Steel Construction 01/2013 Free Sample Copy by Ernst & Sohn

Steel Construction

Volume 6 Februar 2013 ISSN 1867-0520

Design and Research

– Recent changes in U.S. connection design practice – Weld design and fabrication for RHS connections – Cyclic load behaviour of friction T-stub beam/column joints – Design model for composite beam–RC wall joints – Evaluation of RBS beams for moment frames in seismic areas – Thin-walled structural hollow section joints – Selecting materials for fastening screws – Monopile foundations for offshore wind turbines – Assembly of stadium roof structure – Aluminium/polycarbonate roof covering to stadium

01_SC_U1_Titelseite.indd 4

12.02.13 07:46

02_SC_U2_02_SC_U2_2-09.qxd 16.05.12 09:54 Seite 1

Form + Function = DETAN DETAN tension rod systems. Your solution for transparent design.

orm and function are perfectly combined in the DETAN tension rod system. Individual system solutions make it possible to realise even the most complicated designs and aesthetic details both indoors and outdoors.

anchor disk

locking nut seals (optional)

locking nut

tension rod

coupler with sail

connection plate to structure fork end pin circlip

DETAN tension rod structures can consist of one or all of the standard components shown in this example.

Simple With screw connections, no welding is required. Standard tools mean simple and convenient installation. Safe DETAN stands for reliable planning through certification, type testing and dimensioning software. A high load capacity allows safe use in a wide range of applications and gives more design freedom for aesthetic architecture.

of the tension rod system. The high strength of the tension rods makes them thin and saves on material use: cost optimisation and design in a single stroke. Many advantages with one result: HALFEN-DEHA provides safety, reliability and efficiency for you and your customers.

Efficient Materials such as carbon steel S 460 and cast steel for the fork ends have significantly increased the efficiency

HALFEN Engineering Support Tel.: 02173/970-90 35 www.halfen.de HALFEN GmbHGmbH 路 Liebigstr. 14 路 D-40764 Langenfeld 路 Tel.: +49 (0) 2173/970-0 路 www.halfen.com

02_SC_U2.indd 1

01.02.13 08:44

Content

The picture shows the fully assembled steel roof support structure of the football stadium in Gdan´sk. The stadium has a characteristic silhouette – its shape and the colours of the façade resemble a cut block of amber. The steel roof structure has a quasi-elliptical form, with a maximum diameter of 220 m and minimum diameter of 187 m. It is 38 m high and the roof girders extend 48 m over the grandstand below. The roof structure weights 7150 t and was assembled in 226 days (see pp. 54–60)

Steel Construction 1

Volume 6 February 2013, No. 1 ISSN 1867-0520 (print) ISSN 1867-0539 (online)

Editorial 11

Dan Dubina, Daniel Grecea 7th International Workshop on Connections in Steel Structures 2012 – Connections VII Articles

Charles J. Carter, Cynthia J. Duncan Recent changes in U.S. connection design practice

Matthew R. McFadden, Min Sun, Jeffrey A. Packer Weld design and fabrication for RHS connections

Massimo Latour, Vincenzo Piluso, Gianvittorio Rizzano Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

José Henriques, Luís Simões da Silva, Isabel Valente Design model for composite beam-to-reinforced concrete wall joints

Florea Dinu, Dan Dubina, Calin Neagu, Cristian Vulcu, Ioan Both, Sorin Herban Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

Ram Puthli, Jaap Wardenier, Andreas Lipp, Thomas Ummenhofer Thin-walled structural hollow section joints Reports

Wilhelm Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG www.ernst-und-sohn.de

Thomas Misiek, Saskia Käpplein, Detlef Ulbrich Selecting materials for fastening screws for metal members and sheeting

Rüdiger Scharff, Michael Siems Monopile foundations for offshore wind turbines – solutions for greater water depths

Jerzy Ziółko, Alojzy Les´niak Assembly of the steel roof structure for the football stadium in Gdan´sk

Dariusz Kowalski The aluminium and polycarbonate covering to the roof over the stadium in Gdan´sk

Journal for ECCS members

www.wileyonlinelibrary.com, the portal for Steel Construction online subscriptions

03_SC_Inhalt_1-13.indd 1

Regular Features 18 38 60 66 67

People Book Reviews News Announcements ECCS news

Products & Projects

01.02.13 10:08

Products & Projects

Solutions for an avoidable problem: Corrosion protection at expansion joints

footways, or it must be cut free ex post. However a later free cutting will almost always damage the corrosion protection of the edge beams.

It is an embarrassing problem, and it costs money: damages at the corrosion protection of expansion joints, which are already visible before the bridge is opened to the traffic. Maurer Söhne presents two solutions for avoidance: stainless steel or edge profile protection. “We have several years of experience with this issue. It is now upon the owners of the bridge or the concessionaries to request such a possible quality”, explains Rolf Kiy, who is responsible for the work group “Corrosion Protection” at Maurer Söhne.

Edge profile protection keeps the joint at the footway area free Maurer Söhne offers a solution for this problem: the edge profile protection. It consists of a special plastic cap which is being clamped onto the edge beams of the expansion joints. Like a formwork it keeps the space of the bituminous filler in the specified thickness. After the concreting this protection layer is being removed. The bituminous filler is free, and the corrosion protection remains undamaged. We would like to mention here that the responsibility of the protection of the corrosion

In Germany, the combined length of all expansion joints is about 5,000 km. The procurement costs make about 1 % of the total costs of a bridge. In the course of a maintenance check of over 2,000 bridges in the state of Bavaria, the owner which is the Southern Bavarian Expressway Administration noticed that less than 5 % of all damages fall into the category of bridge bearings and expansion joints. However, the major share of these damages is located in the upper part of the expansion joints. Experts detected several main reasons for such damages at the corrosion protection: – Unprotected passing of the expansion joints in construction phase or during the renovation of asphalt. – Corrosion protection being damaged during the reinforcement of the footway section by way of reinforcement bars. – Damage caused by the cut of the bituminous filler after concreting and asphalting – Dirt that has a paint erasing effect.

Fig. 1. Damages at the corrosion protection caused by ”free cutting“ of the expansion joint by way of a saw blade

The objective of Maurer Söhne is to increase the awareness for such types of damages and act preventively. There are 2 possible solutions available: – Hybrid profile with stainless steel head. – Maurer edge profile protection at the concreting phase of the footway section, which is connected with an on site created protection during construction phase.

Hybrid: Upper part in stainless steel The principle of the hybrid profile is a simple one: the upper part of the profile is being made of stainless steel, which employs a long term resistance against corrosion. The hybrid structure of the edge beam is being employed in order that the lower part of the profile which consists of mild steel can ensure an optimum welding connection with the anchorage in the bridge structure. Of course this welding seam has been structurally analysed and is fatigue tested as well as enjoys the General Approval. The electrolyte is being kept away with a side cover in order to prevent crack corrosion. Moreover the hybrid profile can be provided with upper rhombic plates for the purpose of reduction of noise emission.

Fig. 2. Hybrid profile after 5 years of service: looks like new, even under traffic

Already passed the real life test The first hybrid joints were installed in 2006 at the region of Ansbach and thereafter very thoroughly examined. After more than 5 years in service the expansion joints look like new. Thus we can say that this solution is a proven one.

Install safely against corrosion The second approach in terms of corrosion protection aims at the situation during installation. Generally it should be observed that the expansion joint shall be protected during installation, in particular when it comes to passings of job site vehicles. Along the edge of each expansion joint there is a bituminous filler, which is a gap that must be kept free when the expansion joint gets an asphalt connection or a concrete connection at the

Fig. 3. left: The Maurer edge profile protection prevents damages at the corrosion protection, optimises the fast and economic installation, secures expansion joints according RiZ Übe 1 and ZTV-ING and replaces improvised solutions at the job site right: So clean is the look when the Maurer edge profile protection kept the space free for the bituminous filler between concrete and expansion joint. Ex post free cutting is no longer required, and the corrosion protection was not damaged (© Maurer Söhne)

Steel Construction 6 (2013), No. 1

04_SC_A04-A08.indd 4

01.02.13 08:45

Products & Projects of the expansion joint during the construction phase is with the contractor, which must be made clear and ensured by the supervisor. Usually, the required protection covers during construction phase consist of a plastic foil and/or of wood.

Further Information: Maurer Söhne GmbH & Co. KG, Frankfurter Ring 193, 80807 München, Tel. +49 (0)89 – 32394-0, Fax +49 (0)89 – 32394-234, info@maurer-soehne.de, www.maurer-soehne.de

Software for Statics and Dynamics

The 3D Framework Program

Crane Girder Analysis According to EN 1993-6 Another new feature in CRANEWAY is a load case table for the clear representation of load combinations with information about loading, dynamic coefficients and partial safety factors for the design situations ultimate limit state, fatigue, deformation and support forces. The 3D rendering where crane rides can be animated has been improved, too.

If the rail section is taken into account for the determination of cross-section properties, you can define interruptions for its welds. Moreover, the welds ao and au (between web and top/bottom flange) are designed separately when welded cross-sections are used. When you determine the horizontal deformation of the crane girder, you have the option to take account of column heights according to table 7.1 d.

Performance of All Required Designs

The Ultimate FEA Program 3D Finite Elements

CAD/BIM Integration

With the new version 8 of the standalone program CRANEWAY you can now design crane runways not only according to DIN 4132 but EN 1993-6 (Eurocode 3). When calculating the crane girder according to Eurocode 3 it is possible to lay out single- and multi-span beams for bridge as well as suspension cranes (girder with traveling trolley).

Steel Construction

The program performs all designs that are required for the crane girder analysis: – Stress design for crane runway and welds – Analysis of fatigue behavior and fatigue design for crane runway and welds – Deformation analyses – Plate buckling design also local for wheel load introduction

Solid Construction

Cross-Sections

Bridge Construction

Fig. 1. Entering crane loads according to Eurocode in CRANEWAY

Connections

3D Frameworks

Stability and Dynamics

Up-to-Date Information...

Fig. 2. Graphical representation of design results (© Dlubal)

04_SC_A04-A08.indd 5

RSTAB 8 and RFEM 5: new versions CRANEWAY: design of crane girders according to Eurocode 3 and DIN WEBINARS: learn online, learn live WEBSHOP: order quickly and easily

Free Trial Version at

www.dlubal.com

Further Information: Dlubal Engineering Software Am Zellweg 2, D-93464 Tiefenbach Tel.: +49 9673 9203-0 Fax: +49 9673 9203-51 info@dlubal.com www.dlubal.com

01.02.13 08:45

Products & Projects

The fatigue design is performed for up to three cranes operating at the same time, based on the concept of nominal stress according to EN 1993 1-9. In addition to I-shaped rolled cross-sections you can design asymmetrical, welded I-beams in CRANEWAY. Both section types can be combined with L-sections or channels. Furthermore, it is possible to apply the rail or splice so that the crosssection resistance is increased. All designs are shown in clearly-arranged results tables, sorted by different topics. A corresponding cross-section graphic is always displayed together with the table values. Finally, it is possible to integrate descriptive graphics into the printout report representing the documentation for the crane girder design to be prepared for test engineers. It is also possible to specify the amount of data output for individual designs. Further Information: Dlubal Engineering Software, Am Zellweg 2, 93464 Tiefenbach, Tel. +49 (0)9673 – 92 03-0, Fax +49 (0)9673 – 92 03-51, info@dlubal.com, www.dlubal.de

‘Ultimate Cutting performance’ with triple-edged blade. If a nut is jammed, corroded or damaged, it is more often than not, impossible to unscrew it without damage to the bolt or stud. The ultimate solution is to use a nut splitter to free the bolt, by splitting through the nut.

In the energy sectors and other heavy-duty industries, fasteners employed are normally of high-tensile strength (and therefore hard) and will necessitate the use of nut splitters that have highstrength ability and the quality to deal with the materials encountered. The ENS hydraulic nut splitter is specially designed for use on BS1560, ANSI B16.5, API 6A & API 17D flanges. It easily and quickly cracks problematic nuts of up to 5.3/8” or 130 mm across the flats. The shaped high-strength heads are interchangeable with each cylinder (to cover the fastener size range) and screw together offering adjustment and the flexibility to cut a variety of nut sizes per head. The heavy duty, triangular blade design is an important feature as it offers three cutting edges and enables the operator to save time, since he has only to turn the blade 120° to have a new cutting edge. It also provides the assurance of being able to complete the job at hand with minimal interruption or delay. The ENS hydraulic nut splitter is also equipped with a dial-in feature which enables the blade travel to be controlled and therefore split the nut without stud damage, avoiding costly replacement of the stud itself (where applicable). ENS is normally operated with a hand pump/gauge and hose, being a single-acting cylinder with spring return. ENS is also available in a sub-sea version, using a double-acting cylinder. In this case a powered pump unit is employed and connects to a ‘diver control valve’ local to the splitting operation; water depth is immaterial as the pump runs continuously during the process and ensures minimal diver effort and rapid blade retraction. The ENS range has 4 standard models (1-4) each with 2 or 4 interchangeable heads, and covers 3/4–3.1/2 inches, M20–M90 fasteners (1.1/4´´–5.3/8´´, 30 mm–130 mm nuts). Tools have a fixed piston stroke and a built-in adjustable scale to indicate the fastener size; there is a tool selection sheet from which the ideal

Theory of Structures Fundamentals, Framed Structures, Plates and Shells

Online-Bestellung: www.er nst-und-sohn.de

preliminary

■ This book provides the reader with a consistent approach to theory of structures on the basis of applied mechanics. It covers framed structures as well as plates and shells using elastic and plastic theory, and emphasizes the historical background and the relationship to practical engineering activities. This is the first comprehensive treatment of the school of structures that has evolved at the Swiss Federal Institute of Technology in Zurich over the last 50 years. The many worked examples and exercises make this a textbook ideal for in-depth studies. Each chapter concludes with a summary that highlights the most important aspects in concise form. Specialist terms are defined in the appendix. There is an extensive index befitting such a work of reference. The structure of the content and highlighting in the text make the book easy to use. The notation, properties of materials and geometrical properties of sections plus PETER MARTI brief outlines of matrix algebra, tensor calculus and Theory of Structures calculus of variations can be found in the appenFundamentals, Framed dices. This publication should be regarded as a key Structures, Plates and Shells work of reference for students, teaching staff and 2013. approx. 750 pages, practicing engineers. Its purpose is to show readers approx. 600 fig., approx. 30 tab. how to model and handle structures appropriately, Hardcover. to support them in designing and checking the approx. € 98,–* structures within their sphere of responsibility. ISBN: 978-3-433-02991-6

Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG

W i l e y

Tel. +49 (0)6201 606-400 Fax +49 (0)6201 606-184 service@wiley-vch.de

C o m p a n y

0216400006_dp_180x128mm.indd 1

Steel Construction 6 (2013), No. 1

04_SC_A04-A08.indd 6

Kundenservice: Wiley-VCH Boschstraße 12 D-69469 Weinheim

* Der €-Preis gilt ausschließlich für Deutschland. Inkl. MwSt. zzgl. Versandkosten. Irrtum und Änderungen vorbehalten. 0216400006_dp

– Stability analysis for lateral-torsional buckling according to second-order analysis for torsional buckling

26.11.12 11:20

01.02.13 08:45

Products & Projects was held in January at the Vigyan Bhawan, New Delhi in the august presence of dignitaries and representatives from the Government of India and leaders of the corporate world.

Specially designed for use on BS1560, ANSI B16.5, API 6A & API 17D flanges: The ENS hydraulic nut splitter (Foto: SPX Hydraulic Technologies)

head size can be selected to fit each flange. ENS operates at a maximum pressure of 10,000 psi/700 Bar, each tool incorporates an internal pressure-relief valve for protection of the operator and process safety. Further information: SPX Hydraulic Technologies, Albert Thijsstraat 12, NL 6471 WX Eygelshoven, The Netherlands, Tel. +31 (45) 5678877, Fax +31 (45) 5678878, infoeurope@spxhydraulictech.com, www.spxhydraulictech.com

Tata Steel recognized for its exemplary performance in economic, social and environmental facets of Indian business Tata Steel was awarded the ‘CII-ITC Sustainability Prize’ in the ‘Category A’ for Large Independent Company at the CII-ITC Sustainability Awards 2012. The award ceremony

Mr Chanakya Choudhary, Chief Resident Executive- New Delhi, Tata Steel received the prize from The Hon’ble President of India, Shri Pranab Mukherjee. On receiving the Prize Mr H M Nerurkar, MD Tata Steel said, “We feel honoured to be recognized for our corporate leadership in social responsibility and sustainable development initiatives. Tata Steel has built a legacy of achieving business success through responsible social, environmental and economic practices that help build inclusive societies.” He extended his heartfelt thanks to CII-ITC for acknowledging Tata Steel’s efforts. Tata Steel has been awarded the Sustainability Prize (in the ‘Category A’ for Large Independent Company – for companies with turnover of above 500 Crores) earlier in the years of 2006, 2007, 2008 and 2011. The award this year makes it the 5th time for the steel maker since 2006 and for two years in succession, underlining Tata Steel’s ethos built on a commitment for values beyond steel. The annual CII-ITC Sustainability Awards are given to encourage and recognize Indian corporates who embed sustainability and, thereby, champion the cause of intergenerational parity. The awards are conferred to Indian businesses that demonstrate excellent performance in the area of Sustainable Development. Further information: Tata Steel Limited, Registered Office, Bombay House, 24, Homi Mody Street, Mumbai – 400 001, Ph: +91 022 66658282, www.tatasteel.com

Steel Construction 6 (2013), No. 1

04_SC_A04-A08.indd 7

01.02.13 08:45

Performance Based Building Design 1 From Below Grade Construction to Cavity Walls approx. 260 pages, approx. 172 figures, Softcover. approx. € 59,–* ISBN 978-3-433-03022-6

Order online! www.er nst-und-sohn.de

Date of Publication: April 2012

Package: Performance Based Building Design 1 and 2 approx. € 99,–* ISBN: 978-3-433-03024-0 Date of Publication: September 2012

Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG

W i l e y

Customer Service: Wiley-VCH Boschstraße 12 D-69469 Weinheim

Tel. +49 (0)6201 606-400 Fax +49 (0)6201 606-184 service@wiley-vch.de

C o m p a n y

0238100006_dp_180x128mm.indd 1

08.02.12 14:58

Steel Structures

Order online: www.er nst-und-sohn.de

■ The Finite Element Method (FEM) has become a standard tool used in everyday work by structural engineers having to analyse virtually any type of structure. After a short introduction into the methodolgy, the book concentrates on the calculation of internal forces, deformations, ideal buckling loads and vibration modes of steel structures. Beyond linear structural analysis, the authors focus on various important stability cases such as flexural buckling, lateral torsional buckling and plate buckling along with determining ideal buckling loads and second-order theory analysis. Also, investigating cross-sections using FEM will become more and more important in the future. ROLF KINDMANN, For practicing engineers and students in engineerM AT T H I A S K R A U S ing alike all necessary calculations for the design of Steel Structures structures are presented clearly.

Design using FEM April 2011. 542 pages. 365 fig. 90 tab. Softcover.  59,90 *

ISBN 978-3-433-02978-7

Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG

W i l e y

Customer Service: Wiley-VCH Boschstraße 12 D-69469 Weinheim

Tel. +49 (0)6201 606-400 Fax +49 (0)6201 606-184 service@wiley-vch.de

C o m p a n y

0200100006_dp180x128mm.indd 1

04_SC_A04-A08.indd 8

Author information: ■ Univ.-Prof. Dr.-Ing. Rolf Kindmann teaches steel and composite design at the Ruhr University in Bochum and is a partner of the Ingenieursozietät Schürmann-Kindmann und Partner in Dortmund. ■ Dr.-Ing. Matthias Kraus is a research assistant at the same chair.

* Prices are valid in Germany, exclusively, and subject to alterations. Prices incl. VAT. Books excl. shipping. Journals incl. shipping. 0200100006_dp

■ Just like building physics, performance based building design was hardly an issue before the energy crisis of the 1970ies. With the need to upgrade energy efficiency, the interest in overall building performance grew. As the first of two volumes, this book applies the performance rationale, advanced in applied building physics, to the design and construction of buildings. After an overview of materials for thermal insulation, water proofing, air tightening and vapour tightening and a discussion on joints, building construction is analysed, starting with the excavations. Then foundations, below and on grade constructions, typical load bearing systems and floors pass the review to end with massive outer walls insulated at the inside and the outside and cavity walls. Most chapters build on a same scheme: overview, overall performance evaluation, design and construction. The book should be usable by undergraduates and graduates in architectural and building engineering, though also building engineers, who want to refresh their knowledge, may benefit. The level of discussion assumes the reader has a sound knowledge of building physics, along with a background in structural engineering, building materials and building construction.

*€ Prices are valid in Germany, exclusively, and subject to alterations. Prices incl. VAT. Books excl. shipping. Journals incl. shipping. 0238100006_dp

HUGO S.L.C. HENS

18.01.2011 16:08:31 Uhr

01.02.13 08:45

Editorial DOI: 10.1002/stco.201310009

7th International Workshop on Connections in Steel Structures 2012 – Connections VII 4) Connections for structures with hollow sections 5) Bolting and special connection topics 6) Bracing and truss connections

Prof. Dan Dubina

Prof. Daniel Grecea

The International Workshop on Connections in Steel Structures (Connections I), jointly organized by the European Convention for Constructional Steelwork (ECCS) and the American Institute of Steel Construction (AISC) has been held since 1987. It started in Paris (Cachan), and was followed by Connections II in 1991 (Pittsburgh, Pennsylvania), Connections III in 1995 (Trento), Connections IV in 2000 (Roanoke, Virginia), Connections V in 2004 (Amsterdam) and Connections VI in 2008 (Chicago). The success of the “Connections“ series has been confirmed by the number of outstanding scientists and engineers, mainly from Europe and the USA, but also from other areas, who, over the years, have contributed to the workshops with their papers, knowledge and professional experience. All that is included in the series of Proceedings volumes, summarizing 289 scientific papers, all of a very high quality. The recommendations issued at the end of each workshop are regarded as valuable references in codification and practice in Europe, the USA and elsewhere. The 7th Workshop took place from 30 May to 2 June 2012 in the historic city of Timis¸oara, Romania. Hosts were the “Politehnica” University and the Romanian Academy, under the supervision of ECCS and AISC. This time, 44 papers were presented authored by 112 outstanding specialists in structural connectins coming from Europe and USA, but also from Canada, Brazil, Chile, China and Australia. Six topics were covered by these contributions: 1) Structural design, design codes 2) Methods of analysis 3) Connections for seismic effects

At the end of the event, a Concluding Panel, chaired by Prof. Riccardo Zandonini (ECCS) and Dr. Reidar Bjorhovde (AISC), summarized and wrapped up the main contributions collected during oral presentations and open discussions. The Connection VII Proceedings, with the final versions of the papers and conclusions, will be published by the ECCS. Among the papers presented during the Connections VII Workshop, all of great interest for the profession, the following five have been selected for the present issue of “Steel Construction – Design and Research”: 1) Recent changes in U.S. connection design practice, Charles J. Carter, Cynthia J. Duncan, USA 2) Weld design and fabrication for RHS connections, Matthew R. McFadden, Min Sun, Jeffrey A. Packer, Canada 3) Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads, Massimo Latour, Vincenzo Piluzzo, Gianvittorio Rizzano, Italy 4) Design model for composite beam-to-reinforced concrete wall joints, José Henriques, Luís Simões da Silva, Isabel Valente, Portugal 5) Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas, Florea Dinu, Dan Dubina, Calin Neagu, Cristian Vulcu, Ioan Both, Sorin Herban, Romania We express our gratitude to the authors of these papers, indeed to all contributors to Connections VII. Thanks go to the chairs of ECCS and AISC and the Technical Committees of the two bodies involved in organizing and promoting the series of workshops on connections, particularly to Prof. Frans Bijlaard, chairman of ECCS-TC10, and Dr. Charles Carter, vice-president and chief structural engineer at the AISC.

Prof. Dan Dubina Chairman of Connections VII Workshop

Prof. Daniel Grecea Scientific Secretary of Connections VII Workshop

05_001_Editorial_1-13.indd 1

01.02.13 08:46

Articles Charles J. Carter* Cynthia J. Duncan

DOI: 10.1002/stco.201300004

Recent changes in U.S. connection design practice The 2010 AISC Specification for Structural Steel Buildings (AISC 360-10) forms the basis for the 14th edition of the AISC Steel Construction Manual. Both publications reflect changes in connection design requirements and practices. This paper summarizes the most relevant changes in connection design requirements and practices made in these latest versions of these documents.

1 Basic bolt strength increased U.S. practice in the design of bolted joints for shear has long since been based on reducing the basic shear strength to account for conditions in which the shear distribution in the joint is not uniform. For simplicity, this reduction has been applied to all bolted joints so that the bolt shear strength is not usually affected by the number of bolts in the joint. Prior to the 2010 AISC Specification [1], a 20 % reduction was included in the basic strength for joint lengths up to 50 in. (1270 mm). Above that dimension, an additional 20 % reduction was required in the calculations. A re-evaluation of existing data and common joint lengths in modern construction led to a change in the 2010 AISC Specification. A similar approach is used, but the initial reduction is taken as 10 % and the length at which an additional reduction (of 17 %) is taken is 38 in. (965 mm). This new approach is illustrated and compared with the old approach in Fig. 1. In theory, the non-uniform distribution is present only in end-loaded

joints (see Fig. 2). However, for simplicity, the reduction is applied to all joints, and also to account for restraint and behaviour that is customarily ignored in many connection design approaches.

2 Bolt strength groupings established ASTM A325 and A490 bolts are the usual fasteners contemplated for bolted joints in U.S. practice. The twist-off-type tension-control configurations of these products have become prevalent in the U.S. market-

place, and so ASTM standards have been developed to define them: ASTM F1852 is similar to A325, and ASTM F2280 is similar to A490. When added to the other grades that exist in the U.S. marketplace, such as ASTM A354 and A449, and also counting all the metric equivalents that exist for these standards, there are many fastener options and many of those have similar or identical strength levels for design. To simplify the provisions used in the AISC Specification, these products have been grouped as shown in Table 1. One unintended item of confusion has been discovered: group A and B tension and shear strength levels do not have anything to do with the class A and class B faying surface classifications used in slip-critical connection design.

Selected and reviewed by the Scientific Committee of the 7th International Workshop of Connections in Steel Structures, 30 May–2 June 2012, Timişoara, Romania * Corresponding author:

carter@aisc.org

Fig. 1. Comparison of bolt shear strengths in the 2005 and 2010 AISC Specifications

06_002-004_Carter_Duncan (1-4)_cs6.indd 2

01.02.13 08:47

Ch. J. Carter/C. J. Duncan · Recent changes in U.S. connection design practice

Fig. 2. Examples of end-loaded and non-end-loaded joints Fig. 3. Test specimen used in AISC slip-critical joint research

Table 1. Bolt strength levels as grouped in the 2010 AISC Specification Basic strength Group

ASTM

The new equation for calculating slip resistance is given as

Shear

Tension

ksi

MPa

ksi

MPa

ksi

MPa

A325, A325M, F1852, A354 gr. BC, A449

620

372

457

A490, F2280, A354 gr. BD

113

780

457

579

3 Slip-critical connection design simplified and improved Up until the 2005 AISC Specification, the designer was asked to decide if slip was to be prevented as a matter of serviceability or strength. Dubiously buried in the background of this decision was the reality that the actual checks were calibrated to give similar results in common cases, making the choice confusing at best. In 2005 changes were made that created different levels of design between serviceability and strength. However, the strength-level slip checks caused concern in the industry because some joints previously designed for serviceability slip were now re-

quired to be designed with more bolts at the strength-level slip resistance. These included connections with oversized holes or slotted holes parallel with the direction of the load. Large-scale (see Fig. 3) and other research [3], [4], [5] was undertaken almost immediately, and much was learned about slip behaviour and joint design requirements. The results affected the design method, allowing significant simplification and better ways to address the behaviour. The serviceability–strength dichotomy was eliminated, slip coefficients were changed and requirements regarding when to use fillers in the joint were added, among other refinements.

Rn = μDu hfTb Ns The variables Tb and Ns are unchanged. They represent the bolt pretension and number of slip planes respectively. A resistance factor for LRFD or safety factor for ASD is required: – For standard holes and short slotted holes perpendicular to the direction of the load: φ = 1.0 and Ω = 1.50 – For oversized and short slotted holes parallel with the direction of the load: φ = 0.85 and Ω = 1.76 – For long slotted holes: φ = 0.70 and Ω = 2.14 The value of the slip coefficient μ was changed from 0.35 to 0.30 primarily because of the wide variability of the slip resistance of class A “clean mill scale” surfaces. The slip coefficient for class B surfaces was maintained as μ = 0.50 for class B “blast-cleaned” surfaces and “blast-cleaned surfaces with class B coatings”. A reduction applicable to joints in which multiple fillers are used was

Steel Construction 6 (2013), No. 1

06_002-004_Carter_Duncan (1-4)_cs6.indd 3

01.02.13 08:47

Ch. J. Carter/C. J. Duncan · Recent changes in U.S. connection design practice

added; alternatively, additional bolts can be added to develop the fillers. The filler factor hf is determined as follows: – Where bolts have been added to distribute loads in the fillers: hf = 1.0 – Where bolts have not been added to distribute loads in the fillers: hf = 1.0 for one filler between connected parts, and hf = 0.85 for two or more fillers between connected parts

ways. Provisions in section J2.4 (a) and (c) in the 2010 AISC Specification are based on a load–deformation behaviour that is affected by the weld size [7]. Accordingly, these provisions have been clarified to reflect that they are based on fillet weld groups in which the size of the weld is uniform. When the weld group is not of uniform size, section J2.4 (b) can be used to account for size variations.

It also is worth noting that prior to the 2010 AISC Specification, fillers > ¾ in. (19 mm) thick had to be developed. This is no longer the case. A reduction factor still applies to the bolt shear strength when fillers are not developed, but the 2010 Specification recognizes that the reduction factor need not exceed 0.85 regardless of the thickness of the filler.

7 Prying action formulas improved with simple change

4 Base metal design at welds Table J2.5 in the 2010 AISC Specification summarizes the available strengths for welds and base metal and weld metal in welded joints. Base metal strength at welds is now based on the rupture strength rather than the yield strength. Previously, the design was based on yielding in the base metal, which has come to be viewed as conservative and incorrect since the weld itself adjacent to the base metal is designed for a rupture limit state.

5 Directional strength increase extended to out-of-plane loading Prior to 2010 the AISC Specification included the words “in plane” when provisions were given for the directional strength increase for fillet welds, i.e. the provisions were limited to loading in the plane of the weld or weld group. Common usage of the provisions in practice, however, extended these provisions to out-of-plane loading as well. Research [6] was conducted to evaluate that practice and showed that the restriction (the words “in plane”) could be eliminated. Accordingly, they do not appear in the 2010 AISC Specification.

6 Weld group size uniformity requirements added Fillet welds used in groups are generally all of the same size – but not al-

design. As a result, eccentricity requirements re-appeared in the single-plate connection design procedures in the 14th edition of the AISC Manual. Table 2 illustrates the eccentricities that are used in the design of single-plate connections. References

Changes to the bolt shear strength values necessitated a change in the 14th edition of the AISC Steel Construction Manual procedures for single-plate connections. In the 13th edition of the Manual, the 20 % bolt shear strength reduction was used as a convenient way of simplifying the design of single-plate connections. That is, we knew the effect of most eccentricities was less than the 20 % reduction, and we also knew that shear connections are not end-loaded and did not need the 20 % reduction. On this basis it was accepted that most eccentricities in these connections could be ignored. The changes to the 2010 AISC Specification cut the margin on bolt strength to a 10 % reduction, which was no longer enough to offset the impact of eccentricity in the connection

[1] AISC: Specification for Structural Steel Buildings (ANSI/AISC 360-10), AISC, Chicago, IL, 2010. [2] AISC: Steel Construction Manual, AISC Chicago, IL, 2011. [3] Borello, D. B., Denavit, M. D., Hajjar, J. F.: Behavior of Bolted Steel Slip-Critical Connections with Fillers. Report No. NSEL-017, Department of Civil & Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, 2009. [4] Dusika, P., Iwai, R.: Development of Linked Column Frame Lateral Load Resisting System. 2nd Progress Report for AISC and Oregon Iron Works, Portland State University, Portland, OR, 2007. [5] Grondin, G, Jin, M., Josi, G.: Slip-Critical Bolted Connections – A Reliability Analysis for the Design at the Ultimate Limit State. Preliminary Report prepared for AISC, University of Alberta, Edmonton, Alberta, CA, 2007. [6] Kanvinde, A. M., Grondin, G. Y., Gomez, I. R., Kwan, Y. K.: Experimental Investigation of Fillet Welded Joints Subjected to Out-of-Plane Eccentric Loads. Engineering Journal, American Institute of Steel Construction, 3rd Quarter, 2009. [7] Muir, L. S.: Deformational Compatibility in Weld Groups. ECCS/AISC Workshop Connections in Steel Structures VI. 23–24June 2008, Chicago, IL. [8] Swanson, J. A.: Ultimate Strength Prying Models for Bolted T-Stub Connections. Engineering Journal, AISC, 2002, vol. 39, No. 3, 3rd Quarter, AISC, Chicago, IL, pp. 136–147 [9] Thornton, W. A.: Strength and Serviceability of Hanger Connections. Engineering Journal, AISC, 1992, vol. 29, No. 4, 4th Quarter, AISC, Chicago, IL, pp. 145–149.

Table 2. Bolt strength levels as grouped in the 2010 AISC Specification

Keywords: connections; bolts; welds; prying action; slip critical; AISC

Treatment of prying action in the AISC Manual and other sources has traditionally been based on the use of Fy in the calculations. At the same time, it has long since been known that the resulting predictions of the equations for prying action are significantly conservative [8], [9]. To address this in a simple manner, the AISC Manual now uses Fu in place of Fy for prying action checks.

8 Single-plate connection eccentricity calculations revised

n 2–5 6–12

Hole type

e [in.]

max. tp or tw [in.]

SSLT

a/2

none

STD

a/2

db/2 + 1/16

SSLT

a/2

db/2 + 1/16

STD

db/2–1/16

Authors: Charles J. Carter, SE, PE, PhD, Vice-President and Chief Structural Engineer, American Institute of Steel Construction, Chicago, IL, USA, carter@aisc.org Cynthia J. Duncan, Director of Engineering, American Institute of Steel Construction, Chicago, IL, USA, duncan@aisc.org

Steel Construction 6 (2013), No. 1

06_002-004_Carter_Duncan (1-4)_cs6.indd 4

01.02.13 08:47

Articles Matthew R. McFadden Min Sun Jeffrey A. Packer*

DOI: 10.1002/stco.201300006

Weld design and fabrication for RHS connections The 2010 AISC Specification for Structural Steel Buildings has expanded its scope in chapter K “Design of HSS and Box Member Connections” to include a section K4 “Welds of Plates and Branches to Rectangular HSS”. This paper discusses the historical development of the effective weld properties and analyses the structural reliability of the provisions. Additionally, there is a discussion on recent changes in U.S. and Canadian specifications/ codes with regard to the limit states for fillet weld design and the acceptance/rejection of the (1.00 + 0.50 sin1.5θ) term. Finally, there is a discussion of the details of an experimental research programme being performed at the University of Toronto in collaboration with AISC to determine the weld effective length in RHS T-connections under branch in-plane bending moments. In conclusion, it is found that the inclusion of the (1.00 + 0.50 sin1.5θ) term for RHS gapped K-connections as well as T- and X-connections, based on the limit state of shear failure along the effective throat of the weld, may be unsafe for fillet weld design when used in conjunction with the current effective weld length rules.

1 Introduction Two methods are currently available for the design of welded connections between rectangular hollow sections (RHS) [15]: (I) The welds may be proportioned to develop the yield strength of the connected branch wall at all locations around the branch. This approach may be appropriate if there is low confidence in the design forces, uncertainty regarding method (II) or if plastic stress redistribution is required in the connection. This method will produce an upper limit for the weld size required and may be excessively conservative in some situations. (II) The welds may be designed as “fit-for-purpose” and proportioned to resist the applied forces in the Selected and reviewed by the Scientific Committee of the 7th International Workshop of Connections in Steel Structures, 30 May–2 June 2012, Timişoara, Romania * Corresponding author: jeffrey.packer@utoronto.ca

branch. The non-uniform loading around the weld perimeter due to the relative flexibility of the connecting RHS face requires the use of weld effective lengths. This approach may be appropriate when there is high confidence in the design forces or if the branch forces are particularly low relative to the branch member capacity. Where applicable, this approach may result in smaller weld sizes providing a more economical design with increased aesthetic value. The primary focus of this paper is method (II), but it is interesting to com-

pare the results of method (I) for the design of fillet welds in various steel specifications/codes (see Table 1). Clearly, there is quite a disparity. Fillet welds, being the least expensive and easiest type of weld, are the preferred and most common weld type for hollow section connections. The design of fillet welds in structural steel buildings in the USA is governed by Table J2.5 of the AISC Specification [1] and is based on the limit state of shear failure along the effective throat using a matching (or under-matching) filler metal. For a simple 90° RHS T-connection under branch axial tension (see Fig. 1a), the LRFD strength of a single weld is given by

ΦR n = ΦFnw A we

(

)(

= 0.75 0.60FEXX D/ 2 l w

The design of fillet welds in Canada is governed by CSA S16-09 [4] section 13.13.2.2, and although different coefficients are used, an identical resistance is obtained. The prior edition, CAN/ CSA S16-01 [3], included an additional check for shearing of the base metal at the edge of a fillet weld along the fusion face (see Fig. 1b), which frequently governed and thus generally

Table 1. Comparison of fillet weld effective throats required to develop the yield resistance of the connected branch member wall in Fig. 1(a), for ASTM A500 Grade C RHS Specification or code

ANSI/AISC 360-10 Table J2.5 [1]

1.43 tb

AWS D1.1/D1.1M: 2010 section 2.25.1.3 and Fig. 3.2 [2]

1.07 tb

CSA S16-09 section 13.13.2.2 [4]

0.95 tb

CAN/CSA S16-01 section 13.13.2.2 [3]

1.14 tb

CEN (2005) Directional Method [6]

1.28 tb

07_005-010_Packer (1-6)_cs6.indd 5

)

01.02.13 08:47

M. R. McFadden/M. Sun/J. A. Packer · Weld design and fabrication for RHS connections

This IIW document [11] thus specifically acknowledges the effective length concept for weld design.

3 2010 AISC Specification, section K4 weld design procedures

a) 90° RHS-T-connection under branch axial tension

b) Detail of fillet weld cross-section showing assumed failure planes

Fig. 1. Comparison of fillet weld limit state design checks

resulted in larger weld sizes. However, the current fillet weld design requirements for both AISC 360-10 [1] and CSA S16-09 [4] are based solely on the limit state of shear failure along the effective throat.

2 Historical treatment of weld design for RHS connections In 1981 Subcommission XV-E of the International Institute of Welding (IIW) produced its first design recommendations for statically loaded RHS connections, which were updated and revised with a second edition later in that decade [10]. These recommendations are still the basis for nearly all current design rules around the world which deal with statically loaded connections for onshore RHS structures, including those in Europe [6], Canada [14] and the USA [1]. Research at the University of Toronto [8, 9] concerning fillet-welded RHS branches in large-scale Warren trusses with gapped K-connections revealed that fillet welds in that context can be proportioned on the basis of the loads in the branches, thus resulting in relatively smaller weld sizes compared with IIW [10]. It was concluded, simplistically, that the welds along all four sides of the RHS branch could be taken as fully effective when the chordto-branch angle is ≤ 50°, but that the weld along the heel should be considered as completely ineffective when the angle is ≥ 60°. A linear interpolation was recommended when the chord-tobranch angle is between 50° and 60°. Based on this research, the formulae for the effective length of branch member welds in planar, gapped, RHS Kand N-connections, subjected to predominantly static axial loads, were taken in Packer and Henderson [13] as

Section K4 of the AISC Specification [1] contains a detailed design method considering effective weld properties for various RHS connection types.

Le =

2H b + 2B b sin θ

when θ ≤ 50° (1a)

T-, Y- and X-connections under branch axial load or bending Effective weld properties are given by

Le =

2H b + Bb sin θ

when θ ≥ 60° (1b)

Le =

2H b + 2beoi sin θ

Sip =

 H  t w  Hb  + t w beoi  b  (4)   3  sin θ   sin θ 

In a further study by Packer and Cassidy [12], which used 16 large-scale connection tests designed to be weldcritical, new weld effective length formulae for T-, Y- and X-connections (aka cross-connections) were developed. It was found that more of the weld perimeter was effective for lower branch member inclination angles for T-, Y- and X-connections. Thus, the formulae for the effective length of branch member welds in planar T-, Yand X-connections (for RHS members), subjected to predominantly static axial loads, were revised in Packer and Henderson [14] to

Le =

2H b + Bb sin θ

when θ ≤ 50° (2a)

Le =

2H b sin θ

when θ ≥ 60° (2b)

Linear interpolation was recommended between 50° and 60°. The latest (3rd) edition of the IIW recommendations [11] requires that the design resistance of hollow section connections be based on failure modes that do not include weld failure, with the latter being prevented by satisfying either of the following criteria: (I) welds are to be proportioned to be “fit for purpose” and to resist forces in the members connected, taking account of connection deformation/rotation capacity and considering effective weld lengths, or (II) welds are to be proportioned to achieve the capacity of the connected member walls.

(3)

 H  t Sop = t w  b  B b + w B2b sin θ 3  

( )

−

beoi =

(

t w /3 B b − beoi

)

(5)

10  Fy t    B ≤ Bb B/t  Fyb t b  b

(6)

When β > 0.85 or θ > 50°, beoi/2 shall not exceed 2t. This limitation represents additional engineering judgement. In contrast to Eqs. (2a) and (2b), the effective weld length in Eq. (3) was – for consistency – made equivalent to the branch wall effective lengths used in section K2.3 of the AISC Specification [1] for the limit state of local yielding of the branch(es) due to uneven load distribution, which in turn is based on IIW [10]. The effective width of the weld transverse to the chord beoi is illustrated in Fig. 2b. This term beoi was derived empirically on the basis of laboratory tests in the 1970s and 1980s [5]. The effective elastic section modulus of welds for in-plane bending and out-of-plane bending, Sip and Sop respectively (Eqs. (4) and (5)), apply in the presence of the bending moments Mip and Mop as shown in Fig. 2b. Although based on informed knowledge of general RHS connection behaviour, Eqs. (4) and (5) have not been substantiated by tests, and are therefore purely speculative.

Steel Construction 6 (2013), No. 1

07_005-010_Packer (1-6)_cs6.indd 6

01.02.13 08:47

M. R. McFadden/M. Sun/J. A. Packer · Weld design and fabrication for RHS connections

a) Various load cases

Two large-scale simply supported fillet-welded, RHS Warren trusses, comprised of 60° gapped and overlapped

K-connections, spanning 39.4 ft (12.0 m) and 40.0 ft (12.2 m) were tested by Frater and Packer [8, 9]. Quasi-static loading was applied in a carefully controlled manner to produce sequential failure of the tension-loaded, filletwelded connections (rather than connection failures). In addition, a series of weld-critical tests were performed by Packer and Cassidy [12] on four T-connections and 12 X-connections, with the branches loaded in quasi-static, axial tension. The effective leg sizes of the welds (measured along the branch member and chord member respectively) plus the throat sizes were recorded. The measured geometric and mechanical properties of these trusses and welds and the failure loads of all welded connections are subsequently used here to evaluate nominal weld strengths and predicted weld design strengths according to the AISC Specification [1], with weld failure as the only limit state. Table J2.5, section J4 [1] and Eqs. (3), (6), (7) and (8) were used to calculate the nominal strengths (excluding the resistance factor) of the 31 welded

a) Actual strength vs. predicted nominal strength (Rn)

b) Actual strength vs. predicted LRFD strength (0.75Rn)

b) Effective weld length dimensions

Fig. 2. Effective weld length terminology for T-, Y- and X-connections under branch axial load or bending

Gapped K- and N-connections under branch axial load Effective weld lengths are given by:

Le =

(

2 H b − 1.2t b sin θ

) + 2 (B

− 1.2t b

)

when θ ≤ 50° Le =

(

2 H b − 1.2t b sin θ

) + (B

(7a) b

− 1.2t b

)

when θ ≥ 60°

Mn–op = FnwSop

(10)

where Fnw = 0.60FEXX

(11)

4 Evaluation of AISC 2010 specification with experiments on RHS welds under predominantly axial loads

(7b) When 50° < θ < 60°, linear interpolation is to be used to determine Le. Eqs. (7a) and (7b) are similar to Eqs. (1a) and (1b) but the former incorporate a reduction to allow for a typical cold-formed RHS corner radius. For gapped K- and N-connections, the simplified nature of these effective length formulae (Eqs. (7a) and (7b)) was preferred to the more complex ones that would result if the branch effective widths for the RHS walls in AISC Specification section K2.3 [1] were to be adopted. Weld effective length provisions for overlapped RHS K- and N-connections were also provided in AISC Specification section K4 [1], based on branch effective widths for the RHS walls in section K2.3. However, in this case no research data on weld-critical overlapped RHS Kand N-connections were available. The available strength of branch welds is determined – allowing for non- uniformity of load transfer along the line of the weld – as follows by AISC [1]: Rn or Pn = FnwtwLe

(8)

Mn–ip = FnwSip

(9)

Fig. 3. Correlation with test results for gapped K-connections without the inclusion of the (1.00 + 0.50 sin1.5θ) term

a) Actual strength vs. predicted nominal strength (Rn)

b) Actual strength vs. predicted LRFD strength (0.75Rn)

Fig. 4. Correlation with test results for T- and X-connections without the inclusion of the (1.00 + 0.5 sin1.5θ) term

Steel Construction 6 (2013), No. 1

07_005-010_Packer (1-6)_cs6.indd 7

04.02.13 12:00

M. R. McFadden/M. Sun/J. A. Packer · Weld design and fabrication for RHS connections

connections tested by Frater and Packer [8, 9] and Packer and Cassidy [12]. The predicted strength of each welded connection, without a fillet weld directional strength increase of [1.00 + 0.50 sin1.5θ] (discussed in the following section), was determined by adding together the individual weld element strengths along the four walls around the branch footprint and is given as a predicted nominal strength Rn. In order to assess whether adequate, or excessive, safety margins are inherent in the correlations shown in Figs. 3a and 4a, it is possible check to ensure that a minimum safety index of β+ = 4.0 is achieved (as currently adopted by AISC per chapter B of the Specification Commentary). This is done by using a simplified reliability analysis in which the resistance factor Φ is given by Eq. (12) [7, 16]:

(

Φ = mR exp −αβ + COV

)

(12)

where mR mean of ratio of actual element strength to nominal element strength = Rn COV associated coefficient of variation of this ratio α coefficient of separation, taken to be 0.55 [16] Eq. (12) neglects variations in material properties, geometric parameters and fabrication effects, relying solely on the “professional factor”. In the absence of reliable statistical data related to welds, this is believed to be a conservative approach. The application of Eq. (12) produced Φ = 0.959 for welded connections in gapped K-connections and Φ = 0.855 for T- and X-connections. As both of these exceed Φ = 0.75, the weld effective length concepts advocated in section K4 of the AISC Specification [1] can, on the basis of the available experimental evidence, be deemed to be adequately conservative.

AISC does not permit the fillet weld directional strength increase, whereas in Canada, the CSA and CISC do not explicitly disallow it, so designers use it. Adopting this enhancement factor leads to a greater calculated resistance for a fillet weld group in an RHS connection and hence much smaller weld sizes (as demonstrated in Table 1). The correlation plots in Figs. 3 and 4 have been recomputed with weld metal failure as the only limit state and the inclusion of the (1.00 + 0.5 sin1.5θ) term in Figs. 5 and 6. If the (1.00 + 0.5 sin1.5θ) term is taken into consideration in the analysis of the data presented in this paper, the statistical outcomes change to: – For gapped K-connections: mR = 0.999, COV = 0.180 and Φ = 0.673 (using Eq. (12) with β+ = 4.0) – For T- and X-connections: mR = 0.819, COV = 0.164 and Φ = 0.571 (using Eq. (12) with β+ = 4.0)

6 Current research on RHS moment connections

As both of these Φ factors are < 0.75, the effective length formulae, with the (1.00 + 0.50sin1.5θ) term included, may be unsafe for use in fillet weld design.

A further experimental study to determine the weld effective length in RHS T-connections subjected to branch inplane bending moments is being carried out at the University of Toronto. The test specimens have been designed such that they are weld-critical under the application of branch in-plane bending moments (weld failure to precede connection failure). The bending moment at the connection is induced by applying a lateral point load to the end of the branch in a quasi-static manner until the weld fails. Key parameters such as branch-to-chord width ratios (β ratios) of 0.25, 0.50, 0.75 and 1.00 with chord wall slenderness values of 17, 23 and 34 are being investigated. In order to determine the effectiveness of the weld in resisting the applied forces, the non-uniform distribution of normal strain and stress in the branch near the connection will be measured using strain gauges oriented along the longitudinal axis of the branch at numerous locations around the footprint.

a) Actual strength vs. predicted nominal strength (Rn)

b) Actual strength vs. predicted LRFD design strength (0.75 Rn)

Fig. 5. Correlation with test results for gapped K-connections with the inclusion of the (1.00 + 0.5 sin1.5θ) term

5 Introduction of the (1.00 + 0.50 sin1.5θ) term A debate has recently emerged regarding the application of an enhancement factor (of 1.00 + 0.50 sin1.5θ) to the nominal strength of the weld metal for fillet welds loaded at an angle of θ° to the weld longitudinal axis in hollow section connections. In the USA the

Fig. 6. Correlation with test results for T- and X-connections with the inclusion of the (1.00 + 0.5 sin1.5θ) term

Steel Construction 6 (2013), No. 1

07_005-010_Packer (1-6)_cs6.indd 8

01.02.13 08:47

M. R. McFadden/M. Sun/J. A. Packer · Weld design and fabrication for RHS connections

This will give a representative strain and stress distribution around the adjacent weld and hence the effectiveness of the weld can be determined. Based on the results of the experimental programme, the values postulated in Table K4.1 of the 2010 AISC Specification [1] will be verified or adjusted. The specimens were fabricated at Lincoln Electric Co.’s Automation Division in Cleveland, Ohio. An experienced robotic welding technologist controlled a Fanuc Robot Arc-Mate 120iC 10L, adapted to perform the gas metal arc welding process with spray metal transfer (GMAW-P), to weld the connections. For the experimental programme, robotic welding offers several advantages: improved weld quality, excellent weld/base-metal fusion and root penetration, continuous electrodes, consistent travel speeds and the ability to weld in all positions. The welding process parameters used were as follows: 0.035 in. diameter AWS ER70S-6 MIG wire, 23 V, 375 ipm wire feed speed, 90 % Ar – 10 % CO2 shielding gas mixture at 30 to 50 CFH, ¼ to ½ in. contact tube to work distance and varying travel speeds depending on weld type and size. Stepped connections (β ≤ 0.85) were clamped

to a level table and welded in the horizontal position as shown in Fig. 7a. The matched connections (β > 0.85) were mounted in rotating chucks and welded in the flat position using coordinated motion as shown in Fig. 7b, with fillet welds along the transverse branch walls and PJP flare-bevel-groove welds along the longitudinal branch walls. The test specimens are undergoing full-scale testing at the University of Toronto Structural Testing Facilities. The test setup shown in Fig. 8a consists of pin and roller supports for the chord with a 77 kip capacity MTS actuator mounted on a rigid steel frame and attached to a point load application device on the branch member. Fig. 8b shows the typical observed failure mode of weld rupture due to shear failure along the weld effective throat.

a) Stepped RHS connections welded in the horizontal position

b) Matched RHS connections welded in the flat position using coordinated motion

7 Conclusions Design guides or specifications/codes requiring the welds to develop the yield capacity of the branch members produce an upper limit for the weld size required and may be excessively conservative in some situations. Although this is considered to be a sim-

Fig. 7. Automated welding of specimens at Lincoln Electric Co.

plified design method for fillet welds, it is shown that there is quite a disparity in the effective throat size required to develop the branch wall yield capacity. Additionally, the current fillet weld design requirements in both AISC 360-10 [1] and CSA S16-09 [4] are based solely on the limit state of weld metal shear failure along the effective throat, whereas previous versions [3] included an additional check for shearing of the base metal at the edge of a fillet weld along the fusion face, which frequently governed and resulted in generally larger weld sizes. Alternative design methods that consider weld effective lengths could potentially result in a relatively smaller weld size, thus achieving a more economical design with increased aesthetic value. By comparing the actual strengths of fillet-welded joints in weld-critical T-, X- and gapped K-connection specimens with their predicted nominal strengths and design strengths, it has been shown that the relevant effective length design formulae in AISC Specification section K4 [1] – without using the (1.00 + 0.50 sin1.5θ) term for fillet welds – result in an appropriate weld design with an adequate safety level. Conversely, it is shown that the inclusion of the (1.00 + 0.50 sin1.5θ) term for such connections based solely on the limit state of weld failure along the effective throat of a fillet weld may be unsafe for design as it results in an inadequate reliability index. A limitation of this study is that all test specimens were under predominantly axial loading in the branches. However, the weld effective length formulae for T-, Y- and X-connections in AISC Specification Table K4.1 8 [1] also address branch bending. The test data available do not provide an opportunity to evaluate the accuracy of formulae applicable to branch bending loads and therefore the equations postulated are purely speculative. The objective of the research being performed at present at the University of Toronto is to verify or adjust these equations.

Acknowledgements

a) View of test setup

b) Shear failure along weld effective throat

Fig. 8. Full-scale testing at the University of Toronto

The financial and in-kind support of the Natural Sciences & Engineering Research Council of Canada, the Steel Structures Education Foundation, the American Institute of Steel Construc-

Steel Construction 6 (2013), No. 1

07_005-010_Packer (1-6)_cs6.indd 9

01.02.13 08:47

M. R. McFadden/M. Sun/J. A. Packer · Weld design and fabrication for RHS connections

tion, Lincoln Electric Co. and Atlas Tube, Inc. are all gratefully acknowledged. Notation Awe effective (throat) area of weld B overall width of RHS chord member, measured at 90° to the plane of the connection overall width of RHS branch Bb member, measured at 90° to the plane of the connection D weld leg size FEXX filler metal classification strength Fnw nominal stress of weld metal yield strength of hollow section Fy chord member material Fyb yield strength of hollow section branch member material Hb overall height of RHS branch member, measured in the plane of the connection effective length of groove and Le fillet welds to RHS for weld strength calculations Mip in-plane bending moment Mop out-of-plane bending moment Mn-ip nominal weld resistance for inplane bending Mn-op nominal weld resistance for outof-plane bending nominal strength of welded joint Pn nominal strength of welded joint Rn Sip weld effective elastic section modulus for in-plane bending Sop weld effective elastic section modulus for out-of-plane bending beoi effective width of transverse branch face welded to chord weld length lw mR mean of ratio of actual element strength to nominal element strength = professional factor t design wall thickness of hollow section chord member

tb tw α β β+ θ

design wall thickness of hollow section branch member effective weld throat thickness separation factor = 0.55 width ratio = ratio of overall branch width to chord width for RHS connection safety (reliability) index for LRFD and limit states design acute angle between branch and chord (degrees); angle of loading measured from a weld longitudinal axis for fillet weld strength calculation (degrees)

References [1] ANSI/AISC 360-10:2010: Specification for structural steel buildings. American Institute of Steel Construction, Chicago. [2] AWS D1.1/D1.1M:2010: Structural welding code – steel, 22nd ed., American Welding Society, Miami. [3] CAN/CSA-S16-01:2001: Limit states design of steel structures, Canadian Standards Association, Toronto. [4] CSA-S16-09:2009: Design of steel structures, Canadian Standards Association, Toronto. [5] Davies, G., Packer, J. A.: Predicting the strength of branch plate–RHS connections for punching shear. Canadian Journal of Civil Engineering 9 (3), 1982, pp. 458–467. [6] EN 1993-1-1:2005(E): Eurocode 3: Design of steel structures, Part 1-1: General rules and rules for buildings, European Committee for Standardization, Brussels. [7] Fisher, J. W., Galambos, T. V., Kulak, G. L., Ravindra, M. K.: Load and resistance factor design criteria for connectors. Journal of the Structural Division 104 (9), 1978, pp. 1427–1441. [8] Frater, G. S., Packer, J. A.: Weldment design for RHS truss connections, I: Applications. Journal of Structural Engineering 118 (10), 1992, pp. 2784–2803.

[9] Frater, G. S., Packer, J. A.: Weldment design for RHS truss connections, II: Experimentation. Journal of Structural Engineering 118 (10), 1992, pp. 2804– 2820. [10] IIW Doc. XV-701-89:1989: Design recommendations for hollow section joints – predominantly statically loaded, 2nd ed., International Institute of Welding, Paris. [11] IIW Doc. XV-1402-12:2012: Static design procedure for welded hollow section joints – recommendations, 3rd ed., International Institute of Welding, Paris. [12] Packer, J. A., Cassidy, C. E.: Effective weld length for HSS T, Y, and X connections. Journal of Structural Engineering 121 (10), 1995, pp. 1402–1408. [13] Packer, J. A., Henderson, J. E.: Design guide for hollow structural section connections, 1st ed., Canadian Institute of Steel Construction, Toronto, 1992. [14] Packer, J. A., Henderson, J. E.: Hollow structural section connections and trusses – a design guide, 2nd ed., Canadian Institute of Steel Construction. Toronto, 1997. [15] Packer, J. A., Sherman, D. R., Lecce, M.: Hollow structural section connections, AISC steel design guide No. 24. American Institute of Steel Construction. Chicago, 2010. [16] Ravindra, M. K., Galambos, T. V.: Load and resistance factor design for steel. Journal of the Structural Division 104 (9), 1978, pp. 1337–1353. Keywords: welding; connections; joints; rectangular hollow sections

Authors: Matthew R. McFadden, Min Sun Research Assistants, Department of Civil Engineering, University of Toronto, Canada, matthew.mcfadden@mail.utoronto.ca min.sun@utoronto.ca Jeffrey A. Packer Bahen/Tanenbaum Professor of Civil Engineering, University of Toronto, Canada, jeffrey.packer@utoronto.ca

Steel Construction 6 (2013), No. 1

07_005-010_Packer (1-6)_cs6.indd 10

01.02.13 08:47

Articles Massimo Latour Vincenzo Piluso* Gianvittorio Rizzano

DOI: 10.1002/stco.201300007

Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads Eurocode 8 has introduced the possibility of adopting partial-strength joints for seismic-resistant MR frames, provided it is demonstrated that connections perform adequately under cyclic loads. A programme of experiments devoted to investigating the cyclic behaviour of traditional joint details has recently been carried out by the authors. Within this programme, the analysis of the results obtained has revealed that even though connections designed to dissipate the seismic energy in bolted components can provide significant advantages because they are easy to repair after a destructive seismic event, they possess reduced dissipation capacity when compared with RBS connections and traditional full-strength joints. An advanced approach aimed at enhancing the hysteretic behaviour of double split tee (DST) joints and the ambitious goal of preventing joint damage is presented here. The system proposed is based on the idea of using friction dampers within the components of beam-to-column joints. A preliminary set of prototypes has been tested experimentally and the performances of joints under cyclic loading conditions have been compared with those of traditional joint details. The experimental work was carried out at the Materials & Structures Laboratory of Salerno University.

1 Introduction According to the most recent seismic codes [2, 6] steel moment-resisting frames (MRFs) can be designed according to either the full-strength criterion (based on the dissipation of the seismic input energy at the beam ends) or the partial-strength criterion (which concentrates damage in the connecting elements and/or the panel zone). In the former case, which aims to promote yielding of the beam ends, the beam-to-column joint is designed to have an adequate overstrength with respect to the connected beam to account for strain hardening and random material variability effects which affect the flexural resistance actually developed by the beam end. In the lat-

Selected and reviewed by the Scientific Committee of the 7th International Workshop of Connections in Steel Structures, 30 May – 2 June 2012, Timis¸oara, Romania * Corresponding author: v.piluso@unisa.it

ter case, beam yielding is prevented as the joints are designed to develop a bending resistance less than the beam plastic moment, so that dissipation occurs in the connecting elements. In addition, the consequence of this with regard to column design is that the hierarchy criterion has to be applied by making reference to the maximum moment that connections are able to transmit. This design philosophy, as demonstrated by Faella et al. [10], is particularly cost-effective in cases where the beam size is mainly governed by vertical rather than lateral loads, i.e. low-rise/long-span MRFs. Traditionally, the design of MRFs [17], based on the use of full-strength beam-to-column joints, requires only the prediction of the monotonic response of connections [7, 8]. In particular, in order to characterize the behaviour of such joints, only the prediction of the initial stiffness and the plastic resistance is needed, whereas the cyclic behaviour is governed by the width-to-thickness ratios of the plate elements of the connected beam. Conversely, as the energy dissipation

supply of semi-continuous MRFs relies on the ability of connections to withstand a number of excursions into the plastic range without losing their capacity to sustain vertical loads, it is evident that in order to apply partial-strength joints successfully, proper characterization and prediction of the response of connections under cyclic loading conditions [4, 5, 11, 13, 22] are necessary. Therefore, the use of partial-strength joints is allowed, both in AISC and Eurocode 8, provided that the designer demonstrates the “conformance” of the cyclic behaviour of connections adopted in the seismic load-resisting system. As a result, joints have to be pre-qualified accordingly with the ductility class of MRFs. It is for this reason that a set of pre-qualified connections with the corresponding design criteria is suggested [3]. Their cyclic behaviour has been investigated experimentally and demonstrates the development of plastic rotation supplies compatible with the corresponding ductility class. Unfortunately, pre-qualified connections are not suggested in Eurocode 8. Therefore, aiming to provide engineers with the tools they need to predict the cyclic behaviour of joints, new efforts in the development of analytical approaches are needed, unless specific experimental tests are carried out. With this in mind, a number of experimental programmes dealing with the characterization of the cyclic behaviour of beam-to-column connections have been carried out over last two decades. In a recent work by the authors’ research group [12], the behaviour of bolted joints designed to possess the same strength, but detailed to involve different components

08_011-018_Piluso_(1-7)_cs6.indd 11

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

in the plastic range, has been investigated experimentally, pointing out the hysteretic behaviour. In particular, it has been shown that the energy dissipation provided by the whole joint can be obtained as the sum of the energy dissipations due to the single joint components, provided that the joint components are properly identified and their cyclic response is properly measured. This result is very important because it testifies to the applicability of the component approach to the prediction of the joint behaviour under cyclic loads as well [13]. Within the above research programme, due to the significant advantages from the reparability point of view, double split tee (DST) connections were recognized as an interesting solution that can be used in dissipative semi-continuous MRFs. In fact, DST connections can be easily repaired after destructive seismic events and allow joint rotational behaviour (i.e. the rotational stiffness, strength and plastic rotation supply) to govern by fixing the bolt diameter properly and by simply calibrating three geometrical parameters: the width and thickness of the T-stub flange plate and the distance between the bolts and the plastic hinge arising at the stem-to-flange connection [20, 21]. On the other hand, joints involving bolted components in the plastic range also entail several disadvantages. First of all, even though experimental studies have demonstrated that bolted components are able to dissipate significant amounts of energy, it should be recognized that their hysteretic behaviour is less dissipative compared with other joint typologies or the cyclic response of steel H-shaped sections. This is mainly due to contact and pinching phenomena, which usually lead to the quick degradation of strength and stiffness of the tee elements. For this reason, on the one hand, the use of hourglass-shaped T-stub flanges has been recently proposed [15], where, in other words, the dissipative capacity of classic tee elements has been improved by applying the same concepts to the T-stub flanges as are usually developed to design hysteretic metallic dampers, such as ADAS devices [1, 9, 23, 24]. On the other hand, an innovative approach aimed at enhancing the dissipation capacity

of classic rectangular T-stubs by using friction pads has also been proposed [14], with the primary aim of joint damage prevention. This latter approach, which can be considered as an innovative application of the seismic protection strategy based on supplementary energy dissipation, is presented here. The main scope of the work is to investigate the possibility of designing dissipative DST connections by exploiting the cyclic behaviour of friction materials and by simultaneously preventing joint damage. In particular, the aim of the two innovative DST joints shown below is to dissipate the seismic input energy by means of the slippage of the stems of the tees on a friction pad, which is interposed between the tee stems and the beam flanges. In this way, under seismic loading conditions, the structural elements do not undergo any damage provided that rigorous design procedures for failure mode control are applied [16, 18]. However, energy dissipation is assured by the alternate movement of the tee stems on the friction pads, which are preloaded by means of high-strength bolts. Therefore, the present paper proposes adopting a new type of dissipative beam-to-column joint, namely the dissipative DST connection with friction pads, in the seismic design of semi-continuous MRFs. Its behaviour is investigated by means of experimental tests under displacement control in cyclic loading conditions.

2 Experimental tests on friction materials To start with, in order to investigate the frictional properties of different interfaces to be used in DST friction joints, a sub-assemblage comprising two layers of friction material or metal located between three steel plates made of grade S275JR steel was assembled at the Materials & Structures Laboratory of Salerno University (Fig. 1). In order to allow the relative movement of the steel plates on the interposed friction material, one of the inner plates has slotted holes. Conversely, the other inner plate and the two outer plates have round holes. The clamping force was applied by 16 preloaded M20 grade10.9 bolts, and the holes were drilled with a ∅ 21 mm drill bit. With the aim of evaluating the magnitude of the friction coefficient, several different layouts of the sub-assemblage were considered, varying four parameters: the interface, the tightening torque, the number of tightened bolts and the type of bolt washer. The frictional properties of the following five different interfaces have been evaluated (Fig. 2): – Steel on steel – Brass on steel – Friction material M0 on steel – Friction material M1 on steel – Friction material M2 on steel In particular, two different types of washer were employed: circular flat steel washers in the first part of the

Fig. 1. Scheme of the sub-assemblage tested

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 12

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

Fig. 2. Force–displacement curves of interfaces

experimental programme, a packet of steel disc springs interposed between bolt head and steel plate in the second part of the work (Fig. 3). In addition, the experiments were carried out by varying the bolt tightening level in the range between 200 and 500 Nm, thus obtaining different values for the clamping force acting on the sliding surfaces. The main goal of the experimental programme is to obtain the friction coefficients, both static and kinetic, of the materials investigated for normal force values varying in a

range leading to sliding forces suitable for structural applications and for velocity values compatible with seismic engineering applications. In addition, the experimental work is also devoted to evaluating the variation in the sliding force as the number of cycles of the applied loading history increase. In fact, as already demonstrated by Pall and Marsh [19], an interface subjected to cyclic loading conditions can essentially respond in one of two ways. The first type of response provides a monotonically softening be-

haviour; in this case the maximum sliding load is reached during the first cycle, whereas in all subsequent cycles only degradation behaviour is expected. The second type of response is characterized by three phases: first, a hardening response, then a steadystate phase and, finally, a load degradation phase. The tests were carried out with a Schenck Hydropuls S56 universal testing machine. The testing apparatus comprised a hydraulic piston (loading capacity ± 630 kN, maximum stroke

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 13

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

± 125 mm) and a self-balanced steel frame used to counteract the axial loadings. In order to measure the axial displacements, the testing device is equipped with an LVDT, whereas the tension/compression loads are measured by a load cell. The cyclic tests were carried out under displacement control for different displacement amplitudes at a frequency of 0.25 Hz (Figs. 2 and 3). The average values of the static and kinetic coefficients of friction for all the tests were determined with the following expression:

F m n Nb

(1)

where m number of surfaces in contact n number of bolts Nb bolt preloading force F sliding force The values obtained are given in Table 1. Table 1. Values of friction coefficients Interface

mstatic

mdynamic

Steel on steel Brass on steel M0 on steel M1 on steel M2 on steel

0.173 0.097 0.254 0.201 0.158

0.351 0.200 0.254 0.201 0.180

Concerning the behaviour exhibited by the five materials under cyclic loads, the main results of the experimental programme can be summarized as follows: – The steel on steel interface exhibited a high coefficient of friction, but with an unstable behaviour initially characterized by a significant hardening behaviour and, subsequently, by a quick softening behaviour. – The brass on steel interface exhibited a significant hardening behaviour with a low static friction coefficient. – Material M0, a rubber-based material developed for automotive applications, exhibited a very stable behaviour and high energy dissipation capacity, also under high preloading values. – Material M1, a rubber-based material developed for electrical machines, exhibited a cyclic behaviour with some pinching and a low fric-

Fig. 3. Tested specimens

tion coefficient plus a quick degradation behaviour. – Material M2, a hard rubber-based material developed for applications where low wear is necessary, developed a quite low friction coefficient but exhibited a very stable behaviour and high dissipation capacity.

3 Experimental tests on DST joints with friction pads Starting from the component behaviour, i.e. the test results of the sub-assemblage with friction pads presented in the previous section, it was possible to design dissipative DST connections with friction pads, i.e. with interposed layers of friction material between the beam flanges and the stems of the tee elements. The cyclic behaviour of the

proposed innovative DST connections with friction pads can also be compared with the energy dissipation capacity of a traditional double split tee connection tested in a previous work [12], namely TS-CYC 04. Experimental tests were carried out at the Materials & Structures Laboratory of Salerno University. The testing equipment was that already adopted to test traditional beam-to-column connections [12]. Two steel hinges, designed to resist shear actions up to 2000 kN and bolted to the base sleigh, were used to connect the specimens to the reacting system. The specimen is assembled with the column (HEB 200) in the horizontal position, connected to the hinges, and the beam (IPE 270) in the vertical position (Fig. 4). The loads

Fig. 4. Experimental testing equipment

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 14

04.02.13 11:56

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

were applied by means of two different hydraulic actuators. The first one is a MTS 243.60 actuator with a load capacity of 1000 kN in compression and 650 kN in tension and a piston stroke of ± 125 mm, which was used to apply, under force control, the axial load of 630 kN in the column. The second actuator is a MTS 243.35 with a load capacity of 250 kN in both tension and compression and a piston stroke of ± 500 mm, which was used to apply, under displacement control, the desired displacement history at the beam end. The loading history was defined according to ANSI-AISC 341-10 [2]. Many parameters were monitored and acquired during the tests in order to obtain the test machine history imposed by the top actuator and the displacements of the different joint components. With the aim of evaluating the beam end displacements due to the beam-to-column joint rotation only, the displacements measured by the LVDT-equipped MTS 243.35 actuator were corrected by subtracting the elastic contribution due to the beam and column flexural deformability according to the following relationship [12]:

δ j = δ T3 −

FL3b FLcL2b − × 3EI b 12EIc

2   L  6a  c  + ×    Lc + 2a  Lc + 2a   

(2)

where Ib, Ic beam and column inertia moments Lc column length Lb beam length a length of rigid parts due to steel hinges The experimental tests carried out so far concern four specimens (Fig. 5): – TSJ-M1-460-CYC08, TSJ-M2-460CYC09 and TSJ-B-460-CYC11, which are three double split tee connections. The first two are equipped with layers of friction material, namely M1 and M2, and the third one with a brass plate interposed between the tee stems and the beam flanges. The slipping interfaces were clamped by eight M20 grade 10.9 bolts tightened with a torque of 460 Nm. In order to al-

Fig. 5. Geometrical detail and photo of joint being tested

low the relative movement between the stems of the T-stubs and the beam flanges, two slotted holes were provided in the tee stems. The slots were designed to allow a maximum rotation of 70 mrad. The flanges of the T-stubs are fastened to the column flanges by means of eight M27 grade 10.9 bolts located in holes drilled with a ∅ 30 mm drill bit. – TSJ-M2-DS-460-CYC010, which is a double split tee connection with the same characteristics of the other tested joints but with two disc springs interposed between the bolt nut and the beam flange. The identity tag of each test specimen uniquely identifies the connection detail. In particular, the meaning of the letters is: 1 – Joint typology, i.e. tee stub joint (TSJ) 2 – Friction interface, i.e. friction material M1, friction material M2 and brass (B) 3 – Washer typology, if different from the standard flat washer, i.e. disc spring (DS) 4 – Bolt tightening level 5 – Test number, i.e. CYC number

4 Cyclic behaviour of specimens As already mentioned, the main goal of the work presented here is to provide an innovative approach to preventing structural damage in the dissipative zones of MRFs where the main source of energy dissipation is due to beam end damage in the case of fullstrength connections and damage to connecting plate elements in par-

tial-strength connections. To this end, the proposed beam-to-column joint typology is detailed in order to dissipate the seismic input energy through the slippage of the friction material interposed between T-stub stem and beam flange. In particular, hierarchy criteria at the level of the joint components can be established to assure the desired connection behaviour. Therefore, starting from the design bending moment (100 kNm) established with the aim of developing the same degree of flexural strength of the traditional joints already tested in previous research [12], all the remaining joint components (i.e. T-stub flanges, bolts and column panel zone) have been designed to assure an adequate overstrength with respect to the friction resistance. In particular, the friction interface has been designed according to Eq. (1), considering that the force to be transmitted is simply obtained as the ratio between design bending moment and lever arm. Therefore, the desired friction resistance at the sliding interface has been obtained by properly fixing the number of bolts and the tightening force of the bolts fastening the tee stems to the beam flanges. In perfect agreement with the adopted design criteria, none of the experimental tests showed any damage to the joint components, indicating the involvement of the friction pads only. Therefore, the most important result of the experimental programme is that the proposed connection typology can be subjected to repeated cyclic rotation histories, i.e. to repeated earthquakes, by only replacing the friction pads and by tightening

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 15

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano 路 Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

the bolts again to reach the desired preloading level. In addition, the rotation capacity can be easily calibrated by simply determining the length of the slots where the bolts are located. The results of the experimental programme for DST connections with friction pads are in line with the results found by testing the friction component. As expected, this work shows that the cyclic behaviour of the joint is mainly governed by the cyclic behaviour of the weakest joint component (i.e. the friction component in the cases examined). In fact, as verified during the test TSJ-M1-460-CYC08, where material M1 was adopted, the response of the joint is very similar to that discovered during the uniaxial tests investigating the friction interface behaviour. A sig-

nificant pinching and strength degradation behaviour is seen, after which the design resistance of 100 kNm is reached (Fig. 6). This is also due to the premature fracture of the friction pad, which was not observed in component testing. For this reason, this material will be excluded from the forthcoming developments of this research activity. In the case of friction material M2 (TSJ-M2-460-CYC09), a stable cyclic response with a hardening behaviour due to the increase in local stresses caused by the beam rotation and by the rotational stiffness due to the bending of the tee stems has been indicated (Fig. 6). Furthermore, the results show that a minor strength and stiffness degradation begins at high rotation amplitudes, probably due to

the consumption of the friction pads during the sliding motion. The test on brass friction pads, TSJ-B-460-CYC11, also exhibited good behaviour in terms of the shape of the cyclic response. In fact, the cycles obtained are very stable, also for high plastic rotation values. Nevertheless, a bending moment value lower than the design value of 100 kNm was obtained because of poor friction resistance. This result can be justified on the basis of the results obtained from component testing. In fact, in the case of a brass-on-steel interface (Table 1), the value of the static friction coefficient is much lower than the dynamic one and, as a consequence, a bending moment lower than the one expected has been obtained (Fig. 6). For this reason and considering the high cost of this

Fig. 6. Hysteretic curves of joints tested

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 16

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

Fig. 7. Cyclic envelopes and energy dissipation of DST connections tested

material, the use of brass for friction pads will be excluded from the forthcoming research developments. Finally, in order to reduce the problems related to the consumption of the friction material observed during test TSJ-M2-460-CYC09, another test, namely TSJ-M2-DS-460-CYC10, with the same layout but adopting disc springs interposed between the bolt head and the tee web plate, was carried out. Such a washer type is a high-resistance cone-shaped annular steel disc spring that flattens when compressed and returns to its original shape once the compression is relieved. In this way, the wearing of the friction material, which would lead to partial loss of the bolt preload, is compensated for by the action of the disc spring, which restores the force by maintaining the bolt shaft in tension. In fact, the results of test TSJ-M2-460CYC10 have demonstrated the effectiveness of the disc springs adopted. Therefore, higher dissipation capacity and lower strength and stiffness degradation was obtained (Fig. 6). In addition, in order to compare the cyclic behaviour of DST connections with friction pads with the behaviour of a traditional DST partial-strength joint dissipating in the bolted components and characterized by the same resistance, reference has been made to test TS-CYC04 (Fig. 6) [12]. In particular, the envelopes of the cyclic moment–rotation curves are shown in Fig. 7 for all the specimens tested, both innovative and traditional. It can be seen that the bending moment corresponding to the knee of the curve, corresponding to the design value of the joint resistance, is similar

for all the tests with friction materials, but the post-elastic behaviours obtained are quite different with respect to traditional DST connections. In fact, compared with the case of joint TS-CYC04, friction DST joints do not exhibit significant hardening behaviour whose magnitude is limited to the effects due to the bending of the T-stub stems. With reference to tests TS-M2460-CYC09 and TS-M2-DS-460CYC10, it is worth noting that the hysteresis cycles are wide and stable with no pinching. This is the reason why the joints, despite the reduced hardening behaviour, are able to dissipate more energy than connection TSCYC04 (Fig. 7).

5 Conclusions The possibility of enhancing the cyclic behaviour of traditional DST joints dissipating the seismic input energy in bolted components has been analysed in this paper. In particular, the cyclic rotational response of four double split friction tee stub beam-to-column joints adopting different friction materials has been investigated. The response in terms of energy dissipation and the shape of the hysteresis loops of the proposed structural connection details have been compared with those of a traditional DST joint tested in a recent programme of experiments. The results obtained are very encouraging, confirming the merit of the proposed approach. In particular, all the experimental tests have confirmed that the strategy of adopting friction pads between the components of bolted connections can be effective for the ambitious goal

of damage prevention. This is because the proposed DST connection is able to withstand repeated cyclic rotation histories, i.e. repeated earthquakes, by simply replacing the friction pads and retightening the connecting bolts.

Acknowledgements This work was partly supported with the research grant DPC-RELUIS 20102013. References [1] Aiken, I., Nims, D., Whittaker, A., Kelly, J.: Testing of Passive Energy Dissipation Systems. Earthquake Spectra, 9(3), 1993. [2] ANSI/AISC 341-10, American National Standard: Seismic Provisions for Structural Steel Buildings. 22 June 2010. American Institute of Steel Construction, Chicago, Illinois, USA. [3] ANSI/AISC 358-10. American National Standard: Prequalified Connections for Special and Intermediate Steel Moment Frames for Seismic Applications. Including supplement No. 1: ANSI/AISC 358s1-11. American Institute of Steel Construction, Chicago, Illinois, USA. [4] Astaneh-Asl, A.: Experimental Investigation of Tee Framing Connection. AISC, 1987. [5] Bernuzzi, C., Zandonini, R., Zanon, P.: Experimental analysis and modelling of semi-rigid steel joints under cyclic reversal loading. Journal of Constructional Steel Research, 2, 1996, pp. 95–123. [6] CEN, 2005a, Eurocode 8: Design of structures for earthquake resistance – Part 1: General rules, seismic actions and rules for buildings. [7] CEN, 2005b, Eurocode 3: Design of steel structures – Part 1-1: General rules and rules for buildings.

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 17

01.02.13 08:48

M. Latour/V. Piluso/G. Rizzano · Experimental behaviour of friction T-stub beam-to-column joints under cyclic loads

[8] CEN, 2005c, Eurocode 3: Design of steel structures – Part 1-8: Design of joints. [9] Christopoulos, C., Filiatrault, A.: Principles of Passive Supplemental Damping and Seismic Isolation. IUSS PRESS. Pavia 2000, Italy. [10] Faella, C., Montuori, R., Piluso, V., Rizzano, G.: Failure mode control: economy of semi-rigid frames. In: Proceedings of XI European Conference on Earthquake Engineering. Paris, 1998. [11] Faella, C., Piluso, V., Rizzano, G.: Structural Steel Semirigid Connections, CRC Press, Boca Raton, Ann Arbor, London/Tokyo, 1999. [12] Iannone, F., Latour, M., Piluso, V., Rizzano, G.: Experimental Analysis of Bolted Steel Beam-to-Column Connections: Component Identification. Journal of Earthquake Engineering, vol. 15, No. 2, Feb 2011, pp. 214–244(31). [13] Latour, M., Piluso, V., Rizzano, G.: Cyclic Modeling of Bolted Beam-toColumn Connections: Component Approach. Journal of Earthquake Engineering, 15(4), 2011, pp. 537–563. [14] Latour, M., Piluso, V., Rizzano, G.: Experimental Analysis of Innovative Dissipative Bolted Double Split Tee

Beam-to-column Connections. Steel Construction, vol. 4, No. 2, June 2011, pp. 53–64. [15] Latour, M, Rizzano, G.: Experimental Behaviour and Mechanical Modeling of Dissipative T-Stub Connections. Journal of Structural Engineering, 138(2), 2012, pp. 170–182. [16] Longo, A., Montuori, R., Piluso, V.: Theory of Plastic Mechanism Control of Dissipative Truss Moment Frames. Engineering Structures. 37 (2012), pp. 63–75. [17] Mazzolani, F. M., Piluso, V.: Theory and Design of Seismic Resistant Steel Frames, E&FN Spon, an imprint of Chapman & Hall, 1st ed., 1996. [18] Mazzolani, F. M., Piluso, V.: Plastic Design of Seismic Resistant Steel Frames. Earthquake Engineering and Structural Dynamics, vol. 26, No. 2 (1997), pp. 167–191. [19] Pall, A., Marsh, C.: Response of Friction Damped Braced Frames. Journal of the Structural Division, 108(6), 1981, pp.1313–1323. [20] Piluso, V., Faella , C., Rizzano, G.: Ultimate behavior of bolted T-stubs. Part I: Theoretical model. Journal of Structural Engineering ASCE, 127(6), 2001, pp. 686–693.

[21] Piluso, V., Faella , C., Rizzano, G.: Ultimate Behaviour of Bolted T-stubs. Part II. Experimental Analysis, Journal of Structural Engineering, ASCE, vol. 127, No. 6, 2001, pp. 694–704. [22] Piluso, V., Rizzano, G.: Experimental Analysis and modelling of bolted T-stubs under cyclic loads. Journal of Constructional Steel Research, 64, 2008, pp. 655–669. [23] Soong, T., Spencer Jr., B.: Supplemental Energy Dissipation: State-of-the-Art and State-of-the-Practice. Engineering Structures, 24, 2002, pp. 243–259. [24] Whittaker, A., Bertero, V., Alonso, J., Thompson, C.: UCB/EERC-89/02 Earthquake Simulator Testing of Steel Plate Added Damping and Stiffness Elements. Berkeley: College of Engineering University of California, 1989. Keywords: T-stub joints; friction; beam-tocolumn joints; experimental; cyclic

Authors: Massimo Latour, mlatour@unisa.it Vincenzo Piluso, v.piluso@unisa.it Gianvittorio Rizzano, g.rizzano@unisa.it DICIV – Department of Civil Engineering, University of Salerno, Italy

People Professor Jerzy Ziółko – Doctor Honoris Causa On November 28th, 2012 – on the eve of his 78th birthday, the University of Technology & Life Sciences in Bydgoszcz

(Poland) honoured Professor Jerzy Ziółko, the Editorial Board Member of “Steel Construction – Design and Research”, by her doctor honoris causa dignity. Professor Jerzy Ziółko is well known, in his home country and abroad, as a

Prof. Jerzy Ziółko received the doctor honoris causa dignity of the University of Technology & Liefe Sciences in Bydgoszcz, Poland

prominent specialist of steel constructions, the tanks for fluid fuels – in particular. Born 1934 in Radom, he studied civil engineering at the Gdansk University of Technology (GUT) 1952–1957, and worked as engineer for the “Mostostal” enterprise in 1957–1963. Later, he joined GUT to perform teaching, research, and practical engineering. Accordingly, he has been promoted to doctor of engi. neering (dr inz.) and habilitated doctor . of engineering (dr hab. inz.). In 1979 he has been granted the scientific title and the position of the professor. For further details of his career of life, please, see “Stahlbau” 78 (2009), 11, 879–880. Here, it should be mentioned only that he is author of ca. 200 publications – including 14 books, partly as co-author and issued also abroad. He supervised and promoted 15 doctor engineers. He is a long-standing Member of the Committee for Civil Engineering of the Polish Academy of Sciences. In 2004, he became the Foreign Member of the Ukrainian Academy of Construction. Zbigniew Cywin´ski

Steel Construction 6 (2013), No. 1

08_011-018_Piluso_(1-7)_cs6.indd 18

01.02.13 10:08

Articles José Henriques Luís Simões da Silva* Isabel Valente

DOI: 10.1002/stco.201300003

Design model for composite beam-to-reinforced concrete wall joints A design model for composite beam-to-reinforced concrete wall joints is presented and discussed in this paper. The model proposed is the component method extended to this type of joint. The characterization of the active components is therefore performed in terms of force-deformation curves. In this type of joint, special attention is paid to the steel-concrete connection where “new” components, not covered in EN 1993-1-8, are activated. The application of the model allows the designer to obtain the joint properties in terms of the moment-rotation curve. The accuracy of the proposed model is verified by comparing it with available experimental and numerical results. The latter were developed in the FE program ABAQUS and previously validated by experimental results.

1 Introduction Many office- and car park-type buildings use a combination of reinforced concrete structural walls and steel and/or composite members. In such structural systems, the design of the joints is a challenge due to the absence of a global approach. Designers are faced with a problem that requires knowledge of reinforced concrete, anchorages in concrete and steel/ composite behaviour. Owing to the different design philosophies, especially with regard to the joints, no unified approach is currently available in the Eurocodes. The component method is a consensus approach for the design of steel and composite joints which has proved to be efficient. Therefore, a design model extending the scope of the component method to steel-to-concrete, beam-to-wall joints is proposed in this paper. To address the problem, a composite beam-to-reinforced concrete wall joint, tested experimentally within the RFCS research project “InFaSo” [1], was chosen. The joint configuration under analysis was developed to provide a semi-continuous solution, allowing transfer of bending moments between the supported and supporting members. The joint depicted in Fig. 1 may be divided into two zones: I) upper zone, connection between reinforced concrete slab and wall II) bottom zone, connection between steel beam and reinforced concrete wall

In the upper zone, the connection is achieved by extending and anchoring the longitudinal reinforcing bars of the slab (a) into the wall. Slab and wall are expected to be concreted in separate stages and therefore the connection between these members is provided by the longitudinal reinforcing bars only. In the bottom zone, fastening technology is used to connect the steel beam to the reinforced concrete wall. Thus, a steel plate (b) is anchored to the reinforced concrete wall using headed anchors (c). The plate is embedded in the concrete wall flush with the face of the wall. A steel bracket (d) is then welded to the external face of the plate. A second plate (e) is also welded to this steel bracket to create a “nose”. The steel beam with an extended end plate (f) sits on the steel bracket, and the extended part of the end plate and steel bracket “nose” form an interlocked connection to prevent the steel beam slipping off the steel bracket. A contact plate (g) is placed between the beam end plate and the anchor plate at the level of the beam bottom flange. According to the structural demands, the joint configuration can cover a wide range of design load combinations (M-V-N) without the need for significant modifications to the connection between the steel and the concrete parts. The versatility of the joint is illustrated in Fig. 2. Three working situations are possible: I) semi-continuous, with medium/high capacity for hogging bending moment, shear and axial compression II) pinned, for high shear and axial compression III) pinned, for high shear and axial tension

a b c d e f g

– – – – – – –

Longitudinal reinforcement bars Ancor plate Headed anchors Steel bracket Steel plate welded to steel bracket Steel beam end plate Steel contact plate

Fig. 1. Composite beam-to-reinforced concrete wall joint studied in [1]

09_019-026_da_Silva (1-3)_cs6.indd 19

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

Fig. 3. Stress flows in the semi-continuous joint under bending moment and shear

Fig. 2. Versatility of the steel-to-concrete joint for different loading conditions

According to the detailing of Fig. 1, the weakness of the “nose” system means that the sagging bending moment capacity is very limited and heavily dependent on the resistance of the “nose”. For the same reason, the resistance to tensile loading is also reduced. Therefore, the use of this type of detail in conjunction with cyclic loading, e. g. seismic action, is restricted. Pinned behaviour of the joint is very easily obtained by removing the connection between the slab and the wall. Consequently, in terms of erection, this is a very efficient solution; however, for the above reasons, the joint should not be subjected to axial tension. Whenever this is a requirement, adding a fin plate as shown in Fig. 2(iii) provides a straightforward solution. In this case the tension capacity is improved and due to the symmetry of the joint, cyclic loading can be accommodated. Only the semi-continuous joint solution subjected to hogging bending moment is analysed in this paper.

2 Sources of joint deformability and joint model Understanding the behaviour of the joint under bending moment and shear force requires us to identify the mechanics of the joint. The assumed stress flows are shown schematically in Fig. 3. Accordingly, in the upper zone, only tension is transferred via the longitudinal reinforcement. Further, in this region there is no shear, and no tension is assumed to be transferred through the concrete, from the slab to the wall, as the small bond developed is neglected. In the bottom zone the shear load is transferred from the steel beam to the reinforced concrete wall according to the following path: a) from the beam end plate to the steel bracket through contact pressure b) from the anchor plate to the reinforced concrete wall through friction between the plate and the concrete and between the shafts of the headed anchors and the concrete through bearing

In the bottom zone there is also a transfer of compression to the reinforced concrete wall through the contact plate between the beam end plate and the anchor plate. Then, in the reinforced concrete wall the high tension and compression loads introduced by the joint flow to the supports. According to the stress flows described, corresponding to a hogging bending moment, the active components are identified and listed in Table 1 and their locations are shown in Fig. 4a. Please note that the numbering of the joint components used here differs from the usual numbering proposed in [2]. Components 7, 8, 9 and 10 should not control the behaviour of the joint as their activation only results from the out-of-plane deformation of the bottom and top edges of the anchor plate in compression. The anchor row at the bottom part is activated in tension, due to the outof-plane deformation, and acts similarly to a prying force. Component 11, the “joint link”, represents the equilibrium of stresses in the reinforced concrete wall zone adjacent to the joint.

Table 1. List of active components in composite beam-toreinforced concrete wall joint subjected to a hogging bending moment Component ID

Basic joint component

Type/Zone

Longitudinal steel reinforcement in slab

tension

Slip of composite beam

tension

Beam web and flange

compression

Steel contact plate

compression

Anchor plate in bending under compression

bending/ compression

Concrete

compression

Headed anchor in tension

tension

Concrete cone

tension

Pull-out of anchor

tension

Anchor plate in bending under tension

bending/ tension

Joint link

tension and compression

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 20

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

3 Characterization of activated joint components 3.1 Components in tension zone

a) Location of the identified joint components

b) Joint component model

Fig. 4. Application of the component method to a composite beam-to-reinforced concrete wall joint subjected to a hogging bending moment

A representative spring and rigid link model for the components identified is illustrated in Fig. 4b: three groups of springs are separated by two vertical rigid bars. The rigid bars prevent the interplay between tension and compression components, simplifying the joint assembly. Another simplification is introduced by considering a single spring to represent the joint link. Concerning the tension springs, it is assumed that slip and the longitudinal reinforcement are at the same level although slip is observed at the steel beam/concrete slab interface. At the bottom part of the joint in this model, rotational springs (5) are considered in the anchor plate to represent the bending of this plate. In a simplified model, the behaviour of these rotational springs, as well as the effect of the bottom row of anchors, should be incorporated into an equivalent translational spring representing the contribution of the anchor plate to the joint response. Each group of components is discussed in the next section.

In the case of full interaction being achieved between the slab and the steel beam, the longitudinal reinforcement in tension limits the resistance of the tension zone of the joint. This component is common in composite joints where the longitudinal reinforcement is continuous within the joint or its anchorage is assured. In EN 1994-1-1 [3], each layer of longitudinal reinforcement is considered as an additional row of bolts contributing to the resistance of the joint. The longitudinal reinforcement within the effective width of the concrete slab is assumed to be stressed up to its yield strength. In terms of deformation, a stiffness coefficient is provided by the code which takes the following into account: I) the configuration of the joint, double- or single-sided II) the depth of the column III) the area of longitudinal reinforcement within the effective width of the concrete flange IV) the loading on the right and left sides, balanced or unbalanced bending moment No guidance is given with regard to estimating the deformation capacity. Sufficient deformation capacity to allow a plastic distribution of forces should be available if the ductility class of the reinforcing bars is B or C according to EN 1992-1-1 [4]. A more sophisticated model of this component can be found in [5], where the behaviour of the longitudinal reinforcement is modelled taking into account the embedment in concrete and the resistance increases as far as the ultimate strength of the steel. The component is modelled by means of a multi-linear force–displacement curve

Table 2. Analytical expressions for longitudinal reinforcement component Reference

EN 1994-1-1 [3]

Expression Resistance

Fsy = σ y A sr

Stiffness coefficient

ksr =

Deformation capacity

not given

A sr 3,6h

Fs = σ sr,i A sr where Resistance

σ sr1 =

fctm kc ρ

 Es  1 + ρ  Ec  

σ srn = 1,3σ sr1

(

Δ ≤ Δ sry : Δ = ε h + L t

ECCS publication No. 109 [5]

Deformation

εsr1 =

σ sr1 − Δεsr Es

Δεsr =

fctm kc Esρ

)

ρ ≥ 0,8 % : Δ sru = 2L t εsrmu

( ) ρ ≥ 0,8 % and a > L t : Δ sru = ( h + L t ) εsrmu + ( a − L t ) εsrmy ρ ≥ 0,8 % and a < L t : Δ sru = h + L t εsrmu

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 21

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

load is then assumed to be resisted by the next stud deforming elastically until its plastic resistance is reached. Further load is then carried by the next stud and so on. The deformation capacity of the component is then limited by the deformation capacity of the shear connection between the concrete slab and the steel beam. In EN 1994-1-1 [3] the contribution of the slip of the composite beam is taken into account by multiplying the stiffness coefficient of the longitudinal steel reinforcement in tension by a slip factor kslip.

3.2 Components in the compression zone

Fig. 5. Behaviour of the component “longitudinal steel reinforcing bar in tension”

with hardening. This model allows the designer to estimate the deformation at ultimate resistance. The deformation is then assumed to be the deformation capacity of the component. Table 2 summarizes the analytical expressions for both models. Fig. 5 illustrates the force–deformation curves characterizing the behaviour of the components according to these models. In the ECCS model [5], the initial range is very stiff as the concrete is uncracked. Then, as cracks form in the concrete, a loss of stiffness is observed up until the cracking stabilizes. At this stage the response of the longitudinal reinforcing bar recovers the proportionality between stress and strain for the bare steel bar up to yield strength. Finally, the ultimate resistance is achieved assuming that the bars may be stressed up to their ultimate strength. In the Eurocode model, linear elastic behaviour is considered up to yielding of the longitudinal reinforcing bar. In this joint the composite beam is designed assuming full interaction between the steel beam and the RC slab; therefore, no limitation to the joint resistance is expected from component 2, slip of composite beam. Concerning the deformation of this component, as verified in [6], a small contribution to the joint rotation may be observed. According to [7], the slip at the connection depends on the stud nearest to the face of the wall. As the load increases, this stud provides resistance to slip until it becomes plastic. Additional

a) Idealized mechanical model

In the compression zone the beam web and flange in compression and the steel contact in compression are components already covered by EN 1993-1-8 [2] and EN 1994-1-1 [3]. Furthermore, according to the scope of the experimental tests [1], the contribution of these components to the joint response was limited to the elastic range. The reader is therefore referred to [2] and [3] for the characterization of these components. Concerning the anchor plate in compression, this connection introduces the anchorage in the concrete into the problem. As the main loading is compression, the anchorage is not fully exploited. In order to reproduce its behaviour, a sophisticated model of the anchor plate in compression is under development. As illustrated in Fig. 4, several components are activated, carrying tension, compression and bending. Owing to the similarities between the problems, the model under development is an adapted version of that used by Guisse et al. [8] for column bases. In the absence of specific tests on the anchor plate in compression, the model is based on numerical investigations. Fig. 6 depicts the idealized mechanical model and the reference numerical model. The steel/concrete contact is reproduced by considering a series of extensional springs that can only be activated in compression. Owing to the deformation of the anchor plate, the row of anchors on the unloaded side is activated in tension and increases the compressive resistance and stiffness of the anchor plate. For the row of anchors on the unloaded side, a single extensional spring concentrates the response of three components: I) anchor shaft in tension II) concrete cone failure III) headed anchor pull-out failure

b) Reference numerical model

Fig. 6. Anchor plate connection

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 22

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

Table 3. Analytical characterization of the components relevant for the anchor plate in compression Component

Reference

Expression

Guisse et al. [8]

f − E ε j c c2 Fi =   ε2 c2 

Resistance 6 Deformation

δ i = εi hc,eq

Resistance

 πd 2  f Nst = n   4  y

EN 1993-1-8 [2] δ st,y =

Deformation

2  δ   d   i i   + Ec   A c,i  hc,eq   hc,eq   

Nst kst

 πd 2  Ea   4  where kst = hef

Resistance

CEN-TS [9]

 A c,N  Nc =  0  ψ mN0c where N0c = 16,8 0.95fck,cube h1.5 ef  A c,N 

Deformation

–

rigid

Resistance

CEN-TS [9]

NPO = 11fck

Deformation

Furche [10]

(

π d 2h − d 2

9 δ PO

k k = αp a A C1

M y = fy Resistance 5 and 10

conventional

Deformation

Three rotational springs are then considered to reproduce the bending of the plate according to its deformation. The location of these springs is based on the numerical observations (see Fig. 6b). The properties of these components are given in Table 3. For the parameters involved, please refer to the references given in the table. Fig. 7 shows the comparison between the results of the numerical and the analytical models. The results are given in terms of load applied to the anchor plate and deformation in the direction of the load at its point of application. Despite the good accuracy of the analytical model, its full validity has yet to be established, as a parametric study has shown some discrepancies between the models. The final calibrated model should be presented in [11]. The above model aims to reproduce accurately the behaviour of the anchor plate in compression. However, it is perhaps too complex for design purposes. Thus, simplified modelling of the anchor plate in compression is envisaged. Again, owing to the similarities between the problems, a modified version of the T-stub in compression [2] is foreseen as follows: – For resistance and stiffness, the β factor is set to 1 because the use of grout between plate and concrete is not expected.

Mpl = fy φu =

)

4   N   0.95f n A  h ck,cube 

2 bap t ap

6 2 bap t ap

2 × 0.15 t ap

– For stiffness, an exact value of the bearing width c has been determined according to [12] instead of the approximation given in the EN 1993-1-8 [2]. Thus, c is taken to be 1.4t instead of 1.25 t. Consequently, components 5 to 10 are replaced in the joint component model, shown in Fig. 4, by a single equivalent spring representing the T-stub in compression. This is the model used in section 4.

Fig. 7. Comparison of results of analytical and numerical models

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 23

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

3.3 Joint link The joint link is a component to consider the resistance and deformation of the reinforced concrete wall in the zone adjacent to the joint. The loading on this member from the part of the structure above may affect this component. However, only the joint loading is considered in the present study. As for the anchor plate under compression, no specific experimental tests have been performed to analyse this part of the joint. Therefore, a simplified analysis has been performed numerically. Owing to the nature of this part of the joint (reinforced concrete), the model is based on the strutand-tie method commonly implemented in the analysis of reinforced concrete joints. The problem is three-dimensional, which increases its complexity as the tension is introduced with a larger width than the compression, which may be assumed as concentrated within an equivalent dimension of the anchor plate (equivalent rigid plate as considered in T-stub in compression). Thus, a numerical model considering only the reinforced concrete wall and an elastic response of the material has been tested to identify the flow of principal stresses. These show that compression stresses flow from the hook of the longitudinal reinforcing bar to the anchor plate. The strut-and-tie model (STM) depicted in Fig. 8a is idealized in this way. Subsequently, in order to contemplate the evaluation of the deformation of the joint, a diagonal spring is idealized to model the diagonal concrete strut in compression, as illustrated in Fig. 8b. The ties correspond to the longitudinal steel reinforcing bars already considered in the joint model. The properties of this diagonal spring are determined as follows: – Resistance is obtained based on the strut and node dimensions and admissible stresses within these elements. The node at the anchor plate is in a triaxial state of compression. Therefore, high stresses are attained (confinement effect). Concerning the strut, a “bottle shape” is identified. Because of the 3D nature, stresses tend to spread between nodes. Given the dimensions of the wall (infinite width), the strut dimensions should not be critical to the joint. As

stated in [13], as nodes represent “bottlenecks” for stresses, it can be assumed that the concrete strut is safe if the node failure criterion is satisfied. Thus, the node at the hook of the bar is assumed to be the critical component in the joint link. The resistance of the spring is then obtained according to the dimensions of this node and the admissible stresses at the node. The admissible stresses are defined according to EN 1992-1-1 [4]. – With respect to the deformation, the problem is more complex because the strain field within the diagonal strut is highly variable. However, several numerical calculations [11] considering geometrical variations (wall thickness, beam depth, bend radius) have revealed that the shape of the force–deformation curve is independent of these variables. Thus, as a simplification, a mathematical equation is proposed to approximate the horizontal component of the deformation (in mm) of the joint link as a function of the horizontal load on the joint (Fj,h in kN), as expressed in Eq. (1).

(

)

2 + 7.47E −5 F d j,h = 6.48E −8 Fj,h j,h cosθ

(1)

Table 4 gives the admissible stresses for nodes according to EN 1992-1-1 [4]. Node 1, illustrated in Fig. 9, is characterized by the hook of the longitudinal reinforcing bar. The dimension shown is assumed to be as defined in the CEB Model Code [14]. Concerning the width of the node, based on a numerical study [11], Eq. (2) was derived to determine an effective width “under” each reinforcing bar contributing to the node resistance. 0.96 −1.05   d rb   cos θ  s ≥ 80 mm: b = 40.9  12  eff,rb  cos 45°   rb    0.96 0.61 −1.05   d rb   s rb   cos θ  s rb < 80 mm: beff,rb = 40.9   80    cos 45°   12    

(2) where: beff,rb effective width “under” each reinforcing bar spacing of reinforcing bars srb diameter of reinforcing bars drb θ angle of diagonal strut assumed in model Finally, to simplify the assembly of the joint model, the diagonal spring representing the joint link component is converted into a horizontal spring as shown in Fig. 4. The properties of the horizontal spring are obtained directly from the diagonal spring determined as a function of the angle of the diagonal spring.

Table 4. Admissible stresses at STM nodes according to EN 1992-1-1 [4]

a) STM

Fig. 8. Joint link modelling

b) Single diagonal spring

Node

Admissible stress

0.75 vfcd

3 vfcd

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 24

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

Table 5. Summary of the global results of the joint properties and quantification of the approximation with respect to experimental results Approach Analytical

Numerical

Fig. 9. Definition of the dimension related to the hook of the longitudinal reinforcing bar at node 1 according to CEB Model Code [14]

4 Application of the design model to a composite beam-to-reinforced concrete wall joint The model depicted in Fig. 4 (simplified by the use of a modified version of the T-stub in compression model, as described in section 3) is used to obtain the joint properties. The assembly procedure is then direct; no distribution of resistance among rows is required because only one tension row is identified. In order to determine the bending moment and rotation at the joint, it is necessary to define the lever arm hr of the joint. According to the joint configuration, it is assumed that the lever arm is the distance between the centroid of the longitudinal steel reinforcement and the middle of bottom flange of the steel beam. Thus, the smallest resistance of the activated components governs the bending moment resistance and may be expressed as follows:

( )

M j = Min Fi h r

(3)

where Fi represents the resistance of all activated components within the joint under bending moment loading determined as described above. With respect to the joint rotation, it is important to consider the contribution of all components. Again, as only one tension and one compression row is activated, it is easy to obtain the joint rotation. The component governing the resistance controls the rotation capacity of the joint. The joint rotation capacity may be determined as follows:

∑ 1 Δi = n

Δu

(4)

where ΣΔi represents the sum of the deformations of the activated components for a load equal to the resistance of

a) Test 1

Test

Mj/Mj,test

Φj/Φj,test

Sj/Sj,test

0.99

0.90

0.85

1.05

0.92

0.87

0.97

1.27

1.10

1.02

1.46

0.88

the governing component. In the case of the governing component, its entire deformation capacity should be considered. The accuracy of the model described has been assessed using the experimental results performed at the University of Stuttgart within the RFCS research project “InFaSo” [1]. The specimens tested consisted of a cantilever composite beam supported by a reinforced concrete wall. The joint configuration depicted in Fig. 1 was used to connect both members. A vertical load was applied at the free edge of the composite beam up to failure. The load induced a hogging bending moment in the joint. The geometrical and material properties, as well as a detailed discussion of the tests, can be found in [15]. The moment–rotation curves for two of the specimens tested are compared in Fig. 10. The results of a numerical model are also included. The calibration and validation of this numerical model are presented in [16]. The parameter varied between the specimens selected is the diameter of the longitudinal reinforcing bars (percentage of reinforcement within slab): test 1 = 6 No. ∅16 mm; test 2 = 6 No. ∅12 mm. Concerning the analytical model, the ECCS model [5] for the longitudinal steel reinforcement was considered; for the slip of the composite beam, the approach proposed in [7] is used. The curves show a good approximation between analytical, numerical, and experimental results. The accuracy of the analytical and numerical approaches in relation to the experimental results are quantified in Table 5. In terms of resistance, the approximation is excellent. In terms of rotation at maximum bending moment, the results of the analytical approach are interesting when we consider that this parameter is not usually quantified. The resistance of each component according to the analytical model is given in Table 6; the percentage of resistance activated for each component is also included. It can be seen that as in the experimental tests, the longitu-

b) Test 2

Fig. 10. Moment–rotation curves comparing experimental, numerical and analytical results

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 25

01.02.13 08:49

J. Henriques/L. S. da Silva/I. Valente · Design model for composite beam-to-reinforced concrete wall joints

Table 6. Resistance of the components according to the analytical model and percentage of activation Component

Test 1

Test 2

Fr,i [kN]

% active

Fr,i [kN]

% active

811.9

100.0

460.9

100.0

1200.0

67.7

1200.0

38.4

824.9

98.4

824.9

55.9

2562.0

31.7

2562.0

18.0

5 to 10

2017.6

40.2

2017.6

22.8

1224.8

66.3

930.0

49.6

Governing

Component 1

dinal reinforcing bar in tension is the governing component. According to the analytical estimation, in test 1 the beam web and column in compression are close to full activation. On the other hand, in both tests the steel contact plate and the anchor plate in compression are the components with the lowest level of activation compared with their load capacity.

5 Conclusions and general recommendations A design model based on the component method for a composite beam-to-reinforced concrete wall joint is proposed in this paper and compared with experimental and numerical results. Although some of the approaches of the individual components are incomplete, at the current stage the model is demonstrated to be accurate. Based on the results presented and the considerations achieved during this research work, some design suggestions are proposed: I) Designing the longitudinal reinforcement in the composite beam to be the governing component allows better control of the joint response. The characterization of this component can be more accurate in an inelastic range in comparison with the other activated components. Furthermore, if the steel reinforcing bars are class C (according to [4]) a ductile response can be achieved. II) Owing to the complexity of the problem, reducing the joint link component to a single spring is a simplification with practical interests. However, this approach is limited and therefore the failure of the joint in this component should be avoided. References [1] Kuhlmann, U., Eligehausen, R., Wald, F., Simões da Silva, L., Hofmann, J.: New market chances for steel structures by innovative fastening solutions. Final report of RFCS project INFASO, project No. RFSPR-CT-2007-00051, Brussels, 2012. [2] European Committee for Standardization – CEN: EN 1993-1-8. Eurocode 3: Design of steel structures. Part 1-8: Design of joints, Brussels, 2005. [3] European Committee for Standardization – CEN: EN 1994-1-1. Eurocode 4: Design of composite steel and concrete structures. Part 1-1: General rules and rules for buildings, Brussels, 2004. [4] European Committee for Standardization – CEN: EN 1992-1-1. Eurocode 2: Design of concrete structures. Part 1-1: General rules and rules for buildings, Brussels, 2004. [5] European Convention for Constructional Steelwork – ECCS. Design of Composite Joints for Buildings. ECCS publication

Component 1

No. 109, Technical Committee 11, Composite Structures, 1st ed., Belgium, 1999. [6] Aribert, J. M.: Influence of Slip on Joint Behaviour. Connections in Steel Structures III, Behaviour, Strength and Design, 3rd International Workshop, Trento, Italy, 29–31 May 1995. [7] Anderson, D., Najafi, A. A.: Performance of Composite Connections: Major Axis End Plate Joints. Journal of Constructional Steel Research, vol. 31, 1994, pp. 31–57. [8] Guisse, S., Vandegans, D., Jaspart, J.-P.: Application of the component method to column bases: Experimentation and development of a mechanical model for characterization. Research Centre of the Belgian Metalworking Industry, MT195, Liège, 1996. [9] European Committee for Standardization – CEN: CEN/TS 1992-4: Design of fastenings for use in concrete, final draft, Brussels, 2009. [10] Furche, J.: Zum Trag- und Verschiebungsverhalten von Kopfbolzen bei zentrischem Zug. PhD thesis, University of Stuttgart, 1994. [11] Henriques, J.: Behaviour of joints: simple and efficient steelto-concrete joints. PhD thesis, University of Coimbra (to be published). [12] Steenhuis, M., Wald, F., Sokol, Z., Stark, J.: Concrete in compression and base plate in bending. Heron 2008; vol. 53, No. 1/2; pp. 51–68. [13] Schlaich, J., Schäfer, K., Jennewein, M.: Toward a Consistent Design of Structural Concrete. PCI Journal, 32(3), 1987, pp. 74– 150. [14] Comité Euro-International du Béton – CEB: CEB-FIP Model Code 1990: Design Code. Lausanne, 1993. [15] Henriques, J., Ozbolt, A., Žižka, J., Kuhlmann, U., Simões da Silva, L., Wald, F.: Behaviour of steel-to-concrete joints II: Moment resisting joint of a composite beam to reinforced concrete wall. Steel Construction – Design and Research, Volume 4 (No. 3), 2011, pp. 161–165. [16] Henriques, J., Simões da Silva, L., Valente, I.: Numerical modeling of composite beam to reinforced concrete wall joint. Part II: Global behavior. Engineering Structures, 2012 (submitted for publication). Keywords: steel-to-concrete joints; design model; component method; joint components

Authors: José Henriques, Luís Simões da Silva ISISE - Department of Civil Engineering, University of Coimbra, Portugal jagh@dec.uc.pt; luisss@dec.uc.pt Isabel Valente ISISE – Department of Civil Engineering, Engineering School, University of Minho, Portugal, isabelv@civil.uminho.pt

Steel Construction 6 (2013), No. 1

09_019-026_da_Silva (1-3)_cs6.indd 26

01.02.13 08:49

Articles Florea Dinu* Dan Dubina Calin Neagu

Cristian Vulcu Ioan Both Sorin Herban

DOI: 10.1002/stco.201300005

Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas Beams with a span-to-depth ratio < 4 are not very common in the design of momentresisting frames. For such beams, the shear stresses may become a controlling factor in the design, as the moment capacity is influenced by the presence of the shear. This is an important matter when such a beam is part of a seismic resisting system that is designed according to the dissipative concept. In this case the contribution from the shear force affects the dissipation capacity and plastic mechanism. This paper presents the testbased evaluation of moment frames with short beams and reduced beam section (RBS) connections, for the purpose of exploring the application of the plastic hinge model. Fullscale specimens, taken from an 18-storey building, have been tested. The test results and their interpretation are summarized here.

1 Introduction Owing to their inherent ductility, moment-resisting frames are often used in systems resisting seismic forces. Inelastic behaviour is intended to be accommodated through plastic hinges in beams near the beam-to-column connections, and also at column bases. Although considered as deemed-to-comply connections, welded beam-to-column connections have experienced serious damage and even failures during strong seismic events. These failures have included fractures of the beam flange-to-column flange groove welds, cracks in column flanges and cracks through the column section [1]. To reduce the risk of the brittle failure of such connections, either connection strengthening or beam weakening can be applied. The first approach consists of providing sufficient connection overstrength, e. g. by means of haunches or cover plates. The second approach can benefit from the “reSelected and reviewed by the Scientific Committee of the 7th International Workshop of Connections in Steel Structures, 30 May–2 June 2012, Timişoara, Romania * Corresponding author: florea.dinu@ct.upt.ro

duced beam section” (RBS) or “dogbone” concept, initially proposed by Plumier [9] and then developed and patented by ARBED, Luxembourg. (In 1995 ARBED waived all patent and claim rights associated with RBS for the benefit of the structural design community.) Proper detailing of the RBS, including flange cut-outs and beam-tocolumn welds, is needed to ensure the

formation of plastic hinges in the reduced zones. It is economical to keep the width of bays within certain limits because long bays make the structure flexible and therefore increase the drift, which may control the design. On the other hand, short bays can reduce the dissipation capacity due to the presence of large shear forces. As a result, the connection qualification specifies minimum span-to-depth ratios to be used for moment frame connections. When prequalified connections are utilized outside the parametric limitations, project-specific qualification must be performed to permit the prediction of behaviour and acceptance criteria [1]. This paper presents part of a research project that was carried out to check the validity of the moment frame connections of an 18-storey structure. The paper describes the calibration of

Fig. 1. Plan and elevation of building

10_027-033_Dinu_et_al (1-5)_cs6.indd 27

01.02.13 08:50

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban · Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

Fig. 3. Cruciform cross-section of columns

Fig. 2. Moment frame with short beam: a) bolted flush end plate connection, b) shear slip-resistant connection

numerical models for two types of RBS connections using the generalpurpose finite element analysis program ABAQUS [8]. The finite element models were calibrated using experimental tests performed on four fullscale specimens at the Steel Structures Laboratory, “Politehnica” University of Timişoara, Romania. The particular feature of the project is the use of very short bay widths coupled with the use of RBS connections for the moment frame connections. In addition, the project incorporates flush end plate bolted connections for beam splices and therefore it addresses concerns regarding the potential for brittle failure of the bolts.

2 Experimental program 2.1 Specimens and test setup The study is connected with the design of an 18-storey office building located in Bucharest, Romania. The building is 94 m high and the plan dimensions are 43.3 × 31.3 m, see Fig. 1. It is located in a highly seismic area characterized by

a design peak ground acceleration of 0.24 g for a return period of 100 years and soft soil conditions with TC = 1.6 s. The long corner period of the soil is noteworthy, which in this case may affect flexible structures. For the serviceability check, the return period is 30 years, whereas for collapse prevention it is 475 years. The lateral force resisting system consists of external steel framing with closely spaced columns and short beams. The central core has also steel frames with closely spaced columns and short beams. The ratio of beam length to beam depth L/h varies from 3.2 to 7.4, which results in seven different types of beam. Some beams are below the generally accepted lower limit (L/h = 4). The moment frame connections employ reduced beam section (RBS) connections that are generally used for beams loaded mainly in bending (Fig. 2). Circular radius cuts in both the top and bottom flanges of the beams were used to reduce the flange area. The detailing followed the recommendations of AISC 341-05 [1]. Welds

of beam flanges and web to column flanges are complete joint penetration groove welds. Two types of beams, which have the shortest L/h ratio, were selected for the experimental program. Table 1 shows the characteristics of the beams tested experimentally. The first beam, denoted RBS-S, has a clear length of 1450 mm and the lowest span-depth ratio, L/h = 3.2. The second type, denoted RBS-L, has a clear length of 2210 mm and a corresponding span-depth ratio L/h = 4.9 (Fig. 2). The web and flange thicknesses for both beam types are 20 and 14 mm respectively. The designed solution adopted for the splice beam connection was a bolted flush end plate slip-resistant connection. After the first series of tests it was decided to change this solution to a classical shear slip-resistant splice connection (Table 1). The column has a cruciform cross-section made from two hot-rolled profiles (HEA800 and HEA400, see Fig. 3). Beams and columns are both made from grade S355 steel. The base material characteristics were determined experimentally. The measured yield strengths and tensile stresses of the plates and sections were greater than the nominal values. The greatest increase was recorded for the hotrolled profiles, being lower for plates. It should be noted that the ratio between nominal and actual yield stress is limited to 1.25 by seismic design code EN 1998-1 [7].

Table 1. Characteristics of beams tested experimentally Type

h [mm]

b [mm]

L [mm]

fy [N/mm2]

Mp [KNm]

Vp [KN]

Mp/Vp

L/h

Splice connection

RBS-S1, 2

450

250

1450

355

641

1845

0.35

3.2

flush end plate

RBS-L1, 2

450

250

2210

355

641

1845

0.35

4.9

flush end plate

RBS-S3

450

250

1450

355

641

1845

0.35

3.2

gusset-plate

RBS-L3

450

250

2210

355

641

1845

0.35

4.9

gusset-plate

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 28

01.02.13 08:50

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban · Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

2.2 Results

Table 2. Material properties of rolled sections Section

Steel grade

HEA800

S355

HEA400

Element

fy [N/mm²]

fu [N/mm²]

Au [%]

flange

410.5

618.5

15.0

web

479.0

671.2

13.0

flange

428.0

592.0

15.1

web

461.0

614.0

12.8

fy [N/mm²]

fu [N/mm²]

Au [%]

373.0

643

17.0

403.0

599

16.5

S355

Table 3. Material properties of flat steel Section

Steel grade

14 mm

S355

20 mm

S355

Element beam flange beam web

Fig. 4. Test setup

Fig. 5. Loading protocol: a) determination of yielding displacement, b) cyclic loading protocol

Fig. 4 shows the test setup. Specimens were tested under a cyclic loading sequence taken from the ECCS recommendations [4]. Thus, according to the ECCS procedure, the yielding displacement Dy and the corresponding yielding force Fy are obtained from the monotonic force–displacement curve (Fig. 5a). In order to reduce the number of tests, the monotonic test was replaced by the push-over curve

obtained numerically using the general-purpose finite element analysis program ABAQUS. The yielding displacement is then used for establishing the cyclic loading. It consists of generating four successive cycles for the displacement ranges of ± 0.25 Dy, ± 0.5 Dy, ± 0.75 Dy and ± 1.0 Dy followed up to failure by series of three cycles each with a range of ± 2n × Dy, where n = 1, 2, 3,… etc. (Fig. 5b).

Table 4 summarizes the experimental results, with observations regarding the behaviour and failure mode of each specimen. Specimens with longer beams, RBS-L1 and RBS-L2, remained elastic until a drift of 30 mm, or 0.6 % of the storey height. Two failure modes were recorded. The first mode involved the fracture of the top beam flange-to-column flange welds, which afterwards propagated into the beam web. The second failure mode involved the fracture of the bottom flange due to the large tensile forces at ultimate load. Both failures occurred at interstorey drifts > 5 % of the storey height. The plastic behaviour was dominated by the buckling of the flange in compression and out-of-plane buckling of the web. Specimens with shorter beams, RBS-S1 and RBS-S2, remained elastic until a drift of 25 mm, or 0.5 % of the storey height. The visible buckling of the flange in compression was first observed, followed by out-of-plane buckling of the web. Failure of the first short specimen, RBS-S1, involved fracture of the bottom flange due to the large tensile forces at ultimate load, followed by fracture of the beam web. The failure of the second specimen, RBS-S2, involved the fracture of the bolts at the splice connection. The plastic behaviour was dominated by the buckling of the flange in compression and shear buckling of the web. Fig. 6 and Fig. 7 show the evolution of the out-of-plane plastic deformations in the web zone adjacent to the column. Under the increasing lateral force, the plastic mechanism in the web involves both bending moment and shear force. The contribution of the shear force to the overall deformation is more important for the short specimens, RBS-S, and it can be observed following the inclination of the shear buckling waves of the web. Fig. 9 shows the recorded moment-rotation curve for all specimens. The total rotation of the joint has two major components: rotation of the beam (reduced beam section) and distortion of the web panel in the reduced region. Owing to the large stiffness of the columns, the contribution of the column web panel can be neglected. The specimens exhibited good rotation capacity and stable hysteretic behav-

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 29

01.02.13 08:50

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban · Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

Table 4. Results of experimental tests Specimen

Failure mode

RBS-L1

– cracks initiated in top flange welds, fracture propagated in web

– failure at interstorey drift of 5 % – no slip at splice connection – large dissipation capacity, reduced cyclic degradation

RBS-L2

– failure due to fracture of flange in reduced area, then propagation in web

– failure at interstorey drift of 5 % – no slip at splice connection – large dissipation capacity, reduced cyclic degradation

RBS-L3

– cracks initiated in bottom flange-tocolumn welds, fracture propagated in web

– failure at interstorey drift of 4.5 % – no slip at splice connection – large dissipation capacity, reduced cyclic degradation

RBS-S1

– failure due to fracture of flange in reduced area, then propagation in web

– failure at interstorey drift of 5 % – moderate slip at splice connection – large dissipation capacity, reduced cyclic degradation

RBS-S2

– failure due to fracture of bolts at beam splice connection

– failure at large interstorey drift – large slip at splice connection – large dissipation capacity, reduced cyclic degradation

RBS-S3

– failure due to fracture of flange in reduced area, then propagation in web

– failure at interstorey drift of 5 % – no slip at splice connection – large dissipation capacity, reduced cyclic degradation

Details: failure mode and force–displacement curve

Observations

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 30

01.02.13 08:50

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban · Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

Storey drift, %H

Fig. 6. Shear web deformation history, specimen RBS-L1

Fig. 7. Shear web deformation history, specimen RBS-S2

RBS-L1

RBS-L2

RBS-L3

RBS-S1

RBS-S2

RBS-S3

Fig. 8. Moment–rotation relationship for cyclically loaded joints

Fig. 9. Hysteresis curves: a) RBS-S3, b) RBS-L3

iour up to 5 % interstorey drift. This capacity supports the design of the structure which is based on a 2.5 % interstorey drift limitation at the ultimate limit state. The specimens showed reduced degradation in both strength and stiffness. The bolts in first series of tests (with flush-end plate connections) slipped in specimens RBS-S1, RBS-S2, causing some

pinching in the hysteresis curves (Fig. 8). For second series of tests (with shear slip-resistant connections), no slip was recorded (Fig. 8).

3 Numerical investigation 3.1 Description of the numerical model In order to optimize the design of the reduced beam section connections, a

numerical analysis was carried out. For this purpose, a numerical model able to simulate the large post-elastic strain cyclic deformation was calibrated. All the components were modelled using solid elements. In order to have a uniform and structured mesh, some components with a complex geometry were partitioned into simple shapes. The engineering stress-strain curves of the steel grades obtained from tensile tests were computed into true stresstrue plastic strain and used further in the numerical model. The modulus of elasticity was considered to be 210 000 N/mm2 and Poisson’s ratio taken as 0.3. For the cyclic analysis, a combined isotropic/kinematic hardening model was used for the material, containing the cyclic hardening parameters from Dutta et al. [3]. A dynamic explicit type of analysis was used. The

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 31

01.02.13 11:47

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban 路 Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

justment of RBS-S3 model was necessary.

4 Conclusions and general recommendations

c) Fig. 10. Comparison between RBS-L3 and RBS-L3_MOD (FEM): a) cyclic behaviour, b) beam configuration, c) dimensions of flange cut-out

load was applied through displacement control at the top of columns.

3.2 Results The main objective of numerical simulation was to optimize the shape of the cut-out in the beam flanges in the reduced zone. Fig. 9 shows the hysteresis curves of test specimens RBS-S3 and RBS-L3, with a shear slip-resistant connection. As expected, this type of connection prevented bolt slippage and therefore a continuous beam was taken into account within the numerical simulations. It can be seen that

the behaviour anticipated by the numerical simulation is confirmed by the tests. Based on the numerical results, the length of the reduced beam section for RBS-L3 model was shortened from 450 to 300 mm (RBS-L3_MOD) (Figs. 10b, 10c). This new solution did not affect the stiffness but decreased the amount of shear force (Fig. 10a). As can be seen in Fig. 11, the concentration of the plastic deformations shifts from beam end to the reduced beam section zone. The Von Mises stress distribution for the two cases is presented in Fig. 12. No ad-

The main conclusions regarding the experimental tests and numerical analysis of RBS connections for short coupling beams are summarized below: Experimental tests performed on two types of short beam with RBS connections confirmed the design procedure. The specimens exhibited excellent ductility and rotation capacity up to 60 mrad before failure. However, the flush end plate connection exhibited significant slippage, which lead to a reduction in the stiffness. Based on these observations, the splice connection was redesigned using a detail that is more appropriate for the predominant shear stress state at mid-length of the beams. This new connection detail consists of gusset plates on web and flanges and preloaded high-strength bolts. This new configuration can prevent bolt slippage and therefore both the stiffness and axial straightness of the assembly will not be altered. For very short beams, the interaction between the shear and normal stresses causes an inclination of the buckled shape in the web. The plastic rotation capacity has two major components, i. e. rotation of the beam (reduced beam section) and distortion of the web panel in the reduced region.

Fig. 11. Equivalent plastic strain: (a) RBS-L3, (b) RBS-L3_MOD

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 32

04.02.13 12:01

F. Dinu/D. Dubina/C. Neagu/C. Vulcu/I. Both/S. Herban · Experimental and numerical evaluation of an RBS coupling beam for moment-resisting steel frames in seismic areas

a) Fig. 12. Von Mises equivalent stress: (a) RBS-L3, (b) RBS-L3_MOD

Due to the high stiffness of the columns, the contribution of the column web panel can be neglected. Numerical simulation allowed a modified RBS configuration to be calibrated in order to eliminate the stress concentration near the beam-to-column welds. The resulsts obtained for this adjusted model showed a better behavior with the development of plastic deformations in the reduced area only.

Acknowledgments Funding of the project was made in the frame of the contract 76/2011 “Numerical simulations and experimental tests on beam column subassemblies from the structure of a 17 storey steel building in Bucharest, Romania” between the “Politehnica” University of Timis¸oara and DMA ARCHITECTURE & INTERIOR DESIGN LTD, Bucharest, Romania Cristian Vulcu was partially supported by the strategic grant POSDRU/ 88/1.5/S/50783, project ID50783 (2009), co-financed by the European Social Fund – Investing in People,

within the Sectoral Operational Programme Human Resources Development 2007–2013.

References [1] AISC 341-05: Seismic provisions for structural steel buildings. American Institute for Steel Construction, 2005. [2] Johansson, B., Maquoi, R., Sedlacek, G., Müller, C., Beg, D.: Commentary and worked examples to EN 1993-1-5, JRC – ECCS cooperation agreement for the evolution of Eurocode 3, European Commission, 2007. [3] Dutta, A., Dhar, S., Acharyya, S. K.: Material characterization of SS 316 in low-cycle fatigue loading, Journal of Materials Science, vol. 45, No. 7, 2010, pp. 1782–1789. [4] ECCS – European Convention for Constructional Steelwork, Technical Committee 1, Structural Safety and Loadings; Working Group 1.3, Seismic Design: Recommended Testing Procedure for Assessing the Behaviour of Structural Steel Elements under Cyclic Loads, 1st ed., 1986. [5] EN 1993-1-1 Eurocode 3: Design of steel structures. General rules and rules for buildings, CEN, 2005.

[6] EN 1993-1-5 Eurocode 3: Design of steel structures. Part 1-5: General rules – Plated structural elements, CEN, 2006. [7] EN 1998-1. European Committee for Standardization – CEN: Eurocode 8: Design provisions for earthquake resistance of structures. Part 1.1: General rules. Seismic actions and general requirements for structures, Brussels, 2004. [8] Hibbit, D., Karlson, B., Sorenso, P.: ABAQUS User’s Manual, 2007, version 6.9. [9] Plumier, A.: New Idea for Safe Structures in seismic Zones. IABSE Symposium, Mixed structures including new materials, Brussels, 1990, pp. 431–436. Keywords: steel moment frame; reduced beam section; short beam; cyclic test; seismic design

Authors: Florea Dinu, Dan Dubina, Calin Neagu, Cristian Vulcu, Ioan Both, Sorin Herban “Politehnica” University of Timis¸oara, Department of Steel Structures & Structural Mechanics, Romania Dragos Marcu, Popp & Asociatii, Bucharest, Romania, dragos.marcu@popp-si-asociatii.ro

Steel Construction 6 (2013), No. 1

10_027-033_Dinu_et_al (1-5)_cs6.indd 33

01.02.13 08:50

Articles Ram Puthli* Jaap Wardenier Andreas Lipp Thomas Ummenhofer

DOI: 10.1002/stco.201300008

Thin-walled structural hollow section joints Several national design standards allow the use of hollow sections with minimum nominal wall thicknesses down to 1.5 mm. However, the international standards that will be used in the future, based on the work of the International Institute of Welding (IIW-XV-E), still prescribe a minimum thickness of 2.5 mm. This paper presents the background to the IIW-XV-E higher thickness limit and presents strong arguments for extending the scope of these international standards to include structural hollow sections with wall thicknesses down to 1.5 mm.

1 Introduction Normal buildings where structural hollow sections are used, such as industrial buildings, stadiums, shopping centres, etc. are usually designed nowadays using steel grade S355 and higher. The sections in this grade with the smallest wall thicknesses offered by the manufacturers for hot-formed sections to EN 10210 [1] are 21.3 × 2.3 mm (CHS) and 50 × 30 × 2.6 mm or 40 × 40 × 2.6 mm (RHS). For cold-formed sections to EN 10219 [2] the figures are 21.3 × 2 mm (CHS) and 20 × 20 × 2 mm (RHS). However, some specialized industries require very light structures with minimum obstruction from the steel frames and lattice girders, e. g. horticultural greenhouses, so that maximum sunlight is available for the photosynthesis of plants. The horticulture industry primarily makes use of coldformed hollow sections to EN 10219 [2] in steel grades S235 and sometimes S275 because the wall thicknesses of these sections vary from 1.5 to 2.5 mm for the lattice girders. Hot-rolled sections cannot be manufactured with such thin walls. Glass (thermally toughened glass for horticultural greenhouses and laReceived 17 October 2012, revised 7 November 2012, accepted 20 November 2012 * Corresponding author: e-mail puthli@kit.edu

minated or air-insulated laminated glass for garden centres) or specially manufactured plastic film are typical forms of cladding for such greenhouses. These structures normally have lattice girder spans of 6.4 to 12.8 m and depths of 0.4 to 0.55 m. With a glass cladding, the designs usually require 60 × 40 × 2 (or 3) mm RHS chords, depending on the loading, and 25 × 25 × 2 mm RHS braces. For a plastic film cladding, the chords are generally 50 × 30 × 2 mm RHS, the braces 20 × 20 × 1.5 mm RHS. It should be borne in mind that these structures are only used for agricultural production, e. g. vegetables, fruit, plants, bulbs and flowers. Such buildings enclose large areas, up to 500 × 400 m and even more, with total heights varying from 3.0 to 7.0 m. Consequently, a relatively large number of hollow sections are required for such structures, meaning that a 0.5 or 1.0 mm increase in wall thickness could also mean larger section sizes. Not only would this result in the cost of steel almost doubling for the owner, but also a reduction in the amount of sunlight reaching the plants for photosynthesis. Every 1 % reduction in sunlight results in a corresponding 1 % decrease in production. The following discussion presents technical arguments and evidence for reducing the minimum wall thickness from 2.5 to 1.5 mm. In every case described in this paper, the wall thicknesses of the braces

are less than or equal to the chord wall thicknesses, and all < 2.5 mm.

2 Acceptable slenderness limits for hollow sections in joint design The international standards (e. g. section 3.4 in [7]) define cross-section classification as an identification of the extent to which the resistance (to axial compression or bending moment) and rotation capacity of a cross-section are limited by its local buckling resistance. For example, four classes are given in Eurocode 3, Part 1-1 [5] together with three limits for diameter-to-thickness ratio for CHS or width-to-thickness ratio for RHS. Examples of cross-section classification can be found in section 5.5 of Eurocode 3, Part 1-1 [5]. However, for joint design involving hollow sections [6, 7], only class 1 or class 2 hollow sections (described as compact sections in Eurocode 3) – with, additionally, diameter-to-thickness ratio (d/t) limited to 50 for CHS and side length-to-thickness ratio (b/t or h/t) limited to 40 for RHS – are permitted. Class 2 sections are the more slender of the two classes and are defined by d/t ≤ 70 e2 (CHS) and c/t ≤ 38 e (RHS), where c is the flat part of the RHS, equal to (b or h) – 2 t – 2r. Here, b is the width, h the height, t the wall thickness and r the inner corner radius. All designs, also for lightweight structures such as horticultural greenhouses, have to comply with these limits. The material parameter is defined by the non-dimensional factor

235 , where fy is the yield stress. fy

The small hollow sections used in horticultural greenhouses are invaria-

11_034-038_Puthli_(1-8)_cs6.indd 34

01.02.13 08:50

R. Puthli/J. Wardenier/A. Lipp/Th. Ummenhofer · Thin-walled structural hollow section joints

(

Ni,Rd = f β, γ,g'

fy0 t 02

) sin θ

f(n)

For other failure criteria, such as brace effective width, chord shear or punching shear: the diameter, width and/or depth dimensions are also included in the equations in addition to yield stress and thickness. Therefore, from a strength point of view, it is important to consider whether the effect of the dimensions could be different for joints with small sections. According to [1], [2], the tolerances for diameter, width or depth are 1 %, with a minimum of ± 0.5 mm and a maximum of ± 10 mm; for the smaller dimensions this could have some effect on the strength. The minus tolerance for the thickness is –10 % for wall thicknesses below 5 mm and ± 0.5 mm above 5 mm. This thickness tolerance is partly compensated for by the ± 6 % mass tolerances. However, this tolerance results in a larger influence (scatter) on the joint strength for small wall thicknesses. Another factor influencing the joint strength is the weld size, which may have a large influence on the „effective“ width ratio β, especially for smaller width ratios. In many cases, especially for thin-walled sections, the welds were considerably larger than required, contributing considerably to the scatter. The above factors show that for joints with thin-walled sections in particular, we can expect a considerably larger scatter than for the thick-walled sections. It is shown in Vegte van der et al. [14] that the smaller specimens have a larger strength than the larger ones, which is mainly caused by the larger welds. Therefore, numerical analyses as used for the recommenda-

Experimental data from tests on hollow section joints with wall thicknesses < 2.5 mm have been collected for this paper and compared with the international standards [6], [7]. Some of these comparisons with Eurocode 3, Part 1-8 [6] are given in Figs. 1, 2, 3 and 4. The evaluation for [13] and the ISO draft [7] was mainly based on numerical (finite element) analysis, with the experimental database used for validation only. This approach allowed careful scrutiny of the test data, some of which dates from the 1960 s, where the test setup and other experimental details could not be verified any more. Some of the data presented here for comparison with Eurocode3, Part 1-8 [6] have been rejected in the ISO draft [7] evaluation. All the comparisons shown in the figures are made with characteristic resistances excluding γM = 1.1 and not with design resistances. Therefore, NChar,EC3 = 1.1 × Ni,Rd. Figs. 1 to 4 all show that the experimental values are safely above the characteristic values. Fig. 1 shows two values, both with NTest/NCharac,EC3 = 0.99, i. e. the difference between this and the 1.0 value is negligible. For RHS T-joints, only one data point could be recovered from 2,00

Kurobane et al. (1964) Togo (1967)

NTest / NCharac,EC3

1,60 1,40 1,20 1,00 0,80 0,60 0,20

0,40

0,60

2,00

Kanatani (1965)

1,80

Makino (19761979) EN 1993-1-8

1,60 1,40 1,20 1 20 1,00 0,80 0,60 0,20

0,40

0,60

0,80

1,00

Fig. 3. Experimental results plotted against characteristic joint capacities (EC3 [6]) for CHS T-joints with brace wall thicknesses < 2.5 mm

2,00 Wardenier (1976)

1,80

EN 1993-1-8

1,60

1,40 1 20 1,20

1,40 1,20

1,00

0,80

0,60

0,20

0,40

0,60

0,80

1,00

Fig. 1. Experimental results plotted against characteristic joint capacities (EC3 [6]) for CHS X-joints with chord and/or brace wall thicknesses < 2.5 mm

0,2

0,4

0,6

0,8

Fig. 4. Experimental results plotted against characteristic joint capacities (EC3 [6]) for RHS gapped K-joints with chord wall thicknesses of 1.87 mm

Steel Construction 6 (2013), No. 1

11_034-038_Puthli_(1-8)_cs6.indd 35

1,00

Fig. 2. Experimental results plotted against characteristic joint capacities (EC3 [6]) for CHS gapped K-joints with chord and/or brace wall thicknesses < 2.5 mm

Togo (1967)

1,60

0,80

Kanatani (1965)

1,80

Washio et al. (1963)

1,80

NTestt / NCharac,EC3

All the design equations in the standards [6], [7] for chord face failure (or plastification) are a function of the non-dimensional joint parameters β, γ, g’, n and θi and also the yield stress and thickness:

4 Experimental data with wall thicknesses < 2.5 mm used in establishing design formulae in IIW recommendations [12, 13] for international standards

2,00

NTeest / NCharac,EC3

3 Design equations

tions in [7] provide more realistic data than experiments if not all dimensions, including the welds, are properly measured.

NTestt / NCharac,EC3

bly offered in steel grade S235 and sometimes S275, with the material parameter e equal to 1.0 (S235) or 0.92 (S275). They tend to be within the class 1 limit.

01.02.13 08:50

R. Puthli/J. Wardenier/A. Lipp/Th. Ummenhofer · Thin-walled structural hollow section joints

the old tests [15] for wall thicknesses < 2.5 mm. This was for chord and brace wall thicknesses of 2.1 mm, resulting in NTest/NCharac,EC3 = 1.44, so no figure is provided here. As discussed before, the figures all show the expected wide scatter in the data for these thin-walled joints. This is partly because of the relatively large variation in the (not measured) weld throat thickness and in the wall thickness tolerances for thin-walled sections, where nominal values were sometimes used when the actual values were not measured. All this evidence shows that, in general, the joint strength of small specimens and small wall thicknesses exhibits a higher safety margin than the larger specimens with larger wall thicknesses.

5 Minimum wall thickness limits in some standards In Germany, hollow sections have been designed according to DIN 18808 [3], where in Table 3 wall thicknesses of 1.5 mm and more have been allowed since before 1984 for all girder joints. Only in the case of L-joints welded end to end (e. g. for staircases) is the minimum wall thickness 2.5 mm according to Table 6 [3] because of the obvious welding difficulties when welding hollow sections end to end. In Australia, the steel design standard AS 4100 [4] was originally intended for hot-rolled I sections only. About 20 years ago, the scope of the standard was extended to cold-formed sections with wall thicknesses ≥ 3 mm. This was then extended further in 1998 to permit the use of cold-formed hollow sections with wall thicknesses ≥ 1.6 mm, based on further research. A so-called capacity factor f, equivalent to the reciprocal of the partial safety factor in the Eurocodes, is used in this standard. Most fillet welds have f = 0.8, except for longitudinal fillet welds in RHS < 3 mm, where f = 0.7. Longitudinal fillet welds would then mainly correspond to slotted plate connections into a CHS/RHS. From an Australian perspective, welds in a tubular truss (e. g. K, N, etc) would either be butt welds or transverse fillet welds, in which case the f factor for < 3 mm is the same as everywhere else. The IIW recommendations of 1989 [12] formed the basis for Euro-

code 3, Part 1–8 [6], where the following minimum wall thickness was first introduced: “7.1.1 (5) the nominal wall thickness of hollow sections should not be less than 2.5 mm.” The present IIW-XV-E recommendations dating from 2008 [13], which are the basis for the draft ISO standard 14346 [7], also include the following regarding minimum wall thickness: “5. Requirements: – The nominal wall thickness of hollow sections shall be limited to a minimum of 2.5 mm.” These wall thickness limitations were originally included in the 1989 version [12] because of welding and material aspects, such as the possibility of burning through thin walls when welding. Unfortunately, the standards have omitted a proviso for thin walls, i. e. that thinner walls down to a minimum of 1.5 mm are allowed when welders are properly qualified and the welding is appropriately approved and executed according to the relevant welding standards.

6 Contemporary welding methods for thin-walled hollow sections The concerns expressed above were due to problems that could occur during manual metal-arc welding and initially with CO2 welding. However, the modern (semi-)automatic gasshielded metal-arc welding processes employed on present-day truss and

Fig. 6. Typical welded RHS gap joint for girders supporting laminated glass roof (photo: Deforche Construct N. V.)

lattice girder assemblies have overcome such problems. To the authors’ knowledge, in the past two decades, many thousands of tonnes of hollow sections with wall thicknesses between 1.5 and 2.5 mm have been fabricated into lattice girders and erected at various sites by European firms, not only within Europe, but also in North America. Fig. 5 shows a lattice girder being welded on one side, before turning it to weld the other side. Fig. 6 shows a typical detail after fabrication, which shows that the measured weld size “a” is greater than that required by the standards. The gas-shielded metal-arc process used to weld the thin-walled hollow sections includes a process where the consumable electrode is automatically fed at a constant speed while the arc length is maintained essentially constant by the electrical characteristics of the welding power source.

Fig. 5. RHS sections in a template and semi-automatic welding from a suspended solid electrode coil (photo: Deforche Construct N. V.)

Steel Construction 6 (2013), No. 1

11_034-038_Puthli_(1-8)_cs6.indd 36

01.02.13 08:50

R. Puthli/J. Wardenier/A. Lipp/Th. Ummenhofer · Thin-walled structural hollow section joints

Fabricators welding lattice girders for horticultural greenhouses use shielded gas in accordance with EN ISO 14175 [8], using either 100 % CO2 or argon with between 15 and 25 % CO2. The designation for this in the certificate is EN ISO 14175: M21/C1. The designation M21 is for argon shielding gas with 15–25 % CO2, and C1 stands for 100 % CO2. The filler material is 1 mm diameter in accordance with EN ISO 14341 [9] and designated EN ISO 14341-AG3Si1 and AWS A5.18-ER70S-6. “A” is a symbol for the classification by tensile strength and 47 J impact energy, “G” designates a wire electrode and/ or deposit produced by gas-shielded metal-arc welding and 3Si1 is the chemical composition of the wire electrode. The AWS [11] electrode specification A5.18 is for gas-shielded metal-arc welding (GMAW) with the electrode classification ER70S-6 referring to the specification for carbon steel filler metals for GMAW. Finally, although the weld is designed in accordance with the minimum weld size prescribed in the standards, fabricators known to the authors do not provide weld thicknesses < 2 mm, even when the brace walls are 1.5 mm thick. Fig. 5 shows MAG welding (metal-arc welding using active gas) with solid wire electrodes, process number 135, according to EN ISO 4063 [10]. The welding procedures described here and currently used for thin-walled structural hollow sections have no adverse effects on joint strength. These procedures preclude any burning through the parent metal, even when the weld length is oversized. Such oversized weld lengths are common worldwide when welding thin-walled hollow section joints, as described in [16] and [17].

7 Concluding remarks Economic arguments, the scientific background and practical evidence have been presented in this paper in order to reduce the minimum permissible wall thickness from 2.5 to 1.5 mm for hollow sections used in lattice girders. In two international standards [6], [7], this is not yet the case. In the original 1989 version of the IIW-XV-E recommendations [12], the basis for Eurocode 3 [6], as well as the

more recent version of 2008 [13], on which the ISO draft [7] is based, the evaluation of experimental data also included joints with hollow sections with wall thicknesses between 1.5 and 2.5 mm. This article only presents that part of the experimental database where hollow sections < 2.5 mm were used as the basis for both standards, showing that the strength of these thin-walled joints is adequate and safe when applying the formulae in both the international standards. This paper provides evidence to extend the scope of the existing standards to include structural hollow sections with a wall thickness down to 1.5 mm. Moreover, the welding methods used in fabrication shops for lightweight structures with small hollow sections having wall thicknesses of 1.5 and 2 mm are described. The welding is performed in line with the established welding standards, which require properly qualified welders and appropriate approval through the usual channels. Steel fabricators specialized in this area have been producing many thousands of tonnes of lattice girders with such thin hollow sections for several decades.

Symbols and abbreviations CHS circular hollow section RHS rectangular hollow section Ni,Rd design value of joint resistance, expressed in terms of internal axial force in brace member i (i = 1 or 2) overall out-of-plane width of bi RHS member i (i = 0, 1 or 2) overall diameter of CHS memdi ber i (i = 0, 1 or 2) overall in-plane depth of RHS hi member i (i = 0, 1 or 2) nominal yield stress fy nominal yield stress of chord fy0 g’ gap between brace members in K- or N-joint divided by chord wall thickness n maximum stress-to-yield stress ratio in CHS or RHS chord chord wall thickness t0 β ratio of mean diameter or width of brace members to mean diameter or width of chord = d1 /d0 or d1 /b0 or b1 /b0 (for T-, Y- and X-joints) = (d1 + d2)/2d0 or (b1 + b2)/2b0 or (b1 + b2 + h1 + h2)/4b0 (for Kand N-joints)

e γ γM θi f

factor depending on fy half diameter (or half width)-tothickness ratio of chord (γ = d0/2 t0 or γ = b0/2 t0) partial safety factor for joint resistance included angle between brace member i (i = 1 or 2) and chord resistance factor (1/γm)

References [1] EN 10210 (2006): Hot-finished structural hollow sections of non-alloy and fine-grain steels – Part 1: Technical delivery conditions; – Part 2: Tolerances, dimensions and sectional properties. [2] EN 10219 (2006): Cold-formed welded structural hollow sections of non-alloy and fine-grain steels – Part 1: Technical delivery conditions; – Part 2: Tolerances, dimensions and sectional properties. [3] DIN 18808 (1984): Steel structures consisting of hollow sections predominantly statically loaded. [4] AS 4100 (1998): Australian Standard – Steel structures. [5] EN 1993–1–1 (2010): Eurocode 3: Design of steel structures. Part 1-1: General rules and rules for buildings. [6] EN 1993–1–8 (2010): Eurocode 3: Design of steel structures. Part 1-8: Design of joints. [7] ISO 14346 (draft) /IIW doc. XV1329-09 (2009): Static design procedure for welded hollow section joints – Recommendations (IIW doc. XV-140212 and IIW doc. XV-E-12–433). [8] EN ISO 14175 (2008): Welding consumables. Gases and gas mixtures for fusion welding and allied processes. [9] EN ISO 14341 (2011): Welding consumables. Wire electrodes and weld deposits for gas-shielded metal-arc welding of non-alloy and fine-grain steels. [10] EN ISO 4063 (2010): Welding and allied processes. Nomenclature of processes and reference numbers. [11] AWS D1.1/D1.1M (2004): American National Standard. Structural Welding Code – Steel. [12] IIW (1989): Design recommendations for hollow section joints – Predominantly statically loaded. 2nd ed., IIW doc. XV-701-89, IIW Annual Assembly, Helsinki. [13] IIW (2008): IIW static design procedure for welded hollow section joints – Recommendations. IIW doc. XV-128108, IIW Annual Assembly, Graz. [14] Vegte, van der G. J., Wardenier, J., Zhao, X.-L., Packer, J. A.: Evaluation of new CHS strength formulae to design strengths. Proceedings of 12th International Symposium on Tubular Struc-

Steel Construction 6 (2013), No. 1

11_034-038_Puthli_(1-8)_cs6.indd 37

01.02.13 08:50

R. Puthli/J. Wardenier/A. Lipp/Th. Ummenhofer · Thin-walled structural hollow section joints

tures, Shanghai, Tubular Structures XII, Taylor & Francis Group, London, 2008, pp. 313–332. [15] Puthli, R., Herion, S., Veselcic, M.: Static behaviour of joints made of thinwalled hollow sections – Pilot tests. Final report, CIDECT 5BI/5BL-5/01, 2001. [16] Zhao, X.-L., Hancock, G. J.: Butt welds and transverse fillet welds in thin cold-formed RHS members. Journal of Structural Engineering, ASCE, 121(11), 1995, pp. 1674–1682. [17] Pham, L., Bennetts, I. D.: Reliability of fillet weld design, Civil Engineering Transaction, IEAust, CE26(2), 1983, pp. 119–124.

Keywords: Thin walled hollow sections; CHS; RHS; cold-formed; thickness limits

Authors: Prof. Dr. R. Puthli, puthli@kit.edu KIT – Steel & Lightweight Structures, Research Centre for Steel, Timber & Masonry, Karlsruhe Institute of Technology, Otto-Amman-Platz 1, 76131 Karlsruhe

Visiting Professor, Department of Civil & Environmental Engineering, National University of Singapore Dipl.-Ing. Andreas Lipp, andreas.lipp@kit.edu KIT – Steel & Lightweight Structures, Research Centre for Steel, Timber & Masonry, Karlsruhe Institute of Technology Otto-Amman-Platz 1, 76131 Karlsruhe

Prof. Dr. Ir. J. Wardenier, j.wardenier@tudelft.nl Professor Emeritus Structural & Building Engineering, Faculty of Civil Engineering & Geosciences, Delft University of Technology, Netherlands

Prof. Dr.-Ing. Thomas Ummenhofer, thomas.ummenhofer@kit.edu KIT – Steel & Lightweight Structures, Research Centre for Steel, Timber & Masonry, Karlsruhe Institute of Technology Otto-Amman-Platz 1, 76131 Karlsruhe

practitioners. Masons, carpenters, locksmiths, roofers, draftsmen, architects, engineers, contractors, developers, experts, economists and lawyers: they are all there, men and women, both famous and forgotten. Equally present are the forces that have shaped the constructive field: institutions that direct, companies that inno-

vate, work forces that produce and controversies that emerge. These essays bring to life centuries of a history built, recorded, lived and ruined according to the changing temporalities of cities and sensibilities of societies. The themes revolve around three pillars: knowledge, people and objects. Methods and tools are improving through the development of heritage restoration and digital technologies. New historical topics are appearing such as energy, natural and technological risk prevention, material recycling, diffusion and transfer of knowledge in colonial situations, modern re-appropriation of old techniques, legal frameworks, economics and institutions, craftsmen’s tasks, construction site organization, labour in construction, contractor responsibility and public authority involvement in building industries. Two thousand pages covering more than 12 millennia: these 3 volumes are a monumental challenge equal to a constructive venture.

Book reviews Carvais, R., Guillerme, A., Nègre, V., Sakarovitch, J. (eds.): Nuts & Bolts of Construction History – Culture, Technology and Society LIBRAIRIE PICARD, Paris, 2012 3 volumes, 2082 pages, 17 × 24 cm, softcover, 1150 illustrations. ISBN 978-2-7084-0929-3; € 120 This rich collection of 3 volumes presents a state of international research in the history of construction. It is organized through 240 independently constituted elements and defends a history of construction open to all cultures, desiring to balance the engineering sciences with the humanities and social sciences. This book seeks to update existing axes of research by taking into account the profound changes sweeping across our planet through the framework of sustainable development and cohabitation. The act of building is excavated by archaeologists, leafed through by archivists and construed by historians and

Steel Construction 6 (2013), No. 1

11_034-038_Puthli_(1-8)_cs6.indd 38

01.02.13 08:50

Reports DOI: 10.1002/stco.201300009

Selecting materials for fastening screws for metal members and sheeting Dedicated to Prof. Dr.-Ing. Helmut Saal on the occasion of his 70th birthday

Thomas Misiek Saskia Käpplein Detlef Ulbrich

This paper deals with the parameters for choosing the materials for fastening screws used in connections involving thin-walled sections and thin sheeting. Different types of corrosion processes and repeated bending due to thermal elongation are identified as the most important parameters; these are explained in detail here. Based on that, some general recommendations for choice of material are given.

1 Introduction Thin-walled building components such as trapezoidal and corrugated profiled sheeting, cassettes (liner trays) and sandwich panels as well as cold-formed sections are typically fixed by thread-forming screws: on the one hand, self-tapping screws, where pre-drilling is necessary, and on the other hand, self-drilling screws, where drilling and thread-forming are combined in one operation. Most fastening screws (usually referred to simply as “screws”) are made from stainless steel or zinc-plated carbon steel. Since the aforementioned building components mostly involve external walls or roofs exposed to the weather, washers with a scorched EPDM sealing (so-called sealing washers) or EPDM sealing rings are necessary. The metallic part of the sealing washers is made from stainless steel, carbon steel or aluminium. Selecting materials for fastening screws and washers must take into account safety (durability and loadbearing capacity of corroded fasteners, resistance of fasteners to repeated bending caused by thermal movement) and aesthetics aspects. The latter aspect is important because thin sheeting is often used for façades, and corrosion products will affect the appearance.

Fig. 1. Fastening screws: a) self-drilling screws, b) self-tapping screw

This paper provides some guidance on choice of materials for fastening screws in connection involving thin-walled sections and thin sheeting. The guidance is based on [1] and the authors’ own experience as experts in liability cases, amended by the results of tests, some of which have already been published in [2]. The paper can be seen as an addendum to [3] and [4], which do not cover this topic. Although this information primarily concerns screws, in principle, it can be transferred to related types of fasteners such as blind rivets and powder-actuated pins.

2 Fastening screws 2.1 Preliminary remarks Fig. 1 shows several examples of typical fastening screws. Most of them have a sealing washer, one a sealing ring. As it is quite usual, the one with the sealing ring has a mushroom head, whereas all the others have hexagonal heads. The self-drilling screw is also shown with a shank (i. e. unthreaded portion), which is used, for example, for crest fixings or fixing sandwich panels. In the case of sandwich panels, screws with an additional thread under the head are very often used, the intention of which is to help achieve a watertight connection.

2.2 Materials 2.2.1 Carbon steel Carbon steel screws are usually made of case-hardened or heat-treatable steel [5]. Steel grade 1.1147 is a quite common material. Owing to the heat treatment, carbon steel fasteners attain a high strength with surface hardness values of about

Fig. 2. Screws after neutral salt spray test: carbon steel screw (left) and austenitic stainless steel (right)

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 39

01.02.13 08:51

Reports 530 HV0.3 and core hardness values of 320–400 HV10. However, these high hardness values are accompanied by the risk of hydrogen embrittlement. Carbon steel screws must be protected against corrosion. The most common type of corrosion protection is galvanic zinc plating to EN ISO 4042 [6]. It should be pointed out that fasteners listed in a European technical approval usually have a zinc coating specified as A3K, which means that the fasteners are zinc-plated with a coating thickness ≥ 8 µm and passivated; the usual standard is just A2K with a coating thickness ≥ 5 µm. Hot-dip galvanizing of such screws is not possible because the thick layers of zinc would clog the threads. Zinc flake coating systems with products such as Ruspert and Dural are also used. These coatings have an organic or inorganic matrix with dispersed zinc and aluminium particles. Unfortunately, these coatings may be damaged during transport or by abrasion during installation, possibly even scraped off. It is therefore difficult to assess such coatings.

2.2.2 Stainless steel 2.2.2.1 Preliminary remarks Stainless steels for screws can be classified according to the system given in EN ISO 3506-4 [7], which specifies property classes (20H to 40H, corresponding to 200–400 HV10). The property classes that can be achieved depend on the steel group (austenitic, martensitic or ferritic stainless steel). Stainless steels achieve their corrosion resistance through a passive layer of chromium oxide.

2.2.2.2 Austenitic stainless steel Designations such as steel grades A2 and A4 for austenitic stainless steels are quite familiar and refer predominantly to the corrosion resistance. According to EN ISO 3506-4, austenitic stainless steels for screws contain 15– 20 % chromium and 8–19 % nickel. Austenitic stainless steel A4 also contains a significant amount of molybdenum (2–3 %). Austenitic stainless steels cannot be hardened by heat treatment, but by cold working. Increasing the surface hardness by nitriding is also possible. If self-drilling screws for drilling into steel are required, a drill-point made of hardened carbon steel has to be welded to the tip of the screw. After installation, only the stainless steel part of the screw should form part of the loadbearing system, and not the welded drill-point. Drilling into aluminium is possible with drill-points made of stainless steel.

2.2.2.3 Martensitic stainless steel According to EN ISO 3506-4, martensitic stainless steels for screws contain 11.5–18 % chromium. Martensitic stainless steels can be heat treated, which allows the production of self-drilling screws in one piece complete with a drillpoint (i.e. does not have to be welded on). Corrosion resistance is considerably lower than for austenitic stainless steel and their high hardness after heat treatment makes them vulnerable to hydrogen embrittlement and stress corrosion cracking.

2.2.2.4 Ferritic stainless steel According to EN ISO 3506-4, ferritic stainless steels for screws contain 15–18 % chromium. Ferritic stainless steels cannot usually be heat treated. They do not play a significant role in loadbearing connections involving thin-walled building components, so will not be considered in the following.

2.2.3 Aluminium Screws made of aluminium usually consist of wrought aluminium alloys 6000 or 7000. As they are rather soft, applications for aluminium for screws are limited to, for example, self-tapping screws for fixing sheets to timber supporting structures. Aluminium achieves its corrosion resistance through a passive layer. Up to now, no ETA for screws made of aluminium has been published; therefore, they will not be dealt with here.

2.3 State of the art in regulations European technical approvals (ETAs) for screws have been available since 2010. The approvals specify characteristic resistance values for different loading situations depending on type of fastener, material, thickness of components to be connected, etc. Regarding corrosion protection, the information given in the ETAs is rather weak, and expressed as follows: “The intended use comprises fastening screws and connections for indoor and outdoor applications. Fastening screws that are intended to be used in external environments with high or very high corrosion category according to the standard EN ISO 12944-2 should be made of stainless steel.” and “Fastening screws completely or partly exposed to external weather or similar conditions are made of stainless steel or are protected against corrosion. For the corrosion protection the rules given in EN 1090-2:2008+ A1:2011, EN 1993-1-3:2006+AC:2009 and EN 1993-1-4: 2006 are taken into account.” The formulation in the ETAs has to respect the different traditions in European countries. Whereas some countries have banned carbon steel fasteners from applications in external environments or applications with comparable moisture conditions, others have not, or have no regulations at all. The current ETAs do not refer directly to the informative Annex B of both EN 1993-1-3 [8] and EN 1999-1-4 [9], which gives further recommendations on choice of material. But in general, for the corrosion protection of the screws, the information given there should be taken into account. Table 1 gives recommendations for the preferred screw materials depending on the materials of the components and the corrosivity of the environment. The corrosivity of the environment is classified by referring to the corrosivity categories of EN ISO 12944-2 [10] (Table 2), which of course do not take into account the microclimate (locally increased moisture, concentration of salts, etc.). More strenuous requirements might be necessary for specific situations and construction details, e. g. if screws are posi-

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 40

01.02.13 08:51

Reports Table 1. Fastener material with regard to corrosion environment according to EN 1993-1-3 and EN 1999-1-4 (abbreviated and edited) Corrosivity category C1 C2

C5-I

C5-M

Material of fastener Sheet material

Aluminium

Electrolytically galvanized steel, coating thickness ≥ 8µm

Stainless steel, case-hardened, 1.4006 (C1)

Stainless steel, 1.4301 (A2)

A, B, C

D, E, S

–

C, D, E

–

C, E

–

(×)

–

(×)

–

(×)

–

(×)

–

(×)

–

(×)

D*)

–

(×)

–

(×)

D*)

–

(×)

–

Abbreviations and footnotes ×

Type of material recommended from corrosion standpoint

(×)

Type of material recommended from corrosion standpoint under the specified condition only, insulation washer of material resistant to ageing between sheeting and fastener

Aluminium irrespective of surface finish

Uncoated steel sheet

Hot-dip zinc-coated (Z275) or aluzinc-coated (AZ150) steel sheet

Hot-dip zinc-coated plus organic coating

Aluzinc-coated (AZ185) steel sheet

–

Type of material not recommended from corrosion standpoint

Stainless steel

Always check with sheet supplier

Table 2. Atmospheric corrosivity categories according to EN ISO 12944-2 [10] and examples of typical environments Corrosivity category

Corrosivity level

Examples of typical environments in temperature climate (informative) Exterior

Interior

very low

–

Heated buildings with clean atmospheres, e.g. offices, shops, schools, hotels.

low

Atmospheres with low level of pollution. Mostly rural areas.

Unheated buildings where condensation may occur, e.g. depots, sport halls.

medium

Urban and industrial atmospheres, moderate sulphur dioxide pollution. Coastal areas with low salinity.

Production rooms with high humidity and some air pollution, e.g. food-processing, plants, laundries, breweries, dairies.

high

Industrial areas and coastal areas with moderate salinity.

Chemical plants, swimming pools, coastal shipyards and boatyards.

C5-I

very high (industrial)

Industrial areas with high humidity and aggressive atmospheres.

Buildings and areas with almost permanent condensation and with high pollution.

C5-M

very high (marine)

Coastal and offshore areas with high salinity.

Buildings and areas with almost permanent condensation and with high pollution

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 41

01.02.13 08:51

Reports tioned in cavities or voids. This is the case with external wall claddings with a ventilation cavity, or with roofs and façades in the form of multi-layer shells where corrosive agents could accumulate and moisture could infiltrate. Table 1 does not cover screws with zinc flake or similar coatings because of the different properties of the coatings available. Experience with such coated screws for flat roof systems tested according to ETAG 006 [11] has revealed the vulnerability of the coatings. In addition, it must be mentioned that according to these informative annexes, unprotected screws made of steel without zinc plating may be used in corrosivity category C1. In fact, this results in the same application range and assumed corrosion resistance for both zinc-plated and unprotected screws!

3 Types of corrosion 3.1 Atmospheric corrosion (general corrosion) Atmospheric corrosion is the development of a uniform layer of oxide (rust) on carbon steel screws under the influence of neutral water or a humid atmosphere. Pollution will increase the corrosion problem. Atmospheric corrosion can be found in nearly all applications for fastening screws in roofs and walls. Atmospheric corrosion leads to a reduction in cross-section, thus a reduction in the loadbearing capacity of the connection. As rust develops, so the aesthetics are also affected – and not only the screw itself: rusty draining water can also stain façade or roof sheets. In cases where the washers are affected, too, leakage may become a problem. Corrosion by aeration cells is a special case of atmospheric corrosion in areas with oxygen deficiency. Examples are screws passing through wet insulation materials or seam fasteners with capillary moisture between the sheets. Examples of damage can be found in [1].

3.2 Galvanic corrosion (bimetallic corrosion) Galvanic corrosion occurs when two metals with a sufficiently different electrode potential (expressed by their positions in the electrochemical series) come into contact in the presence of an electrolyte (e. g. moisture from rain). An electric current is generated by the difference in the electric potential. The current causes the less electropositive metal (anode) to corrode by the dissolution of ions. Electrons react with hydrogen ions and form atomic or molecular hydrogen, evolving at the cathode. Important parameters are the relative positions of the two metals in the electrochemical series and therefore their potential difference, which depends on the type of electrolyte. It is important to distinguish between the standard potential (determined with the standard hydrogen electrode) and the practical potential, e. g. in seawater or acidic water. Another important parameter is the ratio of surface of cathode and anode. Potential differences become less problematic if the surface ratio of anode to cathode is high. That is why stainless steel or aluminium fasteners shall be used for fixing aluminium sheets or aluminium components in general. If carbon steel fasteners are to be used, the thin layer of galvanic zinc plating would be eroded within a shorts time but the remaining carbon steel fastener would not be affected due to the higher position of steel in

the electrochemical series compared with aluminium. From then on the protective effect of the zinc is missing. On the other hand, using aluminium fasteners for fixing steel sheet is not recommended from the point of view of galvanic corrosion. The surface ratio of cathode (base metal carbon steel) to anode (less noble metal aluminium) is low, leading to an accelerated rate of corrosion. Aluminium fasteners could be used for fixing aluminium parts to aluminium supporting structures. As fasteners made from aluminium are not that common and have other disadvantages, only stainless steel fasteners should be used for fixing aluminium components in areas with at least a slight likelihood of electrolytes occurring. Otherwise, the use of stainless steel fasteners in aluminium components in severe maritime environments leads to severe corrosion of the aluminium adjacent to the fasteners due to the aggressive electrolyte. Here, the connection should be shielded from the influence of seawater by coating, grout, etc. The consequences range from spoiling the appearance of façades and structures to failure of the connection due to a reduction in the cross-section of the corroded fastener if bimetallic corrosion is not taken into account sufficiently.

3.3 Hydrogen embrittlement Hydrogen embrittlement describes the reduction in ductility and the subsequent brittle fracture of metals caused by hydrogen. Recombination of atomic hydrogen diffusing into the metal, especially at the grain boundaries, causes pressure in cavities and tensile stresses in the atomic lattice of the metal matrix. Owing to these stresses, the term “hydrogen-induced stress corrosion cracking” is also used. The term “cathodic stress corrosion cracking” refers to the electrochemical process. Tensile stresses (e. g. residual stresses from cold forming, but also tensile stresses from tightening) increase susceptibility because of lattice deformation, which eases the diffusion of hydrogen into the steel and its most highly stressed parts. Non-alloy steels such as the ferritic carbon steels with tensile strengths of about 1000 N/mm² or hardness values above 320 HV10 are prone to hydrogen embrittlement. These steel grades are typically used for screws. Martensitic steels with high strength values are also prone to hydrogen embrittlement [12], whereas austenitic (stainless) steels are not affected. Sources of hydrogen are production processes such as pickling prior to galvanic plating (primary hydrogen embrittlement, delayed brittle fracture). EN ISO 4024 gives recommendations for mitigating hydrogen embrittlement. The options comprise stress relieving or tempering, which of course can only be applied during production. But the standard also states that complete elimination of hydrogen embrittlement cannot be assured. Hydrogen can also evolve from corrosion processes, e. g. from galvanic corrosion (secondary hydrogen embrittlement). A typical example of a brittle failure of a screw by hydrogen embrittlement is fixing an aluminium sheet to a steel structure with a carbon steel screw. Exposure to weathering leads to galvanic corrosion (rain as electrolyte) with the release of atomic hydrogen. The thin layer of zinc coating is not diffusion-resistant to hydrogen or may even have been already damaged during installation or the corrosion

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 42

01.02.13 08:51

Reports ciple, there are two reasons for pitting corrosion: as local damage to the passive layer of, for example, stainless steel or aluminium, or as galvanic corrosion of single grains or precipitation of a base metal (e.g. an alloying element) in the surrounding noble metal. Usually, pitting corrosion and its consequences are not of importance for the applications of screws discussed here.

3.6 Crevice corrosion

Fig. 3. Fracture surface of a screw after failure due to hydrogen embrittlement

Crevice corrosion is a localized form of attack that is initiated by the differences in the oxygen levels between the creviced and exposed regions. Crevices occur around the threads of the fastening screws and the components being connected. It is not likely to be a problem except in stagnant solutions where a build-up of chlorides can occur. The severity of crevice corrosion very much depends on the geometry of the crevice: the narrower and deeper the crevice, the more severe the corrosion. In principle, pitting and crevice corrosion are similar phenomena, but the attacks start more easily in a crevice than on an open surface.

4 Governing parameters for choosing materials regarding corrosion

Fig. 4. Fracture surface of a screw after overtightening (ductile fracture)

process. The hydrogen diffuses into the screw, especially at the grain boundaries, and accumulates at the most highly stressed parts (usually at the radius between shank and head). Brittle failure will occur in the shank near the head. Fig. 3 shows a typical fracture surface – the splits at the grain boundaries can be seen well. For comparison, Fig. 4 shows a fracture surface after a ductile failure.

3.4 Stress corrosion cracking Stress corrosion cracking (or “anodic stress corrosion cracking”) as it will be discussed here is a form of corrosion of ferritic, austenitic and martensitic stainless steels under the influence of chloride, acidic or oxidizing electrolytes. It also results in a reduction in ductility and resistance, leading to a brittle intergranular or transgranular failure. Mechanical tensile stresses (e. g. residual stresses from cold forming, but also tensile stresses from tightening or from external loads) increase susceptibility. Depending on the electrolyte, increased ambient temperatures might be necessary for failure. Usually, stress corrosion cracking and its consequences are not of importance for the applications of screws discussed here.

3.5 Pitting corrosion Pitting corrosion is a highly localized form of corrosion that can be found in stainless steel or aluminium. In prin-

Fastener materials must be selected depending on the materials of the structural parts to be connected, the stressing by corrosivity of the surroundings and the intended life cycle. The most important point is corrosion which is also influenced by the materials of the parts to be connected. It must be pointed out that the high strength of the fasteners does not have a significant effect on the resistance of the connection because with thin-walled sections and thin sheeting, failure of the building components being connected is the governing parameter. Stresses due to forces therefore do not normally become critical for the choice of material for fasteners. Corrosivity of the surroundings depends on moisture conditions, air pollution (dust, which may dissolve in water, chloride in industrial and marine environments or from road de-icing salts, sulphur dioxide emitted from power plants and traffic, etc.) and period of exposure. Moisture may gain access to the screws by way of weather conditions, but also via condensation at thermal bridges. Some thermal insulation materials such as mineral wool can work like a sponge, absorbing water. If screws are installed through a sandwich panel containing saturated mineral wool, corrosion directly affects the loadbearing part of the screw. It is also important to realize that conditions may get worse as corrosive agents accumulate. One prominent example is the accumulation of both corrosive agents such as de-icing salts and moisture behind external walls with a ventilation cavity. On the other hand, occasional rain may even have a cleaning effect on the screw. So detailing is also a very important parameter.

5 Bending of fasteners under repeated loading Different temperatures in the connected components (e. g. between a trapezoidal profiled sheet panel in a façade heated by the sun and a section of the supporting struc-

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 43

01.02.13 08:51

Reports Table 3. Choice of corrosion resistance class according to German national approval Z-30.3-6 Exposure

Humidity, yearly average value U of humidity

Chloride content of surrounding area, distance M from the sea, distance S from busy roads with road salt application

Exposure to redox-affecting chemicals (e. g. SO2, HOCl, Cl2, H2O2)

pH-value on surface

Location of structural parts

Corrosion resistance class

Exposure class

Criteria and examples

SF0

dry

U < 60 %

SF1

seldom moist

60 % ≤ U < 80 %

SF2

often moist

80 % ≤ U < 95 %

SF3

permanent moist

95 % < U

SC0

low

SC1

medium

SC2

high

M ≤ 1 km S ≤ 0.01 km

SC3

very high

indoor swimming pool, road tunnel

SR0

low

rural, urban

SR1

medium

industrial area

SR2

high

indoor swimming pool, road tunnel

SH0

alkaline (e. g. with contact to concrete)

9 < pH

SH1

neutral

5 < pH ≤ 9

SH2

low acidic (e. g. with contact to wood)

3 < pH ≤ 5

SH3

acidic (exposure to acids)

pH ≤ 3

SL0

indoors

indoors, heated and not heated

SL1

outdoors, exposed to rain

exposed structures

×3)

roofed structures

×3)

SL2

SL3

outdoors, accessible but protected from weather outdoors, non-accessible4), ambient air has access

rural, urban, M > 10 km, S > 0.1 km industrial area, 10 km³ M > 1 km, 0.1 km³ S > 0.01 km

III

× ×

× ×1) ×2) × ×1) ×2)

× × ×

accumulation of pollutants on surface by air pollution, cleaning not possible

Only the exposure leading to the highest corrosion resistance class (CRC) has to be taken into account. No higher requirements result from the coincidence of exposure conditions. Contaminated steel surfaces (e.g. paint, grease, dirt) may lead to lower corrosion resistance. 1)

If accessible structures are cleaned regularly or exposed to rain, corrosion will be much lower and the CRC may be reduced by one class. Otherwise the CRC has to be increased by one class if corrosion-relevant substances can be deposited on and remain on the surfaces of structural parts. 2) If accessible structures are cleaned regularly, corrosion will be much lower and the CRC may be reduced by one class. 3) If the life cycle is limited to 20 years and pitting corrosion up to 100 mm is tolerated, CRC I may be chosen (no visual demands). 4) Structures are classified as non-accessible if an inspection of their condition is extremely difficult and a necessary rehabilitation is very expensive.

ture) and resulting differences in thermal elongation may lead to additional stresses in the fasteners. In fixings where both components are directly adjacent to each other, the stresses will lead to shear forces in the fastener and to an elongation of the holes in the components. Otherwise, the distance between the components being connected will lead to additional bending stresses in the fastener. This is the case for crest fixings in trapezoidal profiled sheeting and for sandwich panel fixings, where the screw passes through

both faces of the panel. If thermal elongation changes repeatedly from day to night, the bending stresses in the fasteners will also be repeated. This has to be taken into account when designing the fasteners, e. g. according to [4]. The resistance of the connection to repeated bending, expressed by the allowable head deflection max. u, depends on the thickness tII of the supporting structure and the corresponding degree of rotational restraint provided by that structure. Whereas for smaller thicknesses tII, local deforma-

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 44

01.02.13 08:51

Reports

Fig. 5. Allowable value for head deflection of austenitic stainless steel screws [2]

Fig. 6. Allowable value for head deflection of carbon steel screws

Table 4. Allocation of steel grades to corrosion resistance classes No.

Steel name1)

Steel grade1)

Steel grade2)

Type of stainless steel3)

X2CrNi12

1.4003

X6Cr17

1.4016

not suitable for fastening elements

X5CrNi18-10

1.4301

X2CrNi18-9

1.4307

A2L

X3CrNiCu18-9-4

1.4567

A2L

X6CrNiTi18-10

1.4541

A A

X2CrNiN18-7

1.4318

X5CrNiMo17-12-2

1.4401

X2CrNiMo17-12-2

1.4404

A4L

X3CrNiCuMo17-11-3-2

1.4578

A4L

X6CrNiMoTi17-12-2

1.4571

X2CrNiMoN17-13-5

1.4439

X2CrNiN23-4

1.4362

X2CrNiMoN22-5-3

1.4462

1.4539

A A

X1NiCrMoCu25-20-5

X2CrNiMoNbN25-18-5-4

1.4565

X1CrNiMoCuN25-20-7

1.4529

X1CrNiMoCuN20-18-7

1.4547

1) 2) 3) 4) 5)

Corrosion resistance class CRC4) I / low

II / medium

III / high

IV / very high

according to EN 10088-1 according to EN ISO 3506 F – ferritic steels; A – austenitic steels; AF – austenitic-ferritic steels For choice of corrosion resistance class (CRC) see Table 3. Actually not covered; therefore the steel grade according to EN 10088-1 should be used.

tion of the structure is the governing effect (allowing for large values u of deformation/deflection), for larger thicknesses tII, full restraint is achieved and the effects of geometry and screw material dominate the resistance. Bending tests with fasteners under repeated loading have shown that stainless steel fasteners behave much better than carbon steel fasteners. Bending tests according to [4] were performed with austenitic stainless steel screws and carbon steel screws. Figs. 5 and 6 show the results of the tests. The maximum allowable value of head deflection max. u divided by the length of the cantilever (equal to the thickness dc of a sand-

wich panel) is plotted against the thickness tk,II, the core sheet thickness of the supporting structure. The results cover tests with different screws, both self-drilling and self-tapping, from different manufacturers. Statistical evaluation leads to the design curves given in the figures. The design curve for austenitic stainless steel fasteners can be written as

max. u = 0.3mm ⋅

dC ≥ 0.07 ⋅ dC t k,II

and for carbon steel fasteners as

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 45

(1)

01.02.13 08:51

Reports max. u = 0.1mm 2 ⋅

dC t 2k,II

≥ 0.023 ⋅ dC

(2)

Whereas max. u for smaller thicknesses tk,II does not depend that much on the material of the screw, material properties become important for larger thicknesses and higher degrees of rotational restraint. With greater thicknesses tk,II, the allowable value max. u for the head deflection of austenitic stainless steel fasteners is three times as high as the value for carbon steel fasteners. The reason for this difference in behaviour is that the high hardness of case-hardened carbon screws prevents a reduction in critical stress peaks through local plastic deformation. Although head deflection is usually not critical for austenitic stainless steel screws, it must be checked carefully if carbon steel screws are to be used.

6 Recommendations and summary The following recommendations are based on the effects described in sections 3 to 5 and are backed up by considerable experience on a national level: – Screws completely or partly exposed to the weather or comparable moisture conditions should be made from austenitic stainless steel. This does not refer to welded drill-points, but it has to be checked that the screw-in length is large enough to ensure that carbon steel parts are not part of the loadbearing system. – The length of the construction period should be taken into account, also with respect to the time of year of the construction works. – Austenitic stainless steels of higher grades (e. g. A4) are necessary in applications where concentrations of corrosive agents may accumulate or in surroundings with higher corrosivity. This may be the case with screws in the ventilation cavities or voids of external walls or otherwise shielded from direct rain and where regular cleaning is not foreseen or not possible. Tables 3 and 4, published in [13], help the designer to select the right stainless steel grade with respect to the corrosion resistance of fastening elements. – Screws made of carbon steel or martensitic stainless steel are not suitable in cases where minimum corrosion resistance requirements exist. Carbon steel screws, including electrolytically galvanized or coated fasteners, and screws made of martensitic stainless steel should only be used where moisture does not affect them. This covers: – fastening the inner shells of multi-shell roof and wall structures (decking profiles or cassettes) around dry and predominantly closed rooms provided the outer shell prevents the entry and accumulation of corrosive agents and rain (outer shell made of sheeting), – fastening the decking profiles of unventilated single shell roofs around dry and predominantly closed rooms with insulation on the outer side (typical flat roof applications with insulation membranes), – ceiling systems over dry and predominantly closed rooms. – Fastening of aluminium sheeting should only be done with screws made of stainless steel (or aluminium) ex-

cept in severe maritime environments, which require additional protection measures for the connections. – Galvanic corrosion should be taken into account when designing and detailing structures. – Although repeated bending of fasteners due to thermal elongation and movement is not usually critical for the design of austenitic stainless steel fasteners, care should be taken when using hardened martensitic or carbon steel screws. References [1] Wieland, H.: Korrosionsprobleme in der Profilblech- und Flachdach-Befestigungstechnik [corrosion problems in roofing and siding], Befestigungstechnik Bau, SFS intec AG, Heerbrugg, 1988. [2] Misiek, T., Käpplein, S., Hettmann, R., Saal, H., Ummenhofer, T.: Rechnerische Ermittlung der Tragfähigkeit der Befestigung von Sandwichelementen [computation of loadbearing capacity of fixings of sandwich panels], Bauingenieur 86 (2011), pp. 418– 424. [3] ECCS TC 7. The Testing of Connections with Mechanical Fasteners in Steel Sheeting and Sections. ECCS pub. No. 124, Brussels, 2009. [4] ECCS TC 7 & CIB W56. Preliminary European Recommendations for testing and design of fastenings for sandwich panels. CIB report pub. 320/ECCS pub. No. 127, CIB/ECCS, Rotterdam/Brussels, 2009. [5] EN ISO 10066:1999: Drilling screws with tapping screw thread – Mechanical and functional properties. [6] EN ISO 4042:1999: Fasteners – Electroplated coatings. [7] EN ISO 3506-4:2009: Mechanical properties of corrosion-resistant stainless steel fasteners – Part 4: Tapping screws. [8] EN 1993-1-3:2006+AC:2009: Eurocode 3: Design of steel structures – Part 1-3: General rules – Supplementary rules for cold-formed members and sheeting. [9] EN 1999-1-4:2007+AC:2009: Eurocode 9: Design of aluminium structures – Part 1-4: Cold-formed structural sheeting. [10] EN ISO 12944-2:1998: Paints and varnishes – Corrosion protection of steel structures by protective paint systems – Part 2: Classification of environments. [11] ETAG 006: Systems of Mechanically Fastened Flexible Roof Waterproofing Membranes. EOTA, Brussels, 2000. [12] Landgrebe, R., Gugau, M., Friederich, H.: Anfälligkeit gewindeformender Schrauben aus korrosionsbeständigen Stählen gegenüber Spannungsrisskorrosion [susceptibility of thread-forming screws made from stainless steels with relation to stress corrosion cracking], Materials and Corrosion 53 (2002), pp. 165–175. [13] German technical approval Z-30.3-6:2009-04: Erzeugnisse, Verbindungsmittel und Bauteile aus nichtrostenden Stählen [products, fastening elements and structural parts made of stainless steels]. Keywords: thin-walled structures; screws; fasteners; connections; corrosion

Authors: Dr.-Ing. Thomas Misiek, Breinlinger Ingenieure Tuttlingen – Stuttgart, Kanalstr. 1–4, 78532 Tuttlingen, thomas.misiek@breinlinger.de Dipl.-Ing. Saskia Käpplein, Versuchsanstalt für Stahl, Holz und Steine Karlsruher Institut für Technologie, Otto-Ammann-Platz 1, 76131 Karlsruhe, saskia.kaepplein@kit.edu Dipl.-Ing. Detlef Ulbrich, Deutsches Institut für Bautechnik (DIBt), Kolonnenstraße 30B, 10829 Berlin, dul@dibt.de

Steel Construction 6 (2013), No. 1

12_039-046_Misiek (2012-1-9)_cs6.indd 46

01.02.13 08:51

Reports DOI: 10.1002/stco.201300010

Monopile foundations for offshore wind turbines – solutions for greater water depths Rüdiger Scharff Michael Siems

A study into the feasibility of monopile foundations at sites in the North Sea with water depths as great as 35 m has been carried out in which various options for the geometric configuration and the selection of material were considered. In parameter studies, simplified design methods were applied to assess the effects of the individual load components at the draft design stage. The fatigue limit state becomes more and more relevant as the water depth increases; therefore, dynamic effects must be examined with special care. Turbine concepts with low RNA mass and low rated speed help to achieve the desired design in the soft-stiff regime. As a result, it can be said that monopile foundations with their great manufacturing advantages can be constructed for water depths beyond the current limits of practical experience if the logistical challenges in handling large masses are solved.

1 Introduction Since the “Alpha Ventus” offshore wind farm successfully went into operation in 2010, further wind farm projects have been implemented or are about to enter their installation phase in the North Sea and Baltic Sea. These projects supply considerable additional experience with different kinds of foundation for wind turbine generators (WTG). The monopile, which consists of a foundation pile and a transition piece, is normally a highly efficient type of foundation and is therefore at present the most popular option at water depths not exceeding 20 to 25 m. At greater depths, which is where the majority of approved wind farms in the Exclusive Economic Zone (EEZ) in the German Bight are located, more complex multi-member designs, such as tripods, jackets or tripiles, are proposed. A concept study was carried out to examine the possibility of using monopile foundations at greater depths of water, too – max. 35 m. Owing to the special advantages of the monopile (it is easy to fabricate and install), the ability to extend the field of application of this type of foundation is of general interest The benefits are to be found in the possibility of using higher-strength fine-grain structural steels and modern post-weld treatment methods. There is a correlation between requirements that follow from the ultimate limit state (ULS), fatigue limit state (FLS) and serviceability limit state (SLS) when designing a wind turbine support structure, and the static and dynamic requirements that wind turbines have to meet. To provide the required system rigidity at increasing water depths, larger

pile diameters and penetration lengths have to be considered. The consequently larger component masses have to be handled as part of the logistics and installation concepts, which are not the focus of this study.

2 Dynamic behaviour 2.1 Soft-stiff design for the support structure Of the different design conditions that have to be accounted for, the required eigenfrequency intervals proved to be the dominating condition in the context of this concept study. For current wind turbine sizes, the established design criterion for the natural vibration characteristics of onshore or offshore wind farms is the so-called soft-stiff design. In this case the lowest natural bending frequencies of the complete system in the longitudinal and transverse directions are adjusted so that they remain above the excitation frequency due to rotor imbalance (1P) and below the excitation frequency due to the blade passing frequency (3P) for the entire operating range of the turbine. Wherever possible, additional safety margins have to be complied with; a partial overlap of eigenfrequencies and excitation intervals is acceptable only when the plant control system also provides for vibration control. Under real conditions, the turbine manufacturer will define the permissible frequency interval (see the example of a Campbell diagram in Fig. 1). As the slenderness of the support structure increases as a result of the greater water depth, so the first eigenfrequencies of the complete system decrease. A design in the “soft-

Fig. 1. Example of a Campbell diagram

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 47

01.02.13 08:52

Reports soft” region below 1P excitation would therefore be a conceivable solution. However, this leads to a very soft system, which will not meet the SLS requirements.

– Monopile and tubular tower diameters – RNA mass – Hub height

2.2 Options for influencing the dynamic behaviour

The graphs below show the results for two geometric configurations. Configuration A has a monopile diameter of 7500 mm and a tower diameter of 5500 mm; the values for configuration B are 8500 mm and 6500 mm respectively. The penetration depths and the plate thicknesses assumed for the given conditions are based on empirical values. The graphs always represent a mean value of the first natural bending frequency depending on water depth for three RNA masses, together with a bandwidth that results from the range of soil parameters. The graphical results are based on a hub height (HH) of 100 m; assessments for other hub heights are shown as numerical values.

To achieve the lower limit of the required frequency interval for the soft-stiff design, the diameters of the monopile, the transition piece and the tubular tower are available as influencing parameters. The pile penetration length has little significance because additional serviceability conditions have to be complied with when determining this factor. The plate thicknesses, which are primarily dimensioned with a view to strength, are also of minor importance for the natural vibration characteristics. Regarding the mass inertia, the mass of the rotor/nacelle assembly (RNA) is of decisive significance, in addition to the dead weight of the support structure. There is a fourth power dependence between the flexural rigidity of the tubular support structure and its diameter, whereas its mass increases with the square of its diameter.

This would mean that the diameter of the structure and the eigenfrequency are directly proportional. However, this does not apply in practice for two reasons: on the one hand, the RNA acts as an additional, large constant summand in the mass term and, on the other, the diameter of the tubular tower is normally determined by the turbine manufacturer, which is why options for variation in the foundation design are limited to part of the cantilever system. This implies that, technically, the intended system eigenfrequency can only be varied via the flexural rigidity of the monopile, to which economical limitations apply. For this reason, the monopile foundation design for great water depths can only be successful in connection with current developments for offshore wind turbines of the multi-megawatt class. There are two trends that have a favourable effect on the problem of the vibration characteristics. At present, developments are focusing on concepts for systems without gear units, which allow the RNA mass of the WTG to be considerably reduced. On the other hand, systems with a larger swept area are entering the market, which reach their rated capacity at lower speeds. The reduced head mass on the one hand and lower 1P excitation frequencies on the other reduce the demands made on the integral stiffness of the loadbearing system. With larger systems, tubular towers with diameters of up to 6.5 m are also becoming common.

3 Optimized geometry for wave action In view of the great water depths, loads from wave action and ocean current dominate the WTG-specific loads at both the ULS and the FLS. Fig. 4 shows the ULS bending moment contributions due to wave action, wind and imperfections using the example of a monopile with a constant

Fig. 2. Graph showing how system eigenfrequencies depend on water depth, RNA mass and hub height – configuration A

2.3 Parameter study of the eigenfrequency characteristics To allow a first assessment of the system eigenfrequencies for preliminary design purposes, an extensive parameter study was carried out for the dynamic properties. The following influencing factors were considered: – Water depth at the location – Bandwidth of typical lateral soil stiffnesses in the North Sea

Fig. 3. Graph showing how system eigenfrequencies depend on water depth, RNA mass and hub height – configuration B

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 48

01.02.13 08:52

Reports

Fig. 4. ULS bending moment contributions for a monopile foundation with a diameter of 8000 mm in a water depth of 35.0 m

8000 mm diameter, a water depth of 35.0 m LAT and a hub height of 105 m LAT. This shows that the wave action accounts for approx. 2/3 of the design moment, in which the load combination of extreme wave and reduced wind velocity are decisive. As has been discussed above, a high system stiffness is the primary objective in the context of an economic foundation design. For a given monopile diameter, a transition has to be provided between the monopile and the smaller diameter of the tubular tower, which is normally conical and should therefore start at a high level. Alternative design options have been examined for this transition which allow a cylindrical pile to extend above the waterline and therefore achieve a high integral stiffness. This positive effect is opposed, in particular, by higher wave loads because of the large area in the waterline region with the highest particle velocities which the waves can attack. Fig. 5 compares the wave loads on a cylindrical and various conical monopile options, in each case with a constant diameter at the interface level. On the left, the overturning moment at the mudline is shown for an increasing diameter at the pile tip; the graph on the right shows the corresponding base shear forces. For the purpose of these investigations, the position of the top end of the cone was

Fig. 6. Diagram showing how eigenfrequency and fatigue behaviour depend on the position of the conical transition

always assumed to be constant, which is why the wave load for the conical pile converges towards a limit value as the diameter increases, but rises almost linearly for the cylindrical pile. This explains why design options with cylindrical foundation piles and a sudden transition between the pile and the smaller tubular tower do not produce economical design results for greater water depths. In the FLS analysis, the effects are even more distinct, which is why for the design of the conical transition, a compromise has to be found between the requirements concerning the dynamic system behaviour and the wave loads. In a parameter study, both target variables were examined as a function of the vertical position of the conical transition. For a monopile with a pile tip diameter of 8500 mm, Fig. 6 shows the first system eigenfrequency as well as the wave loads and the resultant fatigue damage relative to the corresponding values of a cylindrical pile. The damage evidently decreases considerably when the cone is located closer to the foundation, whereas the eigenfrequency only exhibits a moderate decrease. It was assumed that under structural considerations, the conical transition is placed either exclusively in the monopile or in the region of the grouted connection.

4 Design example 4.1 Conditions and load assumptions

Fig. 5. Wave loads acting on cylindrical and conical monopiles

The concept study focused on water depths â&#x2030;Ľ 25 m. At a water depth of 35 m, a preliminary design analysis was carried out for two locations, the results of which are presented below. The designs are based on an assumed new-generation WTG of the 6 MW class with an RNA mass of 375 t and a low rotor speed range, so the required interval of the first eigenfrequency is between 0.21 and 0.27 Hz. The analyses start with results from preliminary load simulations, given as extreme values for the ULS and as a damage-equivalent load range at the interface level for the FLS. The hub height of the system is at 105 m LAT; the diameter of the tubular tower was assumed to be 6.5 m at the bottom and 4.5 m at the top.

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 49

01.02.13 08:52

Reports In order to account for the soil conditions in the German Bight realistically, two representative soil profiles with the necessary geotechnical characteristics were defined. A sand location with good to very good bearing capacity throughout and medium dense to very dense sands is used as a reference value for favourable soil conditions. An alternative clay location with a near-surface cohesive layer down to a depth of 20 m provides the reference value for unfavourable soil conditions. These two locations do not allow absolute limits for the lateral soil stiffness to be represented, which was, however, not the aim of this study. The sea states occurring were defined on the basis of empirical values from different projected locations corresponding to the water depth selected. The extreme wave height with a 50-year return period is Hmax50 = 19.0 m. The long-term distribution of the wave heights was defined with fictitious scatter data, which allow a three-parameter Weibull distribution of the significant wave heights with the following characteristics: – Scale parameter: λ = 1.50 m – Shape parameter: k = 1.40 – Location parameter: h0 = 0.25 m Furthermore, there are currents with design values of 1.0 and 0.4 m/s (50-year return period) for the tide- and wind-induced components respectively. All cases assume that an active scour protection system is installed. Rules and regulations are taken from the relevant national standards. In addition, reference is made to [1] for offshore-specific concerns. The projected service life of the foundation is 25 years.

4.2 Special design aspects An integral beam model consisting of monopile, transition piece including secondary steel, tubular tower and WTG was used for all verification analyses. The lateral loadbearing capacity of the monopile is accounted for by adopting the modulus of subgrade reaction approach with non-linear spring characteristics, which have to be adjusted depending on soil type and depth below mudline. Monopiles with increasing diameter are known to exceed the range of experience of the established p-y method, which is why the results must be examined critically. There are a number of theoretical papers in this context, e. g. Wiemann et. al. [11] or Achmus [10], which mainly conclude that the modulus of subgrade reaction tends to be overestimated at greater depths. Ref. [10] proposes two methodological approaches that also start from the p-y method and can be applied in practice to account for the uncertainties present. The first is the consideration of increased bandwidths for the soil parameter scatter, which are an input for the p-y curves. Secondly, it is proposed that 3D finite element models of the pile-soil system be calibrated with a non-linear material law against the results of the p-y method for comparative piles within the range of experience, and that after simulating the loadbearing characteristics of the monopile actually used, suitable modifications be made for the spring characteristics. Since the primary goal of the concept study was to compare the effects of different design options, rather than

provide an exact representation of a specific soil profile, the former proposal was used, also with a view to the numerical effort required for representing different pile systems. The essential soil parameters – friction angle j, cohesion cu and relative strain ε50 – were varied so that the bandwidth of the initial modulus of subgrade reaction was about ± 30 % at the sand location and about ± 40 % at the clay location. The bandwidths of the ultimate bearing capacities pu are between ± 20 and ± 25 % at both locations. It turned out that these assumptions allow the system stiffness to be estimated fairly accurately, with a satisfactory scatter range for design purposes. For more refined designs, the aforementioned FE model calibration can be used. In connection with the p-y method, care must also be taken that for modal analyses in particular, a suitable secant modulus is used for the linearization of the soil stiffness required instead of an initial tangent modulus to provide a good representation of the load level considered. Some of the p-y curves that are proposed in the literature theoretically supply infinitely high tangent moduli or are evaluated at arbitrary sampling points, which is the case for the soft clay model in [5] and [6], such that unrealistic pile constraint conditions can be produced. The pile penetration length was determined based on deformation criteria. At a decreasing penetration length-topile diameter ratio L/D, stocky piles increasingly behave like rigid bodies in the soil, i. e. their curvature is not decisive for the pile head rotation observed at the mudline. Current projects with monopile foundations therefore do not normally use a vertical tangent as the design criterion for the pile penetration length, but rather the convergence of different deformation parameters, in particular rotation at the mudline under extreme loads (an overview is included in [12], for instance). Fig. 7 illustrates the conditions and the selection of penetration length for the monopile at the sand location. Intensive research is currently being undertaken regarding the effects of the high cyclic loads to which WTG monopile foundations are exposed, regarding the factors strength degradation and deformation accumulation. For an overview, see [13] and [14], for instance. Chapter 13 in [3] lists proposals for practical design work. For the purposes of the present study, no additional specifics arise for the time being. It can be assumed that the possible strength degrada-

Fig. 7. Determination of pile penetration length (example: sand location)

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 50

01.02.13 08:52

Reports tion at the ULS is sufficiently accounted for with the reductions according to [5] and [6]. The values have been calibrated for sand with load tests involving approx. 100 cycles. Experience shows that the equivalent cycle number of the design storm in accordance with [4] can be expected to be lower. The displacement accumulation under cyclic loads can be estimated with the empirical logarithmic approach yzyk = yn = 1 · (1 + t · λnN) according to [3], using the pile head displacement yN = 1 under a static load, after the load spectrum determined has been converted into an equivalent cycle number N for extreme load. A method that can be used for this conversion is, for instance, described in [12]. Under the prevailing conditions, there are some uncertainties regarding the selection of the deformation parameter t. However, the corresponding SLS analysis proved to be uncritical because the monopiles considered with the necessary large diameters have pile head displacements that are about 20 to 50 % lower than the displacements usually calculated for monopiles at a water depth of approx. 20 m in comparable soil conditions. The design analyses for the structural steelwork, including shell stability, were carried out in accordance with DIN EN 1993, with due consideration being given to [2]. Owing to the requirements imposed by the dynamic system behaviour and the fatigue strength, the utilization ratios at the ULS only reach approx. 60 % at the sand location and 70 % at the clay location for the steel grade S355 commonly used. An exception is the cable entry region. For the study, an inner cabling system with an opening in the monopile about 3 m above the mudline was used, which considerably weakens the cross-section at this point. If this opening is designed for fatigue, the ULS analysis governs, and the local wall thickness can be significantly reduced when higher-strength steels of up to S690 are applied. As a next step, the total mass of the monopile can also be considerably reduced because the wall thicknesses in this region are determined on the basis of the notch effect at steps in thickness. Together with higher-strength steels, the inner cabling solution proved to be more economic for great water depths than external cabling because in the former case no wave loads act on the J tube and no support elements are necessary. As has been mentioned above, wave action dominates at the present great water depths when compared with WTG loads – also at the FLS. Several methods are available for the FLS analysis, which differ in their prediction accuracy and numerical requirements; a suitable approach has to be selected from those available. In connection with the concept study, it was not possible to perform an integrated load simulation for the complete model in which the simultaneous stochastic wind and wave actions are considered; it would not have been useful either because it is hardly possible to address the effects of individual geometry variations directly. For this reason, the damage resulting from the two load processes was determined separately and then accumulated on the basis of the equivalent stress ranges to obtain the total damage within the service life using the superposition rule

Fig. 8. Schematic representation of the damage determination process for wave action within the time domain

∆σ eq,sum = ∆σ eq,wave 2 + ∆σ eq,WTG2 (see, for instance, [9]). Preliminary damage-equivalent load ranges were available for the WTGs. The deterministic concept for individual waves and the simulation within the time domain are commonly used in practice to calculate the damage as a result of wave action, disregarding methods in the frequency domain. The time domain simulation requires numerous individual time series to be numerically simulated in order to account for all sea states with the actual dynamic structural response (the individual steps are outlined in Fig. 8). Owing to the great numerical effort, the method cannot be applied in each design step of a study. In contrast, the deterministic concept represents the wave loads from all the sea states within a wave height exceedance diagram, which can be converted into a wave height spectrum with selectable graduation. The damage contribution from each wave height in the spectrum is determined for an individual deterministic wave of constant height and period, and finally summed up to produce the total fatigue damage. For a more exact representation, the reader is referred to [7] and [16], for instance. Owing to the lower computation effort and the fact that the effects of parameter variations are easy to follow, the deterministic concept should be favoured for the task discussed here. If dynamic amplifications in the structural response to the individual waves play a significant role, this is of particular relevance. An assessment of the dynamic amplification based on the theory of the single-mass oscillator, which is often used because of its simplicity, cannot be applied to the support structures dealt with here. Since the wave forces attack regions far below the RNA mass, the structural response does not consist of just one eigenmode. Fig. 9 shows the harmonic response of a WTG support structure at different heights in comparison to the results for a single-mass oscillator. In the wave excitation resonance region especially, substantial errors can occur if the structural response is oversimplified. It is generally advisable to verify the results of the deterministic concept for key elements of a design process using the more exact simulation method.

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 51

01.02.13 08:52

Reports

Fig. 9. Consideration of the dynamic system response in the deterministic concept for the damage calculation under wave action

4.3 Results The two monopiles in the design example have diameters of 8200 and 8500 mm at the more favourable sand location and the less favourable clay location respectively. The required penetration lengths differ by more than 10 m. Further design details are summarized in Table 1. The diameters are selected such that the first system eigenfrequencies at both locations are close to each other, which is desirable for the overall wind farm design. Besides the requirement to achieve the frequency interval for the WTG, a high system rigidity is desired for a further reason, i. e. to limit the dynamic amplifications from the proximity between the first system eigenfrequency and the wave periods in the most frequent sea states according to the Weibull distribution for a successful FLS analysis. This explains why the selection of the most efficient monopile diameter is a highly complex design task. For the sand location, the required wall thicknesses and the design results for ULS and FLS are shown graphically

in Fig. 10. Since the conditions for the monopile at the clay location are similar at approx. 5 to 10 % higher ULS utilization ratios, they are not shown separately. For design purposes, the FLS analysis is of greater relevance than the Table 1. Design results for design example in 35 m water depth Site A (sand)

Site B (clay)

Diameter of monopile

8200 mm

8500 mm

Pile penetration length

37.5 m

48.0 m

Total length of monopile

78.0 m

88.5 m

Total mass of monopile

1340 t

1550 t

Mass of S355

1275 t (95.2 %)

1490 t (96.1 %)

Mass of S460

65 t (4.8 %)

60 t (3.9 %)

Mass of transition piece

295 t

Volume/mass of grout

~ 33 m続/82 t

1st eigenfrequency

0.255 to 0.260 Hz

0.254 to 0.260 Hz

Fig. 10. Design results for the sand location

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 52

01.02.13 08:52

Reports ULS analysis for most parts of the support structure. This is why circumferential seams should be subjected to postweld treatment over a distance of approx. 25 m about the bending moment maximum. More recent post-weld treatment methods, such as the HiFIT method [17], promise to improve the fatigue strength even more than current normative methods, so the discrepancy between ULS and FLS is reduced, permitting an even more economical design.

5 Summary and conclusions A concept study was carried out for the design of monopile foundations in water depths of ≥ 25 m for locations in the German Bight. A major challenge in this context is to achieve sufficient system stiffness for the soft-stiff design, while at the same time minimizing fatigue-inducing wave loads. Both issues were addressed in parameter studies. The development of the first natural bending frequency of the complete system was determined for a broad parameter field, with water depth, ground conditions, hub height and RNA mass, for two geometric configurations. The corresponding graphs can be used for preliminary design purposes. Whereas the arrangement of the transition from the monopile diameter to the more slender tubular tower at a higher level with respect to the waterline achieves a structure with a higher flexural rigidity, this also results in higher wave loads, which strongly affects the fatigue damage. This conflict is quantified with examples. A general conclusion is that WTG concepts with a low nacelle mass and low rotor speeds as well as large tower diameters of 6000 mm and more create favourable conditions for a successful monopile foundation design at greater water depths. At water depths greater than 25 to 30 m LAT, the FLS analysis generally becomes more relevant than the ULS analysis. To mitigate this effect and to enhance the overall system efficiency, post-weld treatment is a useful approach. Attention is drawn to recent developments in post-weld treatment techniques with a high increase in fatigue life. The monopile foundation design possible for a water depth of 35 m is presented in the second part of the paper with two examples for the preliminary design of a fictitious WTG with a nacelle mass of 375 t and a hub height of 105 m, with reference being made to special aspects of the analysis. Owing to the high fatigue loads, the maximum ULS utilization ratios of approx. 70 % are relatively small. However, the use of higher-strength steels of up to S690 can effectively reduce the total mass of the foundation structure at local stress concentrations such as cable entry points. The important aspects of logistics and installation were deliberately left unconsidered since this paper sees its primary objective as taking a closer look at the potential of technical developments in the wake of the general progress in offshore wind energy technology. References [1] Germanischer Lloyd: Guideline for the Certification of Offshore Wind Turbines in Rules and Guidelines, Part IV – Industrial Services, 2005 ed.

[2] Deutsches Institut für Bautechnik: Richtlinie für Windenergieanlagen, Mar 2004 ed. [3] Deutsche Gesellschaft für Geotechnik e.V.: Empfehlungen des Arbeitskreises Pfähle (EA Pfähle), 2nd ed., 2012, Ernst & Sohn, Berlin. [4] German Maritime & Hydrographic Agency: Anwendungshinweise für den Standard “Konstruktive Ausführung von Offshore-Windenergieanlagen des BSH” (rev. ed.), 2012 ed. [5] American Petroleum Institute: Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms – Load and Resistance Factor Design. API-RP 2A-LRFD, 1st ed., Jul 1993. [6] American Petroleum Institute: Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms – Working Stress Design. API-RP 2A-WSD, 21st ed., Dec 2000, with errata & supplement 3, Oct 2007. [7] Hapel, K.-H.: Festigkeitsanalyse dynamisch beanspruchter Offshore-Konstruktionen. Vieweg, Braunschweig, 1990. [8] Gasch, R., Twele, J.: Windkraftanlagen. 7th ed., Vieweg+Teubner, 2011. [9] Kühn, M.: Dynamics and Design Optimisation of Offshore Wind Energy Conversion Systems. PhD Thesis, Delft University, 2001. [10] Achmus, M.: Bemessung von Monopiles für die Gründung von Offshore-Windenergieanlagen. Bautechnik 88 (2011), pp. 602– 616. [11] Wiemann, J., Lesny, K., Richwien, W.: Evaluation of Pile Diameter Effects on Soil-Pile stiffness”; Proc. 7th German Wind Energy Conference (DEWEK), 2004. [12] Krolis, V. D., van der Zwaag, G. L., de Vries, W.: Determining the Embedded Pile Length for Large-Diameter Monopiles. Marine Technology Society Journal, vol. 44, No. 1, 2010, pp. 24–31. [13] Stahlmann, J., Gattermann, J.: Gründung von OffshoreWindenergieanlagen – Stand der Technik? Proc. 8th colloquium “Bauen in Boden und Fels”, Technische Akademie Esslingen, 2012. [14] Tas¸an, E.: Zur Dimensionierung der Monopile-Gründungen von Offshore-Windenergieanlagen. PhD Thesis, TU Berlin, 2011. [15] Dührkop, J.: Zyklisch horizontal belastete Offshore-Monopiles. Workshop “Gründungen von Offshore-Windenergieanlagen”, Karlsruhe Institute of Technology (KIT), Veröffentlichungen des Institutes für Bodenmechanik und Felsmechanik, vol. 172, 2010, pp. 209–223. [16] Schaumann, P., Kleineidam, P., Wilke, F.: Fatigue Design bei Offshore-Windenergieanlagen. Stahlbau 73 (2004), pp. 716–726. [17] Ummenhofer, T., Herion, S., Weich, I.: Schweißnahtnachbehandlung mit höherfrequenten Hämmerverfahren – Ermüdungsfestigkeit, Qualitätssicherung, Bemessung. Stahlbau 78 (2009), pp. 605–612. Keywords: monopile; offshore foundation; eigenfrequency; dynamic; soft-stiff; fatigue; wind turbine

Authors: Dipl.-Ing. Rüdiger Scharff, Dr.-Ing. Michael Siems Ingenieurgesellschaft Peil Ummenhofer mbH, Daimlerstr. 18, 38112 Braunschweig Phone: +49 (0)531 1233100 rscharff@ipu-ing.de Internet: www.ipu-ing.de

Steel Construction 6 (2013), No. 1

13_047-053_Siems-Scharff (2012-1-10)_cs6.indd 53

05.02.13 08:21

Reports DOI: 10.1002/stco.201300001

Assembly of the steel roof structure for the football stadium in Gdan´sk Jerzy Ziółko Alojzy Les´niak

The technology of assembling the roof over the stadium built in Gdan´sk for the EURO 2012 European Football Championship is discussed here. The stadium has a characteristic silhouette – its shape and the colours of the façade resemble a cut block of amber. The steel roof structure has a quasi-elliptical form, with a maximum diameter of 220 m and minimum diameter of 187 m. It is 38 m high and the roof girders extend 48 m over the grandstand below. The roof structure weights 7150 t and was assembled in 226 days.

1 Introduction The most recent European Football Championship tournament took place in June and July 2012. Poland and the Ukraine had been entrusted by UEFA with organizing that event. Each of the hosts was obliged to build four modern stadiums meeting UEFA’s standards. Warsaw, Gdan´ sk, Poznan´ and Wrocław in Poland, and Kiev, Kharkiv, Donetsk and Lviv in Ukraine were the cities selected as host venues. The stadium in Gdan´sk is designed for about 41 000 spectators. It is sited on a plot of 43 650 m2 located between the old town and new port districts. A German architectural practice, RKW Rhode, Kellermann, Wawrowsky GmbH from Düsseldorf, won the competition for the stadium design and subsequently prepared the architectural and conceptual designs. Detailed design work was awarded as follows: – Foundations – Prof. Dr. hab. Eng. Michał Topolnicki, Gdan´sk University of Technology – Concrete for foundations and grandstands – Autorska Pracownia Konstrukcyjna “Wojdak”, Dr. Eng. Ryszard Wojdak (design checked by Prof. Dr. hab. Eng. Tadeusz Godycki-C´ wirko) – Steel roof structure – Konsultacyjne . . Biuro Projektów Zółtowski, Dr. hab. Eng. Krzysztof Zółtowski, professor, Gdan´sk University of Technology, and his team – Steel structure assembly – Martfer Polska Sp. z o.o. – Grandstand structure components, landscape architecture and coordination of all design work – Eilers & Vogel, Hannover The construction work was carried out by the Hydrobudowa Polska/Alpine Polska Consortium. The steel structure for the roof over the grandstand was fabricated and assembled by a consortium composed of:

– – – –

. ENERGOMONTAZ POŁUDNIE S.A., Katowice PBG – Technologia, Wysogotowo (near Poznan´) Martifer Polska Sp. z o.o., Gliwice Ocekon Engineering, Kosice (Slovakia)

Eng. Tomasz Osubniak, MSc (Martifer Polska Sp. z o.o.) was the construction manager for the stadium . roof and Eng. Tomasz Zyska, MSc (ENERGOMONTAZ POŁUDNIE S.A.) was responsible for the final stage of the construction work and dismantling of the supporting lattice beams and erection towers as well as lowering the roof from an auxiliary structure.

2 General characteristics of stadium roof structure The plan form of the stadium is close to that of an ellipse (Fig. 1) with the major axis measuring 220 m and the minor

Fig. 1. Plan of stadium (the numbered boxes are the erection towers supporting the roof structure during assembly and the black squares on the periphery of the stadium show the locations of the bearings for the 82 roof support girders – see Fig. 2)

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 54

01.02.13 08:52

Reports

Fig. 2. Steel roof support girder – diagram and cross-section not to scale (for description see text)

Fig. 4. Roof girder support on the reinforced concrete ring on the floor at a height of about 7.0 m above ground level

Fig. 3. Bracing between neighbouring girders: circular hollow sections for the ring (horizontal) members, round bars for the cross-braces

Fig. 5. Cast steel support bearings for the roof girders and (in the background) erection tower components stored on site prior to erection

axis 187 m. The stadium roof is a welded steel structure composed of 82 sickle-shaped support girders (Fig. 2) plus bracing members and purlins. The support girders (distributed along the stand periphery at a spacing of about 8 m) are in the form of space frames with a trapezoidal cross-section. The spacing of the members of the upper chord varies from 1205 mm at the supports to 4280 mm on the girder axis in the curving zone, whereas the spacing of the lower chord members varies from 405 to 1200 mm respectively. The greatest girder depth (in the curving zone) amounts to 5.8 m. Grade S355J2 steel was used for the girders. The chords (both upper and lower) are made from ∅ 355.6 mm circular hollow sections with wall thicknesses of 10 to 16 mm, with the sections in the lower part of the girder having the thickest walls. All bracing members are made from ∅ 219.1 × 8 mm circular hollow sections. The girders are joined together at each truss joint of the upper chord with horizontal tubular bars and X-type cross-braces made from round bars (Fig. 3). It is these bracing members plus a ∅ 500 × 20 mm linking ring, connecting together the girder ends inside the stadium, that turn the roof structure into a quasi-rigid shell with a hole in the middle. Therefore, it can be independent of the reinforced concrete structure

of the stand, only supported on it at 82 articulated joints on a support ring integrated into the floor at a height of about 7.0 m above ground level (Fig. 4). The pivot bearing for the steel roof girder has been achieved by welding the bottom ends of the upper and lower chords to a horizontal ∅ 500 × 24 mm circular hollow section and inserting this section into a cast steel cradle-type bearing (Fig. 5). The section transferring the support reaction to the bearing originally had a diameter of 508 mm and wall thickness of 28 mm, but was machined on the outside in order to fit the bearing exactly. The tube was subsequently filled with B28 class expanding concrete and closed with circular covers welded on the ends. The total weight of the steel roof structure for the stadium in Gdan´sk amounted to 7150 t.

3 Assembly Cutting of tubular sections for the frame bars for joining together at various angles and chamfering of edges for welding in various positions was carried out at the works in Kosice and partly at cooperating plants in Hungary. The prepared hollow sections were transported to the Martifer Metallic Constructions plant in Gliwice, where they were

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 55

01.02.13 08:52

Reports joined together into units with dimensions suitable for road transport to Gdan´sk. Prior to shipping, a trial assembly on special frames was performed at the Gliwice plant (Fig. 6). A layer of anti-corrosion primer and the first coat of paint were also applied here. Following checking, the girders were dismantled ready to be transported in parts to Gdan´sk on multi-axle trailers (Fig. 7). After arrival at the construction site, the parts of the roof girders were stored along assembly transoms, which were equipped with special equipment consisting of posts with brackets, platforms and grips to facilitate assembling and welding of the structures. Welding was performed outside in good weather, but on rainy days or in strong winds was executed inside special protective tents. The tents could be moved along tracks laid on both sides of the transoms to cover those zones where welding was required (Fig. 8). The dimensions of the protective tent were: span 13 m, height 12.62 m, length up to 35.09 m (depending on the number of 2.92 m long modules). The tent had a steel framework mounted on carriages, which allowed it to be moved along the track. The side walls and roof were made of non-flammable PVC-coated polyester fabric stretched across the steel framework. The end walls, made of the same material, were of folding type. These tents were also used to protect assembled units during repairs to the anti-corrosion paint applied at the factory damaged in the vicinity of site welds or damaged during transportation.

The lower sections of the roof girders (the “façade” sections), have been checked thoroughly for correctness of jointing, were transported directly to the reinforced concrete structure of the grandstand and mounted on the articulated bearings (Fig. 9). The respective roof girders had to be fixed to the reinforced concrete ring beam on the top of the stand until the entire roof structure was assembled. This was accomplished with the aid of a segmented steel ring made of successively mounted I sections. The members of the roof girder lower chord were fixed to the ring using steel U-bolts (Fig. 10). In order to set about assembling the “roof” sections of the support girders it was necessary to erect 22 erection towers inside the stadium to support lattice beams on which the end sections of the roof support girders could be placed. See Fig. 1 for the layout of the erection towers. The towers were in the form of steel frame columns (in axial compression), square in section. Each tower was tied back to the

Fig. 8. Assembling roof girders on site on two parallel assembly transoms (the movable tents for protecting welding works on rainy or windy days can be seen in the background)

Fig. 6. Trial assembly of a roof girder at the plant in Gliwice (photograph courtesy of the investor)

Fig. 7. Transportation of roof girder parts from the plant to the site (photograph courtesy of the investor)

Fig. 9. The first assembled “façade” part of a roof girder (2 April 2010) – temporary attachment of the girder to the segmented steel ring on the reinforced concrete grandstand ring beam

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 56

01.02.13 08:53

Reports

Fig. 10. Roof girder lower chord members connected temporarily with U-bolts to the steel ring on the reinforced concrete grandstand ring beam

Fig. 12. Anchorage for erection tower stays under the pitch (one ground anchor for the stays of two neighbouring towers)

Fig. 13. A “roof” section of a roof support girder (about 48 m long) being transported on a self-propelled multi-axle platform from the paint shop tent to the erection crane

Fig. 11. Anchorage of erection tower stays on reinforced concrete grandstand (stays of two neighbouring towers fastened to one anchor)

reinforced concrete structure of the stand foundation with four stays. Two of the stays were anchored to the reinforced concrete stand structure (Fig. 11), the other two were fastened to ground anchors beneath the pitch (Fig. 12). The “roof” sections of the girders, have been checked thoroughly for correctness of jointing, were transported to the erection crane on special self-propelled multi-axle platforms (Fig. 13). They were fitted with scaffolds necessary for safe work at heights and accessories to facilitate joining them to the “façade” sections of the girders already in place. A LIEBHERR 1350 crawler crane was used for erecting the girder “roof” sections. The crane, with a boom and jib each 42 m long, a working radius of 34 m and total weight of ballast (main + auxiliary) amounting to 163 t, had a lifting capacity of 38 t (Fig. 14). The weight of the roof girder to be erected was 31 t. Erection of the first roof girder is shown in Fig. 15.

Fig. 14. Diagram of erection of “roof” section

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 57

01.02.13 08:53

Reports

Fig. 15. The first assembled “roof” section of a roof support girder (13 May 2010)

Fig. 17. Fittings for adjusting the weld groove width of the girder upper chord joint a)

Fig. 16. Ends of upper chord members of a girder “roof” section fitted with fixtures to facilitate assembly

Assembly joints of girder chord circular sections had been designed as butt joints welded on one side. To facilitate the connection of these sections, guides made of four crossing metal plates were welded inside the tubes of the part to be assembled. Protruding portions of the plates narrowed towards the ends, thus forming truncated cones that centred the pipes being joined (Fig. 16). Adjustment of the welding groove between the tubular sections to be joined was made possible by welding thrust pieces near the edges of both sections (Fig. 17). These pieces were joined together with bolts provided with nuts above and below the thrust plate. The nuts were turned to adjust the tube edge spacing so as to ensure thorough penetration of the site weld. Each support girder, starting with the second one with assembled “roof” section, was linked to the preceding girder by means of circumferential bars. The purpose of this was to create a rigid structure – compare Figs. 18a and 18b. Fig. 19 shows the tips of the first two assembled roof girders, with segments of the linking ring. On completion of the erection of all girders and welding of the missing sections of this ring, it became the main linking ring playing the essential function in the behaviour of the roof. Fig. 20 shows a further stage in the construction work – erection of 12 support girders.

Fig. 18. Diagram showing “roof” girders resting on the supporting lattice beams: a) girder immediately after assembly, b) two adjacent girders linked with circumferential bars prior to erection of the next girder

Dismantling the supporting beams and erection towers was an essential and technically difficult stage of the work. Dismantling was only possible after having welded together and checked the whole main ring connecting all the roof support girders. Once that was done, the contractor had to set about disconnecting the roof support structure from its temporary links to the reinforced concrete ring beam on the top of the stand; the roof support structure had to become independent of the reinforced concrete structure. Cutting of the U-bolts joining the steel girders to

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 58

01.02.13 08:53

Reports

Fig. 19. Tips of two adjacent girders resting on supporting beam – components of the main linking ring measuring ∅ 500 × 20 mm can be seen on the girder tips

Fig. 21. Fully assembled steel roof support structure (31 August 2010 – photograph courtesy of the investor)

Fig. 20. Twelve erected “roof” segments of the roof support structure

the reinforced concrete structure (see Fig. 10) stirred up a lot of emotion as that could not be done for all 82 girders simultaneously. The fear was that successive releasing of the girders from those temporary joints could result in local destressing of the steel structure and hence deformations. However, no such effects took place. The tops of the erection towers, and specifically the middle parts of these top sections, could be shifted vertically within a range of several tens of centimetres with the help of appropriate hydraulic jacks. Such a design made it possible to dismantle the supporting beams and afterwards the erection towers themselves. The procedure adopted can be outlined as follows: – Installation of two hydraulic jacks on each tower, controlled from a central station. – Jacking up the whole steel roof by 30.0 mm. – Removal of the backing pads and supports on which the beams rested. – Lowering of the supporting beams by 650 mm using the hydraulic jacks. That operation was broken down into stages of about 200 mm each and the state of the roof support structure was checked after each stage. – Geodesic measurements made on completion of the final stage showed that the respective support girders had settled at their ends by 260–370 mm. Those results differed from

Fig. 22. The authors on the site: Eng. A. Les´niak (left) and Prof. J. Ziółko (right) (27 August 2010)

Fig. 23. The stadium in Gdan´sk after commissioning (19 July 2011 – photo by A. Jemiołkowski)

the calculated ones by 5–12 %. This was acknowledged as acceptable, bearing in mind unavoidable differences in the fabrication of the respective girders.

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 59

01.02.13 08:53

Reports – The difference in the levels between the curved roof girders and the tops of the supporting beams was in the range of 390 to 280 mm, which was sufficient clearance for dismantling the supporting beams. The operations for raising and lowering the roof support structure were carried out by a specialized Polish company, SLING. The erection towers consisted of three segments bolted together and therefore they could be dismantled by separating the segments after successively unbolting them. The fully assembled roof support structure is shown in Fig. 21, and the authors of this paper can be seen in Fig. 22. Covering of the structure with polycarbonate sheets on aluminium purlins was carried out by another contractor and will be the subject of another paper.

The erection of the steel roof structure to the stadium in Gdan´sk lasted from 2 April to 14 November 2010. Despite the fact that most of the work was performed at substantial heights, no serious accidents occurred. The stadium, commissioned after completion of all works, is shown in Fig. 23. Keywords: steel roof structure; football stadium; assembly

Authors: Prof. Dr. hab Jerzy Ziółko – professor, University of Technology & Life . Sciences, Bydgoszcz; consultant to ENERGOMONTAZ POŁUDNIE S.A., Katowice, for the steel roof structure at the PGE Arena, Gdan´sk (e-mail: jziolko@pg.gda.pl) Eng. Alojzy Les´niak – project manager. for steel roof structure, PGE Arena, Gdan´sk, ENERGOMONTAZ POŁUDNIE S.A., Katowice

News Flat Carbon Europe cuts its energy bills with turbine technology Flat Carbon Europe (FCE) has reduced its energy bills by more than 3 % a year, and cut its CO2-equivalent emissions by around 176000 t a year, thanks to the use of a new technology. The top recovery turbines (TRT) technology reuses high-pressure gases (known as flue gases) from the blast furnaces to drive electricity generators. TRT turbines generate energy by exploiting a property that is common to all gases – they expand as pressure drops. Fine particles are removed from the flue gas using dry and wet scrubbing systems. During the scrubbing process, the gas cools and its pressure drops. Before it can be used in the gas pipe network, the gas needs to be reduced to 0.1 bar. The most energy efficient way to do this is to lead the gas through a turbine, where it activates a generator to produce electricity. TRT does not have an impact on the blast furnace operations. As blast furnace gas is very combustible, it is normally

used in other parts of the plant to generate heat or energy for other processes. With the TRT system, the flue gas generates energy twice – first in the turbine and also when it is burnt for its usual purpose. TRT is a proven technology, with very limited risks: if the system fails, the expanding gas is accommodated in the existing scrubber. This is what happens in any blast furnace that is not equipped with the TRT technology. The technology has great potential, as each TRT has the same capacity as three

to four land-based wind turbines. To date, six blast furnaces at four of our sites have been equipped with such turbines and are generating more than 482 GWh of electricity each year. As a result, FCE’s energy bill has already been cut by more than 3% a year. TRT also provides ArcelorMittal with security over the sustainability of our long-term energy supply, and reduces our exposure to rising energy prices. ArcelorMittal is actively looking for energy partners to help the company increase the amount of electricity we produce via TRT. An additional eight blast furnaces in Europe have been identified as being suitable for conversion. Together they have the potential to produce another 475 GWh/year using existing TRT technology. The company is also looking to expand the technology to its blast furnaces beyond Europe. Significant efforts are already underway at its plants in Brazil and South Africa.

A TRT rotor being readied for installation (Source: ArcelorMittal)

Further information: www.arcelormittal.com

Steel Construction 6 (2013), No. 1

14_054-060_Ziolko_Lesniak (1-1)_cs6.indd 60

01.02.13 10:12

Reports DOI: 10.1002/stco.201300002

The aluminium and polycarbonate covering to the roof over the stadium in Gdan´sk Dariusz Kowalski

This paper presents information about structural elements of the roof covering to the stadium in Gdan´sk built for the 2012 European Football Championship in Poland and the Ukraine. The paper discusses elements of the polycarbonate covering, the supporting structure and the drainage system. It also provides information about tests and research performed prior to construction, which determined the solutions adopted as well as the roof’s present and future condition.

1 Introduction Apart from its shape, the most distinguishing element of the Gdan´sk football stadium built for the 2012 European Football Championship is the colour of its covering. It is this external enclosure element in yellow and brown that reflects the architectural vision originating from amber, a fossil treasure of the Gdan´sk coast (Fig. 1). The oval shape of amber is ideal for the functional system of the grandstand surrounding a football pitch. The sports facility with its oval shape (measuring 235.88 × 203.51 m) became a reflection of the idea (Fig. 2). The authors of the project, Konsorcjum Stadion Gdan´sk, RKW Rhode Kellermann Wawrowski GmbH+Co from Dus-

Fig. 1. An example of polished amber

seldorf, carried out all their design work based on the aforementioned architectural vision. The structure of roof fully reflects the shape and colour of amber [6].

2 General outline of stadium structure The stadium may be divided into two basic parts in terms of both its materials and its functions. The first part is a wholly independent reinforced concrete structure for spectators in which the permanent facilities are housed. This part of the arena contains all the functions connected with the organization of sports events and with meeting spectators’ demands – stands with seats, catering facilities, social amenities and internal circulation. The technical core of the entire facility and its technical infrastructure are also to be found here. The reinforced concrete part provides a base for supporting the roof at an elevation of 6.82 above the level of the football pitch, which is the other, independent part of the stadium – the steel support structure for the stadium roof. This part functions both as a wall enclosure, i. e. façade covering the reinforced concrete part, as well as a roof over the reinforced concrete stands. The basic support structure of the stadium roof consists of 82 steel girders creating a curving ribbed structure with a hole in the middle. The size of this oval hole corresponds approximately to the size of a football pitch, i. e. 122.4 × 90 m. The lattice girders with a system of 20 circumferential circular hollow sections connecting the girders plus the system of bracing members are the elements that determine the basic shape of the stadium. The main body of the stadium and its covering reach a height of approx. 45.2 m above football pitch level. The roof structure is covered with polycarbonate sheets supported on aluminium purlins.

3 Roof covering

Fig. 2. The sports arena with its shape matching the concept of local richness

The roof covering to the stadium is made of strips of flat polycarbonate sheets. Certain strips of the covering are separated from one another with radial roof gutters. The gutters are embedded in the covering and in the roof part they serve to drain rainwater. In the façade part the gutters, over a considerable part of their length, constitute a spacing element between certain strips of the covering made from the colourful sheets. The covering sheets are made up

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 61

01.02.13 08:54

Reports Clamping strip PC sheet

seal

Warsaw [4, 5]. The fixing and sealing deviate from standard solutions commonly used for such building elements (Fig. 3). The tests confirmed that it was correct to use the double set of seals to ensure the tightness of the covering, thermal movement of the sheets and appropriate tolerances for delivery and assembly of loadbearing and covering elements (Fig. 4).

4 Support structure for covering seal Aluminium purlin Fig. 3. Standard method of fixing and sealing the polycarbonate sheets 120 seal

Clamping strip PC sheet 25

seal 25

Aluminium purlin Fig. 4. The fixing and sealing method used for the covering at the Gdan´sk stadium

of modules with a fixed width of 800 mm. Aluminium supporting elements are mounted at this axial spacing. Around the circumference of the facility, the sheet length is determined by a division imposed by the geometrical system of the upper chord of the support girder and the varying distance between girders. This method of coverage made it necessary to build each of its elements individually. A sheet with a specific geometrical format may be mounted in only two places on the stadium structure. The polycarbonate sheets are mounted on two long edges – on aluminium purlins equipped with a set of seals. Along the shorter edges located at radial gutters, the sheets are seated on special elements only, i. e. so-called “parapets” equipped with seals in the sheet support zones. The method of fixing and sealing the sheets was developed following additional tests carried out by the Institute of Building Technology in

The purlins for the polycarbonate covering were made of individually designed aluminium sections with closed cross-sections and various depths ranging from a minimum of 25 mm to a maximum of 225 mm (Fig. 5) [7]. The individual fabrication options for the sections made it possible to select optimum depths for the sections to suit the spans between the support points on the girders and the distribution of loads. The sections were made of aluminium to EN AW-6060 (EN AW-AlMgSi) and PN EN 573-3:2009, and the supply conditions of T64 to PN-EN 515:1996. An extrusion method was employed to produce the sections. This production technology made it possible to supply loadbearing sections together with support elements for the plates and elements used for fixing seals. The small fall on the roof part of the covering meant it was necessary for purlin sections used in this part to be bent to enable rainwater to drain away from the polycarbonate sheets and directly into the radial gutters. The bending radius was 50 m in the section between circumferential gutters Nos. 1 and 2. The elements were bent cold in a fabrication plant using appropriate roller bending machines. Straight elements were used on the façade part. Gutter support elements were made from aluminium sheets in sections with a modular length of 800 mm. The metal gutter is merely a support element, which gives a basic shape to the rainwater drainage elements made from a modified PVC reinforced roof film. The elements of the aluminium structure carrying the external covering are fixed to the main steel roof structure with the help of support elements that allow adjustment of their positions with respect to one another. The adjustment made it possible to mount loadbearing elements for the covering in compliance with the fixed assembly spacing regardless of tolerances for the steel roof structure as built. The connection consists of two parts: a) Fixed part – in the form of steel support brackets welded to the main roof structure. The elements were used for profiling the support surface for the covering. The support plate of the bracket was equipped with elongated holes for adjusting the location of subsequent elements.

Fig. 5. Aluminium sections for roof purlins

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 62

01.02.13 08:54

–25

–14

Reports

–25

Makrofon multi UV 3×/25–25

Fig. 7. Section through polycarbonate sheet Table 1. Comparison of thicknesses of constituents of polycarbonate sheets Standard thicknesses according to technical approval [1]

Thicknesses supplied for the stadium in Gdan´sk [4]

[mm]

upper wall

0.7 ±0.20

1.32 –0.12

lower wall

0.7 ±0.15

1.30 –0.11

rib wall

0.45 ±0.15

0.43 –0.03 upper part 0.63 –0.09 lower part

slanted wall

0.12 ±0.04

0.15 –0.02 upper part 0.11 –0.02 upper part

Sheet constituent

Fig. 6. The support structure for the covering

b) Adjustable part: – in the form of steel unequal angles for the assembly of loadbearing purlins connected with screws and equipped with elongated holes at the contact with the fixed surface of the bracket, – made of aluminium angles for carrying gutters, which were fixed with self-tapping screws at appropriate levels. Elongated assembly holes were also used in the purlins to accommodate assembly tolerances. All the elements of the steel and aluminium structure were painted at the workshop production stage. Furnace-hardened powder paints were used for protection purposes (Fig. 6). The durability of this type of protection was confirmed by tests performed in salt chambers, which create conditions favouring accelerated degradation of the coating and development of corrosion.

5 Covering material – polycarbonate One type of polycarbonate sheet manufactured by Bayer was used: Makrolon Multi extended UV 3×25-25 ES. It is a sheet with a thickness of 25 mm and the chamber structure illustrated in Fig. 7. As seen in the section, the sheet has a lattice structure, which determines its high loadbearing capacity and rigidity. For the purposes of the Gdan´sk design, the geometrical structure of the sheet cross-section was modified by the manufacturer by way of additional thickening of all walls in the element (Table 1). The sheet used is characterized by a greater unit weight (5 kg/m2) [2] compared with boards manufactured as standard and according to the manufacturer’s approval (3.5 kg/m2) [1]. This made it possible to achieve greater loadbearing capacities as well as better rigidity, durability and resistance to impacts from a large 50 kg soft body and small bodies equivalent to hail.

The properties of the entire covering setup consisting of polycarbonate sheets plus aluminium support structure have been confirmed in tests on elements with real measurements as performed in test stations at the Institute of Building Technology in Warsaw [4]. The increase in thickness also improved the fire resistance parameters, which made it possible to classify the sheets according to EN 13501-1:2007 as follows [3]: – B for response to fire (non-flammable product, no flashover) – s2 for smoke emission (average smoke emission) – d0 for occurrence of burning drops/particles (no burning particles) In order to increase protection against the destructive power of ultraviolet radiation, the sheet material was coated with a special 100 mm co-extruded layer during production, which was melted homogeneously into the material. The manufacturer has assured that this solution will increase the life of the material at least three-fold [2], i. e more than 40 years of use in the stadium. However, the modification of the material connected with the thickening of walls creating the sheet structure had a negative impact upon the light transmission index, which decreased by about 10 % and continued to be the value at the intended level of 60 % for colourless boards. This value is particularly significant for issues connected with the growth of grass on the football pitch. Therefore, a considerable part of the canopy was clad with colourless boards with a higher light transmittance. The loadbearing capacity of sheets with a span of 800 mm centre to centre and greatest assembly length of approx. 7 m is at least 7.9 kN/m2, which ensures appropriate reserves of loadbearing capacity for environmental loads

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 63

01.02.13 08:54

Reports Table 2. Bearing capacity of polycarbonate sheets according to [4]

Support conditions

Sheet length

Characteristic loadbearing capacity

[m]

[kN/m2]

> 10

9.2

8.4

8.2

7.9

Simply supported on two long edges with span of 800 mm centre to centre, free rotation at the support

The protective metal elements were additionally sealed with silicone compound where they are in contact with the sheet (Fig. 8).

6 Assembly The assembly of the support structure elements for the covering and the polycarbonate covering itself was carried out using a single element method. The faรงade was assembled from scaffolding erected on the ground and the roof was assembled from scaffolding suspended from the steel roof girders. The main advantage of aluminium and polycarbonate, i.e. their low weight, proved extremely useful during assembly, as the entire process was carried out manually (Fig. 9). Cranes and hoists were only used for transporting

such as wind and snow and maintenance personnel on the roof. It should be emphasized here that the loadbearing capacity of the sheet depends on its length as presented in Table 2. However, it should be pointed that, during use, the material undergoes degradation and a decrease in its strength connected with that degradation. According to the information received from the manufacturer, the initial strength of approx. 60 MPa will decrease over time to approx. 50 MPa. This decrease in strength may occur within 20 years of use. Sheets in five amber colours laid in a pattern specified in the architectural design were used for covering the stadium. The polycarbonate sheets were cut to length in the factory and fitted with channel-type closing elements. Sealing tapes were fitted to the end of each sheet to prevent water ingress into the channels and protective metal fixtures. Full tape Silicone

PC sheet

Silicone

Permeable tape

Aluminium section

Fig. 8. Sealing of the polycarbonate sheet channels

Fig. 9. Assembly of the covering made of polycarbonate sheets

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 64

01.02.13 08:54

Reports boxes of materials. All connections within the covering were supplied as screw connections for ordinary and self-tapping screws. The polycarbonate sheets were fixed to the structure with the use of clamping strips attached with self-tapping screws. The assembly of the support structure for the roof covering started in June 2010, before the main steel roof structure had been completed – after a period of approx. three months following erection of the first steel element. The first elements of the covering of PC sheets were mounted at the end of November 2010 and the work completed at the beginning of May 2011. During this period, the assembly work had to be interrupted due to severe weather conditions in the winter. Some figures connected with the covering are impressive: – approx. 44 000 m2 of covering laid, i. e. approx. 17 500 sheet elements, – over 400 t of aluminium structure used for the loadbearing purlins.

7 Roof drainage Drainage of the entire covering was embedded in the roof covering. The open drainage system includes 144 radial gutters and three circumferential gutters. Both gutter systems were included in both the roof part and façade part. The radial gutters were mounted above each tubular section of the upper chord of the main steel girder. The circumferential gutters constituting the main drainage system for the roof were located as follows: – internal edge of roof, above a roof support member, – at the junction between the façade and roof parts, and – at half the height of the façade part.

Fig. 10. The stadium enclosure during the day and at night

9 Additional elements

The method of assembling and fastening of polycarbonate sheets with the use of clamping strips determined the delivery of the roof part and the bent loadbearing purlins already mentioned. The curvature allows gravity drainage of rainwater from the sheet surface and directly into the adjacent radial gutters on both sides. Water from the radial gutters drains into the circumferential gutters, where drainage inlets are located. Next, the water flows through pipes fastened to the girders. The water drains through the pipes to a channel located at the base of girders, where it is directed to storage containers. The rainwater is used for watering the pitch and for sanitary purposes.

Apart from the basic elements of the covering described above, the following technical equipment was embedded in the covering: – a double row of circumferential snowguards to protect against a thick layer of snow becoming dislodged from the curving part of the covering, – trolley system for maintenance of the covering, – a fall protection system for maintenance personnel working on the covering, – openings in the covering for access and technical equipment, – a weather monitoring system.

8 Illumination

10 Evaluation of the covering

The stadium enclosure serves not only to protect against the weather; it also constitutes a decorative and informative element. The light transmittance through the polycarbonate sheets has been exploited by the designers for the purpose of illuminating the entire facility. The lighting installations for the decorative backlighting of the façade were placed on three levels on the reinforced structure of the facility. The backlit facade emphasizes accurately the shape of the structure supporting the stadium roof and covering as can be seen in the photographs (Fig. 10). The upper line of backlighting reflects the varying shape of the stands, which is also visible on the façade.

After one year of use following commissioning, no cases of damage to sheets or worsening of properties connected with the watertightness of the entire covering were found on the covering to the stadium roof. The experience gained from stadiums worldwide as well as the Gdan´ sk stadium allow us to assume that such solutions will become common in our landscape for both public and housing facilities. It is expected that the polycarbonate covering and its support structure will not require any major renovations over the next approx. 15–20 years of use, which will certainly contribute to lowering the operational costs of the facility.

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 65

01.02.13 08:54

Reports References [1] Technical Approval ITB AT-15-3518/2005: MAKROLON MULTI UV chamber polycarbonate sheets, Bayer Sheet Europe, Institute of Building Technology, Warsaw, 2005. [2] Technical datasheet for Makrolon multi UV 3x/25-25 ES product, Gdan´sk stadium, Bayer, 2010. [3] Classification of response to fire according to EN 13501-1:2007, Institute of Building Technology, Warsaw, 2010. [4] Technical evaluation of MAKROLON MULTI extended UV 3×/25−25 ES sheets in the context of use for the enclosure to the BALTIC ARENA stadium in Gdan´sk, Institute of Building Technology, Warsaw, 2010. [5] Research work and technical opinion regarding to stadium enclosure to the BALTIC ARENA stadium in Gdan´sk, Institute of Building Technology, Warsaw, 2010.

[6] Construction and working design, including technical specifications for supply and acceptance of works, Konsorcjum Stadion Gdan´sk, RKW Rhode Kellermann Wawrowski GmbH+Co. Dusseldorf, 2008. [7] Workshop design – polycarbonate cladding documentation, RH Plus Robert Hulewicz Warszawa, Metalplast Stolarka Sp. z o.o. Bielsko Biała, 2009. Keywords: polycarbonate; aluminium; covering; stadium

Author:

. Dr. inz. Dariusz Kowalski (e-mail: kowdar@pg.gda.pl), 1) Gdan´sk University of Technology, Faculty and Civil & Environmental Engineering Department of Metal Structures & Management in Building 2) Biuro Inwestycji Euro Gdan´sk 2012 Spółka z o.o.

Announcements DFE 2013 – 5th International Conference on Design, Fabrication and Economy of Metal Structures Location and date: Miskolc, Hungary, 24–26 April 2013 The purpose of DFE 2013 is to bring together the scientific communities of the conference topics: – design – fabrication – economy Information and registration: dfe2013conf@uni-miskolc.hu www.dfe2013.uni-miskolc.hu

International IABSE Conference Location and date: Rotterdam, The Netherlands, 6–8 May 2013 Topics: – Load Carrying Capacity and Remaining Lifetime – Assessment of Structural Condition – Modernisation and Refurbishment – Materials and Products – Structural Verification Information and registration: www.iabse.org

STEEL BUILD 2013 - China (Guangzhou) International Exhibition for Steel Construction & Metal Building Materials Location and date: China (Guangzhou), 9–11 May 2013 Exhibition scope: – Steel structure products – Steel structure components – Protective system of steel structure – Technologies of new type residence and fittings, decoration products – Processing equipments and testing equipments of steel structure – Three dimension garage equipments and steel structure door industry – Computer design, analysis, calculation and CAD paint software – Area of construction Design and Real Estate project Planning and Design

perience between different institutions, owners, contractors, bridge designers and constructors, as well as scientific experts. The selected papers to be presented at the Conference are mainly related to the bridges across the Danube and its tributaries, i.e. bridges in the Danube Basin. The conference has also the aims at promoting advances in bridge engineering. Information and registration: http://danubebridges.com

ECCOMAS Thematic Conference Structural Membranes 2013 Location and date: Munich, Germany, 9–11 October 2013

Information and registration: steelbuild@grandeurhk.com www.steelbuildexpo.com

Information and registration: http://congress.cimne.com/ membranes2013

8th International Conference – Bridges in Danube Basin

CWE 2014 -6th International Symposium on Computational Wind Engineering

Location and date: Timis¸ oara, Romania, 4–6 October 2013 Conference scope: The general aim of the conference is the overall exchange of knowledge and ex-

Location and date: Hamburg, Germany, June 8–12, 2014 Information and registration: www.cwe2014.org

Steel Construction 6 (2013), No. 1

15_061-066_Kowalski (1-2)_cs6.indd 66

01.02.13 10:09

ECCS news TC6 – Fatigue & Fracture Chairperson: Dr. M. Lukic Date: 25–26 April 2013, Delft, Netherlands

TC7 – Cold-Formed Thin-Walled Sheet Steel in Buildings Chairperson: Prof. J. Lange Date: 6–7 June 2013, Paris, France

TWG 7.5 – Practical Improvement of Design Procedures

Conferences Second Luso-African Conference on Sustainable Steel Construction

Chairperson: Prof. Bettina Brune Date: 6–7 June 2013, Paris, France

TWG 7.9 – Sandwich Panels & Related Subjects Chairperson: Mr. Paavo Hassinen Date: 18 February 2013, Mainz, Germany Date: 6–7 June 2013, Paris, France

TC8 – Structural Stability Chairperson: Prof. H. H. Snijder Secretary: Dr. Markus Knobloch Date: 21 June 2013, Stuttgart, Germany

TWG 8.3 – Plate Buckling Chairperson: Prof. U. Kuhlmann Secretary: Dr. B. Braun

TWG 8.4 – Buckling of Shells Following the success of the last event, CMM will organize the second LusoAfrican Conference on Sustainable Steel Construction in July 2013. This conference, the main objective of which is the promotion of national steel and composite construction on the African continent, will be held in Maputo, Mozambique. With the purpose of presenting several national and local companies to the African steel construction market, the organizers expect an increase in the number of corporate and individual participants, in a scenario of sustainable growth for this event.

Technical Committees (TC) activities TMB – Technical Management Board Chairperson: Prof. L. Simões da Silva Vice-Chairperson: Prof. M. Veljkovic

PMB – Promotional Management Board Chairperson: Mr. Bertrand Lemoine

TC3 – Fire Safety Chairperson: Prof. P. Schaumann Secretary: Prof. Paulo Vila Real

Chairperson: Prof. J.M. Rotter Secretary: Prof. S. Karamanos

TC9 – Execution & Quality Management Chairperson: Mr. Kjetil Myrhe Date: 6 November, Brussels, Belgium

TC10 – Structural Connections Chairperson: Prof. Thomas Ummenhofer Secretary: Mr. Edwin Belder Date: 11–12 April 2013, Liège, Belgium Date: 10–11 October 2013

TC11 – Composite Chairperson: Prof. R. Zandonini Secretary: Prof. Graziano Leoni Date: 24 May 2013, Munich, Germany

TC13 – Seismic Design Chairperson: Prof. R. Landolfo Secretary: Dr. A. Stratan Date: 27 June 2013, Naples, Italy

TC14 – Sustainability & Eco-Efficiency of Steel Construction Chairperson: Prof. Luís Bragança Secretary: Ms. Heli Koukkari

TC15 – Architectural & Structural Design Chairperson: Prof. P. Cruz

TC News TC6 – Fatigue & Fracture On 31 May 2012 the committee consisted of 25 full, 16 corresponding and 8 guest members – a total of 49 experts from 17 countries. The committee has been chaired by M. Lukić since 2005. The terms of reference guiding the committee in its activities are: 1. Exchange information about relevant research and design issues in fatigue of structures 2. Publications and standards upgrade in relation to the issues within the preceding item 3. Stimulation of partnership in funded research proposals 4. Broadening of competences on fracture 5. Collaboration with other ECCS technical committees and other professional and/or scientific organizations Add. 1&2: There are three active working groups within TC6: – Assessment of existing steel structures (chairman B. Kühn), dedicated to the refinement and extension of existing ECCS-JRC recommendations on this topic. Apart from refining the existing content, new parts are being added on damage cases and their causes, fatigue category differentiation in riveted members, wind power structures, orthotropic bridge decks and crane structures. Additional worked examples will be added, too. – Extended fatigue approach to take into account execution quality, different steel grades and post-weld treatment (chairman H.-P. Günther) acts as the mirror group to the CEN/ TC250/SC3 Evolution Group on EN 1993-1-9 (see below). Initially established as a self-sufficient group prior to the start of the work on revising the Eurocodes, it has gradually mutated to a seminar-like discussion platform for new rules intended to extend those already in place in the standard. – Statistical analysis of fatigue data (chairman A. Nussbaumer) was founded last year in order to deal with the establishment of coherence in the analysis of the results of fatigue tests, nowadays different in different recommendations and standards, thus preventing a direct comparison of the proposed rules. Add. 2&4: The committee is active – in cooperation with CEN/TC250/SC3 – in

Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 67

01.02.13 08:55

ECCS news maintaining and extending EN 1993-1-9 and EN 1993-1-10. Two Evolution Groups have been set up and the meetings have been taking place with attendees from both SC3 and TC6. Short- and long-term tasks have been allotted and deadlines fixed. For EN 1993-1-10 there is still a need for more experts in fracture to become engaged in the work of the committee. Add. 3: Members of the committee were active in RFCS research project BriFaG (Bridge Fatigue Guidance) which ended in June 2011. New (mainly RFCS) research project proposals were discussed at the latest meeting in Berlin and three of them were chosen for submission in September 2012. Add. 5: TC6 is contributing to the organizing and scientific committees for the “Fatigue Design” international conferences. The next conference will take place in 2013.

TWG 7.5 – Practical Improvement of Design Procedures Cold-formed members and structures New developments in the field of coldformed members and structures have been presented and discussed: Mahen Mahendran presented Australian research activities and an innovative type of cold-formed beam, the LiteSteelBeam (LSB). The research covers fullscale bending and shear tests as well as numerical analyses to investigate the section and the member moment capacities of LSBs with and without circular web openings. The applicability of the current Australian design rules for the LSB was investigated. New design equations were proposed and verified by the research results. V. Ungureanu presented a new type of cold-formed steel beam composed of double C-sections as flanges and a corrugated web supported by a supplementary shear panel at the end bearing connected by self-drilling screws. Five fullscale six-point bending tests have been performed to obtain the failure mode, the load capacity and the deformation of the specimen. A numerical model using ABAQUS was developed. A comparison of FE simulations and tests showed good agreement. Future research will focus on optimization, complete solutions (beam-to-column and beam-to-beam connections), proposals for an analytic model and design procedures according to EN 1993-1-5. Furthermore, the use of webs made of trapezoidal sheeting and a new welding technology will be investigated.

Z. Nagy investigated the joint stiffness of bolted single-storey frames. An ABAQUS model for single-storey frames with CFS bolted joints was developed and calibrated to full-scale tests. Good force–displacement agreement between tests and FE results was obtained. Further parametric studies were performed to study other joint typologies. An improved component method provides quite fair values for the initial stiffness, which gives more realistic results in structural analysis. Further developments for simplified design tools are intended.

EN 1993-1-3 Evolution Group Evolution groups for amending the Eurocodes have been established. EVG EN 1993-1-3 is chaired by L. Sokol (F). I. Balaz reported on the outcomes of the last EVG meeting in Paris on 29 March 2012. He presented and commented on the items discussed and asked for further support for TWG 7.5. All TWG 7.5 members are requested to forward further amendments for the revision, extension and simplification of EN 1993-1-3 to the chairman of the EVG, L. Sokol, or to B. Brune. The latest meeting of the EVG took place in Bratislava on 18 October 2012.

TWG 7.5/ERF – Rack structures TWG 7.5 cooperates with the European Racking Federation (ERF) to improve the design and construction of modern rack structures. The ERF representative is Kees Tilburgs. His first presentation dealt with the test procedure for rack uprights in compression according to EN 15512. A new (informative) annex was developed by the ERF based on the latest research of UPC Barcelona (see TWG7.5 – Ithaca, 2012). The annex includes new guidelines for testing and analysing distortional buckling of rack uprights. This comprises a revised test length and a simplified approach for (typically) perforated rack uprights in order to calculate the critical loads with the help of finite strip methods and/or generalized beam theory. The new proposal can be accepted according to the experience of the TWG 7.5. The next report concentrated on the design and testing of pallet beams with special regard to beam stability. In typical pallet rack structures the beams are loaded laterally as well and rotationally restrained by the pallets. But according to EN 15512´s test procedure, the horizontal deflections of the test beam under load should neither be prevented nor amplified. Thus, the horizontal (lateral) instability deformation of the compressed flange will induce a horizontal load that increases progressively with the lateral deformation, resulting in very

conservative test results that are neither realistic nor economical. The basic scientific findings of the loadbearing behaviour of rack beams loaded and supported by pallets (specifying the influence of the variable parameters) are still lacking. This is an issue for future research. The results of an ongoing research programme to develop a design procedure for perforated industrial storage rack columns were presented by M. Casafont. The design procedure includes the effects of distortional buckling. The types of columns studied include those commonly used in Europe and some that are used in the USA. The procedure is being developed based on finite element studies verified by physical testing. It involves the use of the current US rack column design approach with an extension to distortional buckling with finite strip method (FSM) solutions and direct strength method (DSM) approaches. The approach originally developed for individual columns is being studied for the behaviour of such columns in a rack frame. D. Dubina presented experimental and numerical investigations of rack uprights in compression, analysing perforated as well as non-perforated sections. The studies focus on coupled instabilities with special regard to the interaction of distortional buckling and global member stability failure. A new ECBL approach was developed to adapt the European buckling curves for cold-formed sections. Based on a new factor ψ, which defines the coupling erosion, the imperfection factor α was improved and calibrated to tests and numerical calculations. The new ECBL approach also helps to determine the critical combinations of imperfections.

Benchmark examples TWG 7.5 agreed to prepare benchmark examples in order to demonstrate verified and accepted design by FE analysis. This time, two new benchmark examples were presented by M. Casafont and A. Belica. The benchmark example of A. Belica summarizes research results dealing with the restraint of Z-purlins on Z-bearing frames. Tests were carried out at the Klockner Institute in Prague. FE simulations were performed at the Czech Technical University of Prague and calibrated to the test results. The load–displacement curves from testing and ANSYS analysis show good agreement. M. Casafont presented a benchmark example of trapezoidal sheeting in bending. FE simulations using ANSYS have been calibrated to full-scale bending tests. The comparison of test results and FE

Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 68

01.02.13 08:55

ECCS news analysis showed good agreement with respect to the moment capacity and the strains at mid-span. However, the bending stiffness cannot be predicted by numerical calculations as the loadbearing behaviour at the end support could not be reproduced up to now. M. Casafont thus intends to revise the benchmark example.

Any other business B. Brune received an enquiry from the Ernst & Sohn publishing house, which intends to publish a special edition of the journal “Steel Construction – Design and Research” (official journal for ECCS members) dealing with cold-formed sections and structures. TWG 7.5 has agreed to cooperate. B. Brune will contact the chief editor, Dr. Kurrer, to obtain more information concerning timetable and management. All TWG 7.5 members are requested to prepare abstracts of interesting subjects worth publishing and send them to B. Brune. TWG 7.5 will discuss all contributions at the next meeting to be held in Timis¸ oara, Romania. The date of the meeting will be fixed as soon as the timetable has been specified.

TWG 7.9 – Sandwich Panels & Related Subjects Technical Working Group ECC TWG 7.9 together with CIB Commission CIB W056 constitutes the European Joint Committee on Sandwich Constructions. The next meeting of the Joint Committee will be held on 18 February 2013 in Mainz, Germany. The annual meeting of TC7 together with the meetings of the Technical Working Groups ECCS TWG 7.5 and the Joint Committee on Sandwich Constructions will be held on 6–7 June 2013 in Paris. The Joint Committee on Sandwich Constructions is currently preparing European recommendations on the stabilization of steel structures using sandwich panels. The aim is to complete the manuscript by the next meetings and to publish the document during the second half of 2013.

Due to its valuable members from production companies and universities, TC11 represents a permanent observatory for research advances in composite construction and contributes to bridging the gap between research and practice. During the past semester, TC11 activity was aimed at preparing proposals for funding new applied research (RFCS projects).

pacity-design rules”, “design of concentrically braced frames”, “dual structures”, “drift limitations and second-order effects”, “new structural types” and “lowdissipative structures”.

TC 13 – Seismic Design

Steel LCA

The next meeting will be held in Naples on 27 June 2013. This meeting will be organized jointly with the international workshop within the HSS-SERF project coordinated by Prof. Dubina. The workshop will take place in Naples on the afternoon of 27 June and throughout the whole day on 28 June 2013. A new TC13 publication “Assessment of EC8 Provisions for Seismic Design of Steel Structures” (ed. Prof. Landolfo) is ready. This publication describes and discusses the aspects and issues in EN 1998-1:2004 that need clarification and/ or further development. This book is the result of the activities carried out within the framework of Technical Committee “Seismic Design” (TC13) of the European Convention for Constructional Steelwork (ECCS) in the field of codification and technical specifications. The publication is organized into 12 sections and one annex. The basic topics discussed in the text are “material overstrength”, “selection of steel toughness”, “local ductility”, “design rules for connections in dissipative zones”, “new links in eccentrically braced frames”, “behaviour factors”, “ca-

As part of the latest ECCS iAPPs, ECCS has recently launched version 1 of Steel LCA in Apple’s App Store for iPad. It can be downloaded for free – just search for “ECCS”, “LCA” or “Steel”. The aim of the Steel LCA application is to perform simplified life cycle analysis (LCA) of hot-rolled I-sections and hollow sections. In addition, it provides a database of hot-rolled I-sections and cold-formed

Software

Goal and scope

Inventory analysis

Impact assessment

Normalisation and weigthing

Interpretation Fig. 1. Scheme of the environmental life cycle analysis

TC11 – Composite Prior to the last meeting held in Valencia on 19 October 2012, three new members joined the TC11: Wioleta Barcevicz and Slawomir Labocha from Poland, and Renata Obiala from Luxembourg. Twentytwo full members and seven corresponding members, from 15 European countries, Australia and New Zeeland, currently constitute TC11.

Fig. 2. Input for the application Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 69

01.02.13 08:55

ECCS news 1. Choose the cross-section. 2. Enter the values of the required parameters according to the case analysed (length of member, steel grade). 3. Enter the lifespan of the analysis (the period of time considered for the analysis, in years). 4. Select the scope of the analysis; different options are available according to the scope of the analysis (Fig. 2). – In option (i): the user can select a coating system for the section from a list of available products. – In option (ii): in addition, the user can select a recycling rate and a reuse rate for steel and the corresponding transportation system. – And in option (iii): in addition, the user may select a transportation system for steel (from the gate of the factory to the construction site) and a maintenance strategy for the section taking into account the lifespan of the analysis. 5. The results of the LCA are obtained in the results section; a detailed calculation report is automatically generated that can be sent by e-mail.

Fig. 3. User configurations

Finally, in the configuration options, the user may select the default type of analysis as well as the desired outputs (Fig. 3). Similarly to the ECCS EC3 steel member calculator iApp, companies are welcome to supply information on their products for inclusion in the ECCS products database simply by clicking the “Add your company” button in the main menu of the application (Fig. 4). ECCS welcomes feedback to support the improvement of these tools (ECCSapps@steelconstruct.eu).

ECCS EC3 steel member calculator

Fig. 4. iApp interface and “Add your company” button and hot-finished hollow sections and Steel LCA functions according to ISO standards 14040:2006 and 14044:2006. Moreover, the analysis is performed taking into account the modular concept of European standard EN 15804:2012. According to this set of standards, the evaluation comprises four main steps (goal and scope, inventory analysis, impact assessment, interpretation), as shown in Fig. 1. The application provides two additional, optional steps: normalization and weighting. These two steps are considered to be optional in ISO standards, although they play a relevant role in the decision-making pro-

cess. Thus, the complete flowchart for the environmental life cycle analysis is represented in Fig. 1. The application enables three different scopes for the LCA: Option (i): a cradle-to-gate analysis (module A according to EN 15804:2012) Option (ii): a cradle-to-gate analysis plus end-of-life recycling (modules A and D according to EN 15804:2012) Option (iii): a cradle-to-grave analysis plus end-of-life recycling (modules A to D according to EN 15804:2012) Five main steps are needed to obtain LCA results in the application:

As part of the latest ECCS iAPPs, which aim to provide databases of product information and offer support on the use of the structural Eurocodes for steel-intensive applications plus guidance on the sustainability assessment of steel construction, version 2 of the first ECCS Eurocode 3 app “ECCS EC3 Steel Member Calculator” has been released in Apple’s App Store for iPad (the iPhone version is about to be launched). It can be downloaded for free – just search for “ECCS”, “EC3”, “Steel” or “Steel Member”. It provides a database of hotrolled I-sections and cold-formed and hot-finished hollow sections and gives the safety verification of steel beams and columns with such sections according to EC3-1-1. An additional module may be purchased within the application for €8.99, providing the option of verifying

Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 70

01.02.13 10:38

ECCS news Choice of section

User input

Producer information

Section main properties

Summary of results Interface of the application (for beam-columns)

User inputs for columns

User inputs for beams

User inputs for beam-columns

Bending moment distributions available

Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 71

01.02.13 08:56

ECCS news the safety of steel beam-columns. The “ECCS EC3 Steel Member Calculator” is a practical tool with an extremely user-friendly interface that provides a quick check of the member resistance. Just a few clicks are all that are needed to get an overview of a column resistance on site, or evaluate the safety of a beam during a briefing. The first version of “ECCS EC3 Steel Member Calculator” was launched in October 2011 and so far has been downloaded more than 7000 times worldwide. Version 2 is greatly improved, providing the following additional features: – Calculation of the resistance of beamcolumns under axial force and uniaxial bending (paid module, in-App purchase) with a very simple input scheme: – Option to choose between previously defined bending moment distributions and calculation of internal forces along the length of the member: – Option of up to four arbitrarily spaced weak-axis internal restraints and different end conditions – New, improved interface – Automatic calculation of the elastic critical moment Mcr – Extended database: circular, square and rectangular hollow sections to EN 10210 and EN 10219 – Extended database: additional I-section shapes – Local SAVE of reports – Improved user configurations:

General user configurations

“Add your company” template

A summary of results is given in the main menu and a detailed report is automatically generated and may be sent by e-mail, printed out or saved locally. New, improved versions are constantly being developed to keep extending the possibilities for design according to the Eurocodes. The launch of the safety verification of rolled steel members is planned for the beginning of 2013, with class 4 cross-sections and the inclusion of an extended database of cross-sections such as elliptical, L-, T- and U-shaped sections. Multilingual platforms and localized national specific rules are planned so that Nationally Determined Parameters and NCCI guidance are addressed adequately. Cross-platform mobile application platforms (iOS, Android, etc.) will be launched in due course. Companies are welcome to supply information on their products for inclusion in the ECCS products database simply by clicking the “Add your company” button in the main menu of the application. ECCS welcomes feedback to support the improvement of these tools (ECCSapps@steelconstruct.eu).

Steel Construction 6 (2013), No. 1

16_067-072_ECCS_News 1-13.indd 72

01.02.13 08:56

The international journal “Steel Construction – Design and Research” publishes peer-reviewed papers covering the entire field of steel construction research and engineering practice, focusing on the areas of composite construction, bridges, buildings, cable and membrane structures, façades, glass and lightweight constructions, also cranes, masts, towers, hydraulic structures, vessels, tanks and chimneys plus fire protection. “Steel Construction – Design and Research” is the engineering science journal for structural steelwork systems, which embraces the following areas of activity: new theories and testing, design, analysis and calculations, fabrication and erection, usage and conversion, preserving and maintaining the building stock, recycling and disposal. “Steel Construction – Design and Research” is therefore aimed not only at academics, but in particular at consulting structural engineers, and also other engineers active in the relevant industries and authorities. “Steel Construction – Design and Research” is published four times a year. Except for manuscripts, the publisher Ernst & Sohn purchases exclusive publishing rights. Ernst & Sohn accepts for publication only those works whose content has never appeared before in Germany or elsewhere. The publishing rights for the pictures and drawings made available are to be obtained by the author. The author undertakes not to reprint his or her article without the express permission of the publisher Ernst & Sohn. The “Notes for Authors” regulate the relationship between author and editorial staff or publisher, and the composition of articles. “Notes for Authors” can be obtained from the publisher or via the Internet at www.ernst-und-sohn.de/zeitschriften. The articles published in the journal are protected by copyright. All rights are reserved, particularly those of translation into foreign languages. No part of this journal may be reproduced in any form whatsoever without the written consent of the publisher. Brand-names or trademarks published in the journal are to be considered as protected under the terms of trademark protection legislation, even if they are not individually identified as such. Manuscripts are to be sent to the editorial staff or http://mc.manuscriptcentral.com/stco. If required, offprints or run-ons can be made of single articles. Requests should be sent to the publisher. Current prices The journal “Steel Construction – Design and Research” comprises four issues per year. In addition to “Steel Construction – Design and Research print”, the PDF version “Steel Construction – Design and Research online” is available on subscription through the “Wiley Online Library” online service. Prices

print + online

single issue

personal personal personal

161 € 268 sFr 254 $

187 € 309 sFr 292 $

46 € 76,91 sFr

institutional institutional institutional

544 € 904 sFr 855 $

625 € 1040 sFr 984 $

167 € 259,88 sFr

Members of the ECCS – European Convention for Constructional Steelwork receive the journal Steel Construction as part of their membership. Prices exclusive VAT and inclusive postage. Errors and omissions excepted. Subject to change without notice. Prices are valid from 1st September 2012 until 31st August 2013. Personal subscriptions may not be sold to libraries nor used as library copies. A subscription lasts for one year. It can be terminated in writing at any time with a period of notice of three months to the end of the subscription year. Otherwise, the subscription extends for a further year without written notification. Bank details Dresdner Bank Weinheim, A/C No: 751118800 Bank sort code: 67080050, SWIFT: DRESDEFF670 Periodical postage paid at Jamaica NY 11431. Air freight and mailing in the USA by Publications Expediting Services Inc., 200 Meacham Ave., Elmont NY 11003, USA POSTMASTER: Send changes of address to “Steel Construction – Design and Research” c/o Wiley-VCH, 111 River Street, Hoboken, NJ 07030, USA

Publisher Wilhelm Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG Rotherstrasse 21 10245 Berlin Germany Phone +49 (0)30 47031-200 Fax +49 (0)30 47031-270 info@ernst-und-sohn.de www.ernst-und-sohn.de Journal for ECCS members The ECCS – European Convention for Constructional Steelwork and Ernst & Sohn have agreed that from 2010 onwards Steel Construction is the official journal for ECCS members. www.steelconstruct.com Editorial staff Editor-in-chief: Dr.-Ing. Karl-Eugen Kurrer, Ernst & Sohn Phone +49 (0)30 47031-248, Fax -270 E-mail: karl-eugen.kurrer@wiley.com Editorial board Chair: Luís Simões da Silva (Portugal) • Guo-Qiang Li (P. R. China) • Darko Beg (Slovenia) • Richard Liew (Singapore) • Frans Bijlaard • Mladen Lukic (France) (The Netherlands) • Enrique Mirambell (Spain) • Luís Bragança (Portugal) • Kjetil Myhre (Norway) • Dinar Camotim (Portugal) • Kim Rassmussen (Australia) • S. L. Chan (P. R. China) • John Michael Rotter (UK) • Paulo Cruz (Portugal) • Peter Schaumann (Germany) • Dan Dubina (Romania) • Bert Snijder • László Dunai (Hungary) (The Netherlands) • Morgan Dundu • Thomas Ummenhofer (Rep. of South Africa) (Germany) • Markus Feldmann (Germany) • Ioannis Vayas (Greece) • Dan Frangopol (USA) • Milan Veljkovic (Sweden) • Leroy Gardner (UK) • Pedro Vellasco (Brazil) • Richard Greiner (Austria) • Paulo Vila Real (Portugal) • Jerome Hajjar (USA) • Frantisek Wald • Markku Heinisuo (Finland) (Czech Republic) • Jean-Pierre Jaspart (Belgium) • Riccardo Zandonini (Italy) • Ulrike Kuhlmann (Germany) • Jerzy Ziółko (Poland) • Akimitsu Kurita (Japan) • Raffaele Landolfo (Italy) Advertising manager Norbert Schippel, Ernst & Sohn Phone +49 (0)30 47031-252, Fax -230 E-mail: norbert.schippel@wiley.com Services for customers and readers Wiley-VCH Customer Service for Ernst & Sohn Boschstrasse 12, D-69469 Weinheim Tel.: +49 (0)800 1800 536 (with Germany) Tel.: +44 (0)1865476721 (outside Germany) Fax: +49 (0)6201 606184 cs-germany@wiley.com Quicklink: www.wileycustomerhelp.com Layout and typesetting LVD I BlackArt, Berlin Production NEUNPLUS1 GmbH, Berlin © 2013 Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG, Berlin

Inserts in this issue: Verlag Ernst & Sohn GmbH & Co. KG, 10245 Berlin

Steel Construction 6 (2013), No. 1

18_Impressum_Vorschau_1-13.indd 1

04.02.13 12:02

Preview

Steel Construction 2/2013 Karsten Kathage New technical approvals for pallet rack structures

Teoman Peköz, Bettina Brune Design of cold-formed steel members – EN 1993-1-3 compared to the Direct Strength Method

Karsten Kathage, Joachim Lindner, Thomas Misiek, Sivo Schilling Proposal to adjust the design approach for diaphragm action of shear panels according to Schardt and Strehl to European Regulations

Dinar Camotim Local and global buckling analysis using Generalized Beam Theory: Fundamentals, State of the Art and Future Perspectives Kees Tilburgs Demands and developments of the European Racking industry

Ram S. Puthli, Jeffrey A. Packer Structural Design using cold-formed hollow sections

Philip Leach Axial capacity of perforated steel columns Francesc Roure et al. Determination of the beam-to-column connection characteristics in pallet rack structures – a comparison of the EN and RMI methods and analysis of the influence of the shear to moment ratios Zolt Nagy, Lucian Gilia, Robert Ballok Romanian application of cold-formed steel beams of screwed corrugated webs

Reports Ewa Maria Kido, Zbigniew Cywin´ski The new steel-glass architecture of buildings in Japan One report in Steel Construction 2/2013 will focus on the relevant architectural representations in Japan, e.g. buildings of commercial and public use – with special emphasis on their ultramodern design character. The picture shows the “Prada Omotesando” in Tokyo

Subscription

Esther Pfeiffer, Andreas Kern Modern production of heavy plates for constructional applications – Control of production process and quality (subject to change without notice)

Fax +49 (0)30-47 03 12 40

 Yes, we would like to read the journal Steel Construction – Design and Research     

Single issue / year Annual subscription personal Annual subscription personal Annual subscription institutional Annual subscription institutional

4 issues 4 issues 4 issues 4 issues

__ / ____ print print + online print print + online

€ 46 € 161 sFr 268 € 187 sFr 309 € 544 sFr 904 € 625 sFr 1040

$ 254 $ 292 $ 855 $ 984

www.steelconstruct.com

A subscription runs for one calendar year / 4 issues. It can be terminated in writing at any time with a notice period of three months to the end of the subscription year. Without written notification, the subscription extends for a further year. Subscription print + online includes online access to the PDF version via Wiley Online Library. Members of the ECCS – European Convention for Constructional Steelwork receive the journal for free. Free sample copy of all Ernst & Sohn specialist journals: www.ernst-und-sohn.de/journals

 Privat

 Business Customer No.

Company

VAT-No.

Titel, Last name, First name

Street / No.

Department

Country / Zipcode / City

E-mail

Phone

0163510016_pf

Invoice and delivery address:

We guarantee you the right to revoke this order within two weeks, please mail to Verlag Ernst & Sohn, WILEY-VCH, Boschstr. 12, D-69469 Weinheim, Germany. (Timely mailing shall suffice.)

 Datum

Signature

Abo-Coupon-2012-13_181x100.indd 8

18_Impressum_Vorschau_1-13.indd 2

Prices: net-prices excl. of VAT but incl. postage and handling, valid until August 31, 2013, subject to alterations. Other currencies and bulk discounts are available on request. €-Prices are valid in Germany exclusively. 01.08.2012 16:20:48

01.02.13 08:57

Jürgen Grünberg,

Bernd Hillemeier, Claus Flohrer,

Joachim Göhlmann

Jürgen Krell, Gabriele Marquardt,

Concrete Structures for Wind Turbines

Jeanette Orlowsky, Michael Raupach, Karsten Schubert,

Series: Beton-Kalender Series

Stephanie Schuler

Series editor: K. Bergmeister, F. Fingerloos, J.-D. Wörner (eds.)

Reconditioning and Maintenance of Concrete Structures

The wind energy industry in

Series: Beton-Kalender Series

Germany has an excellent global

Series editor: K. Bergmeister,

standing when it comes to the

F. Fingerloos, J.-D. Wörner (eds.)

development and construction of

approx. 230 pages, Softcover. approx. € 49,90 ISBN: 978-3-433-03041-7 Publication date: approx. June 2013

wind turbines. Germany currently

The rehabilitation and mainte-

represents the world´s largest

nance of concrete structures takes

market for wind energy. The ongoing development of ever more powerful wind turbines plus additional requirements for the design and construction of their offshore foundation structures

approx. 210 pages, Softcover. approx. € 49,90 ISBN: 978-3-433-03043-1 Publication date: approx. September 2013

place in the tense atmosphere of the different interests of client, building authorities, construction companies and material manufacturer. All involved parties

*€ Prices are valid in Germany, exclusively, and subject to alterations. Prices incl. VAT. Books excl. shipping. Journals incl. shipping. 0257100006_dp

Since it was founded in 1906, the Ernst & Sohn “Beton-Kalender” has been supporting developments in reinforced and prestressed concrete. The aim was to publish a yearbook to reflect progress in “ferro-concrete” structures until – as the book‘s first editor, Fritz von Emperger (1862-1942), expressed it – the “tempestuous development” in this form of construction came to an end. However, the “Beton-Kalender” quickly became the chosen work of reference for civil and structural engineers, and apart from the years 1945-1950 has been published annually ever since.

should be interested in a general

exceeds the actual experiences

logical guaranteed concept of

gained so far in the various

repair and maintenance measures

disciplines concerned.

necessary for a specified lifetime.

Rüdiger Meiswinkel, Julian Meyer,

Rainer Mallée, Werner Fuchs,

Jürgen Schnell

Rolf Eligehausen

Design and Construction of Nuclear Power Plants Series: Beton-Kalender Series

Design of Fastenings for Use in Concrete – the CEN/TS 1992-4 Provisions

Series editor: K. Bergmeister,

Series: Beton-Kalender Series

F. Fingerloos, J.-D. Wörner (eds.)

Series editor: K. Bergmeister, F. Fingerloos, J.-D. Wörner (eds.)

Online-Order: www.er nst-und-sohn.de/books

Despite all the efforts being put

approx. 144 pages, 70 fig., 14 tab., Softcover. approx. € 49,90 ISBN: 978-3-433-03042-4 Publication date: approx. April 2013

into expanding renewable energy

The European pre-standard

sources, large-scale power

CEN/TS 1992-4 for the design of

stations will be essential as part

fastening by means of headed

of a reliable energy supply

studs, anchor channels as well as

strategy for a longer period. Given that they are low on CO2 emissions, many countries are moving into or expanding nuclear energy to cover their baseload supply. Building structures required for

approx. 130 pages, 81 fig., 4 tab., Softcover. approx. € 49,90 ISBN: 978-3-433-03044-8 Publication date: approx. April 2013

post-installed mechanical and chemical anchors is ready for use. The background and interpretation of the provisions related to the determination of actions and resistance based on limit state design, durability, fire resistance,

nuclear plants whose protective

fatigue and earthquake actions as

function means they are classified

required by CEN/TS 1992 are

as safety-related, have to meet

described in detail.

particular construction requirements more stringent than those involved in conventional construction.

Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG

W i l e y

Tel. +49 (0)6201 606-400 Fax +49 (0)6201 606-184 service@wiley-vch.de

C o m p a n y

0257100006_dp_181x262mm.indd 1

19_SC_U3.indd 1

Customer Service: Wiley-VCH Boschstraße 12 D-69469 Weinheim

24.01.13 11:14

01.02.13 09:01

HOLGER SVENSSON

KLAUS IDELBERGER

Cable-Stayed Bridges 40 Years of Experience Worldwide. With Live Lectures on DVD

The World of Foobridges From the Utilitarian to the Spectacular 2011. 183 pages, 351 figures. Softcover. € 69,–* ISBN 978-3-433-02943-5

2011. 458 pages, 1265 fig., Hardcover. € 129,–* ISBN 978-3-433-02992-3

Also available in German

Fuß- und Radwegbrücken Beispielsammlung

Schrägkabelbrücken 40 Jahre Erfahrung weltweit. Mit DVD: Vorlesungen live

ISBN 978-3-433-02937-4

ISBN 978-3-433-02977-0

■ The need for large-scale bridges is constantly growing due to the enormous infrastructure projects around the world. This book describes the fundamentals of design analysis, fabrication and construction, in which the author refers to 250 built examples to illustrate all aspects. International or national codes and technical regulations are referred to only as examples, such as bridges that were designed to German DIN, AASHTO, British Standards, etc. The chapters on cables and erection are a major focus of this work as they represent the most important difference from other types of bridges.

■ „Footbridges“ is a treasure trove for structural engineers. This book contains 85 examples of footbridges built worldwide over the past three decades and includes open pedestrian and cycle bridges, utility bridges and skywalks in many different environments. The collection is arranged according to load bearing system and span length. There is a brief description of the location and structural system of each bridge illustrated by photographs, plans, elevations and in some cases construction details.

JOACHIM SCHEER

M A R T I N P. B U R K E J R .

Failed Bridges Case Studies, Causes and Consequences

Integral and SemiIntegral Bridges 2009. ca. 272 pages. Hardcover. € 105,–* ISBN 978-1-4051-9418-1

2010. 321 pages. 120 fig. 15 tab. Hardcover. € 79,90* ISBN 978-3-433-02951-0

Online-Order: www.er nst-und-sohn.de

■ When bridges fail, often with loss of human life, those involved may be unwilling to speak openly about the cause. Yet it is possible to learn from mistakes. The lessons gained lead to greater safety and are a source of innovation. This book contains a systematic, unprecedented overview of more than 500 bridge failures assigned to the time of their occurrence in the bridges‘ life cycle and to the releasing events. Primary causes are identified. Many of the cases investigated are published here for the first time and previous interpretations are shown to be incomplete or incorrect. A catalogue of rules that can help to avoid future mistakes in design analysis, planning and erection is included. A lifetime‘s work brilliantly compiled and courageously presented – a wealth of knowledge and experience for every structural engineer.

■ Worldwide, integral type bridges are being used in greater numbers in lieu of jointed bridges because of their structural simplicity, economy, and durability. Written by a practicing bridge design engineer from the USA who has spent his career involved in the origination, evaluation and design of such bridges, this book shows how the analytical complexity due to the elimination of movable joints can be minimized to negligible levels so that most moderate length bridges can be easily and quickly modified or replaced with either integral or semiintegral bridges.

Ernst & Sohn Verlag für Architektur und technische Wissenschaften GmbH & Co. KG

W i l e y

Customer Service: Wiley-VCH Boschstraße 12 D-69469 Weinheim

Tel. +49 (0)6201 606-400 Fax +49 (0)6201 606-184 service@wiley-vch.de

C o m p a n y

0196300006_dp_181x262mm.indd 1

20_SC_U4.indd 1

* E-Prices are valid in Germany, exclusively, and subject to alterations. Prices incl. VAT. Books excl. shipping. Journals incl. shipping. 0196300006_dp

Literature for bridge building by Ernst & Sohn

29.05.12 16:43

12.02.13 07:46