balkema, 1992, 217-236. Miyata, T., Historical View of Long-Span Bridge Aerodynamics, Proc. of Engineering Symposium to Honour Alan G. Davenport for his 40 Years of Contributions, Ontario, Canada, 2002. Miyata, T. and Yamada, H., New Idea on Aero-elastic-coupled Flutter Control for Very Long Span Bridges, Bridge Aerodynamics, Larsen & Esdahl (eds.), Balkema, 1998, 153-163. Miyata, T. and Yamaguchi, K., Aerodynamics of Wind Effects on the Akashi Kaikyo Bridge, J. of Wind Engineering and Industrial Aerodynamics, 48, 1993, 287-315. Miyata, T., Tada, K., Sato, H., Katsuchi, H. and Hikami, T., New Findings of Coupled-Flutter in Full

Model Wind Tunnel Tests on the Akashi Kaikyo Br., Proc. Conf. Cable-Stayed and Suspension Bridges, Vol. 2, Deauville, France, 1994. 6.

Miyata, T., Full Model Testing of Large CableSupported Bridges, A State of the Art in Wind Engineering, Proc. of 9th Int. Conf. Wind Eng., New Delhi, pp.249-280, 1995.

Bleich, F., Dynamic Instability of Truss-Stiffened Suspension Bridges Under Wind Action, ASCE, Vol. 74, No. 8, 1948; Vol. 75, No. 3 and No. 6, 1949.

Scanlan, R. H. and Tomko, J. J., Airfoil and Bridge Deck Flutter Derivatives, ASCE, Vol. 104, No. EM4, 1971.

Simiu, E. and Scanlan, R. H., Wind Effects on Structures (Third Edition), Wiley, 1996.

10. Miyata, T. and Yamada, H., Mode Shape Control Concept for Improving Aerodynamic Instability of Very Long Span Suspension Bridges, Proc. IV Coloquio Int. Sobre el Enlace Fijodel Estrecho de Gibraltar (Sevilla), pp.73-79, May 1995. 11. Miyata, T., Significance of Aero-elastic Relationship in Wind Resistant Design of Long-span Bridges, Journal of Wind Engineering and Industrial Aerodynamics, Volume 90, Issues 12–15, December 2002, pp.1479-1492.

OBITUARY The Indian National Group of the IABSE express their profound sorrow on the sad demise of Late Shri Parkash Chander Bhasin on the 6th December, 2017 at New Delhi. He was the Secretary and Chairman of the Indian National Group (ING) and was closely associated with various activities of this Group since 1962. He was a member of numerous Technical Committees in India and abroad and also member of the Permanent Committee of the International Association for Bridge and Structural Engineering. Shri Bhasin was a man of great ability. His contribution to the activities of engineering profession and Group, will remain as landmark in the history of this group. Shri Parkash Chander Bhasin was well known for his dedication in the profession. Indian National Group of the International Association for Bridge and Structural Engineering sincerely appreciates his contribution to the Group and deeply mourns his untimely death. The Group prays the almighty God to grant strength and courage to the bereaved family to bear the loss. May his soul rest in peace. The Indian National Group of the IABSE express their profound sorrow on the sad demise of Late Dr CV Kand, Chairman, CV Kand Consultants Pvt Ltd on the 24th December, 2017 at Bangalore. He was an active member of the Indian National Group of the IABSE. The Group prays the almighty God to grant strength and courage to the bereaved family to bear the loss. May his soul rest in peace.

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

Wind tunnel Testing of Long Span Cable-Stayed Bridge Models

M. Keerthana

S. Selvi Rajan

P. Harikrishna

G. Ramesh Babu

A. Abraham

Scientist CSIR-SERC, Chennai TN, India

Chief Scientist CSIR-SERC, Chennai TN, India sselvi@serc.res.in

Sr. Principal Scientist CSIR-SERC, Chennai TN, India hari@serc.res.in

Civil Engineering from SVU College of Engineering, Tirupati

Sr. Scientist CSIR-SERC, Chennai TN, India abraham@serc. res.in

Dr. S. Selvi Rajan obtained her Bachelor’s degree in Civil Engineering from IIT-Madras, Master’s degree in Hydraulics and Water Resources from Anna University, Chennai and Ph.D. in Structural Engineering from Anna University, Chennai. She is currently working as Chief Scientist and Head, Wind Engineering Laboratory at CSIRSERC. She is member of many professional societies and BIS codal committees.

Dr. P. Harikrishna obtained his Bachelor’s degree in Civil Engineering from Andhra University, Master’s degree in Structural Engineering from Anna University, Chennai and Ph.D.in Structural Engineering from IIT-Madras. He is currently working as Senior Principal Scientist at CSIRSERC. He is member of many professional societies and BIS codal committees

G. Ramesh Babu obtained his Bachelor’s degree in Civil Engineering from SVU College of Engineering, Tirupati, Master’s degree in Structural Engineering from J.N.T.U College of Engineering, Anantapur. He is pursuing his Ph.D. at IIT-Madras. He is currently working as Principal Scientist at CSIR-SERC. He is member of many professional societies and BIS codal committees.

A. Abraham obtained his Bachelor’s degree in Civil Engineering from Coimbatore Institute of Technology, Coimbatore and Master’s degree in Structural Engineering from Anna University, Chennai. He is pursuing his Ph.D.in Structural Engineering from Anna University, Chennai. He is currently working as Senior Scientist at CSIR-SERC.

keerthana@serc.res.in

M. Keerthana obtained her Bachelor’s degree in Civil Engineering from College of Engineering, Guindy, Anna University, Chennai and Master’s degree in Engineering of Structures from Academy of Scientific and Innovative Research (AcSIR). She is pursuing her Ph.D. in AcSIR. She is currently working as Scientist at CSIRSERC.

Summary Wind tunnel experiments have been conducted on scaled rigid sectional model of four different bridge decks having different cross-sectional shapes/ configurations. The models were subjected to three different wind speeds of about 8, 11 and 13.5 m/s, measured at the height of the bridge deck level. The deck was rotated and tested for 0º, ±3º, ±6º, ±9º, ±12º and ±15º angles of wind attack, for simultaneous measurement of force and pressure. The acquired force and pressure data from the static sectional models were further analysed for each of the tested angles between -15º and +15º. The maximum of mean drag, lift and moment coefficients for the four bridge deck configurations under study was found. Based on Glauert-Den Hartog criterion, aerodynamic stability of the cross sections of the bridge decks investigated was assessed and all the deck sections studied were

The Bridge and Structural Engineer

S. Chitra Ganapathi Civil Engineering from GCE, Salem chitrag@serc.res.in S. Chitra Ganapathi obtained her Bachelor’s degree in Civil Engineering from GCE, Salem and Master’s degree in Structural Engineering from KCE, Coimbatore. She is currently working as Scientist at CSIR-SERC. She is pursuing her Ph.D. at IIT-Madras

found to be aerodynamically stable around 0º angle of wind attack.

Introduction

Bridges are an indispensable part of civil infrastructure. Construction of bridges has an enormous impact on economy and the environment. Recently, cable-stayed bridges have been adopted in increasing numbers in India owing to numerous advantages offered by them. These have high structural stiffness and are economical for bridges of medium to long-span. Wind loads are one of the important loads that govern the design of such bridges. The well-known collapse of Tacoma Narrows Bridge in 1940 clearly underlined the importance of wind loading effects on the performance of long-span bridges. Hence, aerodynamic studies have become an indispensable part of the structural design process of bridges([1], [2], [3]). Volume 47 │ Number 4 │ December, 2017

The cross-sectional shape of the deck plays a predominant role in aerodynamics of bridges[4]. One of the possible solution paradigms in the improvement of the stability of a bridge is to modify/ tailor the aerodynamic shape of the bridge deck cross-section[5]. In early stages of design, it is essential to evaluate the aerodynamic properties of the cross-section shape of the bridge deck through wind tunnel testing. Bridge aerodynamics can be considered to be relatively a new discipline in India with respect to Boundary Layer Wind Tunnel (BLWT) testing. At CSIR-Structural Engineering Research Centre (CSIR-SERC), such studies were initiated in 1999, and, aerodynamic stability was checked for various bridge deck configurations using wind tunnel tests. This paper describes studies carried out on four different deck configurations for checking their aerodynamic stability using the BLWT facility at CSIR-SERC, Chennai. The wind tunnel has a total length of 52 m with test section dimensions of 18 m (Length) x 2.5 m (Width) x 1.8 m (Height). A maximum wind speed of about 55 m/s can be generated in the wind tunnel. Further details of the wind tunnel facility of CSIR-SERC available in the website http://serc.res.in/wel-major-facilities/.

Wind Tunnel Experiments for Bridge Aerodynamics

Wind tunnel testing of bridges may be conducted in one of the following forms: (i) Full Aeroelastic Bridge Model Test (ii) Taut Strip Model Test (iii) Sectional Model Test A brief overview of these methods is presented below: 2.1 Full Aeroelastic Bridge Model Test In this type of testing, a scaled model of the fullscale bridge is tested and model similarity laws interpret the results. They offer a more complete picture of turbulence effects at high wind speeds, three-dimensional influences and construction stages. However, full aeroelastic bridge models are expensive to build and require a large wind tunnel, similarity of atmospheric turbulence, and similarity of mass distribution, reduced frequency and mode shapes. Moreover, the modelling of finer details of the bridge is cumbersome, since nonlinear behaviour of cables that support the decks is difficult to achieve 78

Volume 47 │ Number 4 │ December, 2017

in simplified wind tunnel models. However, full aeroelastic bridge model tests are the most reliable and constitute a generally accepted practical method to analyse the dynamic and aerodynamic action of wind. 2.2 Taut Strip Model Test The “taut strip” wind tunnel model had been developed as a simplified modeling technique for determining the response of long span bridge decks under turbulent wind. It is intended to model the mode shape of the deck of the bridge alone. In this method, only the stiffening girder of the bridge is modelled by the dynamic properties which are controlled by the tautstrings inside the model. When tested in a properly scaled turbulent boundary layer flow, the responses of the individual modes can be measured by selectively filtering the overall response in the neighborhood of the respective natural frequencies. Expressed in a suitable non-dimensional form, these modal responses can then be used to reconstruct the response of the real bridge having similar mode shapes. 2.3 Sectional Model Test Sectional model of a bridge deck is the basic tool to measure a number of parameters used in the development of wind loading on bridges. Sectional model being rigid are very simple, easy to construct and inexpensive. They represent a typical section of the bridge span, built to a model scale usually between 1/100 and 1/25. Aerodynamics of towers and cables are not represented in the sectional model studies. Sectional models do not account for varying vibration amplitude and aerodynamic forces across the length of a span nor do they account for the effects of turbulence. This type of models are intended to be useful in adjusting a deck section for a proposed bridge in minimising the vortex-induced response and flutter. Measurements of aerodynamic admittance function, joint acceptance function and aerodynamic derivatives are made using this type of model test. Along with relatively routine testing of turbulent wind spectra and model dynamic response, these tests provide the building blocks needed for the development of an equivalent static wind load methodology for design, based on experimentally measured parameters. CSIRSERC has put-in lot of effects in conducting wind tunnel experiments on this type of models of different cross-sections, using load cell to measure drag, lift forces and load cell to measure moment component exclusively, in addition to developing dynamic test rig to measure flutter derivatives under free and forced

The Bridge and Structural Engineer

vibrations. Details of these in-house developed load cells are explained elsewhere([6], [7]). In addition to these testing methods, topographic model studies, in which a large model of the topography is placed in the wind tunnel, is also carried out in specific cases. The effect of topography on the behaviour of long span bridges has been shown to be significant and in some case. It is beneficial effect on bridges with relatively poor aerodynamic stability. Several bridges have locations nearer to hills. If the wind is from the direction closer to a steep hill, the bridge is effectively screened by such steep hills. This produces a significant asymmetry in the flow and has an important effect in reducing the tendency for aerodynamic instability. Among these testing methods, the results of sectional model tests conducted in the wind tunnel facility of CSIR-SERC has been presented in this paper. Further, the stability of any bridge cross-section against transverse galloping can be assessed based on the results of sectional model test in wind tunnel using quasi-steady assumption. The formulation of the stability criterion has been presented in the succeeding section.

⋅

where m is the mass per unit length, h is the velocity corresponding to vertical displacement, ωh and ξh are natural frequency and damping ratio. For the system, the expressions in Eq. (2) are valid.

The force acting in the vertical direction (Fy(θ)) for is given by, ... (3) Writing the forces in terms of the non-dimensional aerodynamic coefficients, ... (4) Using relation between U and Ur from Eq. (2)

... (5)

For small rotation, that is, in the vicinity of

... (6)

Further, expanding CFy(θ) through power series and applying Eq. (4),

2.4 Transverse Galloping – Derivation on GlauertDen Hartog Criterion Galloping is a large amplitude oscillation which occurs in a non-circular or a bluff section in a direction normal to the direction of wind flow. When an elastically sprung body is immersed in a fluid flow, the force exerted by the fluid on the body doesn’t depend either on absolute velocity of the fluid (U) or the moving body ( ), but on the relative velocity (Ur), between them (Fig. 1).

... (2)

... (7)

Differentiating the above Equation w.r.t. θ,

... (8)

The equation of motion takes the form: ... (9) The R.H.S. of the above equation contributes to the overall damping of the system[8]. The divergent oscillation is critical or the deck is unstable when the term

becomes negative. This

is the well-known Glauert-Den Hartog criterion for transverse galloping instability. Fig. 1 : Direction of Resultant Wind

With reference to moving body, the equation of motion in the vertical direction is:

The Bridge and Structural Engineer

... (1)

Bridge Deck Configurations Studied

As mentioned earlier, wind tunnel tests on static sectional model of bridge decks have been initiated two decades ago at CSIR-SERC. A few results based on these wind tunnel studies conducted on four Volume 47 │ Number 4 │ December, 2017

different cross-sections are compiled in this paper to provide information about the aerodynamic force coefficients and their variations with change in cross-section of the deck. The geometric details and description of the four configurations have been presented in this section. 3.1 Configuration 1 The overall length of the cable stayed bridge is about 235 m and its width is about 13.65 m. The sectional details of the girder are shown in Fig. 2 schematically. Since the cross-section of the bridge deck has members with sharped edges, the effect of Reynolds number scaling between model and prototype is not expected to be significant. Thus, based on convenience of workability in fabricating the model as well as blockage ratio consideration, a model scale of 1:50 is selected. The wooden sectional model is fabricated in-house using good quality teak wood. The length of the sectional model is 98 cm (L) and its width (B) is 27.3 cm. The model includes all the important details such as kerbs, handrails, camber of the road, longitudinal girders and cross-girders of the prototype bridge. The model was instrumented with 30 numbers of pressure taps across a section at 0.25L, in addition to force measurement. A view of the model under test in the BLWT is shown in Fig. 3.

3.2 Configuration 2 The main span of the cable-stayed bridge is 550 m, with side span of 175 m each. The width of the deck is 21 m. The sectional model of the bridge deck with configuration 2 is fabricated in-house using good quality (imported) acrylic material. The length of the sectional model is about 100 cm and its width (B) is 42 cm and the overall depth (D) is 6 cm. The model includes all the important details such as anti-crash barriers, camber of the road, longitudinal trapezoidal cross-sectional girders with inclined struts of the prototype bridge, as shown in Fig. 4. The model was provided with 60 numbers of circumferential pressure ports at mid-section. The two segments of length 65 cm were fabricated and provided on either sides of the main model as dummy to ensure 2-D flow condition by covering the sectional model across full width of the test section. A view of the model under test in the BLWT is shown in Fig. 5.

Fig. 4 : Schematic Diagram of Bridge Cross-Section (Configuration 2)

Fig. 2 : Schematic Diagram of Bridge Cross-Section (Configuration 1)

Fig. 5 : View of the Bridge Deck Model Inside Wind Tunnel (Configuration 2)

3.3 Configuration 3

Fig. 3 : View of the Bridge Deck Model Inside Wind Tunnel (Configuration 1)

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

The full-scale cable stayed bridge having a main span of 210 m and side spans of 70 m on either side was chosen for wind tunnel studies. The bridge had a box type girder with total width of 13 m at the deck level and a depth of 4.58 m with two symmetric wings

The Bridge and Structural Engineer

on either side. The sectional model bridge deck was geometrically scaled down to a ratio of 1:50 and a physical rigid model was in-house fabricated using acrylic material. The dimensions of the scaled model are 2.34 m (L) × 26 cm (B) × 9.16 cm (D). The model was instrumented for both pressure measurement and force measurement. For the pressure measurement, the bridge deck was instrumented with circumferential pressure taps of 30 numbers at each section at three sections viz. 0.25L, 0.5L and 0.75L, as shown in Fig. 6. A view of the model mounted at both the ends of the wind tunnel using in-house fabricated strain gauge based load cells is shown in Fig. 7. End plates were also provided at on either side of the model.

2-D flow situation, two dummy models were placed on either side of the main model of 1.0 m length, as shown in Fig. 9.

Fig. 8 : Schematic Diagram of Bridge Cross-Section (Configuration 4)

Fig. 6 : Schematic Diagram of Bridge Cross-Section (Configuration 3)

Fig. 9 : View of the Bridge Deck Model Inside Wind Tunnel (Configuration 4)

Experimental Program

The wind tunnel testing of the static sectional model of the bridge deck of all 4 configurations was carried out for three different wind speeds viz., about 8 m/s, 11 m/s and 13.5 m/s and for ±15º of wind incidence with an increment of ±3º. Pressure data were acquired using high speed pressure scanners and in-house fabricated load cells/factory calibrated force balance were used to obtain force data. Fig. 7 : View of the Bridge Deck Model Inside Wind Tunnel (Configuration 3)

3.4 Configuration 4 A cable stayed bridge with a main span of 640 m and a width of 27 m was under study. The deck consists of two deep I-Beams and three shallow I-Beams and is provided with cross-beams at an interval of 4 m c/c, as shown in Fig. 8. The rigid sectional model of deck section was geometrically scaled down to 1:60. 3-D drawings of the model were prepared accordingly to the scale and the model was fabricated using 3D printing technology. The model was placed on the static test rig, which was specially fabricated in accordance to the dimensions of the deck structures. The model was provided with 59 numbers of pressure ports circumferentially at mid-section. To accommodate to the full width of the tunnel to create

The Bridge and Structural Engineer

Data Analysis

5.1 Evaluation of Aerodynamic Coefficients For each configuration, MATLAB program has been developed for the analysis and integration of measured pressures to evaluate various aerodynamic coefficients, viz., pressure, drag and lift coefficients, in addition to moment coefficients. Mean pressure and various mean drag, lift and moment coefficients (CD, CL, CM) are evaluated for each wind tunnel test case as given in the following section. 5.2 Pressure Coefficients The mean pressure , of a pressure trace p(t), taken over a period (duration) of time, Td is expressed as:

Volume 47 │ Number 4 │ December, 2017

... (10)

All the mean pressure coefficients were subsequently deduced with respect to the reference pressure, generally measured at height of the model, i.e, 1 m, using a pitot static tube. The mean pressure coefficient is calculated as a ratio of mean pressure ( ) to reference pressure ( ref) at model height (H). Thus,

... (11)

where,

factors (obtained experimentally using static loads, prior to conducting the experiment) to obtain the forces/moment, which were further converted to non-dimensional aerodynamic force coefficients using Eq. (12). Based on the analysis of data from measured pressure/force, the mean coefficients were thus derived for each configuration from each experiment. Variation of mean drag, lift and moment coefficients for all the four configurations are given in Figs. 10, 11 and 12, respectively.

(N/m2)

= Mean velocity at height of the model (m/s) 5.3 Force Coefficients 5.3.1 Based on pressure measurement With reference to a fixed set of structure axes, orientations for forces in drag, lift and moment are fixed as shown in the respective figures (Figs. 2, 4, 6 and 8). Further, drag and lift directions corresponding to angle of wind incidence ‘θ’ are also define in these Figures. The aerodynamic forces per unit width along the body fixed axes are computed by integrating the measured pressures along with the respective tributary widths. By resolving these forces in the direction of wind and perpendicular to the direction of wind, the drag force (FD) and the lift force (FL) are evaluated using the angle of wind incidence (θ). The mean force coefficients in drag, and lift directions are obtained as given below:

Fig. 10 : Variation of Mean Drag Coefficient with Angle of Wind Attack

... (12) where, = mean drag and lift force, respectively

Fig. 11 : Variation of Mean Lift Coefficient with Angle of Wind Attack

= mean drag and lift force coefficients, respectively = mean moment coefficient B = characteristic dimension (width of the bridge deck) 5.3.2 Based on force measurement The mean value of the forces/moment measured from the force balance are used to calculate the aerodynamic force coefficients, viz, drag, lift and moment coefficients using Eq. (12). In case of data from strain gauge based in-house load cell, the mean values of the readings were multiplied by the calibration 82

Volume 47 │ Number 4 │ December, 2017

Fig. 12 : Variation of Mean Moment Coefficient with Angle of Wind Attack

The Bridge and Structural Engineer

Results and Summary

Based on the analysis of experimental data, the variations of mean drag, lift and moment coefficients with angle of wind attack are shown in Figs. 10, 11 and 12, respectively. The slopes of lift coefficient around 0º are positive for all the four configurations and the values of CD are positive, as expected. Hence, all the deck cross-sections have been observed to be stable around 0º wind incidence based on Glauert-den-Hartog criterion for transverse galloping. This is based on drag coefficient at 0º angle of attack and gradient of lift coefficient with respect to θ at 0º angle of attack.

The maximum values of mean drag, lift and moment coefficients are extracted for each configuration and provided in Table 1. From Figs. 10, 11 and 12, it is observed that the aerodynamic coefficients do not follow a unique variation with reference to the respective cross-sectional shape. Configurations 1 and 4 have more members with sharp edges, whereas, Configurations 2 and 3 have trapezoidal cross-section, acting as bluff structure to wind flow. This suggests that wind tunnel testing is necessary for every configuration, which can be effectively used as a tool to assess the aerodynamic stability of cross-sectional shape of the bridge decks.

Table 1 : Maximum Value of Mean Aerodynamic Coefficients Maximum value Max. mean CD Max. mean CL Max. mean CM

Config.1 0.762 0.547 -0.303

Config.2 0.358 -0.913 0.160

Config.4 0.461 -0.803 -0.160

Bridges”, ASME Journal of applied Mechanics, 2011, 78, 1-11.

References 1.

Config.3 0.570 0.463 -0.145

Miyata T., “Historical view of Long-Span Bridge Aerodynamics”, Journal of Wind Engineering and Industrial Aerodynamics, 36, 2003, 517–538.

Einar N Strommen, “Theory of Bridge Aerodynamics”, Springer, Netherlands, 2006.

Selvi Rajan S., “Investigations into the Aerodynamics of Bridges Through Wind Tunnel Testing”, Ph. D. Thesis, 2008.

Hjorth-Hansen E., “Section Model Tests”, Proc. Aerodynamics of Large Bridges, Balkema, Rotterdam, 1992, pp. 95-112.

Taylor Z.J., Gurka R., and Kopp G.A., “Geometric Effects on Shedding Frequency for Bridge Sections”, Proceedings of the 11th Americas Conference on Wind Engineering, Puerto Rico, USA.

Selvi Rajan S., Jaya K.P., and Lakshmanan N., “Development of Load Cells for Simultaneous Measurement of Drag, Lift, and Moment for Section Models of Bridge Decks Under Wind Load”, Volume 33, Issue 6, Experimental Techniques, SEM, USA, (November/December 2009) pp. 38-45.

Simiu E., and Scanlan R.H., “Wind Effects on Structures - Fundamentals and Applications to Design”, 3rd Edition, John Wiley and Sons, NY 1996.

Graham J.M.R., Limebeer D.J.N., and Zhaox., “Aeroelastic Control of Long-Span Suspension

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

Sustainable engineering for Indian metros with U shape viaducts

Pankaj Kumar JAIN

Anand PANDEY

Serge MONTENS

Project Manager/Lead Structural Engineer Systra MVA Consulting (India) Pvt. Ltd. Faridabad, India pkjain@systra.com

Director-Infrastructure SYSTRA MVA Consulting (India) Pvt. Ltd. Faridabad, India apandey@systra.com

Respons. Pole/Secteur/Service SYSTRA Paris, France smontens@systra.com

Pankaj Kumar Jain Received his Bachelor’s Degree in Civil Engineering from M.B.M. Engineering College, Jodhpur and Master’s Degree in Structural Engineering from Delhi College of Engineering, Delhi. He joined SYSTRA India in 2006 as Structural Engineer and Presently he is working as Project Manager/Lead Structural Engineer in SYSTRA India

Anand Pandey Received his Bachelor’s Degree in Civil Engineering Honours from IIT Roorkee and Master’s Degree in Structural Engineering from IIT Roorkee. He joined Boarder Road Organization in 1992 as Assistant Executive Engineer and presently he is working as DirectorInfrastructure in SYSTRA India

Serge Montens Received his Bachelor’s degree in economics, University Paris, 1977.Engineer, Special School Public Works, Paris, 1978. Engineer, CHEBAP, Paris, 1979. He is presently chief of the bridge division, with in the Civil Engineering Department of SYSTRA company in France. He is an expert in bridge design and structural dynamics. He is a Member French Association for Civil Engineering, French Association for Earthquake Engineering and International Association for Bridges and Structural Engineering

Summary SYSTRA has developed a system of U shape prestressed concrete decks for metro viaducts. The many advantages of this system are presented, along with the design concept: system integration and deck design, architectural design, deck computation. Many examples of applications in India are described.

Introduction

With the huge development of large cities worldwide, the need for efficient grade separated metro systems is increasing significantly. Metro owners often decide to put transit systems on long viaducts. However, environmental considerations require that

Volume 47 │ Number 4 │ December, 2017

great attention is paid to landscaping, architectural appearance, and noise impact for these infrastructures, which are built amidst an urban environment. The U shape prestressed concrete deck described below, developed by SYSTRA, is an economically efficient answer to requirement of urban development. This deck has many advantages compared to conventional box-girder type deck.

Advantages of concrete decks

U-shape

prestressed

2.1 General presentation This concept was developed in the 1990 in order to get a metro viaduct type that can blend in urban

The Bridge and Structural Engineer

environment. The U shape prestressed concrete deck is composed of two side beams and a bottom slab between them. The tracks are supported by the bottom slab, which can be transversely prestressed if necessary. The longitudinal prestressing is located in the bottom slab and in the webs. The deck can carry either one track (section called “small U”), or two tracks (section called “wide U”). This deck section is generally used for simple spans viaducts, with spans below 36 m. But continuous spans have already been built also, and the same shape can be completed with other structural elements in order to achieve much longer spans. 2.2 Lowering the line longitudinal profile The longitudinal profile of metro lines on viaducts is mostly governed by the road traffic clearance below the deck. The rail level is derived from this clearance, plus the depth between the deck bottom and the rail level. With a box-girder section, the deck depth is about 1.25 m to 2.00 m. With a U shape section, the structural depth below the track is equal to the bottom slab thickness. Therefore it is possible to lower the project rail level by 1 m to 1.80 m. This has several consequences: - the level of the applied horizontal forces is lowered, therefore reduced bending moments are applied to piers and foundations, and thereby this decreases their quantities and cost,

stations can also be reduced. This of course reduces their cost, since wind and seismic forces applied to the piers and foundations are reduced. This is also beneficial for the passengers who have to climb from the road level to the platform level. In some cases, this allows elimination of escalators and their associated investment and maintenance costs. 2.4 Integration of the stations in the typical viaduct design In the station, the tracks can be supported by the same deck cross section as for the typical viaduct. The typical ‘U’ passes through the station and its construction can be independent from the station construction. This allows standardisation of the deck section all along the line, stations included. The remaining components (roof, ticketing room, technical rooms, etc.) of the station can be built out of the project construction critical path. 2.5 Full integration of the system The U shape structure was first conceived by a system wide multidisciplinary value engineering team of Systra experts. It is a value engineered structure as it integrates economically all components of the system in a very convenient and practical way, considering both installation and maintenance. This feature of the U shape section is covered by a SYSTRA international patent.

- the visual impact of the viaduct in the urban landscape is reduced. This minimized visual impact is probably one of the main advantages of U-shape prestressed concrete decks.

Fig. 2 : Example of full Integration of the System

2.6 Economical design in term of material quantities

Fig. 1 : Reduction of Visual Impact

2.3 Lowering the stations level Due to the above mentioned lowering of the longitudinal profile of the line, the level of the

The Bridge and Structural Engineer

The deck concrete section is reduced compared to a typical box-girder section. Furthermore, thanks to the lowering of the line longitudinal profile, the bending moments applied to the piers and foundations are reduced, and therefore the material quantities for the piers and foundations are also reduced, both for the typical viaduct and for the stations. The global reduction of quantities and cost is between 15 and Volume 47 │ Number 4 │ December, 2017

30 %. So, this reduces environmental impacts of the construction, such as CO2 emissions and energy consumption. 2.7 Use of side beams for various purposes The side beams are used for structural purpose, but they can also serve for three other functions. They are used as noise barriers. They are also capable of retaining the train on the bridge in case of derailment or severe earthquake, which is not the case for a standard noise barrier. Tests performed for Guangzhou metro have demonstrated this capability. Then, the top flange of the side beams can be used as emergency and/or maintenance walkways, because they are located approximately at the level of the train car floor.

Design concepts

3.1 System integration and deck design For all the possible transportation systems of the metro type, the system components can be quite different. For these various systems, the number of each type of cable is different (high voltage, mean voltage, low voltage, signalling, telecommunication, centralized control), which means that it is not possible to place them systematically at the same location of the deck transverse section for every project. The main characteristics of the rolling stock itself, influence also the deck design. The horizontal clearance obviously impacts the deck width, and then the thickness of the bottom slab which spans transversely between the webs. In the case of curved alignment, it is possible, either to make the deck follow exactly the curve, or to make straight spans with an increased width. The first solution is generally used when the span is made from assembled precast segments, whether the second one is generally used in case of fully precast spans, although it is also possible to build curved fully precast spans. The height of the deck section is generally such that the top of the top flanges is at the same level as the train car floor. The maximum cant value influences the deck width, because it has to be considered together with the clearance. 3.2 Architectural design By its intrusive character in the urban landscape, a metro viaduct needs a great architectural attention. Each project has to be adapted to the site. This must be understood in terms of both urban environment, and cultural environment. 86

Volume 47 │ Number 4 │ December, 2017

3.3 Deck computation In some aspects, the mechanical behaviour of a U shape section is different from what it is for a conventional box-girder. The U shape section is an open section. So, the torsion due to unsymmetrical live load for a deck supporting two tracks, or due to dead loads and live loads in case of a curved alignment, is carried by a combination of Saint-Venant torsion (like in a conventional box-girder) and differential bending of the two webs. The proportion of the two behaviours depends on many parameters, among them the torsion restraint of the webs above the supports. Another characteristic of the U shape section is that the torsion centre of the section is located outside from the section, below the section. It is necessary to take this into account when studying the general torsion due to horizontal transverse loads, such as earthquake or centrifugal force. For these reasons, a careful threedimensional analysis using a finite elements model is required. 3.4 Construction methods For long viaducts, it is obviously interesting to use precasting. In the case of a single track section (“small U”), and moderate span lengths (up to 30m), the weight of a complete span is much reduced (below 220 t), so that full spans can be precast. In this case, longitudinal pre-tensioning is generally provided, since it is more economical than post-tensioning. Two spans, i.e. four precast U beams, can be assembled per day (generally during the night), using two cranes (average 1 hour per U girder including moving and adjustments of the cranes). Alternatively, precast spans can be transported at the deck level and erected with an overhead launching girder, particularly if the transport of the elements at grade is a problem (extremely congested cities). In the case of a two tracks section (“wide U”), the section can be built with match-cast precast segments. The typical segments can be precast with the long bench method, while the pier segments are precast with the short cell method. Two typical segments can be built per day for each long bench. The assembling of the glued segments is generally performed using a self-launching steel gantry beam. The segments are brought either by truck below the gantry beam, or on the already built viaduct. The assembling work is easier than with a conventional box-girder, because of the better work conditions of an open U section. The construction cycle is two to three days per span.

The Bridge and Structural Engineer

Examples in India

4.1 Delhi Metro Line n° 3 The elevated portion of Line n° 3 stretches over 21.68 km. It includes 21 elevated stations. The architect for the viaduct was B+M Architecture. The U spans carry two tracks of the mass-transit system and are mainly designed as per Indian structural rules. The six standard spans include a 25m span (common span length) and a 31m span (longer span length). The deck depth is 2.15 m, and the deck width is 10.75 m. The segments are not prestressed transversally except the pier segments. In elevation, the top flange level is adjusted for passenger emergency evacuation. The same deck is used in stations. Station pier heads include a prestressed cross-girder that carries the platforms (supported by steel box girders) in addition to the station viaduct. A load test was successfully performed on a full-scale test span 25 m long. Construction schedule was tight, with only 20 months to complete the spans assembling of 22 km. Four gantries have been used.

Fig. 4 : Pragati Maidan Extradosed Bridge (Delhi)

4.2 Airport Metro Express Line - Delhi Metro The elevated portion of Airport Line stretches over 7.22 km. It includes 1 elevated station. The twin - U shape structures carry two tracks of the masstransit system and are mainly designed as per Indian structural rules. The standard spans include a 15m span to 25m span (common span length). The deck depth is 1.85 m, and the deck width is 10.09 m. The U-girder are pretension precast full span type. A post-tensioned precast pier cap supports twin U-girder viaduct. A load test was successfully performed on a full-scale test span 25 m long. Construction schedule was tight, with only 24 months to complete the spans assembling of 7.22 km.

Fig. 3 : Delhi Metro Line n° 3 – General view

The Pragati Maidan extradosed bridge on the Delhi metro line 3 extension, has a central span of 93 m. The deck has the same U shape as the typical spans, but with thickened webs to place extradosed cables anchorages. The extradosed cables are covered by a concrete beam that allows considering them as internal prestressing cables. This beam increases also the stiffness of the main span. Thanks to the use of precast segments and only typical construction materials, an innovative extradosed bridge was built in less than one year. This bridge is the first of its kind to be erected using the cantilever construction method, and it is the first railway extradosed bridge ever built in India.

The Bridge and Structural Engineer

Fig. 5 : Delhi Airport Line – General view

4.3 Delhi Metro Faridabad extension The elevated portion of Faridabad extension stretches over 16.06 km. It includes 11 elevated stations. The twin - U shape structures carry two tracks of the masstransit system and are mainly designed as per Indian structural rules. The standard spans include a 15m span to 27m span (common span length). The deck depth is 1.87 m, and the deck width is 10.15 m. Volume 47 │ Number 4 │ December, 2017

The U-girder are pretension precast full span type. A post-tensioned precast pier cap supports twin U-Girder viaduct.

The U-girder are pretension precast full span type. A post-tensioned precast pier cap supports twin U-girder viaduct.

Some load test was successfully performed for serviceability loads for span 27 m, 23 m & 20 m long U-girder.

Some load test was successfully performed for serviceability loads for span 27m, 23m & 20m long U-girder.

Fig. 6 : Delhi Metro Faridabad Extension – General view

Fig. 8 : Noida Greater Noida Metro – General view

The Depot Line has been constructed for connection to main viaduct & depot. It has a structure of U-girder & I-girder with cast in-situ portal & foundation. Extended pier cap is used for crossing road junction with standard U-girder span. Extended pier cap is post-tensioned cast in-situ type 13.5 m long pier cap. Two no. of 40 m span is used for crossing Nallah & Drain. It is steel composite girder type deck. Cross over/turn out spans have a superstructure made of post-tensioned precast I-girders.

The Depot Line has been constructed for connection to main viaduct & depot. It has a structure of U-girder & I-girder with cast in-situ portal & foundation. Cross over / turn out span have a superstructure made of post-tensioned precast I-girder. 4.5 Mumbai Metro Line 1 – Versova-AndheriGhatkopar The elevated portion of Mumbai metro Line 1 stretches over 11.25 km. It includes 9 elevated stations. The twin - U shape structures carry two tracks of the masstransit system and are mainly designed as per Indian structural rules. The standard spans include a 15 m span to 30 m span (common span length). The deck depth is 1.87 m, and the deck width is 10.15 m. The U-girder are pretension precast full span type. A post-tensioned precast pier cap supports twin U-girder viaduct. Some load test was successfully performed for serviceability loads for span 30 m long U-girder.

Fig. 7 : Delhi Metro Faridabad Extension - Extended Pier Caps

4.4 Noida Greater Noida Metro – Sector-71 Noida Delta-1 Greater Noida The elevated portion of Line stretches over 29.7 km. It includes 22 elevated stations. The twin - U shape structures carry two tracks of the mass-transit system and are mainly designed as per Indian structural rules. The standard spans include a 11m span to 27m span (common span length). The deck depth is 1.87 m, and the deck width is 10.15 m. 88

Volume 47 │ Number 4 │ December, 2017

Fig. 9 : Mumbai Metro Line 1 – General view

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

DMRC Airport Metro Express Line

Delhi Metro Rail Corporation (DMRC)

Client

Kanpur Metro

Double U - Full Lucknow Metro Span Precast Rail Corporation (LMRC)

Chennai Metro Rail (CMRL)

Double U - Full Lucknow Metro Span Precast Rail Corporation (LMRC)

Lucknow Metro

Chennai Metro Double U - Full Span Precast

Double U - Full Delhi Metro Span Precast Rail Corporation (DMRC)

Faridabad + Extn - India

Mumbai Metro One Pvt Ltd (MMOPL)

Double U - Full Delhi Metro Span Precast Rail Corporation (DMRC)

Big U Segmental

Superstructure Type

Mumbai Line 1 Double U - Full - VAG Corridor Span Precast

DMRC Dwarka Line 3

Project

S. No.

Ongoing - Tender Completed

Completed

Design Status

15.16

6.50

18.72

16.06

11.25

7.22

Elevated Section Length (km)

150

140

150

167

134

16 T

17 T

16 T

17 T

Standard Standard Axle Span Span Load (m) Weight (t)

Details of U-Shape Deck Metro Viaduct Projects in India

Standard Gauge 1435 mm

Broad Gauge 1676 mm

Gauge

Tender Design Completed + Construction yet to start

Design Completed + Construction yet to start

Design Completed & Priority 1 of 8.5 Km is operational and in Priority 2 of 10 km - Construction going on

12.5 km Operational + for Remaining 3.5 km (Design Completed + Construction Completed)

Operational

Remark

Volume 47 â&#x201D;&#x201A; Number 4 â&#x201D;&#x201A; December, 2017

The Bridge and Structural Engineer

Mumbai Metro Double U - Full Delhi Metro Line 7 Span Precast Rail Corporation (DMRC) / Mumbai Metropolitan Mumbai Metro Double U - Full Region Line 2A Span Precast Development Authority (MMRDA) Mumbai Metro Double U - Full Line 2B Span Precast

Mumbai Metro Double U - Full Line 4 Span Precast

Ghaziabad Double U - Full Delhi Metro Metro Span Precast Rail Corporation Superstructure (DMRC)

Client

Double U - Full Delhi Metro Span Precast Rail Corporation (DMRC) / Noida Metro Rail Corporation (NMRC)

Superstructure Type

Noida Greater Noida (NGN)

Project

S. No.

Completed

32.50

23.70

16.50

Completed

9.50

29.70

Elevated Section Length (km)

Completed

Design Status

145

162

150

17 T

16 T

Standard Standard Axle Span Span Load (m) Weight (t)

Standard Gauge 1435mm

Gauge

Design Completed + Construction yet to start

Design Completed + Construction going on

Design Completed + Construction Completed

Remark

The Depot Line has been constructed for connection to main viaduct & depot. It has a structure of U-girder & I-girder with cast in-situ portal & foundation. Cross over/turn out span have a superstructure made of post-tensioned precast I-girder. Extended pier cap is used for crossing road junction with standard U-girder span. Extended pier cap is post-tensioned cast-in-situ type 18.0 m long pier cap. The WEH bridge is an extradosed bridge with a U-shape prestressed concrete deck which carries both tracks. The main span is 86m. The pylons are made from laterally inclined reinforced concrete shafts. It was designed in order to cross a busy crossroad above a highway bridge.

Fig. 10 : Mumbai Metro Line 1 - WEH Extradosed Bridge

Conclusion

Starting from 1992, more than 300 km of U shape precast prestressed concrete decks for metro viaducts have been successfully designed and built, and are in operation. U-shape prestressed concrete deck is an interesting concept for metro viaducts. It is very cost efficient. Due to the lowering of the line longitudinal profile and stations level, it facilitates blending of the structure in an urban environment. This concept has demonstrated its interest in terms of functionality, economy, and speed of construction. It is a good example of sustainable engineering.

The Bridge and Structural Engineer

References

Dutoit D., Montens S., Chuniaud J.-C., Arnaud P., “U shape Prestressed Concrete Decks for LRT/MRT Viaducts”, IABSE Symposium, Shanghai, 2004. 2. Gauthier Y., Montens S., “U Shape Prestressed Concrete Decks for LRT/MRT Viaducts”, fib Symposium, Naples, 2006. 3. Gauthier Y., Montens S., Kataria R., Touat A., “Pushing the Limits of U Shaped Viaducts”, IABSE Symposium, Weimar, 2007. 4. Gauthier Y., Montens S., Arnaud P., Paineau T., “U Shape Viaducts for Dubai Metro”, fib Symposium, Amsterdam, 2008. 5. Singh M., Kataria R., Mhedden A., Mohammad S., Bajpai P., “Full Span Precast Pre-Tensioned Decks: the Future of Elevated Urban Metro Viaducts”, IIBE Delhi, 2008. 6. Duclos T., “Behaviour of the Flanges of U Shape Girders under Compressive Loads”, fib Symposium, London, 2009. 7. Montens S., Moine P., “150 km of U Shape Prestressed Concrete Decks for LRT Viaducts”, Railway Railway Engineering, London, 2009. 8. Vion P., Joing J., “Fabrication and Erection of U-trough Section Bridges”, Structural Engineering International, 4/2011. 9. Cheikh Mhamed A., Kashani M., Stack M., Montens S., “Construction of a Metro Line Serving the Holy Sites of Mekkah in the Kingdom of Saudi Arabia”, IIBE Mumbai, 2012. 10. Duclos T., Ketfi M., Amami H., Aubazac C., “A Signature Bridge in A Congested Urban Area”, IABSE Symposium, Seoul, 2012. 11. Akraa M., Paumier A., Pruvost E., Montens S., « U Concept Viaduct – Precast Segmental Application to Uijeongbu LRT Project (SouthKorea) », ICUTS, Paris, 2013.

Volume 47 │ Number 4 │ December, 2017

Indian National Group of the IABSE Office Bearers and Managing Committee - 2017 Chairman 1.

Shri D.O. Tawade, Chairman, ING-IABSE & Member (Technical), National Highways Authority of India

Vice-Chairmen

Members of the Executive Committee 13. Shri N.K. Sinha, Former Director General (Road Development) & Special Secretary 14. Shri A.K. Banerjee, Former Member (Technical), NHAI

Shri B.N. Singh, Additional Director General, (I/C) Ministry of Road Transport and Highways

15. Shri G. Sharan, Former Director General (Road Development) & Special Secretary

Shri Alok Bhowmick, Managing Director, B&S Engineering Consultants Pvt. Ltd.,

16. Shri A.V. Sinha, Former Director General (Road Development) & Special Secretary

Shri A.K.S. Chauhan, C.O.O., GR Infraprojects Ltd.,

Shri Vinay Gupta, Chief Executive Officer, Tandon Consultants Pvt. Ltd.,

Honorary Treasurer 6.

The Director General (Road Development) & Special Secretary to the Government of India, Ministry of Road Transport and Highways

Honorary Members 7.

Shri Ninan Koshi, Honorary Member, IABSE & Former Director General (Road Development) & Addl. Secretary

Prof. S.S. Chakraborty, Honorary Member & Past Vice-President, IABSE

Persons represented ING on the Executive Committee and Technical Committee of the IABSE 9.

Dr. Harshavardhan Subbarao, Vice President & Member, Technical Committee of IABSE & Chairman and Managing Director, Construma Consultancy Pvt. Ltd., Past Member of the Executive Committee and Technical Committee of IABSE

10. Prof. S.S. Chakraborty, Past Vice-President, IABSE 11. Dr. B.C. Roy, Past Vice President & Member, Technical Committee, IABSE

17. Shri R.P. Indoria, Former Director General (Road Development) & Special Secretary 18. Dr. Lakshmi Parameswaran, Chief Scientist, Bridges & Structures Div., CSIR-Central Road Research Institute 19. Shri Ashwinikumar B. Thakur, Group Engineer, Atkins India 20. Shri Sarvagya Kumar Srivastava, Engineer-inChief (Projects), Govt of Delhi 21. Dr. Mahesh Kumar, Engineer Member, Delhi Development Authority 22. Shri R.K. Jaigopal, Consultant, Concrete Structural Forensic Consultant 23. Prof Mahesh Tandon, Managing Director, Tandon Consultants Pvt. Ltd.

Past Secretary of the Society, for a Period of Two Years, after they Vacate their Secretaryship 24. Shri R.K. Pandey, Member (Projects), National Highways Authority of India

Secretariat 25. Shri I.K. Pandey, Additional Director General (RD), Ministry of Road Transport and Highways,

Honorary Secretary

26. Shri Ashish Asati, General Manager, National Highways Authority of India

12. Shri I.K. Pandey, Additional Director General (RD), Ministry of Road Transport & Highways

26. Shri K.B. Sharma, Under Secretary, Indian National Group of the IABSE

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

MEMBERS OF THE MANAGING COMMITTEE – 2017 Rule-9 (a): A representative of the Union Ministry of Road Transport and Highways 1.

Shri Manoj Kumar, Director General (Road Development) & Special Secretary

Rule-9 (b): A representative each of the Union Ministries/Central Government Departments making annual contribution towards the funds of the Indian National Group of IABSE as determined by the Executive Committee from time to time 2.

CPWD - nomination awaited

Shri D.O. Tawade, Member (Technical), National Highways Authority of India

Ministry of Railways - nomination awaited

20. Govt of Maharashtra – nomination awaited 21. Shri K Radhakumar Singh, Commissioner (Works), Govt of Manipur 22. Shri PR Marwein, Chief Engineer (Standards), PWD (Roads) Govt of Meghalaya 23. Shri Lalmuankima Henry, Chief Engineer (Buildings), Govt of Mizoram 24. Govt of Nagaland – nomination awaited 25. Govt of Orissa – nomination awaited 26. Shri Anil Kumar Gupta, Suptd Engineer, Central Works Circle, Govt of Punjab 27. Govt of Sikkim – nomination awaited 28. Govt of Tamil Nadu – nomination awaited

Rule-9 (c): A representative each of the State Public Works Departments/Union Territories making annual contribution towards the funds of the Indian National Group of IABSE as determined by the Executive Committee from time to time

29. Govt of Tripura – nomination awaited

Govt of Andhra Pradesh – nomination awaited

Govt of Arunachal Pradesh – nomination awaited

33. Union Territory Chandigarh – nomination awaited

Govt of Assam – nomination awaited

Govt of Bihar – nomination awaited

Govt of Chattisgarh – nomination awaited

10. Govt of Delhi – nomination awaited 11. Govt of Goa – nomination awaited 12. Govt of Gujarat – nomination awaited 13. Govt of Haryana – nomination awaited 14. Govt of Himachal Pradesh – nomination awaited 15. Govt of Jammu & Kashmir – nomination awaited 16. Govt of Jharkhand – nomination awaited

30. Govt of Uttar Pradesh – nomination awaited 31. Govt of Uttarakhand – nomination awaited 32. Govt of West Bengal – nomination awaited

Rule-9 (d): A representative each of the Collective Members making annual contribution towards the funds of the Indian National Group of IABSE as determined by the Executive Committee from time to time 34. Major V.C. Verma, Director (Mktg), Oriental Structural Engineers Pvt. Rule-9 (e): Ten representatives of Individual and Collective Members 35. Shri A.K. Banerjee, Former Member (Technical), NHAI 36. Shri G. Sharan, Former DG (RD) & Special Secretary

17. Govt of Karnataka – nomination awaited

37. Shri A.V. Sinha, Former DG (RD) & Special Secretary

18. Shri S Saju, Joint Director (EE) & Highways Design, DRIQ, Govt of Kerala

38. Shri R.P. Indoria, Former DG (RD) & Special Secretary

19. Shri LD Dube, Chief Engineer (Bridge Zone), Govt of Madhya Pradesh

39. Shri Rakesh Kapoor, General Manager, Holtech Consulting Pvt Ltd

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

40. Shri Ashwinikumar B Thakur, Group Engineer, Atkins India

53. Shri Bageshwar Prasad, CEO (Delhi Region), Construma Consultancy Pvt. Ltd.

41. Prof Mahesh Tandon, Managing Director, Tandon Consultants Pvt. Ltd.

54. Shri G.L. Verma, Proprietor, Engineering and Planning Consultants

42. Shri R.S. Mahalaha, Advisor, ITL

Rule-9 (i): Honorary Treasurer of the Indian National Group of IABSE

43. Shri R.K. Jaigopal, Consultant, Concrete Structural Forensic Concrete 44. Shri Inderjit Singh Ghai, CEO, Consulting Engineers Associates Rule-9 (f): Four representatives of Bridge and Structural Engineering Firms 45. Shri V.N. Heggade, Director & President (Engineering), Gammon Engineers & Contractors 46. Shri A.K.S. Chauhan, C.O.O., GR Infraprojects Ltd. 47. Shri Shishir Bansal, Chief Project Manager, Delhi Tourism & Transportation Development Corporation 48. Shri S.P. Singla, Managing Director, SP Singla Constructions Ltd Rule-9 (g): Two representatives of the Engineering Colleges / Technical Institutes / Universities / Research Institutes 49. Dr Lakshmi Parameswaran, Chief Scientist, Bridges & Structures Div., CSIR – Central Road Research Institute 50. Shri VL Patankar, Former Director, Indian Academy of Highway Engineers Rule-9 (h): Four representatives Engineering Firms

Consulting

51. Shri N.K. Sinha, President, ICT Pvt. Ltd. 52. Shri Alok Bhowmick, Managing Director, B&S Engineering Consultants Pvt. Ltd.

Volume 47 │ Number 4 │ December, 2017

55. The Director General (Road Development) & Special Secretary to the Govt of India Rule-9 (j): Past-Chairman of the Society, for a period of three years, after they vacate their Chairmanship Rule-9 (k): Secretary of the Indian National Group of IABSE 56. Shri I.K. Pandey Rule-9 (l): Persons who have been awarded Honorary Membership of the Parent Body 57. Shri Ninan Koshi 58. Prof S.S. Chakraborty Rule-9 (m): Persons represented ING on the Executive Committee and Technical Committee of the IABSE 59. Dr. Harshavardhan Subbarao Rule-9 (n): Past Members of the Executive Committee and Technical Committee of the IABSE 60. Prof S.S. Chakraborty 61. Dr. B.C. Roy Rule-9 (o): Past Secretary of the Society, for a period of three years, after they vacate their Secretaryship 62. Shri R.K. Pandey

The Bridge and Structural Engineer

HIGHLIGHTS OF THE ING-IABSE WORKSHOP ON “APPLICATION OF IRC:112 - CODE OF PRACTICE FOR CONCRETE ROAD BRIDGES” HELD AT RAIPUR ON 13TH AND 14TH OCTOBER, 2017 The Indian National Group of the IABSE in cooperation with Public Works Department, Govt of Chhatisgarh successfully organised a two day Workshop on “Application of IRC:112 - Code of Practice for Concrete Road Bridges” at Raipur on 13th and 14th October 2017. The Workshop was well attended by more than 250 delegates from various Govt Departments as well as other private and public organizations.

of Chhattisgarh, Shri Anil Rai, Managing Director, CGRDC & OSD, Govt of Chhattisgarh, Prof Mahesh Tandon, Chairman, Scientific Committee, Shri D.K. Pradhan, Engineer-in-Chief, PWD Govt of Chattisgarh and Shri K.K. Pipri, Chief Engineer, National Highways, PWD, Govt of Chhattisgarh as well as other dignitaries.

The aim of the workshop was to provide a unique opportunity to the Engineers of the State PWD, the practicing engineers and the students to interact with experts for dissemination of knowledge and experiences relating to the latest techniques in design of bridges and other structures using the “Code of Practice for Concrete Road Bridges IRC:112: 2011”. Participation of delegates in floor intervention and discussions was very encouraging.

Shri Subodh Kumar Singh, Principal Secretary, Public Works Department, Govt of Chhattisgarh extended warm welcome to the participants of the Workshop. Shri I.K.Pandey, Secretary, ING-IABSE delivered his address during the Inauguration and spoke on the idea behind organizing this Workshop at Raipur. Shri K.K.Pipri, Chief Engineer (National Highways), Govt of Chhattisgarh, proposed Vote of Thanks. On this occasion, Chief Guest of the Seminar also released the Souvenir.

The Workshop was inaugurated by Shri Rajesh Munat, Hon’ble Minister of Public Works Department, Chhattisgarh by lighting the traditional lamp in the presence of Shri I.K.Pandey, Secretary, ING-IABSE, Shri Subodh Kumar Singh, Principal Secretary, Public Works Department, Govt

The Workshop on “ Application of IRC:112 - Code of Practice for Concrete Road Bridges” was addressed by the following eminent experts covering the following themes who were either involved in the preparation of the Code or have used in extensively since its publication.

13th October, 2017 Session-1 Presentation-1

Overview & Scope – Mahesh Tandon

Presentation-2

Basis of Design – Alok Bhowmick

Presentation-3

Actions and their Combinations with Worked Examples – Alok Bhowmick

Presentation-4

Material Properties and their Design Values – Worked Example for Stress-Strain Curves and Creep Calculations – Mahesh Tandon

Discussions for Session-1 Session-2 Presentation-5

Analysis Methods: (With worked Examples on Modelling) – Vinay Gupta

Presentation-6

Serviceability Limit State with Worked Examples – Vinay Gupta

Presentation-7

ULS of Linear Elements for Bending and Axial Forces With Worked Example – Umesh Rajeshirke

Presentation-8

Complete Worked Example IRC:112 – Umesh Rajeshirke

The Bridge and Structural Engineer

for

RCC

Superstructure

Design

Volume 47 │ Number 4 │ December, 2017

with

Presentation-9

Worked Example for Well Foundation Design with IRC 112 – Harpreet Singh

Discussions for Session-2 14th October, 2017 Session-3 Presentation-10

ULS of Shear, Punching Shear and Torsion – JS Pahuja

Presentation-11

ULS of Induced Deformations – Worked Examples – VN Heggade

Presentation-12

Durability and Deterioration of Concrete Structures – Vinay Gupta

Presentation-13

Prestressing Systems – Worked Examples – Alok Bhowmick

Discussions for Session-3 Session-4 Presentation-14

Detailing Requirements Including Ductility Detailing – Alok Bhowmick

Presentation-15

Worked Example of a PSC Segmental Box Girder with IRC 112 – Harpreet Singh

Discussions for Session-4 The concluding remarks of the Workshop were presented by Prof Mahesh Tandon, Chairman, Scientific Committee on 14th October 2017. He expressed the hope that the outcome of the Workshop would have enriched the delegates. The delegates who attended the Workshop mentioned that the subject matter of the Workshop is very timely. Shri K.K.Pipri, Chief Engineer (NH), Govt of Chhattisgarh proposed Vote of Thanks. A light music with dinner was organized in the evening of 13th October 2017 for the participants who rejoiced the evening. The Workshop was a great success.

Photo-1 A view of the Dais during the Inaugural Function

Volume 47 │ Number 4 │ December, 2017

Photo-2 Shri Rajesh Munat, Hon’ble Minister, Public Works Department lighting the traditional Inaugural Lamp along with high dignitaries

The Bridge and Structural Engineer

Photo-3 Shri Rajesh Munat, Hon’ble Minister, Public Works Department, Chhattisgarh Delivering his Address during Inaugural Function

Photo-4 A view of the Audience during the Inauguration

Photo-5 Group Photograph of the Workshop Speakers

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

HIGHLIGHTS OF THE INTERNATIONAL SEMINAR ON “REPAIR, REHABILITATION AND RETROFITTING OF BRIDGES AND STRUCTURES” HELD AT JAIPUR (RAJASTHAN) ON 15TH AND 16TH DECEMBER, 2017

The Indian National Group of the IABSE in Association with Indian Roads Congress and Public Works Department, Government of Rajasthan successfully organized a two day International Seminar on “Repair, Rehabilitation and Retrofitting of Bridges and Structures” at Jaipur on 15th and 16th December, 2017. The Seminar was well attended by more than 250 delegates from various Govt Departments as well as other private and public organizations. This Seminar covered all aspects of repair/ rehabilitation and retrofitting of bridges and structures, provided an insight into the development of new techniques of repairs, new materials and discussed the available methods both in India and abroad to carry out repair and rehabilitation and retrofitting of structures to enhance their serviceability during the design life. During this two day Seminar, several Indian and Foreign experts in the field of Repair, Rehabilitation and Retrofitting of bridges and structures made presentations on the different aspects of bridge repair and rehabilitation and presented some interesting case studies for benefit of the participants from Rajasthan, Public Works Department and other state PWDs and Consultants associated with the topic. Participation of delegates in floor intervention and discussions was very encouraging. The Seminar was inaugurated on Friday, the 15th December, 2017 by Shri Alok, IAS, Principal Secretary, Government of Rajasthan by lighting the traditional lamp. Other dignitaries, Shri D.O. Tawade, Chairman, ING-IABSE and Member (Technical, National Highways Authority of India, Shri Manoj Kumar, Director General (Road Development) and Special Secretary, Ministry of

Volume 47 │ Number 4 │ December, 2017

Road Transport and Highways, Dr Harshavardhan Subbarao, Vice President of IABES, Shri A.K. Banerjee, Chairman, Scientific Committee, Shri S.K. Nirmal, Secretary General, IRC and Shri Anil Garg, Chairman, Local Organising Committee and Chief Engineer (National Highways), Government of Rajasthan also graced the occasion. Hon’ble Union Minister of Road, Transport, Highways, Shipping and Water Resources, River Development and Ganga Rejuvenation, Shri Nitin Gadkari was the Chief Guest in Special Plenary Session held on Friday, 15th December 2017 evening. Smt Vasundhra Raje, Chief Minister of Rajasthan presided over the function. Hon’ble PWD Minister, Shri Yoonus Khan was Guest of Honour for the Session. Shri Manoj Kumar, Director General (Road Development) and Special Secretary, Ministry of Road Transport and Highways, Shri Alok, IAS, Principal Secretary, Government of Rajasthan, Shri D.O. Tawade, Chairman, ING-IABSE and Member (Technical), National Highways Authority of India, Shri S.K. Nirmal, Secretary General, IRC, Shri C.L. Verma, Chief Engineer & Additional Secretary, PWD, Government of Rajasthan, Shri Anil Garg, Chairman, Local Organising Committee and Chief Engineer (National Highways), Government of Rajasthan also shared the dais for special Plenary Session. Shri D.O. Tawade, Chairman, ING-IABSE welcomed the Chief Minister and also spoke on the idea behind organizing this International Conference at Jaipur. The Seminar on “Repair, Rehabilitation and Retrofitting of Bridges and Structures” was addressed by the following eminent experts.

The Bridge and Structural Engineer

Friday, 15th December 2017 Session 1

“Structural Health Monitoring & Development of BMS

Session Chair

Shri A.D. Narain, Former DG(RD) & AS, MoRT&H

Session Co-Chair

Shri R.K. Pandey, Member (Projects), NHAI

Keynote Address – A.K. Banerjee, Chairman, Scientific Committee Presentation-1 Guiding and Targeting Investigations on Prestressed Structures –Marc Brouxel Presentation-2 Approach for Rehabilitation of Bridges within IBMS – Sachin Joshi Presentation-3 Structural Health Monitoring System of The Queensferry Crossing –Henrik Gjelstrup Presentation-4 Bridge Management in U.K with Case Studies –Saprava Bhattacharya Session - 2

Materials and Techniques for Repair, Rehabilitation & Retrofitting

Session Chair

Shri N.K. Sinha, Former DG(RD) & SS, MoRT&H

Session Co-Chair

Shri S.K. Nirmal, Secretary General, IRC

Presentation-5 Euro Code EN 1504 – Use of Composites and UHPFRC – Harshavardhan Subbarao Presentation-6 Innovative Construction Materials – Surendra K Manjrekar Presentation-7 New Materials and Techniques for Corrosion Protection of Bridges & Flyovers – M.K. Kamat Presentation-8 Use of New Materials – Samir Surlekar Session – 3A

Repair, Rehabilitation & Retrofitting and Case Studies

Session Chair

Shri G. Sharan, Former DG(RD) & SS, MoRT&H

Session Co-Chair

Shri Anil Garg, Chief Engineer (NH), PWD Rajasthan

Presentation-9 Repair, Upgrade and Strengthening of Structures – Fernando Guedes De Melo Presentation-10 Repair and Rehabilitation of Bridges in India and Abroad – Niels Bitsch Presentation-11 Modular Joints for Rehabilitation of McDonald’s Bridge at Halifax – Gianni Moor Saturday, 16th December, 2017 Session – 3B Repair, Rehabilitation & Retrofitting and Case Studies

Session Chair

Shri Manoj Kumar, DG(RD) & SS, MoRT&H

Session Co-Chair

Shri M.G. Maheshwari, MD, RSBCC Ltd

Presentation-12 Rehabilitation of Varsova Bridge across Vasai Creek – Dhananjay Bhide Presentation-13 Repair and Retrofitting of a Bow String Arch Bridge in Goa – Alok Bhowmick Presentation-14 Minimally Invasive Maintenance, Refurbishment and Replacement of Expansion Joints – Pascal Savioz Presentation-15 Rehabilitation of Sharavathy Bridge – P.G. Venkatram

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

Session – 3C Repair, Rehabilitation & Retrofitting and Case Studies

Session Chair

Session Co-Chair -

Shri A.K. Banerjee, Former Member (Tech), NHAI Shri C.L.Verma, CE & Addl Secretary, PWD Rajasthan

Presentation-16 Anti Seismic Devices for Bridges and Structures – Enzo Lu Presentation-17 Review of Rapid Construction Techniques for Rehabilitation & Retrofitting of Bridges – Rama Raju P Presentation-18 Added pro value State in Replacement by State-of-the-Art Bridge Bearings and Expansion Joints Solutions – Peter Gunther Presentation-19 Rehabilitation of M.G. Setu at Patna – Umesh Rajeshirke Presentation-20 Rehabilitation of Narmada Bridge – Vinay Gupta Presentation-21 Rehabilitation of Steel Bridges and Case Study – N Bandyopadhyay Session - 4

Performance Monitoring post Repair/Rehabilitation

Session Chair

Shri R.P. Indoria, Former DG(RD) & SS, MoRT&H

Session Co-Chair

Shri A.K. Singh, Member (Projects),NHAI

Presentation-22 Structural Health Monitoring using Digital Imaging Technique – Janardhan Sundaram Presentation-23 Long Term Performance Lakshmy Parameswaran

Monitoring

Second

Thane

Creek

Bridge

–

Presentation-24 Life Cycle Monitoring and Maintenance of Bridges – Chinmoy Ghosh The concluding remarks of the Seminar were presented by Shri AK Banerjee, Chairman, Scientific Committee on 16th December 2017. He expressed the hope that the outcome of the Seminar would have enriched the delegates. The delegates who attended the Seminar mentioned that the subject matter of the Seminar is very timely. Shri Anil Garg, Chairman, Local Organising Committee and Chief Engineer (NH), PWD Rajasthan proposed Vote of Thanks. A light music with dinner was organized in the evening of Friday, the 15th December, for the participants who rejoiced the evening. The Seminar was a great success.

Photo-1 A view of the Dais during the Inaugural Function

100

Volume 47 │ Number 4 │ December, 2017

Photo-2 Shri Alok, IAS, Principal Secretary, Government of Rajasthan lighting the traditional Inaugural Lamp along with high dignitaries

The Bridge and Structural Engineer

Photo-3 Shri D.O. Tawade, Chairman, ING-IABSE Delivering his address

Photo-4 Shri Manoj Kumar, Director General (Road Development) & Special Secretary, Ministry of Road Transport & Highways Delivering his address

Photo-5 Shri Alok, IAS, Principal Secretary, Government of Rajasthan Delivering his address during Inaugural Function

Photo-6 A view of the audience during the Inauguration

Photo-7 Another view of the audience during the Inauguration

Photo-8 Hon’ble Union Minister Shri Nitin Gadkari addressing the gathering during the Special Plenary Session

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

101

Photo-9 Smt Vasundhra Raje, Hon’ble Chief Minister of Rajasthan addressing the gathering during the Special Plenary Session

Photo-10 A view of the Cultural Evening during the Seminar

Photo-11 Shri AK Banerjee, Chairman, Scientific Committee Delivering his address

Photo-12 Shri Yoonush Khan, Hon’ble PWD Minister of Govt of Rajasthan addressing the gathering

Photo-13 View of the audience during the Valedictory Session

Photo-14 Another view of the audience during the Valedictory Session

102

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

&URUHV ,15 DV RQ VW 'HF

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

103

With best compliments from:

CRAPHTS Consultants (I) Pvt. Ltd. (Infrastructure Consultants) (An ISO 9000 – 2015 Organization)

Spectrum of Services : ● ● ● ● ● ● ● ● ● ● ● ●

Traffic Engineering & Transportation Planning Highway Engineering Road Safety Policy, Road Safety Audits and Safety Engineering Construction Supervision & Corridor Management Transportation Master Plan for New Towns Terminal Area Planning, Intermodal Integration and Parking Studies Traffic Management and Control Transport Corridor Development Urban Planning & Design Architectural Planning Structural Engineering Environmental Planning

Address : 14/3, Mathura Road, Faridabad – 121 003, Haryana, Telephone – 91 – 129 – 2278813 /0129-2278813, Fax – 91 – 129 – 2256392, E-mail :info@craphts.com, www.craphts.com

104

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer

Volume 47 │ Number 4 │ December, 2017

105

106

Volume 47 │ Number 4 │ December, 2017

The Bridge and Structural Engineer