Journal of Mechanical Engineering 2012 7-8 by Darko Svetak

http://www.sv-jme.eu

58 (2012) 7-8

Since 1955

Contents Papers

431 444 453 462 470 482 492

Javier López, Jose Breñosa, Ignacio Galiana, Manuel Ferre, Antonio Giménez, Jorge Barrio: Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation Andrej Lebar, Luka Selak, Rok Vrabič, Peter Butala: Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary Mateja Dovjak, Masanori Shukuya, Aleš Krainer: Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings Mohammad Rabiey, Christian Walter, Friedrich Kuster, Josef Stirnimann, Frank Pude, Konrad Wegener: Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser Frane Majić, Ralph Voss, Zdravko Virag: Boundary Layer Method for Unsteady Transonic Flow Jure Marn, Jurij Iljaž, Zoran Žunič, Primož Ternik: Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved Adem Çiçek, Turgay Kıvak, Gürcan Samtaş, Yusuf Çay: Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis

101010101010101010101010101010101010101010101010101010101010101 10101010101010101 01010101010101010 1010 01010101010101 101010101010101010 1010101010101010 101010101 101010101 10101 1010101010101010101 101010101 1010101 0 01010101 0 10101010101 101010 1 0 0 1010101 1010 1010 1 010101 1010 1 0101010 0101010101 101010 1010101 0101010101 10101010 10 01010101 10 0101010101 101010101 0101010101 1010101010101 10101010101010101 101010101010101010101010101010101010101010101010101010101010101

Journal of Mechanical Engineering - Strojniški vestnik

Strojniški vestnik Journal of Mechanical Engineering

7-8 year 2012 volume 58 no.

101010101010101010101010101010101010101010101010101010101010101 101010101010101010101010101010101010101010101010101010101010101 1010101 010101010101 010101010 101010101 1010 0101010 0101 010101 10 0101 10 0101 10 1 1 101 1 1 1 101 10 1 010 0101 101 01 010101010 010101 10101010101010101010101010 101010101010101010101010101010101 101010101010101010101010101010101010101010101010101010101010101 101010101010101010101010101010101010101010101010101010101010101

Technical Editor Pika Škraba University of Ljubljana Faculty of Mechanical Engineering, Slovenia Editorial Office University of Ljubljana (UL) Faculty of Mechanical Engineering SV-JME Aškerčeva 6, SI-1000 Ljubljana, Slovenia Phone: 386-(0)1-4771 137 Fax: 386-(0)1-2518 567 E-mail: info@sv-jme.eu, http://www.sv-jme.eu

Print Tiskarna Knjigoveznica Radovljica, printed in 480 copies

Founders and Publishers University of Ljubljana (UL) Faculty of Mechanical Engineering, Slovenia University of Maribor (UM) Faculty of Mechanical Engineering, Slovenia Association of Mechanical Engineers of Slovenia

http://www.sv-jme.eu

58 (2012) 7-8

Chamber of Commerce and Industry of Slovenia Metal Processing Industry Association

Since 1955

Journal of Mechanical Engineering - Strojniški vestnik

Breñosa, Ignacio Galiana, Manuel Ferre, Antonio Giménez,

Optimization for Multi-Finger Haptic Devices Applied Manipulation Selak, Rok Vrabič, Peter Butala: Analysis, and Remote Recording of Welding Parameters y anori Shukuya, Aleš Krainer: Conventional and Low Exergy Systems for Heating Zero Energy Buildings Christian Walter, Friedrich Kuster, Josef Stirnimann, Wegener: Bond CBN Wheels Using Short-Pulse Fiber Laser Voss, Zdravko Virag: thod for Unsteady Transonic Flow ž, Zoran Žunič, Primož Ternik: od Flow around Healthy and Regurgitated Aortic Valve d Flow Involved Kıvak, Gürcan Samtaş, Yusuf Çay: Forces in Drilling of AISI 316 Stainless Steel ral Network and Multiple Regression Analysis

Strojniški vestnik Journal of Mechanical Engineering

7-8 2012 58

no. year

volume

Cover: Modular multi-finger haptic device for virtual object manipulation: Mechanical structures are based on one module per finger and can be scaled up to three fingers. Mechanical configurations for two and three fingers are based on the use of one and two redundant axes, respectively. The location of redundant axes and link dimensions have been optimized in order to guarantee a proper workspace, manipulability, force capability, and inertia for the device. Image courtesy: Javier López, University of Almería, Mechanical Engineering Area, Spain

International Editorial Board Koshi Adachi, Graduate School of Engineering,Tohoku University, Japan Bikramjit Basu, Indian Institute of Technology, Kanpur, India Anton Bergant, Litostroj Power, Slovenia Franci Čuš, UM, Faculty of Mech. Engineering, Slovenia Narendra B. Dahotre, University of Tennessee, Knoxville, USA Matija Fajdiga, UL, Faculty of Mech. Engineering, Slovenia Imre Felde, Bay Zoltan Inst. for Mater. Sci. and Techn., Hungary Jože Flašker, UM, Faculty of Mech. Engineering, Slovenia Bernard Franković, Faculty of Engineering Rijeka, Croatia Janez Grum, UL, Faculty of Mech. Engineering, Slovenia Imre Horvath, Delft University of Technology, Netherlands Julius Kaplunov, Brunel University, West London, UK Milan Kljajin, J.J. Strossmayer University of Osijek, Croatia Janez Kopač, UL, Faculty of Mech. Engineering, Slovenia Franc Kosel, UL, Faculty of Mech. Engineering, Slovenia Thomas Lübben, University of Bremen, Germany Janez Možina, UL, Faculty of Mech. Engineering, Slovenia Miroslav Plančak, University of Novi Sad, Serbia Brian Prasad, California Institute of Technology, Pasadena, USA Bernd Sauer, University of Kaiserlautern, Germany Brane Širok, UL, Faculty of Mech. Engineering, Slovenia Leopold Škerget, UM, Faculty of Mech. Engineering, Slovenia George E. Totten, Portland State University, USA Nikos C. Tsourveloudis, Technical University of Crete, Greece Toma Udiljak, University of Zagreb, Croatia Arkady Voloshin, Lehigh University, Bethlehem, USA President of Publishing Council Jože Duhovnik UL, Faculty of Mechanical Engineering, Slovenia General information Strojniški vestnik – Journal of Mechanical Engineering is published in 11 issues per year (July and August is a double issue). Institutional prices include print & online access: institutional subscription price and foreign subscription €100,00 (the price of a single issue is €10,00); general public subscription and student subscription €50,00 (the price of a single issue is €5,00). Prices are exclusive of tax. Delivery is included in the price. The recipient is responsible for paying any import duties or taxes. Legal title passes to the customer on dispatch by our distributor. Single issues from current and recent volumes are available at the current single-issue price. To order the journal, please complete the form on our website. For submissions, subscriptions and all other information please visit: http://en.sv-jme.eu/. You can advertise on the inner and outer side of the back cover of the magazine. The authors of the published papers are invited to send photos or pictures with short explanation for cover content. We would like to thank the reviewers who have taken part in the peerreview process.

ISSN 0039-2480 © 2011 Strojniški vestnik - Journal of Mechanical Engineering. All rights reserved. SV-JME is indexed / abstracted in: SCI-Expanded, Compendex, Inspec, ProQuest-CSA, SCOPUS, TEMA. The list of the remaining bases, in which SV-JME is indexed, is available on the website. The journal is subsidized by Slovenian Book Agency.

Strojniški vestnik - Journal of Mechanical Engineering is also available on http://www.sv-jme.eu, where you access also to papers’ supplements, such as simulations, etc.

Instructions for Authors All manuscripts must be in English. Pages should be numbered sequentially. The maximum length of contributions is 10 pages. Longer contributions will only be accepted if authors provide justification in a cover letter. Short manuscripts should be less than 4 pages. For full instructions see the Authors Guideline section on the journal’s website: http://en.sv-jme.eu/. Announcement: The authors are kindly invited to submitt the paper through our web site: http://ojs.sv-jme.eu. The Author is also able to accompany the paper with Supplementary Files in the form of Cover Letter, data sets, research instruments, source texts, etc. The Author is able to track the submission through the editorial process - as well as participate in the copyediting and proofreading of submissions accepted for publication - by logging in, and using the username and password provided. Please provide a cover letter stating the following information about the submitted paper: 1. Paper title, list of authors and affiliations. 2. The type of your paper: original scientific paper (1.01), review scientific paper (1.02) or short scientific paper (1.03). 3. A declaration that your paper is unpublished work, not considered elsewhere for publication. 4. State the value of the paper or its practical, theoretical and scientific implications. What is new in the paper with respect to the state-of-the-art in the published papers? 5. We kindly ask you to suggest at least two reviewers for your paper and give us their names and contact information (email). Every manuscript submitted to the SV-JME undergoes the course of the peer-review process. THE FORMAT OF THE MANUSCRIPT The manuscript should be written in the following format: - A Title, which adequately describes the content of the manuscript. - An Abstract should not exceed 250 words. The Abstract should state the principal objectives and the scope of the investigation, as well as the methodology employed. It should summarize the results and state the principal conclusions. - 6 significant key words should follow the abstract to aid indexing. - An Introduction, which should provide a review of recent literature and sufficient background information to allow the results of the article to be understood and evaluated. - A Theory or experimental methods used. - An Experimental section, which should provide details of the experimental set-up and the methods used for obtaining the results. - A Results section, which should clearly and concisely present the data using figures and tables where appropriate. - A Discussion section, which should describe the relationships and generalizations shown by the results and discuss the significance of the results making comparisons with previously published work. (It may be appropriate to combine the Results and Discussion sections into a single section to improve the clarity). - Conclusions, which should present one or more conclusions that have been drawn from the results and subsequent discussion and do not duplicate the Abstract. - References, which must be cited consecutively in the text using square brackets [1] and collected together in a reference list at the end of the manuscript. Units - standard SI symbols and abbreviations should be used. Symbols for physical quantities in the text should be written in italics (e.g. v, T, n, etc.). Symbols for units that consist of letters should be in plain text (e.g. ms-1, K, min, mm, etc.) Abbreviations should be spelt out in full on first appearance, e.g., variable time geometry (VTG). Meaning of symbols and units belonging to symbols should be explained in each case or quoted in a special table at the end of the manuscript before References. Figures must be cited in a consecutive numerical order in the text and referred to in both the text and the caption as Fig. 1, Fig. 2, etc. Figures should be prepared without borders and on white grounding and should be sent separately in their original formats. Pictures may be saved in resolution good enough for printing in any common format, e.g. BMP, GIF or JPG. However, graphs and line drawings should be prepared as vector images, e.g. CDR, AI. When labeling axes, physical quantities, e.g. t, v, m, etc. should be used whenever possible to minimize the need to label the axes in two languages. Multi-curve graphs should have individual curves marked with a symbol. The meaning of the symbol should be explained in the figure caption. Tables should carry separate titles and must be numbered in consecutive numerical order in the text and referred to in both the text and the caption as Table 1, Table 2, etc. In addition to the physical quantity, e.g. t (in italics), units

(normal text), should be added in square brackets. The tables should each have a heading. Tables should not duplicate data found elsewhere in the manuscript. Acknowledgement of collaboration or preparation assistance may be included before References. Please note the source of funding for the research. REFERENCES A reference list must be included using the following information as a guide. Only cited text references are included. Each reference is referred to in the text by a number enclosed in a square bracket (i.e., [3] or [2] to [6] for more references). No reference to the author is necessary. References must be numbered and ordered according to where they are first mentioned in the paper, not alphabetically. All references must be complete and accurate. All non-English or. non-German titles must be translated into English with the added note (in language) at the end of reference. Examples follow. Journal Papers: Surname 1, Initials, Surname 2, Initials (year). Title. Journal, volume, number, pages, DOI code. [1] Hackenschmidt, R., Alber-Laukant, B., Rieg, F. (2010). Simulating nonlinear materials under centrifugal forces by using intelligent crosslinked simulations. Strojniški vestnik - Journal of Mechanical Engineering, vol. 57, no. 7-8, p. 531-538, DOI:10.5545/sv-jme.2011.013. Journal titles should not be abbreviated. Note that journal title is set in italics. Please add DOI code when available and link it to the web site. Books: Surname 1, Initials, Surname 2, Initials (year). Title. Publisher, place of publication. [2] Groover, M.P. (2007). Fundamentals of Modern Manufacturing. John Wiley & Sons, Hoboken. Note that the title of the book is italicized. Chapters in Books: Surname 1, Initials, Surname 2, Initials (year). Chapter title. Editor(s) of book, book title. Publisher, place of publication, pages. [3] Carbone, G., Ceccarelli, M. (2005). Legged robotic systems. Kordić, V., Lazinica, A., Merdan, M. (Eds.), Cutting Edge Robotics. Pro literatur Verlag, Mammendorf, p. 553-576. Proceedings Papers: Surname 1, Initials, Surname 2, Initials (year). Paper title. Proceedings title, pages. [4] Štefanić, N., Martinčević-Mikić, S., Tošanović, N. (2009). Applied Lean System in Process Industry. MOTSP 2009 Conference Proceedings, p. 422-427. Standards: Standard-Code (year). Title. Organisation. Place. [5] ISO/DIS 16000-6.2:2002. Indoor Air – Part 6: Determination of Volatile Organic Compounds in Indoor and Chamber Air by Active Sampling on TENAX TA Sorbent, Thermal Desorption and Gas Chromatography using MSD/FID. International Organization for Standardization. Geneva. www pages: Surname, Initials or Company name. Title, from http://address, date of access. [6] Rockwell Automation. Arena, from http://www.arenasimulation.com, accessed on 2009-09-07. EXTENDED ABSTRACT By the time the paper is accepted for publishing, the authors are requested to send the extended abstract (approx. one A4 page or 3.500 to 4.000 characters). The instructions for writing the extended abstract are published on the web page http://www.sv-jme.eu/ information-for-authors/. COPYRIGHT Authors submitting a manuscript do so on the understanding that the work has not been published before, is not being considered for publication elsewhere and has been read and approved by all authors. The submission of the manuscript by the authors means that the authors automatically agree to transfer copyright to SV-JME and when the manuscript is accepted for publication. All accepted manuscripts must be accompanied by a Copyright Transfer Agreement, which should be sent to the editor. The work should be original by the authors and not be published elsewhere in any language without the written consent of the publisher. The proof will be sent to the author showing the final layout of the article. Proof correction must be minimal and fast. Thus it is essential that manuscripts are accurate when submitted. Authors can track the status of their accepted articles on http://en.svjme.eu/. PUBLICATION FEE For all articles authors will be asked to pay a publication fee prior to the article appearing in the journal. However, this fee only needs to be paid after the article has been accepted for publishing. The fee is 220.00 EUR (for articles with maximum of 10 pages), 20.00 EUR for each addition page. Additional costs for a color page is 90.00 EUR.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8 Contents

Contents Strojniški vestnik - Journal of Mechanical Engineering volume 58, (2012), number 7-8 Ljubljana, July-August 2012 ISSN 0039-2480 Published monthly

Papers Javier López, Jose Breñosa, Ignacio Galiana, Manuel Ferre, Antonio Giménez, Jorge Barrio: Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation Andrej Lebar, Luka Selak, Rok Vrabič, Peter Butala: Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary Mateja Dovjak, Masanori Shukuya, Aleš Krainer: Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings Mohammad Rabiey, Christian Walter, Friedrich Kuster, Josef Stirnimann, Frank Pude, Konrad Wegener: Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser Frane Majić, Ralph Voss, Zdravko Virag: Boundary Layer Method for Unsteady Transonic Flow Jure Marn, Jurij Iljaž, Zoran Žunič, Primož Ternik: Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved Adem Çiçek, Turgay Kıvak, Gürcan Samtaş, Yusuf Çay: Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis

431 444 453 462 470 482 492

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443 DOI:10.5545/sv-jme.2011.141

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J. Javier López1* – Jose Breñosa2 – Ignacio Galiana2 – Manuel Ferre2 – Antonio Giménez1 – Jorge Barrio2 1 University of Almería, Mechanical Engineering Area, Spain 2 Technical University of Madrid, Centre for Automation and Robotics, Spain

This paper describes the design of a modular multi-finger haptic device for virtual object manipulation. Mechanical structures are based on one module per finger and can be scaled up to three fingers. Mechanical configurations for two and three fingers are based on the use of one and two redundant axes, respectively. As demonstrated, redundant axes significantly increase workspace and prevent link collisions, which is their main asset with respect to other multi-finger haptic devices. The location of redundant axes and link dimensions have been optimized in order to guarantee a proper workspace, manipulability, force capability, and inertia for the device. The mechanical haptic device design and a thimble adaptable to different finger sizes have also been developed for virtual object manipulation. Keywords: haptic, multifinger, virtual manipulation

0 INTRODUCTION Haptic devices are mechatronic systems that allow the user to interact with virtual or remote environments. These kinds of devices are typically integrated into multimodal interfaces that provide haptic, visual, and audio information concerning the manipulation performed by the user. Haptic devices are required to read the user’s hand- (or finger-) position and display forces that represent interaction with the virtual or real environment. In recent years, haptic interfaces have undergone remarkable developments, including the creation of commercialized equipment [1] that has been used for several applications in different fields such as telerobotics [2] and [3], medical surgery [4], medical rehabilitation [5], industry [6], training and education [7], and entertainment, among others. In order to improve the performance of haptic interfaces, it is crucial that the user is provided with a mechanical device that is as “transparent” as possible. Ideally, the teleoperation system would be completely transparent so that operators would feel as if they interacted directly with the remote or virtual

task [8]. Several transparency measures are defined in the literature, the most common being: i) A system is considered transparent if the master and slave's position and force responses are identical respectively, no matter what the object dynamics are [9]; ii) A transparent system requires that the impedance transmitted to or “felt” by the operator equals the environmental impedance the human operator is interacting with [10]. To achieve perfect transparency, haptic devices should have neither inertia nor friction, and infinite bandwidth. Unfortunately, these features are unachievable and compromise each other. Single finger devices are suitable for simple haptic applications designed for palpation or object border exploration. However, multiple fingers are required to perform advanced virtual manipulation tasks such as grasping, screwing, etc. Therefore, multi-finger haptic devices are useful for improving the user’s sense of immersion in virtual environments, and allow more realistic object manipulation and task performance. At this time, there is no cost effective system on the market. For this reason, it is necessary to improve

Fig. 1. Examples of multi-fingered haptics devices; a) Two Phantoms [12]; b) CyberGrasp [13]; c) Spydar [14]; and d) Hiro III [15] *Corr. Author’s Address: University of Almería, Mechanical Engineering Area, Ctra. Sacramento s/n 04120, La Cañada de San Urbano (Almería), Spain, javier.lopez@ual.es

431

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

the design of new multi-finger haptic devices that can provide effective haptic interactions while offering a compromise between cost and device complexity. In this paper, modularity, scalability, and redundancy concepts have been articulated, thereby resulting in devices that comply with the application requirements. Based on this goal, this paper describes the design of a modular multi-finger haptic device for virtual object manipulation wherein the mechanical structure is based on one module per finger that can be scaled up to three fingers. The mechanical design has been optimized based on workspace, manipulability, and force feedback capabilities. Two different configurations for two and three fingers are proposed. Some examples of multi-finger haptic developments are based on the use of several single finger commercial haptic devices [11]. These applications integrate information from the corresponding devices within the same virtual scenario manipulation. One of the most popular single finger haptic devices used for these types of applications is the Phantom from SensAble Technologies, as shown in Fig. 1a. This is a simple solution that relies on high precision at the end-effector by fixing one point for each finger within the workspace. The main drawback of this configuration results from the collisions between the links of both devices, which translates into a significantly reduced workspace. In contrast, a single device that includes several contact points has also been utilized in many applications. In this case, specific haptic devices have been developed to offer higher manipulation dexterity [16] to [18]. These solutions offer better manipulation performance and a significantly increased workspace. Fig. 1b shows the CyberGrasp device, which is an exoskeletal structure that allows separate control over each of the five fingertip contact points. This device only reflects normal forces on fingertips, without any tangential component; therefore the user can penetrate an object tangentially without any force feedback. With this device, a reference point in the wrist is required in order to calculate its 3D location. Fig. 1c shows a haptic device based on a parallel cable structure, called Spydar. The current version implements contact points for all fingers. As shown, in this type of configuration the device is advantageous in relation to accuracy and bandwidth, but workspace orientation is restricted in order to prevent the tangling of wires. This greatly limits operations involving bimanual and cooperative tasks. Finally, other complex solutions, such as the haptic robot Hiro III, also provide contact points for five fingers. However, they are very sophisticated and 432

require the use of a robotic arm, which increases their cost. Fig. 1d shows the Hiro III specular configuration. Like the Spydar, this set-up can also be inconvenient when performing bimanual tasks since they have a very limited workspace for bimanual works. This paper is outlined as follows: Section 1 provides a description of the requirements concerning mechanical implementation and the end-effector design. Section 2 summarizes certain relevant topics concerning performance measures on mechanical structures. These indexes were selected for their application to haptic devices. Section 3 presents the implemented single finger module mechanical structure and a description of the end-effector. The design optimization process of the mechanism is provided in Section 4. Section 5 defines the configuration of the two-finger device. Section 6 presents the structural design of a three-finger device. Finally, Section 7 offers conclusions concerning the design of the two- and three- fingered haptic interfaces. 1 DESIGN REQUIREMENTS FOR VIRTUAL MANIPULATION This section describes the main requirements for multi-finger haptic devices applied to grasping virtual objects, particularly concerning mechanical and endeffector performance. 1.1 Mechanical Requirements The design of haptic devices is a complex task since it implies a trade-off between most requirements. For example, a wider workspace implies greater inertia and decreased rigidity. Moreover, the design’s complexity increases with respect to multi-finger haptic devices, principally in relation to the achievement of a large enough and collision-free common workspace. The following requirements for the design of multi-finger haptic devices must be taken into consideration: 1. It must be easily scalable so that, starting from one basic module (one finger), a number of modules can be easily integrated in order to carry out multi-finger manipulation tasks. The basic module structure must be as simple and compact as possible. 2. It must be able to exert forces to the fingertip in any direction. This implies three active DoF. Otherwise, it would be possible to pass through an object in certain directions without any force contact being perceived by the user.

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

3. A useful workspace means that manipulation tasks may be undertaken with one or more fingers in a natural manner. 4. The apparent inertia of the interface must be as low as possible. Most of the haptic interfaces use parallel or series-parallel mechanical structures [19]. This configuration allows actuators to be located at (or as near as possible to) the base in order to reduce the resulting inertia of the mechanism. 5. The rigidity of the structure must serve to prevent excessive deflections in the end-effector. The series-parallel configuration is a solution that is used broadly in order to achieve rigid and light mechanisms. 6. The device must be capable of exerting a continuous force of at least 3 N on the each user’s finger in any direction [20]. 1.2 End-Effector Requirements When designing haptic devices, special attention must be paid to the end-effector, as it is the part that is in contact with the user. There are different types of endeffectors for haptic devices and they can be classified according to their functionality in the following way: tools or thimbles. Tool end-effectors allow the user to grasp certain tools for manipulating the virtual environment. These kinds of end-effectors usually take the shape of the tool used for a certain operation (scalpel, screwdriver, stylus, etc.). Telesurgery [21] and medical applications [22] provide good examples of these tool-like devices. The main requirements of the tool-like devices are defined by the specific task for which they are designed. The ideal tool-like device is designed in accordance with the tool for which the real task is performed. With the second type, thimble end-effectors, the user inserts his or her finger into a thimblelike structure in order to manipulate the virtual environment in a natural way. This approach is most suitable to multipurpose virtual manipulation and allows natural and direct exploration of virtual objects with the user’s fingers. The thimble design for a multifinger haptic device must comply with the following requirements: 1. It must be optimized so that it adjusts to different finger sizes. 2. It must be ergonomic so that the user feels comfortable when using the haptic device. 3. The fixing force to the user’s finger should be sufficient to ensure that the finger does not come

loose but also low enough that the force does not affect the user’s perception. 4. As this thimble is located at the end part of the device, it must be as light as possible. Otherwise, the inertia of the device will increase and thus affect the user’s perception. 5. The thimble must be attached to the haptic device so that only forces can be exerted to the user (without reacting torques that would apply a twisting sensation to the finger) and to allow three passive DoF rotations to orient the finger within the scenario. 2 PERFORMANCE MEASURES The mechanical design of these devices must prioritize achievement of a large workspace, low mass and inertia, high stiffness, high payload capability, force and motion isotropy, null (or near zero) backlash, low friction, absence of singularities inside the workspace, and high bandwidth [23] and [24]. Several indexes are proposed in the literature to improve kinematics and dynamics performance. Multi-criteria optimization methodologies have been used that take into account several indexes in the design process. 2.1 Kinematic Measures Most of the kinematic performance indexes can be expressed in terms of the Jacobian matrix, J. The velocity equation of the device can be written as:

p = J ω , (1)

where ω and p are the actuators velocity and the end-effector velocity vectors, respectively. Based on the Jacobian matrix some manipulability measures have been proposed and broadly used. The Yoshikawa manipulability index [25] is related to the volume of the manipulability ellipsoid and is expressed as:

w = det( JJ T ). (2)

Another measure of the manipulability and isotropy of the device is the condition number κ of the Jacobian matrix. If the condition number is close to 1, the Jacobian matrix will be a well-conditioned matrix, and the haptic interface will have an isotropic configuration. The 2-norm has been considered for the condition number [26]: σ κ = max , (3) σ min

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

433

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

where σmax and σmin are the maximum and minimum singular values of the Jacobian matrix, respectively. The inverse of the condition number 1/κ is the local conditioning index, LCI, which depends on the position within the workspace. With the objective of evaluating the global behavior of the entire workspace, the global condition index, GCI, is proposed [26] and expressed as: 1 ∫W κ dW GCI = . (4) ∫ dW W

Taking the velocity vector of the actuators as the unit, the maximum and minimum values of the endeffector velocity will be the maximum and minimum singular values of the Jacobian matrix, respectively. These maximum and minimum values define the ellipsoid of manipulability for the given work position. Another global performance index used to quantify the kinematic isotropy of the mechanism is the global isotropy index, GII [27], which measures the global worst-case kinematic performance, and is calculated as the ratio of the minimum and maximum singular values inside the workspace. Other important measures are the minimum and maximum of the maximum forces in all directions for every position of the end-effector, defined by:

Fmax =

λF max , (5)

Fmin =

λF min , (6)

where λF max and λF min are the eigenvalues of the matrix J-1= (J-1)T. In order to evaluate the payload capability inside the workspace, the GPI [28] is expressed as:

GPI max =

GPI min =

∫W ∫W

Fmax dW

∫W dW Fmin dW

∫W dW

, (7)

, (8)

where GPImin is the most useful parameter for evaluating the device payload capability. Since both motors can be independently actuated with their maximum values (assuming as unit), the 2-norm is unrealistic [29]. This indicates that the appropriate norm in this case is the infinite norm, which states that the absolute value of the exerted torque for both motors are independently bound to 1. 434

Others kinematic performance indexes have been proposed but their use is less extended; these include the global stiffness index, GSI, global velocity index, GVI, and maximal inscribed workspace, MIW [26] and [30]. 2.2 Dynamic Measures Dynamic performance indexes are measured based on the mass matrix properties, where the relation between actuator torque and end-effector acceleration are taken into account. The goal for improving dynamic performance is to minimize inertia effects that conflict with high acceleration demands [31]. The performance indexes considered are the effective inertia matrix and generalized inertia ellipsoid [32], dynamic manipulability index [33], and global dynamic index, GDI [27]. A number of factors affect the mechanical device’s bandwidth such as stiffness, inertia, damping, friction, drive-train backlash, actuator limiting, contact, sensor/ actuator collocation, gains, and operator impedance [34]. The device’s design must optimize these factors in order to achieve high bandwidth. From a human perception perspective, voluntary limb movements have a bandwidth of less than 10 Hz. Force feedback perception depends on the frequency stimulus. Kinesthetic stimulus mainly relates to compressive stress (≈10 Hz) and skin motion (≈30 Hz). Tactile stimulus corresponds to vibration patterns that include higher frequencies of around one hundred Hz signals, and some mechanoreceptors even can perceive up to 10,000 Hz [35]. 2.3 Multi-Criteria Methods Several performance indexes must be considered for an optimum mechanical design. All of these criteria imply a trade-off, meaning that one cannot be improved without deteriorating others. To overcome this problem the multi-criteria optimization method has been broadly used [31] and [36] to [38]. Due to the difficulty of optimizing a haptic device that takes into account a large number of performance parameters, only a small number of them are commonly used in the optimization process. The most widely used performance indexes for both the multi-criteria and the single objective methods are the GCI, GPI, GII, and large workspace [19], [36] and [39] to [41]. The following sections describe the design for the proposed haptic device. Since the mechanical structure design has been selected to achieve low inertia and friction, only kinematic measures such as

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

device workspace, GCI, and GPI are considered in the optimization process. 3 SINGLE FINGER SCALABLE BASIC MODULE DESIGN The proposed design is based on a modular and scalable configuration in which the basic unit is one finger. This module has a simple and compact structure that is easily scalable, so that it is possible to join several of these modules in order to produce a multifinger haptic device that prevents collisions between modules. Each finger interacts with a single module that has its own mechanical and electronic components to assure scalability. This paper focuses on the mechanical design. Nonetheless, details about control and the electronic components can be found at [42]. 3.1 Mechanical Structure Design The modular unit has six DoFs (Fig. 2a). The first three DoFs are actuated and configured in a seriesparallel structure. The last three DoFs are passive and allow any orientation for the end-effector. The first DoF is in serial with the second and third DoFs, which are in a parallel structure. The parallel structure has been designed as a 5-bar mechanism. Fig. 2b shows the kinematic model of the 5-bar mechanism. According to the orientation of links l3 and l4, two configurations are allowed: “elbow-outside” and “elbow-inside”. “Elbow-outside” provides higher performance than “elbow-inside”. According to the bar lengths, different kinematics, workspaces, forces, and manipulability maps are obtained. The following equations show the direct kinematics of the parallel mechanism. The position of the end-effector is described such that M2 and M3 joins (θ and φ, respectively) as follows:

 xP = l1 cos ϕ + l cos(ϕ + ψ ) . (9)   yP = l1 sin ϕ + l sin(ϕ + ψ )

Orientations of l1 and l4 links are defined by φ and θ angles, respectively. The γ1 angle is obtained by calculating β and applying the cosines theorem to links l2, l3 y l5.

y −y  β = tan −1  2 1  , (10)  x1 − x2 

l5 = ( x1 − x2 ) 2 + ( y1 − y2 ) 2 , (11)

cos γ 1 =

l32 − l52 − l22 . (12) 2l5l2

Finally, ψ is obtained as:

ψ = γ 1 − β − ϕ , (13)

and consequently:

 l sin ϕ + l2 sin(ϕ + ψ ) − l4 sin θ − d  α = tan −1  1  . (14)  l1 cos ϕ + l2 cos(ϕ + ψ ) − l4 cos θ  The workspace of the 5-bar structure is shown in Fig. 3, and the lengths of the links are calculated in following section. These lengths are computed according to the optimization indexes. The Jacobian matrix is obtained by applying static equilibrium in the mechanism equations. The transpose of the Jacobian matrix relates torques in the actuators to the end-effector forces.

M = J T ⋅ F , (15) J J T =  11  J 21

J12   , (16) J 22 

a) b) Fig. 2. a) The mechanical configuration for each finger with 3 actuated DoFs and 3 other passive DoFs, b) kinematic model of the 5-bar mechanism Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

435

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

 sin (ϕ − α ) sin (ϕ − ψ )  J11 = l1  + sin ϕ  , (17) A  

 sin (ϕ − α ) cos (ϕ − ψ )  J12 = −l1  + cos ϕ  , (18) A  

J 21 = −

J 22 =

l4 cos (π / 2 + α − θ ) sin (ϕ − ψ )

l4 cos (π / 2 + α − θ ) cos (ϕ − ψ ) A

where

l2 cos (ϕ −ψ + π / 2 − α ) l

, (19)

thicker at the bottom where the distal phalanx meets the beginning of the middle phalanx. The thimble designed is shown in Fig. 4. This thimble is designed to fit different finger sizes by means of a screw system that adapts to the sides of the finger. Finally, Velcro is used to hold the finger to the thimble. A technique called stereolithography rapid prototyping (stereolithography) is used in the production of the thimble. An epoxy resin is used for this process. The total weight of the thimble is 76 grams, which makes it ideal for haptic applications.

, (20)

. (21)

The Jacobian expression obtained geometrically is simpler than other formulations [43]. This formulation allows for a 75% reduction in the computation calculus. a)

Fig. 3. Workspace of the 5-bar structure

3.2. End-Effector Design The 5-bar mechanism is linked to the end-effector by a gimble that enables any orientation in the workspace; these rotations are measured by encoders. Fig. 3b shows the gimble, which is made up of two links and three rotational axes that intersect at the fingertip. This implies that only the forces are reflected to the user, without torque components. The end-effector has a thimble shape and is ergonomically designed to allow realistic object exploration and grasping. The thimble design complies with most of the requirements described in Section 2.2. The geometry of the inside of the thimble is similar to that of the human finger in order to obtain similar touch responses [44] and [45]. The thimble is cone-shaped, narrow at the top and somewhat 436

b) Fig. 4. a) The CAD design of the thimble; two screws are used to adapt the thimble to different finger sizes, b) thimble and gimbal system: the three gimbals’ axes intersect at the fingertip in order to reflect forces without a torque component

4 OPTIMIZATION OF THE 5-BAR MECHANISM In order to reduce the inertia of the device, all actuators are located next to its base. Furthermore, since actuator inertia is multiplied by the square of the transmission ratio, a planetary gear with a low ratio (14:1) is used. The material and shape of the links have also been taken into account in this study. As shown in this section, the design of the single finger module has very low inertia and friction, the device dynamics effects is neglected, and the mechanical architecture is optimized with respect to mainly

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

kinematic performance measures such as device workspace, payload capability, and isotropy. Both, the workspace and performance of the 5-bar mechanism are conditioned by the length of the links [46] and [47]. Link lengths are optimized in accordance with the dextrous workspace and load capabilities required. These conditions are implemented in the global conditioning index in order to optimize the mechanical design. 4.1 Dextrous Workspace In grasping tasks the dextrous workspace concerns the common space for all fingers. For two-finger grasping this common workspace is determined by the intersection of the areas covered by the 5-bar mechanism in both fingers. As shown in Section 5, the common dextrous workspace for the user’s two fingers is represented as an ellipse of 21×16 cm axes. This common ellipse is also used in 3-finger grasping tasks. The 5-bar mechanism has been optimized to take into account the workspace within this ellipse. Fig. 5 shows the location of the ellipse in the global workspace for the single finger device.

it must be transferred into a common domain, which ranges from 0 to 1 [36]. The GCI is expressed within a common domain as:

 = GCI − GCI min , (23) GCI GCI max − GCI min

and the GPI index as:

GPI min − GPI min  min min GPI = , (24) min GPI min − GPI max min

Each equation index in Eq. (22) is defined so that a value close to 1 implies good performance of the device. Therefore, the set of lengths with the maximum DI, is the set of optimal lengths. In Eqs. (23) and (24), GCImax and GPI min max are the maximum values obtained for all possible combinations of lengths, and GCImin and GPI min min the minimum values, respectively. In Eq. (22), the weight given for GCI is 0.75, and 0.25 for GPI, thereby giving more importance to manipulability than to payload capability. The optimization of the 5-bar mechanism takses place on the previously mentioned ellipse. The discrete parameter space and optimal solutions are shown in Table 1. In order to reduce the apparent inertia of the interface, the actuators of the 5-bar mechanism are located as close as possible to each other, such that the length d (Fig. 2b) is kept constant and equal to 4 cm. It has been observed that link lengths are very similar for optimum GCI, optimum GPI, and optimum DI. 4.3 Device Performance

Fig. 5. Global workspace of the 5-bar structure and dexterous workspace in grasping tasks

4.2 Kinematic Optimization Multi-criteria design optimization is implemented in order to obtain optimal design solutions. A multicriteria methodology is used that takes into account the optimization of both GCI and GPImin performance indexes. This multi-criteria Design Index, DI, is expressed as:  min  + C GPI DI = C1 GCI , (22) 2 where Ci is the weight given to each performance index. Since the physical meaning and the dimensional units of the two performance indicexes are different,

Link lengths for optimal DI were implemented in the haptic device. Fig. 5 showed the workspace of the 5-bar mechanism and the area in which optimization occurred. The LCI and the minimum of the maximum force reflected Fmin of all directions for every point of the workspace are shown in Figs. 6a and b, respectively. The variation of these parameters between close points inside the workspace is an undesirable behaviour. The values in Fig. 6b represent the minimum value when taking into account all the points within the ellipse of the maximum force in all directions exerted by a torque of 1 Nm torque on each actuator for every point in the workspace. Therefore, real exertable forces are obtained by multiplying the obtained force values by the real torque value transmitted by the actuators. According to the design requirements, the nominal torque of the actuators was

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

437

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

Table 1. 5-bar parameter space Parameter l1 l2 l3 l4 l d

Min. [cm] 8.0 2.0 2.0 2.0 16.0 4.0

Max. [cm] 20.0 8.0 16.0 16.0 22.0 4.0

Step [cm] 1 0.5 1 1 0.5 GCI GPI DI Total discrete samples

selected to ensure an exertable force of 3 N in every direction and at every point of the workspace. The ellipse of manipulability for several workspace points is shown in Fig. 6c. Ellipses close to a circumference imply similar manipulability in any direction. In contrast, ellipses whose principal axes have different values imply that the manipulability changes depending on the direction of the movement. If the minor ellipse axis is near to zero, it indicates the proximity to a singular point.

Total 12 12 14 14 12 -

Optimum GCI 14.0 6.0 11.0 4.0 20.0 4.0 0.432 5.17 0.971

Optimum GPI 12.0 5.0 9.0 3.0 19.0 4.0 0.407 5.81 0.955 7,726,587

a) 8

10 9

Table 2. Main specifications of the haptic device 0,1 8

5 6

2 3 4

8 5 6 7

Property Value Maximum exertable force * 3N Inertia (apparent mass at tip) ** Without encoder gimbal 52 g With encoder gimbal 128 g Stiffness ** 1.54 N/mm Weight of the device 650 g Bandwidth 8 Hz *Minimum of the maximum force in all directions for the nominal position. ** Nominal position (midpoint of the workspace).

Optimum DI 13.0 6.0 10.0 4.0 20.0 4.0 0.431 5.35 0.979

7 -0,1 6 9

7 -0,2

5 4 3

-0.2

-0.1

0.1

0.2

0.3

0.2

438

0.1

0 Y

The specifications of the single finger module are given in Table 2. In the plane of the 5-bar mechanism, the apparent mass in the middle position of the workspace (20 cm from the motor axis) is only 39 g, of which 30% is due to the rotor inertia. The apparent mass about the axis of the serial motor is 52 g. Additionally, 76 g of the thimble weight must be added to the apparent inertia. The mechanical stiffness is 1.54 N/mm and was estimated using Finite Element Analysis (FEA) with Solidworks simulation software, where the stiffness of the device is calculated by simulating an external force at the end-effector and measuring the displacement obtained. The bandwidth of the proposed system was measured by commanding a sinusoidal trajectory and

-0.1

-0.2

-0.1

0.1

0.2

0.3

X c) Fig. 6. a) LCI of the 5-bar mechanism, b) the minimum of maximum forces (in N) for all directions reflected within the workspace, calculated for a normalized torque of 1 Nm on the actuators, c) elipses of manipulability for various points in the workspace; the closer to a circle, the higher the manipulability

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

acquiring the encoder position of the motors. These measurements were taken at the plane of the 5-bar mechanism in the middle position of the workspace (20 cm from the motor axis). The results of this experiment are shown in Fig. 7. As can be seen, the resulting frequency (for -3 dB) is approximately 8 Hz. The current bandwidth is limited by the selected actuator planetary gear.

the rotating redundant DoF is allowed. The wide workspace obtained by the device’s upper plane reflects an inverted position highly appropriate to certain bimanual manipulation tasks [42] in that it provides the user with a wider collision-free space.

Fig. 8. 2-finger haptic device; the user inserts the thumb and index finger in the corresponding thimbles in order to perform virtual object manipulation; the redundant axis significantly increases the device workspace Fig. 7. Measured bandwidth of the proposed 5-bar mechanism; the bandwidth is limited to 8 Hz mainly due to the chosen actuators

5 STRUCTURAL DESIGN FOR A 2-FINGER HAPTIC DEVICE A mechanical structure was designed and developed for the two-finger haptic device in order to enable object manipulation within a virtual scenario [48]. The 2-finger haptic device design uses two scalable haptic modules joined by means of an additional redundant axis. This device was designed specifically for grasping objects with the thumb and index finger. It allows the user to interact with virtual environments and undertake grasping tasks in an easy and comfortable manner. Both modules are connected to the base of the interface by means of an additional and redundant degree of freedom. Fig. 8 shows the 2-finger haptic device designed, where all its actuators are located at the base of the structure in order to lower the inertia. Fig. 9 shows the reachable and dexterous workspaces covered by the 5-bar mechanism of both fingers. As presented in Section 4, this dexterous workspace can approximate an ellipse of 21×16 cm axes. The redundant DoF significantly increases the workspace of this haptic interface. Fig. 10a shows the dexterous workspace supplied by the first DoF of each module and without the redundant DoF. In Fig. 10b the workspace increases considerably when

Fig . 9. Dextrous workspace of the 2-finger haptic device

This 2-finger haptic device has been used in several applications for virtual object manipulations. Fig. 11 provides an example of bimanual virtual manipulation where haptic devices are placed in an inverted position. In this case, a box is grasped and raised by one or two hands in order to study the human weight discrimination for virtual manipulation tasks [49]. 6 STRUCTURAL DESIGN FOR A 3-FINGER HAPTIC DEVICE A three-finger haptic device was developed which applies the previously described concept of scalability.

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

439

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

to achieve an optimized workspace. The resulting 3-finger haptic device design is shown in Fig. 12. The two redundant axes enable wrist rotation without causing collisions between modules, and a wide workspace.

Fig. 12. Design of the 3-finger haptic device, which is made up of three haptic devices and two redundant rotational axes

b) Fig. 10. a) 3D workspace of the 2-finger haptic device without a redundant DoF; b) 3D workspace of the 2-finger haptic device with a redundant DoF

As described in Section 3, this device uses three scalable haptic modules and two redundant axes to ensure a proper workspace. This section focuses on the placement of the two redundant axes in order

The 3-finger haptic device is designed to allow the user to manipulate the virtual environment using the thumb, index finger, and middle or ring fingers depending on the kind of task to be performed. The middle or ring fingers can be used depending on the hand gesture required by the manipulation [50]. The thumb, index finger, and middle fingers are used for performing precise manipulations such as writing. In contrast, the thumb, index finger, and ring finger are more suitable for powerful manipulations such as carrying a suitcase. This second configuration is more stable and adequate for handling heavy loads. The middle or ring finger is selected according to the

Fig. 11. Example of bimanual manipulation using a 2-finger haptic device for each hand; the set-up is optimized for grasping virtual objects; the same object is uni- or bi- manually grasped

440

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

hand gesture required [51] and [52] and based on the precision/power needed for the task or the prismatic/ circular manipulation. A large workspace is obtained for the 3-finger haptic device. It has an almost toroidal shape, as shown in Fig. 13a. The first redundant DoF enables rotation with respect to the device base. It defines the toroidal area and allows the user’s hand long movements. The second DoF represents the rotation of the user’s wrist and it prevents collisions between the module links when the human wrist rotates. This movement is shown in Fig. 13b, where the optimal ellipse achieved by the 5-bar mechanism rotates around the redundant DoF.

to the device base without a significant increase in collisions and, therefore, an increase in the device’s inertia. 7 CONCLUSIONS This paper describes the mechanical design of a two- and three- finger haptic device. The design was optimized for grasping tasks in which virtual objects are manipulated by one or two hands. The main requirements considered for the mechanical design of multifinger haptic devices have been: low inertia, the force reflected, scalability, and an adjustable and ergonomic thimble. The main contribution of this paper concerns the design of multifinger haptic devices to achieve a proper workspace. An optimized, scalable mechanical structure for one finger has been defined. Specifically, we ensure a proper workspace for the resultant multifinger haptic device by means of an addition of a redundant axis for joining modules. As we have shown, one or two redundant axes are adequate for providing a proper workspace for 2-finger or 3-finger haptic interfaces. These redundant axes extend the workspace area and prevent link collisions when hand orientation changes. This solution offers a significantly broader workspace in comparison with the use of several single-finger haptic interfaces together, while being similar in complexity. In contrast, certain multifingered haptic devices exist that offer a larger workspace, but at the expense of greater complexity, since a robot-like device must be used to transport the haptic device. 8 ACKNOWLEDGEMENTS

b) Fig. 13. a) Effective workspace of the 3-finger device with two redundant DoF for grasping tasks; b) The second redundant DoF prevents collisions between the modules

The mechanical design described for 2-finger and 3-finger haptic devices may not be suitable when extended to four and five fingers. The main problems concern the concentration of actuators in the device base and the thimble design four and five fingers, 14 and 17 actuators are required, respectively. This high number of actuators cannot be properly located close

This work has been partially funded by the Spanish “Ministerio de Ciencia e Innovación” under the projects “TEMAR” (DPI 2009-12283) and “DAVARBOT” (DPI 2011-22513), the European Commission Immersive multi-modal interactive presence (IMMERSENCE) under Grant FP6IST-027141, the “Personal Investigador en Formación” by UPM, and the Andalusia Regional Government under the Grant FPDU 2008 co-financed by the European Union through the European Regional Development Fund (ERDF). 9 REFERENCES [1] Salisbury, J.K., Srinivasan, M.A. (1997). Phantombased haptic interaction with virtual objects. IEEE

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

441

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

Computer Graphics Applications, vol. 17, no 5, p. 6-10, DOI:10.1109/MCG.1997.1626171. [2] Škorc, G., Zapušek, S., Čas, J., Šafarič, R. (2010). Virtual user interface for the remote control of a nanorobotic cell using a haptic-device. Strojniški vestnik – Journal of Mechanical Engineering, vol. 56, no. 7-8, p. 423-435. [3] Peer, A., Buss, M. (2008). A New admittance-type haptic interface for bimanual manipulations. IEEE/ ASME Transactions on Mechatronics, vol. 13, no. 4, p. 416-428, DOI:10.1109/TMECH.2008.2001690. [4] Waldron, K.J., Tollon, K. (2003). Mechanical characterization of the immersion corp. haptic, bimanual, surgical simulation interface. 8th International Symposium on Experimental Robotics, vol. 5, p. 106-112. [5] Ferre, M., Galiana, M., Wirz, R., Tuttle, N. (2011). Haptic device for capturing and simulating hand manipulation rehabilitation. IEEE/ASME Transactions on Mechatronics, vol. 16, no 5, p. 808-815, DOI:10.1109/TMECH.2011.2159807. [6] Strolz, M., Groten, R., Peer, A., Buss, M. (2011). Development and evaluation of a device for the haptic rendering of rotatory car doors. IEEE Transactions on Industrial Electronics, vol. 58, no. 8, p. 3133-3140, DOI:10.1109/TIE.2010.2087292. [7] Coles, T.R., Meglan, D., John, N.W. (2011). The role of haptics in medical training simulators: A survey of the state of the art. IEEE Transactions on Haptics, vol. 4, no. 1, p. 51-66, DOI:10.1109/TOH.2010.19. [8] Raju, G.J., Verghese, G.C., Sheridan, T.B. (1989). Design issues in 2-port network models of bilateral remote teleoperatio. IEEE International Conference on Robotics and Automation, p. 1317-1321. [9] Yokokohji, Y., Yoshikawa, T. (1992). Bilateral control of master-slave manipulators for ideal kinesthetic coupling-formulation and experiment. IEEE Transactions on Robotics and Automation, vol. 10, no. 5, p. 605-620, DOI:10.1109/70.326566. [10] Lawrence, D.A. (1993). Stability and transparency in bilateral teleoperation. IEEE Transactions on Robotics and Automation, vol. 9, no. 5, p. 624-637, DOI:10.1109/70.258054. [11] Melder, N., Harwin, W.S. (2005). Force shading and bump mapping using the friction cone algorithm. Eurohaptics Conference and Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems, p. 573- 575, DOI:10.1109/WHC.2005.58. [12] Droll, J.A., Hayhoe, M.M., Triesch, J., Sullivan, B.T. (2005). Task demands control acquisition and storage of visual information. Journal of Experimental Psychology: Human Perception and Performance, vol. 31, no. 6, p. 1416-1438, DOI:10.1037/00961523.31.6.1416. [13] CyberGlove Systems LLC. CyberGlove II, CyberTouch, CyberGrasp, and CyberForce California, from http://www.cyberglovesystems.com, accesed on 2012-03-30.

442

[14] Kohno, Y., Walairacht, S., Hasegawa, S., Koike, Y., Sato, M. (2001). Evaluation of two-handed multi-finger haptic device SPIDAR-8. Precision and Intelligence Laboratory, Tokyo Institute of Technology. Nagatsuta, Yokohama. [15] Endo, T., Kawasaki, H., Mouri, T., Ishigure, Y., Shimomura, H., Matsumura, M., Koketsu, K. (2011). Five-fingered haptic interface robot: HIRO III. IEEE Transactions on Haptics, vol. 4, no. 1, p. 14-27, DOI:10.1109/TOH.2010.62. [16] Kawasaki, H., Mouri, T., Osama, M., Sugihashi, Y., Ohtuka, Y., Ikenohata, S., Kigaku, K., Daniulaitis, V., Hamada, K. and Suzuki, T. (2005). Development of Five-Fingered Haptic Intereface: HIRO II. Proceedings of ICAT, Christchurch, p. 209-214, DOI:10.1145/1152399.1152437. [17] CyberGlove Systems LLC. User’s Guides, from http:// cyberglovesystems.smartsupportapp.com/subjects/ user-guides, accesed on 2012-03-30. [18] Raytheon Company. Big Arm, from http://www. raytheon.com/businesses/rids/businesses/gis/strategic_ solutions/robotics/bigarm/index.html, accessed on 2012-03-30. [19] Lee, S.S., Lee, J.M. (2003). Design of a general purpose 6-DOF haptic interface. Mechatronics, vol. 13, p. 697722, DOI:10.1016/S0957-4158(02)00038-7. [20] O’Malley, M., Goldfarb, M. (2002). The effect of force saturation on the haptic perception of detail. IEEE/ ASME Transactions on Mechatronics, vol. 7, no. 3, p. 280-288, DOI:10.1109/TMECH.2002.802725. [21] Guthart, G.S., Salisbury, J.K. (2000). The IntuitiveTM telesurgery system: overview and application. IEEE International Conference on Robotics & Automation, p. 618-621. [22] Moog Inc. Moog Simodont Dental Trainer, from http://www.moog.com/literature/ICD/Moog-HapticsSimodontDentalTrainer-Datasheet-en.pdf, accessed on 2012-03-30. [23] Fischer, P., Daniel, R., Siva, K.V. (1990). Specification and design of input devices for teleoperation. Proceedings of the 1990 IEEE International Conference on Robotics and Automation, Cincinnati, p. 540-545. [24] Hayward, V. Astley, O.R. (1996). Performance measures for Haptic Interfaces. 7th International Symposium in Robotics Research, p. 195-207. [25] Yoshikawa, T. (1991). Translational and rotational manipulability of robotic manipulators. Industrial Electronics Conference Proceedings, p. 1170. [26] Gosselin, C., Angeles, J., (1991). Global performance index for the kinematic optimization of robotic manipulators. Journal of Mechanisms, Transmissions, and Automation in Design, vol. 113, p. 220-226. [27] Stocco, L., Salcudean, S.E., Sassani, F. (1998). Fast constrained global minimax optimization of robot parameters. Robotica, vol. 16, no. 6, p. 595-605, DOI:10.1017/S0263574798000435. [28] Ozaki, H., Wang, H., Liu, X., Gao, F. (1996). Atlas of the payload capability for design of 2-DOF planar

López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 431-443

parallel manipulators. Proceedings of the 1996 IEEE International Conference on Systems, Man and Cybernetics, Beijing, p. 1483-1488. [29] Merlet, J.P. (2006). Jacobian, manipulability, condition number, and accuracy of parallel robots. Journal of Mechanical Design, Transactions of the ASME, vol. 128, p. 199-206, DOI:10.1115/1.2121740. [30] Liu, X.J., Wang, J., Pritschow, G. (2006). Kinematics, singularity and workspace of planar 5R symmetrical parallel mechanisms. Mechanism and Machine Theory, vol. 41, no. 2, p. 145-169, DOI:10.1016/j. mechmachtheory.2005.05.004. [31] Unal, R., Kiziltas, G., Patoglu, V. (2008). A multicriteria design optimization framework for haptic interfaces. Symposium on Haptics Interfaces for Virtual Environment and Teleoperator Systems Proceedings, p. 231, DOI:10.1109/HAPTICS.2008.4479949. [32] Asada, H. (1983). Geometrical representation of manipulator dynamics and its application to arm design. Journal of Dynamic Systems, Measurement and Control, Transactions of the ASME, vol. 105, no. 3, p. 131-135. [33] Yoshikawa, T. (1985). Dynamic manipulability of robot manipulators. Journal of Robotic Systems, vol. 2, no. 1, p. 113-124. [34] Wall, S.A., Harwin, W. (2001). A high bandwidth interface for haptic human computer interaction. Mechatronics, vol. 11, no. 4, p 371-387, DOI:10.1016/ S0957-4158(00)00024-6. [35] Brooks, T.L. (1990). Telerobotic response requirements. Systems, Man and Cybernetics, Conference Proceedings of IEEE International Conference, p.113120. [36] Lee, J.H., Yi, B.J., Oh, S.R., Suh, I.H. (2001). Optimal design and development of a five-bar finger with redundant actuation. Mechatronics, vol. 11, p. 27-42, DOI:10.1016/S0957-4158(99)00089-6. [37] Castejón, C., Carbone, G., Prad, J.C.G., Ceccarelli, M. (2010). A multi-objective optimization of a robotic arm for service tasks. Strojniški vestnik – Journal of Mechanical Engineering, vol. 56, no. 5, p. 316-329. [38] Kucuk, S., Bingul, Z. (2006). Comparative study of performance indices for fundamental robot manipulators. Robotics and Autonomous Systems, vol. 54, no. 7, p. 567-573, DOI:10.1016/j.robot.2006.04.002. [39] Ryu, D., Song, J.B., Cho, C., Kang, S., Kim, M. (2010). Development of a six DOF haptic master for teleoperation of a mobile manipulator. Mechatronics, vol. 20, no. 2, p. 181-191, DOI:10.1016/j. mechatronics.2009.11.003. [40] Yoon, J., Ryu, J. (2001). Design, fabrication, and evaluation of a new haptic device using a parallel mechanism. IEEE/ASME Transactions on Mechatronics, vol. 6, no. 3, p. 221-233, DOI:10.1109/3516.951360. [41] Stocco, L.J., Salcudean, S.E., Sassani, F. (2001). Optimal kinematic design of a haptic pen. IEEE/ASME

Transactions on Mechatronics, vol. 6, no. 3, p. 210220, DOI:10.1109/3516.951359. [42] Garcia-Robledo, P., Ortego, J., Ferre, M., Barrio, J., Sanchez-Uran, M.A. (2011). Segmentation of bimanual virtual object manipulation tasks using multifinger haptic interfaces. IEEE Transactions on Instrumentation and Measurement, vol. 60, no. 1, p. 69-80, DOI:10.1109/TIM.2010.2065690. [43] Monroy, M., Oyarzabal, M., Ferre, M., Campos, A., Barrio, J. (2008). MasterFinger: Multi-finger haptic interface for collaborative environments. Proceedings of the EuroHaptics, vol. 5024, p. 411-419. [44] Ferre, M., Galiana, I., Aracil, R. (2011). Design of an affordable thimble-like sensor for haptic application based on force-sensing resistors. Sensors, vol. 11, p. 11495-11509, DOI:10.3390/s111211495. [45] Monroy, M., Ferre, M., Barrio, J., Eslava, V., Galiana. I. (2009). Sensorized thimble for haptic applications. IEEE International Conference on Mechatronics, p. 1-6, DOI:10.1109/ICMECH.2009.4957201. [46] Giachritsis, C.D., Ferre, M., Barrio, J., Wing, A. (2011). Unimanual and bimanual weight perception of virtual objects with a new multi-finger haptic interface. Brain Research Bulletin, vol. 85, no. 5, p. 271-276, DOI:10.1016/j.brainresbull.2011.03.017. [47] Cervantes-Sánchez, J.J., Hernández-Rodríguez, J.C., Rendón-Sánchez, J.G. (2000). On the workspace, assembly configurations and singularity curves of the RRRRR-type planar manipulator. Mechanism and Machine Theory, vol. 35, p. 1117-1139, DOI:10.1016/ S0094-114X(99)00061-0. [48] Galiana, I., Bielza, M., Ferre, M. (2010). Estimation of normal and tangential manipulation forces by using contact force sensors. Lecture Notes in Computer Science Springer, Eurohaptics, vol. 6191, p. 65-72. [49] García-Robledo, P., Ortego, J., Barrio, J., Galiana, I., Ferre, M., Aracil, R. (2009). Multifinger haptic interface for bimanual manipulation of virtual objects interaction between two hands and virtual objects with MasterFinger. IEEE International Workshop on Haptic Audio Visual Environments and Games, Lecco, p. 3035. [50] Liu, X.J., Wang, J., Pritschow, G. (2006). Performance atlases and optimum design of planar 5R symmetrical parallel mechanisms. Mechanism and Machine Theory, vol. 41, p. 119-144, DOI:10.1016/j. mechmachtheory.2005.05.003. [51] Cobos, S., Ferre, M., Sánchez-Urán, M.A., Ortego, J., Aracil, R. (2010). Human hand descriptions and gesture recognition for object manipulation. Computer Methods in Biomechanics and Biomedical Engineering, vol. 13, no. 3, p. 305-317, DOI:10.1080/10255840903208171. [52] Cutkosky, M.R. (1989). On grasp choice, grasp models, and the design of hands for manufacturing tasks. IEEE Transactions on Robotics and Automation, vol. 5, no. 3, p. 269-279, DOI:10.1109/70.34763.

Mechanical Design Optimization for Multi-Finger Haptic Devices Applied to Virtual Grasping Manipulation

443

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452 DOI:10.5545/sv-jme.2012.341

Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary Lebar, A. – Selak, L. – Vrabič, R. – Butala, P. Andrej Lebar* – Luka Selak – Rok Vrabič – Peter Butala University of Ljubljana, Faculty of Mechanical Engineering, Slovenia

In certain domains of production engineering we are faced with very small batch production as it is the case in the production of heavy hydro energy equipment. In this domain manual welding is one of the most time consuming operations. Monitoring of the welding process is essential from the point of work organization as well as from the point of process control. In this paper a novel concept of data acquisition and recording of welding parameters to the welding diary is presented. Several considerations on signal acquisition, sampling rate, processing, data aggregation, wireless information transfer, and presentation are discussed. Implementation of the concept is discussed on laboratory and industrial examples. Keywords: arc welding, monitoring, wireless sensor networks, ZigBee communication

0 INTRODUCTION In the production of water turbines and heavy hydro energy equipment, welding is one of the most time consuming operations. Due to the combination of small batch production and different workpiece sizes, high flexibility of welding equipment is a prerequisite. Consequently, a high level of manual labour is needed, which unavoidably leads to lower quality and longer waiting times between operations [1]. Monitoring of manual arc welding is important since it correlates with productivity and influences welders’ skills and stress, which in turn influence the quality. This paper reports on a concept of welding process monitoring based on a microcontroller platform capable of data acquisition, aggregation, and wireless transfer to a data server and presentation in the form of a welding diary. The main welding parameters necessary to be recorded and stored are regulated by the ISO 15609-1 standard [2], which requires, among others, information on material thickness, parent materials, welding consumables, and electrical parameters. Based on this standard, a special document, i.e. a welding procedure specification (WPS) also known as a welding diary is to be used. The aim of the welding diary is to enable traceability of welds. In the welding diary the actual welding parameters and consumables for each run are to be written by the welder. The welding diary incorporates several details: run sequence, the welding process, the size of filler material, the type of current, wire feed speed, run out length, travel speed, and especially electric voltage and current. A note should be made that the term ‘electric current’ such as is used throughout this paper, 444

is frequently substituted with a ‘welding current’, when it is used in the domain of welding technology. However, filling of the welding diary in a paper form is tedious and time consuming for the welder. Moreover, it was found that up to 15% of information in the welding diary is not filled in correctly. In order to improve the reliability of the welding diary creation, methods of online process monitoring should be applied. Research work covered by scientific papers focuses mainly on the measurement and analysis of electric voltage, current, emitted light and sound [3]. Sforza and Blasiis focus on the Fourier analysis of optical signals of visible, infrared and ultraviolet emission of the plasma. They managed to evaluate the quality of welding [4]. A system based on photo diodes to detect a wide variety of weld defects is presented by Mirapeix et al. [5]. Acoustic methods are mainly useful to assess welding process stability and to detect severe deviations in arc behaviour [6], [7]. An alternative approach of the groove geometry measurement based on laser triangulation is presented by Bračun et al. [8]. Rak et al. perform quality analysis immediately after welding by means of cyclographs and autocorrelation [9]. Due to the commercial importance of the welding domain, a lot of information related to monitoring and data transmission of welding process parameters is available in patent databases. For wireless data transmission two technologies are used. The first one uses transmission at frequency 131 kHz [10], while the second one uses ZigBee communication protocol at frequency 2.4 GHz [11]. Regarding real time monitoring of the welding process there are several patented solutions [12] to [14]. Signal processing can be performed at the site of data acquisition [12], which enables additional

*Corr. Author’s Address: University of Ljubljana, Faculty of Mechanical Engineering, Aškerčeva 6, 1000 Ljubljana, Slovenia, andrej.lebar@fs.uni-lj.si

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

flexibility, or the data is transferred to a data server where the processing is done [13] and [14]. The presented concept is suitable for most types of welding, however this paper focuses on the shielded metal arc welding process (SMAW). 1 ONLINE MONITORING CONCEPT Welding stations are often repositioned, even several times a day due to specific requirements in production. Any kind of monitoring constrained to a particular location is therefore, not possible. Furthermore, if wireless technologies are used, transfer rates are limited in harsh industrial environments. This situation fits in the frame of the paradigm of smart environment, networked sensors, and ubiquitous manufacturing. The proposed concept of wireless welding monitoring is presented in Fig. 1 by means of a functional diagram. The key idea is that every welding station is equipped with a custom microcontroller, capable of acquisition, analogue to digital conversion (A/D) and digital data processing. Signal processing and aggregation is performed on-chip in order to avoid transfer rate limitations. The processed data are then transferred to a data server by means of ZigBee wireless communication. Four key steps are conceived: (1) signal acquisition and conditioning, where SMAW process signals, such as voltage and electric current are acquired and digitalised, (2) signal processing, where online aggregation of signals takes place, (3) information transfer by means of ZigBee communication, and (4) information presentation,

where the results of signal acquisition are merged with the corresponding manufacturing execution system (MES) and enterprise resource planning (ERP) system data. In turn, they are presented in the form of a welding diary. In the remainder of this section, the steps are described in detail, focusing on the specifics of SMAW signals. 1.1 The SMAW Process In the SMAW process a consumable electrode with a dry flux coating and a metal core is used. An electric arc is established between the electrode and the workpiece. Energy released in the arc melts the workpiece, the solid metal core of the electrode, and the flux coating. As the flux coating burns away, an inert gas which protects the materials from chemical reaction with ambient air, is released. In addition to the gas, a layer of slag covering and protecting the welded area is produced from the burnt flux coating [15]. The elements of the SMAW process are schematically shown in Fig. 2. Slag which is liquid and lighter than the molten metal rises to the surface forming a protective layer over the hot metal. While the slag is cooling, it protects the weld from the influence of the ambient atmosphere and also slows down the cooling rate of the weld. A typical electric current signal observed in welding with inverter type power supply source is shown in Fig. 3. Burning of stable electrical arc is achieved by constant average consumption of electric

Fig. 1. Functional diagram of the SMAW process monitoring concept Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary

445

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

current, which is in the modern inverter type of power supply achieved by a built-in controller.

1.2 Signal Acquisition and Conditioning The welding process should be observed by means of continuous monitoring of the electric current and voltage. In the paragraphs below, materials and methods to acquire these quantities are described. Electric Current Acquisition Electric current can be measured by various methods, but due to high electric current, electric isolation, required resolution, and dynamical response characteristics, a Hall effect sensor is proposed for the presented application. A wire, which carries the current, runs through a magnetic flux collector, which is in the form of a wound core. The Hall sensor itself is integrated in this core (Fig. 5).

Fig. 2. The SMAW welding process

Fig. 3. Welding current pulse occurs at the moment of contact between melted metal droplet and base material

It can be observed in Fig. 3, that for the particular example short pulses with an amplitude of nearly 30 A are superpositioned on rather constant signal of electric current with value 70 A. The pulses can be explained by the process of metal droplets detachment, which occurs when the surface tension is no longer able to keep the drop attached to the electrode as can be observed in Fig. 4. The forces which determine the detachment process are gravity, force caused by electromagnetic induction, force caused by aerodynamic drag, and a momentum force accounting for the change of mass in the drop [15].

Fig. 4. Molten material drop formation

Monitoring of process parameters can provide better insight into the process and, in the case of manually operated machining, also into the skills of a particular operator. 446

Fig. 5. Hall effect sensor

The Hall effect semiconductor sensor is based on the phenomenon that drifting charge carriers are deflected by a magnetic field, which results in an output voltage signal UH proportional to the external magnetic field B, originating in the measured electric current I. Voltage Modification The voltage signal should be digitised with the best possible resolution, therefore it’s range must be adapted to the input range of an A/D converter optimally. For this, a voltage divider followed by an isolation amplifier is proposed. A suitable electric circuit is shown in Fig. 6. The circuit incorporates a voltage divider, an isolation amplifier and power supply electronics. The isolation amplifier provides electrical isolation and acts as an electrical safety barrier. The voltage division is chosen according to the A/D input range.

Lebar, A. – Selak, L. – Vrabič, R. – Butala, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

Fig. 6. Simplified electronic circuit for voltage modification

Considerations on Sampling Rate The concept of the welding monitoring system was developed gradually. At the beginning laboratory equipment was tested. Later, it was compared to a system based on a microcontroller, dedicated for implementation in industrial environment. During the simultaneous comparison of both systems an issue emerged that despite calibration some differences between signals acquired with the National Instruments laboratory equipment (INI) and the Arduino microcontroller based system (IARD) were identified. In order to explain the differences between INI IARD, a numerical simulation was created. The results of the simulation can be compared to the experimental example the Fig. 7. The simulated reference signal (Fig. 7a) was downsampled to two signals with different sampling rate frequencies, simulating e.g. the faster (INI) and the slower (IARD) DAQ system. The difference between simulated signals can be observed in Fig. 7b. Parameters used were low sampling rate frequency fL = 2 kHz, high sampling rate frequency fH = 5×fL and reference signal with frequency fU = 100×fH = 500×fL. In the phase of monitoring system development a similar comparison was made experimentally with identical sample rate frequencies fL and fH as in simulation (Fig. 7c) The difference between experimental signals can be observed in Fig. 7d. Although the average difference between the signals measured with both DAQ systems in Fig. 7d is zero, deviations arise from the rise time and fall time of the imperfect square wave signal.

Fig. 7. a) Simulated and c) measured signal; corresponding difference between high rate and low rate sampled signal in case of b) simulation and d) measured signal

1.3 Signal Processing In order to achieve automatic welding diary generation during the welding process, the acquired signals should be processed fast enough. This means, that the measured signals should be mapped to statistically good conditioned estimators. Aggregation of acquired signals is necessary in order to achieve an acceptable data transmission rate for wireless transmission. An additional constraint is imposed by hardware capabilities. The number of data units, typical microcontrollers are able to store in random access memory (RAM) which is as low as a few thousand integer values, therefore the proper selection of numerical methods is of utmost significance.

Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary

447

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

There is no ambiguity in the calculation of average signal values, which are calculated by the following equation: xn = xn −1 +

xn − xn −1 , n

(1)

where xn is the new estimate for average value, xn−1 is the old estimate for average value, xn is the current measurement and n is the number of measurements in the current time series. In probability theory the importance of standard deviation σ originates from the central limit theorem, which states that the mean of sufficient large number of independent random variables will tend to be normally distributed and σ is a parameter of normal distribution. Usually, it is calculated using the following equation:

σ ref =

1 n ∑ ( xi − x )2 , n − 1 i =1

(2)

but this requires two passes through the data as can be observed in the following pseudocode: Function: calculate σref X - set of measurements xi - current measurement while experiment running append xi to X end calculate x for x ϵ X accumulate (x - x )2 end calculate σref acc. to Eq. 2.

to accumulate quantities given by mathematical expressions Mk and Qk (Table 1) and calculate the standard deviation as:

σB =

Qn . n −1

(4)

Table 1. Mathematical expressions for Mk and Qk [17] M1 = x1

M i = M i −1 +

Q1 = 0

Qi = Qi −1 +

xi − M i −1 i

(i − 1)( xi − M i −1 ) 2 i

i = 2, ..., n

This principle can be implemented in the pseudocode as follows: Function: calculate σB xi - current measurement while experiment running renew Q value acc. to xi renew M value acc. to xi end calculate σB acc. to Eq. 4.

In the first pass the average value x is calculated and in the second pass the squares of deviations from the mean value are added up. With large data sets two passes can be a serious issue if the available time or computer memory is limited. Unfortunately, this is the very case in the real time situations. Alternatively, a formula, which requires only one pass through data can be used [17]:

σW =

2 1  n 2 1 n   ⋅  ∑ xi −  ∑ xi   . n − 1  i =1 n  i =1  

(3)

The formula performs poorly if rounding errors are present [17], which is the case if single precision variables are used. In the literature it is advised that instead of accumulating ∑i xi and ∑i xi2 it is favorable 448

Fig. 8. a) Signal of electric current I, comparison of differences between b) standard deviations calculated by σW and c) σB

Lebar, A. – Selak, L. – Vrabič, R. – Butala, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

It can be clearly seen in Fig. 8 that the difference σW - σref (Eq. (3)) is significantly larger than the difference sB - σref (Eq. (4)). In both cases the reference standard deviation sref is calculated by standard Matlab function std() with double precision, whereas expressions are calculated with single precision. The double precision calculations are at least three orders better with respect to results. 1.4 Information Transfer The aggregated parameters of each weld need to be stored for future use. In turn, they have to be transferred to a data server and saved into a database. Several challenges are associated with this. Firstly, welding is done by both static and hand-held machine tools. The latter are movable and frequently change location, which makes data transfer by cable impractical. Wireless networks, which have the potential to solve this, have to deal with reliability and robustness issues in harsh industrial environments. Furthermore, if parameters are acquired from several welding stations, synchronisation of data has to be achieved. As a solution to the above mentioned challenges, ZigBee wireless protocol is proposed. ZigBee is based on IEEE 802.15.4 standard for low-rate wireless personal area networks and is targeted at radiofrequency applications that require low data rate, long battery life, and secure networking. What differs ZigBee from competing standards (Bluetooth, WiFi) is that newer solutions, such as ZigBee Pro, offer self-healing network capabilities, extended singlehop range of up to hundreds of metres using 2.4 GHz frequency band, fragmentation, frequency agility, and advanced support for networks of thousands of devices. These properties make ZigBee usable even in industrial environments, where problems of radiofrequency interference are common. The main limitation of the proposed protocol is that the data transfer rate is limited to 250 kbit/s in ideal conditions. Therefore, data aggregation of the previous step plays a very important role. The data are transferred by the ZigBee interface to the data server. The communication algorithm works as follows: when the welding node is turned on, the time of the module is synchronised with the time of the computer. When welding starts the module records welding start time stamp and starts calculating average welding current and voltage. When the welding stops, the welding node records the welding stop timestamp. These four pieces of data are then

transferred to the data server and to the database. The procedure is presented in the pseudocode below: Function: monitor welding synchronize time while true if welding save time of start end while welding update averages update st. deviations end save time of end send data to server end

1.5 Information Presentation The standard [2] specifies technical content of welding procedure specification, which takes the form of a welding diary in practice. In order to fill the welding diary form, welding parameters information needs to be combined with manufacturing execution system (MES) data, which includes information about work orders, welding consumables, and operators. For example, average travel speed can be calculated by dividing the weld length specified in work order data with the welding time. The welding monitoring system is responsible for current, voltage, travel speed and heat input measurements. For information about welding consumables as well as welder data and documentation, MES may be used. ERP system can provide only high-level information, such as manufacturer data. By combining all the information sources, a complete picture of the process can be obtained. 2 CASE STUDY Two case studies are presented to illustrate the approach. Firstly, laboratory experiments are performed in order to verify the viability of the concept. A microcontroller is selected and compared to laboratory measurement equipment, which shows that a low cost microcontroller solution is adequate for SMAW signal acquisition and processing. Results of online signal analysis are discussed. Secondly, an industrial implementation of the concept is presented.

Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary

449

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

2.1 Laboratory Experiments Initially, a comparison between the microcontroller and the laboratory measuring system was made. Measurements were performed on the experimental setup schematically shown in Fig. 9. As a power source, a Fronius Magic wave 2000 was used, the consumable wire was electrode Jadran S (ϕ = 2.5 mm with 0.08% C, 0.35% Si, 0.5% Mn). Structural steel with specimen thickness 4 mm was used in the experiment. The electric current was approximately 70 A, whereas the voltage supply was 21 V. The arc movement speed was 4 mm/s. Arduino microcontroller platform was selected for the development of welding monitoring system. The platform is based on open-source hardware and software. It uses ATmega328 microcontroller with 16 MHz clock speed, six analog inputs (10 bit) and 1 KB of EEPROM memory. The signal was sampled with a sampling rate of 2 kS/s and transmitted to a computer by USB serial communication. The time stamp was set by the microcontroller in microseconds. As a reference, data acquisition device National Instruments USB 6221 with a sampling rate of 10 kHz per channel and 16-bit resolution was employed. LabVIEW data acquisition software was used.

measuring range of LEM HTFS 400 is ±600 A. Response characteristics dI/dt of both sensors is 100 A/µs. Let us refer back to Fig. 7c. The figure shows that electric current measured during welding by laboratory measuring system is in good correlation with the current measured by the Arduino microcontroller platform. Fig. 7d shows the difference between the signals. The difference is in the range of ±5% of the measured signal, therefore it may be confirmed that the Arduino platform is fast and accurate enough for the given task. Online Signal Analysis The algorithm for online standard deviation calculation was tested for a range of possible signals. For this, the influence of slag on electric current was studied, as slag influences the occurrence of electric current pulses (Fig. 3). The experiment was conducted as follows: on a clean steel surface three separate weld beads were deposited. After the deposition of the beads, the remaining slag should be completely cleaned, but in the experiment the first bead was left unaltered, the second was partially cleaned of slag and the third bead was completely cleaned. After the cleaning operation, the second welding pass was applied over all three beads. It was expected that due to the remaining slag the distance between the electrode and base material varies more apparently. This should be reflected in the occurrence of electric current pulses.

Fig. 9. Experimental setup of the SMAW process monitoring system during laboratory experiments

Electric current was measured simultaneously by two Hall effect sensors. The measurement range of LEM LF 1005-S sensor is ±1000 A, while the 450

Fig. 10. Three signals of electric current corresponding to experimental SMA welds - second pass.; a) slag not removed b) slag partially removed and c) slag removed

Lebar, A. – Selak, L. – Vrabič, R. – Butala, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

During the second welding pass the signals were measured by laboratory measuring equipment and analysed online by the Arduino microcontroller platform. Signals were processed by the algorithms presented in section 1.3. The results were saved to EEPROM after each second of analysis. After the welding, the data were transmitted to the central computer by means of ZigBee wireless communication protocol. Despite the fact that signals look rather similar in the time domain (Fig. 10), their distributions (Figs. 11a to c) reveal significant differences which can be used to discriminate a stable and unstable process. The average value of the standard deviation is significantly different between not removed slag σa = 5.17 A and totally removed slag σc = 6.57 A. By using this method it is also possible to distinguish the share of stable and unstable process, e.g. if the arc is discontinued during welding. 2.2 Implementation in Industrial Environment The concept proposed in Fig. 1 is industrially implemented on the welding power station Lincoln

Idealarc DC - 1500 (USA). It is a multi-process DC arc welding power source for automatic welding applications. The machine is designed for submerged arc welding with one welding wire. The operator can set the electric current, voltage, and travel speed of the welding head. The implemented system is based on the same Arduino microcontroller platform it was used for the laboratory experiments. XBee-PRO 60 mW ZigBee device (from Digi) is used for signal transmission and MSSQL database for data storage. The system was in operation in the industrial environment for half a year. During this time 1005 welds were made. The average welding time was 951 s. The total welding time was 263 h and 3.75 MWh of electric energy were used. The start and stop times of welding, average welding process voltage, and electric current are measured. The average voltage and current are calculated according to Eqs. (2) and (5). The measured data is combined with MES data. The results are shown in Fig. 12, where process parameters are combined with operator and warehouse events.

Fig. 11. Standard deviation of electric current signals (d) for three welds presented in Fig. 10. Accompanying distributions are presented in; a) slag not removed, b) slag partially removed, and c) slag removed

Fig. 12. Welding diary graphical presentation Online Monitoring, Analysis, and Remote Recording of Welding Parameters to the Welding Diary

451

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 444-452

Each voltage spike represents one welding run. Together with the operator and warehouse actions, this gives a complete overview of the welding operations. Energy consumption as well as wire and sand consumption are monitored, which enables accurate cost tracking. In this way, the welding diary is completely digitalised, allowing for a detailed analysis of welding on the process and operation levels. 3 CONCLUSION In this paper a new concept of monitoring, analysis and remote recording to the welding diary is presented. Based on the concept, a pilot system has been developed, tested in laboratory conditions and implemented in industrial environment. The system is based on the Arduino microcontroller platform, online signal processing algorithms, and ZigBee wireless communication. It has been shown that the presented system is appropriate for the monitoring task where the complete concept is capable of substituting the tedious manual fulfillment of the welding diary. Furthermore, it has been shown, that the system can be upgraded with a user friendly graphical presentation, which surpasses the conventional welding diary. 4 ACKNOWLEDGEMENTS This work is supported by the EUREKA: Pro factory UES (E! 4177), Slovenian Ministry of science (L22001), and Grant No. 1000-09-310150. Authors wish to thank to the members of Laboratory for Welding for providing assistance in performing and interpretations of measurements. 5 REFERENCES [1] Groover, M.P. (2010). Fundamentals of modern manufacturing, 4th Edition. John Wiley & Sons, Inc, New York. [2] ISO 15609-1:2004 (2004). Specification and qualification of welding procedures for metallic materials - Welding procedure specification. International Organisation for Standardization, Geneva. [3] Chen, Q., Wang, Y. (2002). Online quality monitoring in plasma-arc welding. Journal of Material Processing Technology, vol. 120, p. 270-274, DOI:10.1016/S09240136(01)01190-6.

452

[4] Sforza, P., Blasiis, D. (2002). On-line optical monitoring system for arc welding. NDT&E International, vol. 35, p. 37-43, DOI:10.1016/S0963-8695(01)00021-4. [5] Mirapeix, J., Ruiz-Lombera, R., Valdiande, J.J., Rodriguez-Cobo, L., Anabitarte, F., Cobo, A. (2010). Defect detection with CCD-spectrometer and photodiode-based arc-welding monitoring systems. Journal of Materials Processing Technologies, p. 21322139. [6] Grad, L., Grum, J., Polajnar, I., Slabe, J.M. (2003). Feasibility study of acoustic signals for on-line monitoring in short circuit gas metal arc welding. Journal of Machine Tools&Manufacture, vol. 44, p. 555-561. [7] Horvat, J., Prezelj, J., Polajnar, I., Čudina, M. (2011). Monitoring gas metal arc welding process by using audible sound signal. Strojniški vestnik - Journal of Mechanical Engineering, vol. 57, no. 3, p. 267-278. [8] Bračun, D., Polajnar, I., Sluga, A. (2010). An approach to adaptive control of GMAW based on laser profile measurement. 63rd Annual Assembly of IIW, paper XII2004-10. [9] Rak, I., Vuherer, T., Krepek, R., Köveš, A. (1999). Quality analysis by on-line monitoring of welding process during welding. Strojniški vestnik – Journal of Mechanical Engineering, Special Edition: Design to Manufacture in Modern Industry, p. 491-501. [10] Albrecht, B. (2008). Wireless system control and inventory monitoring for welding-type devices. United States patent US20080061049. [11] Yan, C., Jianfen, C., Hongxi, H., Gaoping, L., Jiayin, W., Guohoi, Z., Junjie, Z. (2010). Dynamic monitoring system of grouping welding equipment. Chinese patent CN 201010191688 A. [12] Hidrajama, T., Okumura, S., Takaoka, K., Ohsawa, N. (2000). Arc welding monitoring device. European patent EP 1027951 B1. [13] Ivkovich, S. (2001). Method and system for weld monitoring and tracking. United States patent US 6583386 A. [14] Davidson, R.R., Flank, J.K., Cleveland, P.W. (2009). Automatic weld arc monitoring system. United states patent US 350301 A. [15] Davies, A.C. (1993). The Science and Practice of Welding, vol. 2: The Practice of Welding, 10th ed., Cambridge University Press, Cambridge. [16] Donald, K. (1998). The Art of Computer Programming, vol. 2: Seminumerical Algorithms, Addison-Wesley, Boston. [17] Higham, N.J. (2002). Accuracy and stability of numerical algorithms. Society for Industrial and Applied Mathematics, Philadelphia, DOI:10.1137/1.9780898718027.

Lebar, A. – Selak, L. – Vrabič, R. – Butala, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461 DOI:10.5545/sv-jme.2011.158

Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings Dovjak, M. – Shukuya, M. – Krainer, A. Mateja Dovjak1,* – Masanori Shukuya2 – Aleš Krainer1 1 University of Ljubljana, Faculty of Civil and Geodetic Engineering, Slovenia 2 Laboratory

of Building Environment, Tokyo City University, Japan

The purpose of the study is to compare two heating and cooling (H/C) systems regarding individual thermal comfort conditions and rational building energy use. Real test room is firstly equipped with low exergy (LowEx) system (i.e. heating-cooling ceiling radiative panels) and secondly with a conventional system (i.e. electric heaters, cooling split system with indoor unit). Additional case presents a thermally noninsulated room equipped with a conventional system. Individual thermal comfort conditions are analyzed through the simulation of human body exergy balance (hbExB), human body exergy consumption (hbExC) rates and predicted mean votes (PMV) index. Measurements of energy use and control of temperature conditions are performed on an integrated control system (ICsIE) based on fuzzy logic. The results confirm that both systems create comfort conditions if the room is thermally well insulated. In case of non-insulated room there appears cool radiant exergy that often leads to discomfort conditions. More acceptable comfortable conditions (PMV closer to 0) do not always result in a lower hbExC rate. Individual characteristics with experimental conditions have a significant influence on separate parts of hbExB. LowEx system connected with ICsIE enables to set air temperature and mean radiant temperature and creates optimal thermal comfort conditions for individual user. The measured energy use for heating was by 11 to 27% lower for LowEx system than for the conventional system. The energy use for cooling was by 41 to 62% lower for LowEx system. The presented approach of reciprocal consideration of individual thermal comfort conditions and building energy use is important for the future design of H/C systems and for their application in near zero energy buildings. Keywords: human body exergy, building heating and cooling, low exergy system, conventional system, building energy use, individual thermal comfort

0 INTRODUCTION To reach the aim of near zero energy buildings according to the Directive 2010/31/EU [1] a new approach to solving problems related to high energy use has to be defined and realized. The most effective is a holistic approach that includes interventions in the building envelope together with efficient heating and cooling systems [2]. On the level of system efficiency besides energy use, thermal comfort of an individualj should also be regarded. Moreover, thermal comfort and productivity are more important than efficient energy use. The same is true when it comes to labour cost and operational costs of buildings. Rant [3] explained the difference between the terms energy and exergy back in 1955. The exergy concept has already been well applied in the analysis of the efficiency of thermal processes [4]. However, the use of exergy concept in a built environment related to thermal comfort is still relatively new. The exergy concept can be derived from two fundamental concepts, energy and entropy, and the concept of the environmental temperature. It can explicitly indicate how much exergy is consumed in a variety of natural phenomena going on inside and outside a built environment [5] and [6]. The study by Dovjak et al. [2] compared the results of exergy and energy analysis of the whole heating system in Slovenian buildings and defined the most effective solutions towards lower

building energy use. Exergy analysis helps us to make connections among processes inside the human body and processes in a building [6] and leads to the design of heating and cooling systems that provide both thermal comfort conditions and rational building energy use [7]. In most public and residential buildings, conventional active systems using high value nonrenewable energy sources are used for heating and cooling (H/C). Nevertheless, on the market there also exist active low exergy systems (LowEx) that use low value energy sources such as renewable and other sustainable sources. They present lowtemperature heating and high-temperature cooling systems. Currently there are many different LowEx technologies available that could be classified as surface heating and cooling systems (i.e. floor, ceiling, wall H/C systems), air heating and cooling systems (e.g. recuperators), generation/conversion of cold and heat (e.g. heat pump, sun collectors), thermal storage (e.g. seasonal storage wall), distribution (e.g. district H/C) [8] to [11]. The paper is focused on large surface H/C systems (H/C radiative panels). LowEx systems have many advantages in comparison with conventional H/C systems. The most important advantages are the improvement of comfort conditions [12] to [14], the reduction of energy use for heating and cooling of buildings [12] to [15] and improved indoor air quality due to higher relative humidity

*Corr. Author’s Address: University of Ljubljana, Faculty of Civil and Geodetic Engineering, Jamova cesta 2, 1000 Ljubljana, Slovenia, Mdovjak@fgg.uni-lj.si

453

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

of air, a higher number of air changes, and lower concentration of mites [6], [14], [16] and [17]. Experiences of architects and engineers working with the design of comfort conditions in winter show that higher surface temperature (Tmr) and lower air temperature (Tai) can result in more comfortable conditions. This coincides with the fact that comfort conditions seem to lead to lower exergy consumption of human body. The relation was first investigated by Isawa et al. [18], Shukuya [19] and [21], Shukuya et al. [20], and Prek [22] and [23]. Simone et al. [24] studied the relation between the human body exergy consumption rate and the human thermal sensation. The results [24] showed that the minimum human body exergy consumption rate was related to the thermal sensation votes close to thermal neutrality, tending to a slightly cooler side of thermal sensation. The whole human body exergy balance under typical summer conditions in hot and humid regions was analyzed by Iwamatsu and Asada [25] and Shukuya et al. [20]. Tokunaga and Shukuya [26] investigated the human-body exergy balance calculation under un-steady state conditions. Schweiker and Shukuya [27] compared the predicted mean vote approach, the adaptive comfort model and the calculation of the human body exergy consumption rate by three distinctive procedures. So far only the human body exergy consumption for an average man or a larger group have been considered, but not of an individual user. Indeed, there is no study where the comparison between LowEx and conventional system is based upon the whole human body exergy balance and measured energy use in building. Moreover, the presented study presents an upgrade of the study by Dovjak et al. [2]. From the exergetic point of view, the comparison between conventional and LowEx systems regarding simulated thermal comfort conditions for individual user and measured building energy use will be investigated. The presented approach of reciprocal consideration of individual thermal comfort conditions and building energy use is important for the future design of H/C systems and for their application in near zero energy buildings. 1 METHOD

of exergy, consumption of exergy, entropy generation and entropy disposal. The general form of exergy balance equation for the human body as a system is expressed as follows [20]:

[Exergy input] – [Exergy consumption] = = [Exergy stored] + [Exergy output]. (1)

To maintain comfort and healthy conditions, it is important that exergy consumption and stored exergy are at optimal values with a rational combination of exergy input and output. The exergy input consists of five components: 1) warm exergy generated by metabolism; 2) warm/cool and wet/dry exergies of the inhaled humid air; 3) warm and wet exergies of the liquid water generated in the core by metabolism; 4) warm/cool and wet/dry exergies of the sum of liquid water generated in the shell by metabolism and dry air to let the liquid water disperse; 5) warm/cool radiant exergy absorbed by the whole skin and clothing surfaces. The exergy output consists of four components: 1) warm and wet exergy contained in the exhaled humid air; 2) warm/cool and wet/dry exergy contained in resultant humid air containing the evaporated sweat; 3) warm/cool radiant exergy discharged from the whole skin and clothing surfaces; and 4) warm/ cool exergy transferred by convection from the whole skin and clothing surfaces into the surrounding air [20]. 1.2 Experimental Set Up

1.1 Exergy-Entropy Processes in Human Beings

Fig. 1. Plan of test room with positions of conventional system (oil-filled electric heaters and split system with indoor A/C unit) and LowEx system (heating and cooling panels)

All natural or human made processes, such as biochemical processes inside the human body or any technological process, present exergy-entropy processes. Their main characteristics are generation

Test room (163.4 m3) 15 m2 glazed window; located at the Chair for Complexes, Faculty

454

Dovjak, M. – Shukuya, M. – Krainer, A.

has one exterior wall with other walls are interior. It is Buildings and Constructional of Civil and Geodetic

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

Engineering, University of Ljubljana, and equipped with LowEx and conventional system for heating and cooling with time separation. The conventional system includes 3 oil filled electric heaters and split system with indoor A/C unit. LowEx system includes six low-temperature-heating and high-temperaturecooling ceiling radiative panels (Fig. 1). The final layer of four panels is a contact stuck 1.25 cm gypsum board, and two panels have stone plates (one is a compact 2 cm marble plate and another a composite 12 mm Al-honeycomb with 3 mm tick stone plate). All panels are fixed with 4 steel screws into ceiling construction (Fig. 2).

Fig. 2. H/C radiative ceiling panels with panel assembly

Panels are connected into H/C system with valves for switching off every panel separately, and with a pump on thermostatic mixing valve. Switching between hot and cool water entering into the panels is manual. Plan and section of panels are presented in Fig. 3.

Fig. 3. Plan and section of H/C LowEx panels; in the plan the grooves for the piping are shown

The systems were compared regarding simulation of individual thermal comfort conditions and measured energy use. Two virtual users (hereinafter called users) were simulated for the analysis of individual thermal comfort conditions.

Users’ characteristics are presented in Table 1. In the simulation users were exposed to experimental conditions based on in-situ real-time measurements (Table 2). In the case of conventional system Tai was equal to Tmr. In the case of LowEx system Tai differed from Tmr. All combinations of Tai and Tmr result in the same To (22.5 °C). An additional case for simulation is a hypothetical test room with non-insulated building envelope. Skin temperature (Tsk), body core temperature (Tcr) and clothing temperature (Tcl) were calculated considering experimental conditions. RHin was set at 60%. Experimental conditions in Table 2 present one of the selected conditions for simulation. However, LowEx connected with an integrated control system of internal environment on the basis of fuzzy logic (ICsIE) enables to set up different combinations between Tai and Tmr as it will be presented in continuation. Individual thermal comfort conditions are analyzed by calculated human body exergy balance (hbExB), human body exergy consumption (hbExC) rates and predicted mean votes (PMV) index with spread sheet software developed by Hideo Asada Rev 2010 [20] and [25]. The human body is treated as a thermodynamic system of core and shell, based on exergy-entropy processes. The calculation procedures follow the human body exergy model by Shukuya et al. [20]. The software enables to calculate the whole hbExB (input, output, stored and consumed exergy rates, PMV, Tcl, Tsk, Tcr) based on input data. Input data are outdoor and indoor experimental conditions (Tai, Tao, Tmr, RHin, RHout), individual data (met, clo), and room dimensions. For exergy calculations, the reference environmental temperature (the outdoor environmental temperature, Tao) and RHout are set at to be equal to Tai and RHin . Energy use for H/C, Tao, Tai, Tmr, RHin and RHout are continuously monitored with ICsIE system. ICsIE system was developed by Trobec-Lah [29] and upgraded by Košir [28], Košir et al. [30], and Kristl et al. [31]. It enables the control of indoor air temperature, CO2 and illuminance under the influence of outdoor environment and users` requests. In such a way, LowEx system connected with ICsIE enables to set up different combinations of Tai and Tmr. The ICsIE system is divided in three parts: the sensor network system, the regulation system and the actuator system. The basic architecture of the system is presented in Fig. 4. Since energy use was measured for the same space equipped with LowEx and conventional system in different periods, approximately the same conditions were selected for the systems’ comparison (equal set-point T, time period, Tao and Tai variate among systems ±1.0 K; 0.8% assumed error).

Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings

455

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

Table 1. User characteristics User Gender Age

Adu [m2]

Met. Eff. clo. rate ins. [met] [clo]

Male

1.9

1.1

0.6

Male

2.0

0.6

Tcr [°C]

Tsk [°C]

Tcl [°C]

36.8 to 36.9 36.9 to 36.8

33.4 to 34.3 34.3 to 33.5

28.6 to 28.3 28.3 to 29.0

measurements of energy use. It can be combined with the exergy and energy analysis of the whole heating system presented in a previous study by Dovjak et al. [2]. Additionally, exergy analysis enables us to consider the effect of room conditions created with different H/C systems on separate parts of hbExB. And as the final result, the most efficient system from user and building point of view is selected. 2 RESULTS AND DISCUSSION

Table 2. Experimental conditions H/C system Conventional LowEx Conventional, non-insulated case

Tai [°C] 22.5 18 27

Tmr [°C] 22.5 27 18

RHin [%] 60 60

va [m/s] 0.1 0.1

22.5

0.1

The presented methodology combines the exergy analysis of individual thermal comfort conditions and

2.1 Individual Thermal Comfort Conditions Table 3 presents the results of simulation of individual thermal comfort conditions for two users exposed to conditions created with conventional and LowEx system. The data show that hbExC rates and PMV index vary among individuals for both systems, even if they are exposed to the same environmental

Fig. 4. Basic architecture of the ICsIE system; the presented sensor array consists of the following sensors: Tai is room air temperature, Tao outdoor air temperature, RHin internal relative air humidity, RHout external relative air humidity, Ilin1 & Ilin2 internal work plane illumination (workplace 1 and 2), Ilout external illumination, CCO2 concentration of CO2, Irgo direct solar radiation, Irdo reflected solar radiation, Wp wind speed, Wd wind direction, Pe precipitation detection, Cheat energy use for heating, Ccool – energy use for cooling [28]

456

Dovjak, M. – Shukuya, M. – Krainer, A.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

conditions. Generally, more acceptable comfortable conditions (PMV closer to 0) can be created with LowEx system with higher surface temperatures than air temperatures (Tmr 27 °C, Tai 18 °C) compared with the conventional system. Table 3. Human body exergy consumption (hbExC) rate and PMV index for two individual users in experimental conditions created with conventional and LowEx systems User

Tai [°C]

1 2

22.5 22.5

1 2 1 2

18.0 18.0 27.0 27.0

Tmr [°C] hbExC rate [W/m2] Conventional system 22.5 2.38 22.5 5.50 LowEx system 27.0 2.23 27.0 5.69 18.0 3.52 18.0 5.56

PMV index

with PMV ‒0.8 and 0 (Table 4). The whole hbExB for both users exposed to conditions created with conventional and LowEx system will be considered in the sequel. Table 4. HbExC rate and PMV index for two individual users and non-insulated case User 1 2

Tai [°C] Tmr [°C] hbExC rate [W/m2] Conventional system, non-insulated case 22.5 10 4.49 22.5 10 6.87

PMV index -0.8 0

-0.2 0.3 -0.1 0.1 -0.2 0.5

However, more acceptable comfort conditions created with LowEx system (Tmr 27 °C, Tai 18 °C) do not always result in lower hbExC rates (User 1: 2.23 W/m2; User 2: 5.69 W/m2) than in the case of conventional system (User 1: 2.38 W/m2; User 2: 5.50 W/m2), as it was proven in previous studies on average test subjects (i.e. a 30 year old male, weighing 70 kg, and 1.75 m tall, 1 met, 0.6 clo), or groups [21] to [27]. Individual differences and experimental conditions have significant influence on hbExC rate. For example, User 1 (1.1 met) has in conditions created with LowEx system (Tai<Tmr) lower hbExC rate and, vice versa, in conditions with Tai>Tmr or Tai=Tmr higher hbExC rate. Opposite situation appears for User 2 (2 met), where Tai<Tmr results in higher hbExC rate. The same conclusions were proven in studies [32] to [37]. Comfort conditions (PMV closer to 0) result in lower hbExC rate only in case of users with metabolic rate around 1 met. Additionally, PMV values do not differ much between systems. They do not give us enough information to select which system creates more acceptable comfort conditions. All created conditions result in PMV between ‒0.5 and +0.5, which is generally acceptable as comfortable conditions. However, the difference in PMV values is more significant if individual differences are considered and not systems. Additionally, PMV values are not intended for individuals, but for groups, and do not have to match AMV (actual mean vote) as it was proven in study [38]. Additional calculation for thermally non-insulated test room with conventional system (Tai 22.5 °C, Tmr 10°C) shows similar results. In those conditions users have the highest hbExC rates (4.49 W/m2, 6.87 W/m2)

2.1.1 Human Body Exergy Balance in Conditions Created with Conventional System Fig. 5 shows a numerical example of the whole hbExB for User 1 in experimental conditions created with the conventional system (22.5 °C Tai, 22.5 °C Tmr, 60%). Input exergy presents thermal radiative exergy exchange between the human body and the surrounding surfaces and it influences the thermal comfort. Cool and warm radiant exergy absorbed by the whole skin and clothing surfaces is zero because Tai is equal to Tmr. Exergy of the inhaled humid air is also zero, because RHin and Tai are equal to RHout and Tao, respectively. The main input exergy (100%) is presented by warm exergy generated by metabolism. This means that 3.32 W/m2 of thermal exergy are generated by bio-chemical reactions inside the human body. It is influenced by the difference between Tcr and Tai and by the difference between Tsk and Tai. It is important to keep the body structure and function and to get rid of the generated entropy. Thus, 3.32 W/m2 have to be released into ambient by radiation, convection, evaporation and conduction and present output exergy. As the moisture contained in the room air is not saturated, the water secreted from sweat glands evaporates into the ambient space. Warm/cool and wet/dry exergy contained in the resultant humid air containing the evaporated sweat is 0.20 W/m2 (6.1% of output and consumed exergies). In our case it appears as warm and wet exergy, because skin temperature is higher than Tai and skin RH is higher than RHin. Warm radiant exergy discharged from the whole skin and clothing surfaces emerges because of higher Tcl than Tai and presents 0.29 W/m2 (8.6% of output and consumed exergies). Exergy of 0.45 W/m2 (13.5% of output and consumed exergies) is transferred by convection from the whole skin and clothing surfaces into the surrounding air, mainly due to the difference between Tcl and Tai. Exergy consumption that presents the difference between

Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings

457

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

exergy input, exergy stored and exergy output is 2.38 W/m2 (71.7% of output and consumed exergies) for User 1 in the case of conventional system. Cool/warm rad. 0.00 Exhalation, Breath air Exg consumption sweat 0.20 0.00 2.38 Warm rad. 0.29

Cool/warm conv. 0.00

from inner part

3.32

Warm conv. 0.45

*Unit [W/m2] PMV=-0.2 [ ] Tcr=36.8°C Tsk=33.4 °C Tcl=28.6°C

Fig. 5. Exergy balance of the human body for User 1 in experimental conditions created with the conventional system (22.5 °C Tai and 22.5 °C Tmr , 60% RHin) Cool/warm rad. 0.00 Exhalation, Breath air Exg consumption sweat 0.37 0.00 5.50 Warm rad. 0.26

Cool/warm conv. 0.00

from inner part

6.86

Warm conv. 0.73

*Unit [W/m2] PMV=0.3 [ ] Tcr=36.9°C Tsk=34.3 °C Tcl=28.3°C

Fig. 6. Exergy balance of the human body for User 2 in experimental conditions created with the conventional system (22.5 °C Tai and 22.5 °C Tmr , 60% RHin)

The whole human body exergy balance for User 2 in experimental conditions created with the conventional system differs from that for User 1 (Fig. 6). Higher metabolic level (2 Met) results in doubled value of metabolic thermal exergy as input exergy (6.86 W/m2), mainly due to larger difference between Tcr and Tai than in the case of User 1 in experimental conditions created with conventional system. Larger difference between Tcr and Tai results also in higher output exergies with exhalation and evaporation of sweat (0.37 W/m2) and warm convection transferred from the whole skin and clothing surfaces (0.73 W/m2). Lower warm radiation discharged from the whole skin and clothing surfaces results due to lower 458

difference between Tcl and Tai. Exergy consumption rate for User 2 (5.50 W/m2) is much higher than for User 1 (2.38 W/m2), mainly due to the doubled value of input exergy by metabolic thermal exergy. An additional analysis of individual hbExB in thermally non-insulated test room with conventional system (Tai 22.5 °C, Tmr 10 °C) shows that lower surface temperatures result in lower temperatures of the skin and clothing surfaces and cause cool radiant exergy rate (for User 1 and User 2: 1.12 W/m2) that is absorbed by the whole skin and clothing surfaces. Such conditions lead to discomfort [2] and [20]. 2.1.2 Human Body Exergy Balance in Conditions Created with LowEx System HbExB does not depend just on individual characteristics. The experimental conditions created with conventional or LowEx systems also have significant influence. For example, if User 1 was exposed to conditions (18.0 °C Tai, 27.0 °C Tmr, 60% RHin) created with LowEx system that resulted in the same operative temperature (22.5 °C) as in conventional system, the whole hbExB differed. Larger differences between Tcr and Tai, and Tsk and Tai result in higher metabolic thermal exergy than in the case of conventional system (4.28 W/m2, 87.6%). Higher Tmr than Tai results in 0.61 W/m2 (12.4%) of warm radiant exergy absorbed by the whole skin and clothing surfaces as input exergy. Larger difference between Tcl and Tai causes higher output exergies with warm radiant exergy discharged from the whole skin and clothing surfaces (0.91 W/m2, 18.6%), higher warm exergy transferred by convection (1.45 W/m2, 29.6%), and slightly higher exhalation and evaporation of sweat (0.30 W/m2, 6.0%). Much higher output exergies and stored exergy result in lower hbExC rate (2.23 W/m2, 45.7%) than in the conventional system (Fig. 7). For User 2 in experimental conditions created with LowEx system similar conclusions can be made as for User 1: larger differences between Tcr and Tai and Tsk and Tai result in higher metabolic thermal exergy (8.44 W/m2) than in conventional system (6.86 W/m2). Higher Tmr than Tai causes 0.61 W/m2 of warm radiant exergy absorbed by the whole skin and clothing surfaces and higher output exergies than in conventional system (due to Tmr > Tai). Warm exergy transferred by convection from the whole skin and clothing surfaces into the surrounding air is 2.10 W/m2 (higher than in the conventional system due to larger difference between Tcl and Tai), exhalation and evaporation of sweat are 0.54 W/m2 (higher due to

Dovjak, M. – Shukuya, M. – Krainer, A.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

larger difference between Tcr and Tai, and Tcl and Tai) and warm radiant exergy discharged from the whole skin and clothing surfaces is 0.72 W/m2 (higher due to larger difference between Tcl and Tai). The difference between input exergies, stored exergy and output exergies is 5.69 W/m2 and presents the hbExC rate for User 2 in the case of experimental conditions created with LowEx system (it is the highest because of the highest input exergies). Warm rad. 0.61 Exhalation, Breath air Exg consumption sweat 0.30 0.00 2.23 Warm rad. 0.91

Cool/warm conv. 0.00

from inner part

4.28

Warm conv. 1.45

*Unit [W/m2] PMV=-0.1 [ ] Tcr=36.8°C Tsk=33.5 °C Tcl=29.0°C Fig. 7. Exergy balance of the human body for User 1 in experimental conditions created with the LowEx system (18 °C Tai and 27 °C Tmr , 60% RHin)

emitted from LowEx system and absorbed by the whole skin and clothing surfaces has positive effect on thermal comfort conditions, as it was proven in studies [32] to [37]. LowEx system connected with ICsIE system enables to set up the combination between Tai and Tmr, which results in optimal conditions for the individual user. 2.2 LowEx vs. Conventional System and Measured Energy Use Energy use was measured for the same test room equipped with LowEx and conventional system in various time periods. Energy use for heating was measured for winter period (H winter: 5.03. to 23.03.2010, 447 heating hours) and summer period (18.06. to 24.06.2010, 81 heating hours) and it presents overall 528 heating hours. Heating was performed also for summer period (H summer: 81 heating hours), because in many occupied spaces with special procedures higher temperatures may be required. Energy use for cooling was measured for summer period 10.06. to 24.06.2010 and 5.07. to 10.07.2010 and presents overall 453 cooling hours. The measured energy use is presented in MJ for heating or cooling as shown in Table 5. Table 5. Measured energy use for heating and cooling [MJ]

Warm rad. 0.61

H/C system

Exhalation, Breath air Exg consumption sweat 0.54 0.00 5.69 Warm rad. 0.72

Cool/warm conv. 0.00

from inner part

8.44

Warm conv. 2.10

*Unit [W/m2] PMV=0.1 [ ] Tcr=36.9°C Tsk=34.1 °C Tcl=27.8°C Fig. 8. Exergy balance of the human body for User 2 in experimental conditions created with the LowEx system (18 °C Tai and 27 °C Tmr , 60% RHin)

The exergy analysis of thermal comfort conditions shows that hbExB significantly differs between individuals and systems. Conditions created with conventional and LowEx system are comfort conditions for both users. Thermally not insulated case results in cool radiant exergy and discomfort conditions. Such conditions often appear in buildings with conventional systems. Warm radiant exergy

H winter H summer C summer C summer H winter H summer C summer C summer

Tao [°C] Tai [°C] Conventional system -2.6 23.7 13.2 25.4 20.0 25.3 26.4 20.9 LowEx system -2.6 23.3 13.5 25.5 19.3 25.7 26.5 20.7

Energy use [MJ] 2.95 1.33 3.24 3.42 2.63 0.97 1.20 2.02

The measured energy use for heating was by 11 to 27% lower for LowEx system than for conventional system. The energy use for cooling was by 41 to 62% lower for LowEx system. The overall energy use for the whole measured cooling period (453 cooling hours) was 518 MJ (1.15 MJ average energy use per cooling hour) for LowEx system, and 1314 MJ (2.90 MJ average energy use per cooling hour) for conventional system. The reason for a relatively low efficiency of the LowEx system during the experiment was a small ceiling surface area that was heated and cooled, i.e. 25% of ceiling surface. The calculated energy use for heating in case of four times larger

Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings

459

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

surface area of panels was 40% lower energy use than for conventional system.

Engineering; Ministry of Higher Education, Science and Technology, Republic of Slovenia.

3 CONCLUSIONS

5 REFERENCES

The results of the simulation of individual thermal comfort conditions show that human body hbExC rates and PMV index vary between experimental conditions created with conventional and LowEx systems. PMV index does not give us enough information to conclude which system is more efficient from the thermal comfort point of view. The results of the whole hbExB give detailed information on the effect of individual characteristics and experimental conditions on individual parts of hbExB. Previous studies on average test subjects show that comfort conditions result in lower hbExC rate. The results of our simulation on individual users showed that more comfortable conditions (PMV closer to 0) do not always result in lower hbExC rate. Individual characteristics affect considerably every part of hbExB. Both systems can create comfort conditions, but the room has to be thermally insulated. LowEx system connected with ICsIE enables the creation of thermally comfortable conditions with regulated individual parts of hbExB with set Tai and Tmr. The measured energy use for heating was by 11 to 27% lower for LowEx system than for conventional system. The energy use for cooling was by 41 to 62% lower for LowEx system. The reason for relatively low efficiency of LowEx system during the experiment was small ceiling surface area that was heated (25%). For applications in the real environment a large part of the ceiling has to be covered with panels. The calculated energy use for heating in the case of a four times larger surface area of panels is 40% lower energy use. The presented approach of reciprocal consideration of individual thermal comfort and building energy use is important for the future design of heating and cooling systems for the application in near zero energy buildings.

[1] Directive 2010/31/EU. (2010). Directive of the European Parliament and of the Council on the energy performance of buildings (recast). European Commission, Brussels. [2] Dovjak, M., Shukuya, M., Olesen, B.W., Krainer, A. (2010). Analysis on exergy consumption patterns for space heating in Slovenian buildings. Energy Policy, vol. 38, no. 6, p. 2998-3007, DOI:10.1016/j. enpol.2010.01.039. [3] Rant, Z. (1955). Energy value and pricing. Strojniški vestnik - Journal of Mechanical Engineering, vol. 1, no. 1, p. 4-7. [4] Galović, A., Ferdelji, N., Mudrinić, S. (2010). Entropy generation and exergy efficiency in adiabatic mixing of nitrogen and oxygen streams of different temperature and environmental pressures. Strojniški vestnik Journal of Mechanical Engineering, vol. 56, no. 12, p. 817-822. [5] Takahashi, H. (1979). Exergy. Applied Physics, vol. 48, no. 8, p. 49-54. [6] Shukuya, M., Hammache, A. (2002). Introduction to the concept of exergy for a better understanding of low temperature heating and high temperature cooling systems. IEA Annex 37, from http://www.vtt.fi/inf/pdf/ tiedotteet/2002/T2158.pdf, accessed on 2011-06-15. [7] Dovjak, M., Shukuya, M., Olesen, B.W., Krainer, A. (2010). Innovative design of renewable energy technology systems for heating and cooling in sustainable buildings. Renewable Energy, Conference Proceedings, p. 1-4. [8] International Energy Agency. (2006). Low Exergy heating and Cooling of Buildings, Annex 37, from http://www.vtt.fi/rte/projects/annex37/Index.htm, accessed on 2011-06-15. [9] Poredoš, A., Kitanovski, A. (2002). Exergy loss as a basis for the price of thermal energy. Energy Conversion and Management, vol. 43, no. 16, p. 21632173, DOI:10.1016/S0196-8904(01)00156-X. [10] Poredoš, A. (2000). The energy efficiency of district cooling for space conditioning. Strojniški vestnik Journal of Mechanical Engineering, vol. 46, no. 8, p. 557-563. [11] Ljubenko, A., Poredoš A., Zager, M. (2011). Effects of hot-water-pipeline renovation in a district heating system. Strojniški vestnik - Journal of Mechanical Engineering, vol. 57, no. 11, p. 834-847, DOI:10.5545/ sv-jme.2010.227. [12] Krainer, A., Perdan, R., Krainer, G. (2007). Retrofitting of the Slovene ethnographic museum. Bauphysik, vol. 29, no. 5, p. 350-365, DOI:10.1002/bapi.200710045. [13] Olesen, B.W. (2008). Radiant floor cooling systems. ASHRAE Journal, vol. 9, p. 16-22.

4 ACKNOWLEDGMENTS COST Action C24 Analysis and design of innovative systems with LowEx for application in build environment, International Energy Agency; Energy Conservation In Buildings And Community Systems; Annex 49; Laboratory of Building environment, Tokyo City University; Department of Civil Engineering, Technical University of Denmark; University of Ljubljana, Faculty of Civil and Geodetic 460

Dovjak, M. – Shukuya, M. – Krainer, A.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 453-461

[14] Košir, M., Krainer, A., Dovjak, M., Perdan, R., Kristl, Ž. (2010). Alternative to the Conventional Heating and Cooling Systems in Public Buildings. Strojniški vestnik - Journal of Mechanical Engineering, vol. 56, no. 9, p. 575-583. [15] Sakulpipatsin, P. (2008). Exergy Efficient Building Design. Doctoral thesis. Delft University of Technology, Delft. [16] Schata, M., Elixman, J.H., Jorde, W. (1990). Evidence of heating systems in controlling house-dust mites and moulds in the indoor environment. Indoor Air ’90 Conference Proceedings, p. 577-581. [17] Lengweiler, P., Nielsen, P.V., Moser, A. (1997). Deposition and resuspension of Particles. Healthy Buildings Conference Proceedings, p. 501-502. [18] Isawa, K., Komizo T., Shukuya, M. (2003). The relationship between human-body exergy consumption rate and a combination of indoor air temperature and mean radiant temperature. Transactions of Architectural Institute of Japan, no. 570, p. 29-35. [19] Shukuya, M. (2006). Comfortable high-performance and low exergy built environment. The International Energy Agency, Energy Conservation in Buildings and Community Systems Conference Proceedings, p. 893898. [20] Shukuya, M., Saito, M., Isawa, K., Iwamatsu, T., Asada, H. (2010). Human body exergy balance and thermal comfort. Working Report of The International Energy Agency, Energy Conservation in Buildings and Community Systems, Annex 49. Fraunhofer. [21] Shukuya, M., (2009). Exergy concept and its application to the built environment. Building and Environment, vol. 44, no. 7, p. 1545-1550, DOI:10.1016/j. buildenv.2008.06.019. [22] Prek, M. (2004). Exergy analysis of thermal comfort. International Journal of Exergy, vol. 1, no. 2, p. 303315, DOI:10.1504/IJEX.2004.005559. [23] Prek, M., Butala, V. (2010). Principles of exergy analysis of human heat and mass exchange with the indoor environment. International Journal of Heat and Mass Transfer, vol. 53, no. 25/26, p. 5806-5814, DOI:10.1016/j.ijheatmasstransfer.2010.08.003. [24] Simone, A., Kolarik, J., Asada, H., Dovjak, M., Schellen, L., Iwamatsu, T., Shukuya, M., Olesen, B.W. (2011). A relation between calculated human body exergy consumption rate and subjectively assessed thermal sensation. Energy and Buildings, vol. 43, no. 1, p. 1-9, DOI:10.1016/j.enbuild.2010.08.007. [25] Iwamatsu, T., Asada, H. (2009). A calculation tool for human-body exergy balance. The International Energy Agency, Energy Conservation in Buildings and Community Systems Annex 49 Newsletter, no. 6, p. 4-5. Fraunhofer. [26] Tokunaga, K., Shukuya, M. (2011). Human-body exergy balance calculation under un-steady state

conditions. Building and Environment, vol. 46, no. 11, p. 2220-2229, DOI:10.1016/j.buildenv.2011.04.036. [27] Schweiker, M., Shukuya, M. (2012). Adaptive comfort from the viewpoint of human body exergy consumption. Building and Environment, in press, DOI:10.1016/j.buildenv.2011.11.012. [28] Košir, M. (2008). Integrated regulating system of internal environment of the basis of fuzzy logic use. Doctoral thesis. University of Ljubljana, Ljubljana. [29] Trobec-Lah, M. (2003). Harmonisation of thermal and daylight fluxes with fuzzy logic. Doctoral thesis. University of Ljubljana, Ljubljana. [30] Košir, M., Krainer A., Kristl, Ž. (2012). Integral control system of indoor environment in continuously occupied spaces. Automation and Construction, vol. 21, p. 199209, DOI:10.1016/j.autcon.2011.06.004. [31] Kristl, Ž., Košir, M., Trobec Lah, M., Krainer, A. (2008). Fuzzy control system for thermal and visual comfort in building. Renewable Energy, vol. 4, no. 33, p. 694-702, DOI:10.1016/j.renene.2007.03.020. [32] Dovjak, M., Shukuya, M., Olesen, B.W., Krainer, A. (2010). Innovative design of renewable energy technology systems for heating and cooling in sustainable buildings. Renewable Energy, Conference Proceedings, p. 1-4. [33] Dovjak, M., Asada, H., Iwamatsu, T., Shukuya, M., Olesen, B.W., Krainer, A. (2011). Lowex vs. conventional systems: User/building/environment. Exergy, Life Cycle Assessment, and Sustainability 2 Conference Proceedings, p. 1-8 [34] Dovjak, M., Simone, A., Kolarik, J., Asada, H., Iwamatsu, T., Schellen, L., Shukuya, M., Olesen, B.W., Krainer, A. (2011). Exergy analysis: The effect of relative humidity, air temperature and effective clothing insulation on thermal comfort. Exergy, Life Cycle Assessment, and Sustainability 2, Conference Proceedings, p. 1-8. [35] Dovjak, M., Shukuya, M. (2011). Integral control of hospital environment. IEEE, Conference Proceedings, p. 1-4 [36] Dovjak, M., Shukuya, M., Krainer, A. (2011). Towards zero energy buildings with conventional or renewable energy technology systems? International Renewable Energy Conference Proceedings, p. 1-6. [37] Dovjak, M., Shukuya, M., Krainer, A. (2011). Solar heating and cooling system for thermal comfort conditions and lower building energy use. The International Solar Energy Society Conference Proceedings, p. 1-10. [38] Ealiwa, A., Taki, A. H., Howarth, A.T., Seden, M.R. (2001). An investigation into thermal comfort in the summer season of Ghadames, Libya. Building and Environment, vol. 36, no. 2, p. 231-237, DOI:10.1016/ S0360-1323(99)00071-2.

Exergy Analysis of Conventional and Low Exergy Systems for Heating and Cooling of Near Zero Energy Buildings

461

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469 DOI:10.5545/sv-jme.2011.166

Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser

Rabiey, M. – Walter, C. – Kuster, F. – Stirnimann, J. – Pude, F. – Wegener, K. Mohammad Rabiey1,* – Christian Walter2 – Friedrich Kuster2 – Josef Stirnimann1 – Frank Pude1 – Konrad Wegener1,2 1

inspire AG, ETH Zurich, Switzerland 2 IWF, ETH Zurich, Switzerland

A systematic research analysis has been applied to study the effect of dressing parameters on grinding forces, work piece roughness and wheel wear of a hybrid bond CBN grinding wheel. This paper presents some of the results achieved by the comparison between a conventional SiC dressed wheel and a grinding wheel dressed by means of a short-pulse fiber laser. The results show high technological potential for the laser dressing method compared to conventional dressing. Lower grinding forces and specific energy, with relatively the same surface roughness and lower total grinding wear (wear by dressing plus wear by grinding) are the biggest advantages of the laser dressing method over the conventional method. However, the economic aspects of laser dressing (investment on laser source and associated add-ons) at the moment, cannot justify the integration of such systems on the grinding machine for all types of applications. The next challenge is optimization of the laser dressing process to increase the efficiency of the process and expand the possible applications both from a technical and commercial point of view. Keywords: Laser, Dressing, Grinding, Conditioning, CBN, Hybrid bond

0 INTRODUCTION Hybrid bound grinding wheels, as a newly developed bonding system, have great potential in industrial applications since they combine the advantages of both a metal bond (low wear and high-strength) and a vitrified bond (porosity and good chip pockets) [1] and [2]. However, the problem of dressing such highstrength bonds, which is also characteristic of metal bond wheels, has yet to be effectively and efficiently solved. Currently, to condition hybrid bond wheels, a conventional mechanical dressing process with a SiC wheel is used. Laser dressing seems to be an appropriate alternative dressing method. Dressing using a laser can be categorized by two main types of processes. The first is laser dressing whereby the laser beam without using any other dressing tool is applied to the surface of the grinding wheel (tangentially, radially or at an angle) and dresses the wheel [3]. The second method is laser assisted dressing in which the laser beam is applied to the grinding wheel to locally heat up the wheel thereby reducing the strength of the bond and simultaneously removing material by means of a conventional dressing tool [4]. In this research, the first method has been used. 1 LASER DRESSING, STATE OF THE ART The application of laser for dressing of Aluminum Oxid was firstly reported from IIT in 1989. Babu et al [5] and [6] investigated on the effect of Nd: YAG laser on Al2O3 and SiC grains, vitrified bond grinding 462

wheel. The laser beam was applied radially to the grinding wheel. The preliminary study showed craters formed by the laser irradiation which were surrounded by a resolidified layer exhibiting multiple cracks. The damages have a direct relation with laser intensities. Laser technology for conditioning of superabrasives was first proposed by Westkaemper [7] for dressing and truing of resin bond CBN grinding wheels. The results show diamond topographies with sufficient chip clearance compared to conventional dressing methods. Timmer [3] utilized a Nd:YAG laser to dress diamond and CBN grinding wheels of different bond types, both radially and tangentially to the wheel surface. Kang et al [8] conducted a study of truing resin and metal bond diamond wheels by utilizing pulsed Nd:YAG laser radiation tangentially to the grinding wheel surface. It was observed that the resin bond material started to decompose, the bronze bond either melted or vaporized, and the slightly damaged diamond grains were removed due to sputtering effects of the bond material. Yung et al. [9] used an acousto-optical Q-switched Nd:YAG laser to dress a resin bond CBN wheel. A 10 to 15% reduction in grinding force compared to conventional dressing as well as good topography of the wheel without destroying or damaging the CBN grains was reported. Furthermore, the Q-switched laser showed a lower heat accumulation on the wheel surface than continuous wave laser processing. Hosokawa [10] also used Nd:YAG laser radiation and reported successful dressing of metal bond wheels. By precise control of the laser parameters and wheel rotation speed, the appropriate grain protrusion height can be

*Corr. Author’s Address: inspire AG, ETH Zurich, Tannenstrasse 3, 8092 Zurich, Switzerland, rabiey@gmail.com

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

generated. Detectable damage to diamond particles, such as micro-cracks or graphitization was not observed below a critical laser power setting. Kunidea et al. [11] used the third harmonic wavelength of a Nd:YAG laser for conditioning of ultrafine grit resin bond diamond wheels. This resulted in a higher cutting edge density compared to wheels treated using a conventional conditioning method with a cup truer. Chen et al. [12] used an acousto-optical Q-switched Nd:YAG laser for the conditioning of bronze bond diamond wheels. 2 EXPERIMENTAL CONDITIONS Grinding experiments are carried out with a series of hybrid bond CBN grinding wheels (dwheel = 150 mm, b = 25 mm) having an average grain size of 126 µm on a commercial surface grinding machine tool. The work piece is hardened steel type 100Cr6 (100×60×30 mm) with a hardness of 60 HCR.

Fig. 1. Setup and machine tool for grinding and conventional dressing experiments

A dynamometer (Kistler, type: 9256B) is mounted under the clamping system of the work piece to measure the tangential force Ft as well as normal force Fn during the grinding operation. An emulsion (3%) with a pressure of 10 bar is used as a cooling lubricant. Fig. 1 shows the experimental setup for grinding and conventional dressing. 2.1 SiC Dressing Conventional dressing experiments are carried out by means of a vitrified SiC wheel (dwheel = 200 mm, b = 30 mm) with the specification 31C-100-L-11310-V138-12. Table 1 shows the selected dressing parameters used in the experiments. In this case, (+)

means down dressing. The total depth of dressing set on the machine for each dressing experiments is 100 µm. The wear of the grinding wheel and the dressing wheel is measured after each dressing process. The wear of the CBN wheels after this dressing operation is about 30 µm. The wear of the SiC wheel is about 70 µm. Table 1. Dressing parameters for experiments Dressing Parameters Dressing speed ratio Dressing depth of cut Dressing feed rate

+4 10 µm 1000 mm/min

2.2 Laser Dressing A schematic of the setup used for laser dressing experiments is depicted in Fig. 2. The beam source is a commercial Q-switched fiber laser system (11) in a master oscillator-power fiber amplifier (MOPFA) configuration. The system has a maximum output power of P = 50 W at pulse energies of up to ep = 1 mJ, and a pulse width between tp = 125 to 150 ns. The emission wavelength of the laser is λ = 1064 nm. For beam delivery, a monomode fiber (10) is employed, while the light is collimated by outcoupling optics (5) at the fiber exit. Finally, the beam is focused through the focal lens (f = 150 mm of a commercial laser processing head (6). The sample wheel (9) is mounted on a precision rotary axis (1) where the rotary speed is controlled via an external power supply (2). The rotary axis and processing head are integrated into a specially designed three-axis machine tool (3). The focal position can be varied by traveling in z-direction. To position the beam on the sample, the entire C-axis is moved by the CNC-axis in the x-y-plane. The laser ON/OFF and the x-y-z-motion are controlled by the CNC (13) of the machine, while the setting of the laser parameters is realized via an additional PC (12) connected to the laser. Via the gas inlet (7), assist gas can be fed in the process through a nozzle coaxial to the laser beam. In this study, compressed air at a pressure of p = 5 bar is used. The diameter of the nozzle is dnozzle = 1 mm and the focal plane is located 1 mm in front of the nozzle. A digital camera (4) is mounted on the laser head to observe the process through the beam path. Before dressing the actual grinding wheel, the ablation behavior of abrasive grains and bond material is studied on samples of the abrasive layer. For this purpose, the laser system is employed in a different setup according to Fig. 3.

Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser

463

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

process parameters for a selective ablation of the bond material with low damage of the abrasive grains (Fig. 4).

Fig. 2. Setup for the laser dressing experiments (1) rotary axis, (2) power supply, (3) three-axis machine tool, (4) digital camera system, (5) fiber outcoupling optics, (6) laser processing head, (7) assist gas inlet, (8) coaxial nozzle, (9) grinding wheel, (10) laser delivery fiber, (11) laser system, (12) PC and (13) CNC unit

Fig. 4. Principle of laser dressing superabrasive grinding wheels

2.3 Grinding Conditions To compare grinding forces, wear rate and surface roughness of the ground work piece, for both laser dressing and conventional mechanical dressing, the surface grinding parameters indicated in Table 2 are used. All tests are done by down grinding. Table 2. Grinding parameters for experiments Grinding Parameters Grinding speed Depth of cut Feed rate

Fig. 3. Setup for laser ablation study (1) laser system, (2) laser delivery fiber, (3) fiber outcoupling optics, (4) mirrors, (5) x, y scanning unit, (6) abrasive layer sample, (7) manual z-axis, (8) x, y motorized linear stages, (9) PC

The sample (6) is placed on a motorized x-ytable (8) combined with a manual z-axis (7) for focal adjustment. The laser beam is deflected on the sample by a x-y scanning unit with a f-theta lens. The ablation is studied by processing reference areas of the abrasive layer with different laser parameters. For the parameter study, laser pulse energy is varied between ep = 0.1 to 1 mJ, pulse repetition rate between fp = 10 to 100 kHz and scanning speed between vs = 0.5 to 5 m/s.The lowest grinding forces where achieved after laser dressing with an average laser power of 50 W at 50 kHz pulse repetition frequency. The processed areas are analyzed by optical 3D microscopy, scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) to find 464

50 m/s 10 to 300 µm 1 to 6 m/min

The grinding speed is limited to 50 m/s due to the small grinding wheel diameter as well as the maximum revolution constraint by the machine spindle. The surface roughness of the ground work pieces are measured by a Taylor-Hobson FormTalysurf roughness tester. 2.4 Measurement Conditions Sinceselective removal of the bond material results in an increase of the surface roughness of the abrasive layer, roughness parameters are used to characterize material removal. A 3D optical microscope (Alicona Infinite Focus) is utilized to measure the surface topography directly within reference areas of 2×2 mm before and after laser processing. Furthermore, the topography of the dressed wheels is measured with this system directly and using the replica method. Due to the complex surface topography (abrasive grains, bond material and pores), not all surface

Rabiey, M. – Walter, C. – Kuster, F. – Stirnimann, J. – Pude, F. – Wegener, K.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

parameters are suitable for abrasive layer topography (Figs. 5 and 6). In this study, the arithmetical mean height Sa and the core roughness depth Sk of the surface is used, since these parameters provide information about the roughness amplitude and showed good reproducibility during experimentation [13].

Fig. 7. SEM image of the abrasive layer surface after laser processing (tp = 125 ns, ep = 0.6 mJ, fp = 70 kHz, vs = 2.5 m/s)

3 EXPERIMENTAL RESULTS The results of the parameter study confirm that a selective removal of the bond material is possible. Fig. 5. SEM image of the unprocessed abrasive layer

Fig. 6. 3D surface measurement of the laser processed abrasive layer

In the parameter study, the change in roughness values is averaged over five reference areas for each combination of laser parameters that is investigated. Furthermore, SEM imaging is used to analyze the damage of abrasive grains in the processed areas. For the dressing experiments, the surface roughness of the grinding wheel is measured after each dressing operation at six positions of equal distance on the circumference of the grinding wheel. Additionally, the wheels surface is analyzed with a conventional optical microscope.

Fig. 8. Laser pulse impact craters on a CBN abrasive grain (tp = 125 ns, ep = 0.5 mJ, fp = 50 kHz)

However, for pulse energies above ep = 0.5 mJ, sporadic grain damage is observed on the processed samples (Fig. 8). With increasing pulse energy and increasing pulse overlap, grain damage becomes more evident and frequent. Since the material removed is relatively small at low pulse energies, laser dressing experiments are carried out at higher pulse energies up to ep = 1 mJ to generate a sufficient grain protrusion in only

Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser

465

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

Fig. 9. a) Optical microscope image of conventional SiC dressed, and b) laser dressed grinding wheels surface (tp = 150 ns, ep = 1 mJ , fp = 50 kHz, nD = 70 /min)

one dressing pass (Fig. 9). Although increased grain damage on the grinding wheel occurs in this case, grinding experiments do not reflect a negative effect in terms of grinding forces and wheel wear. By comparing the grinding forces of both conventional and laser dressed wheel by grinding the material 100Cr6 (60 HRC), it is possible to evaluate the grinding performance and efficiency of the process. Fig. 10 shows the effect of increasing the depth of cut on the specific tangential and normal grinding forces for conventional and laser dressed wheel with constant feed rate and grinding speed. It can be seen that the grinding forces are lower for the laser dressed CBN wheel compared to the conventional dressed wheel. The lower grinding forces may be due to better chip pockets produced by laser ablation on the wheel

surface. It is also possible that some of the active CBN grains became flatted during conventional dressing because of direct contact with SiC grains. These flatted grains cause an increase in force and generate more heat during the process. An investigation on the grain morphology after the conventional dressing as well as laser dressing is still in progress and will be published in the near future. Fig. 11 demonstrates again the tangential and normal grinding forces, in this case versus feed rate with constant depth of cut and grinding speed. Hereby, the results show lower grinding forces by laser dressed compared to the conventional one. In both Figs. 10 and 11, it is observed that the difference between the normal forces of laser and conventional dressed wheel is higher than the

Fig. 10. Specific grinding forces vs. depth of cut by constant feed rate for laser and conventional (SiC) dressed wheel

466

Rabiey, M. – Walter, C. – Kuster, F. – Stirnimann, J. – Pude, F. – Wegener, K.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

Fig. 11. Specific grinding forces vs. feed rate by constant depth of cut for laser and conventional (SiC) dressed wheel

Fig. 12. Specific grinding forces vs. feed rate by constant material removal rate for laser and conventional (SiC) dressed wheel

Fig. 13. Specific tangential grinding forces vs. specific material removal for laser and conventional (SiC) dressed wheel

tangential forces which proves the more effective micro-cutting process, and lower micro-plowing and micro-rubbing in case of the laser dressed wheel. As the material removal rate for both cases is the same,

the result is a lower specific grinding energy when using the laser dressed wheel. On the one hand, increasing the feed rate at constant depth of cut and increasing the depth of cut

Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser

467

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

at constant feed rate causes an increase in specific grinding forces. On the other hand, the only way to increase the material removal rate is either increasing the depth of cut or feed rate (or both). Fig. 12 shows the grinding forces at constant material removal rate. With increasing feed rate, the depth of cut is reduced, so that a constant material removal rate is achieved. It can be seen that with higher feed rate and lower depth of cut, lower grinding forces are measured. A large difference between the normal forces of conventional and laser dressed wheel is evident, however, only little difference between tangential grinding forces is observed. Fig. 13 illustrates the change in the specific tangential grinding force with increasing specific material removal. These experiments are used to find the specific material removal after which dressing of the wheel may be necessary. It can be seen that the tangential grinding forces of the laser dressed wheel

remain lower than tangential grinding force of the conventional dressed wheel. This shows lower heat generation and good long-term stability of grinding conditions for the laser dressed wheel. Fig. 14 shows the wear of the laser and conventional dressed grinding wheels under the same grinding conditions, where the specific material removal is increased. A higher radial wear of the laser dressed wheel can be seen in comparison to the conventional dressed wheel. The radial wear rate for both dressing methods in the beginning is rather high, while it then stabilizes at an almost constant rate with increasing specific material removal. It is important to mention that by laser dressing, the maximum radial wear of 15 µm (depth of cut by laser) is applied to the CBN grinding wheel. Compared to the 30 µm radial wear by dressing using SiC wheel, the total wear rate of laser dressed wheel (dressing and grinding) is still lower than for conventional dressed wheel.

Fig. 14. Radial wheel wear vs. specific material removal for laser and conventional (SiC) dressed wheel

Fig. 15. Surface roughness of ground work piece vs. specific material removal for laser and conventional (SiC) dressed wheel

468

Rabiey, M. – Walter, C. – Kuster, F. – Stirnimann, J. – Pude, F. – Wegener, K.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 462-469

Considering the results in Figs. 13 and 14, the reason for higher wear in grinding with the laser dressed wheel can be interpreted by a higher self-sharpening effect. However, more research is necessary to confirm this interpretation of the results. Fig. 15 shows a comparison of the surface roughness of the ground work piece for a laser and conventional dressed grinding wheel. It can be seen that the work piece surface roughness Ra with the laser dressed wheel is slightly higher than with the conventional dressed wheel. However, the difference is negligible for most practical applications as it is less than 0.05 µm. Therefore, it can be stated that with laser dressed grinding wheels, it is possible to produce parts and components with comparable surface quality to the parts ground with conventional dressed wheel, while significantly lowering grinding forces. 4 CONCLUSION The potential of laser dressing of the hybrid bond wheel is shown for high productive grinding. The hybrid bond can be dressed by laser without any negative effects on the cutting performance of the CBN grits. Generally, lower grinding forces and consequently lower specific energy and heat generation result from dressing with a laser compared to dressing with SiC. For laser dressing, no wear of the dressing tool has to be considered. These are the most important advantages of the laser dressing method. However, a higher wear of the grinding wheel by grinding and higher surface roughness of ground surface should be taken into account. A reasonable efficiency as well as an effective and reliable laser dressing process can be achieved using a proper process parameter set. More investigations need to be done to study the effect of different laser parameters on the efficiency and quality of the dressing process. The economic aspect of using laser dressing in terms of capital investment of the apparatus and dressing time have to be precisely evaluated. This is the next step which is under investigation. The advantages of the laser method must justify the increase in capital investment and related technical matter of integration of the laser system into a grinding machine. 5 ACKNOWLEDGMENT The authors wish to gratefully acknowledge the financial support which was granted by the Swiss Innovation Promotion Agency as well as the

technical support provided by Meister Abrasives AG (Andelfingen), Mägerler AG (Fehraltdorf), and Lasag AG (Thun), all located in Switzerland. 6 REFERENCES [1] Beyer, P. (2005). HPB technology for Vit CBN grinding tools. Industrial Diamond Review, vol. 1, p. 46-48. [2] Beyer, P. (2005). vDD technology for Vitrified bond diamond dressers. Industrial Diamond Review, vol. 2, p. 34-39. [3] Timmer, J.H. (2001). Laserkonditionieren von CBNund Diamant-Schleifscheiben. Doctoral Dissertation, Braunschweig University, Braunschweig. [4] Jackson, M.J., Robinson, G.M., Dahotre, N.B., Khangar, A., Moss, R. (2003). Laser dressing of vitrified aluminium oxide grinding wheels. British Ceramic Transactions, vol. 102, no. 6, p. 237-245, DOI:10.1179/096797803225009346. [5] Babu, N.R., Radhakrishnan, V., Murti, Y.V. (1989). Investigations on laser dressing of grinding wheels-Part I: Preliminary study, Transaction of ASME, vol. 111, p. 244-252, DOI:10.1115/1.3188756. [6] Babu, N.R., Radhakrishnan, V. (1989). Investigations on laser dressing of grinding wheels—Part II: Grinding performance of a laser dressed aluminum oxide wheel. Transaction of ASME, vol. 111, p. 253-261, DOI:10.1115/1.3188757. [7] Westkaemper, E. (1994). Grinding assisted by Nd:YAG Laser. Annals of the CIRP, vol. 44, p. 317-320, DOI:10.1016/S0007-8506(07)62333-6. [8] Kang, R.K., Yuan, J.T., Zhang, Y.P., Ren, J.X. (2001). Truing of diamond wheels by Laser. Key Engineering Materials, vol. 202-203, p. 137-142, DOI:10.4028/ www.scientific.net/KEM.202-203.137. [9] Yung, K.C., Chen, G.Y., Li, L.J. (2003). The laser dressing of resin-bonded CBN wheels by a Q-switched Nd:YAG laser. International Journal of advanced Manufacturing, vol. 22, p. 541-546. [10] Hosokawa, A., Ueda, T., Yunoki, T. (2006). Laser dressing of metal bonded diamond wheel. Annals of the CIRP, vol. 55, no. 1, p. 329-332, DOI:10.1016/S00078506(07)60428-4. [11] Kunidea, Y., Matsuura, H., Kodama, S., Yoshihara, N., Yan, J., Kuriyagawa, T. (2006). Development of a new laser conditioning method for ultra-fine grit diamond wheels. Key Engineering Materials, vol. 329, p. 175180, DOI:10.4028/www.scientific.net/KEM.329.175. [12] Chen, G., Mei, L., Zhang, B., Yu, C., Shun, K. (2010). Experiment and numerical simulation study on laser truing and dressing of bronze-bonded diamond wheel. Optics and Lasers in Engineering, vol. 48, no. 3, p. 295-304, DOI:10.1016/j.optlaseng.2009.11.006. [13] Blunt, L., Jiang, X. (2003). Advanced techniques for assessment surface topography. Elsevier Science & Technology, p. 19-62.

Dressing of Hybrid Bond CBN Wheels Using Short-Pulse Fiber Laser

469

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481 DOI:10.5545/sv-jme.2011.170

Boundary Layer Method for Unsteady Transonic Flow Majić, F. – Voss, R. – Virag, Z. Frane Majić1,* – Ralph Voss2 – Zdravko Virag1 1 Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, Croatia 2 Institute for Aeroelasticity, German Aerospace Center, Germany

A numerical method for determination of unsteady loads in a 2-D transonic flow, with the occurrence of a shock wave, on a pitching airfoil is demonstrated. The method implements the Euler equations for inviscid region and integral boundary layer equations for the viscous region near the airfoil. The viscous-inviscid interaction method is employed using the transpiration velocity concept on the airfoil contour. The Euler solution is calculated by using the Van Leer flux-vector splitting method on the body-fitted C-grid. The boundary layer model is calculated applying Drela’s model of integral boundary layer equations for the laminar and turbulent flow. The transition from the laminar to the turbulent flow is predicted by the en method. The viscous-inviscid interaction method is made in the direct mode. The results obtained by this method are comparable with the calculated RANS and experimental results, while time and computational costs were lower than for RANS calculations. Generally, the pressure coefficient distribution results showed good agreement with the RANS and experimental results. The method predicted the position of a shock wave to be slightly shifted towards the leading edge of the airfoil with respect to the position obtained by the RANS and experimental results. This indicates that the boundary layer model has a strong influence on the inviscid part of the flow. Keywords: unsteady transonic flow, viscous-inviscid coupling, airfoil, transpiration velocity, transition prediction

0 INTRODUCTION The phenomenon of aircraft flutter, which has to be investigated for each new aircraft design or structural modification of existing aircraft, is still one of the important research topics in aeroelasticity, particularly in transonic speeds. This phenomenon is an aeroelastic problem determined by the interaction of the elastic, damping and inertial forces of the structure and the unsteady aerodynamic forces generated by oscillatory motion of the structure itself. Such oscillatory motion can lead to a progressive increase in the amplitude of vibration, ending in the disintegration of the structure. For a given configuration of an aircraft structure the unsteady aerodynamic forces increase rapidly with the flight speed, while the elastic, damping and inertia forces remain unchanged. Because of that there exists a critical flight speed (flutter speed) above which flutter occurs. The flutter speed represents a critical speed at which the structure sustains oscillations following some initial disturbance. Below this speed the oscillations are damped, whereas above it one of the modes becomes negatively damped and unstable oscillations occur unless some form of nonlinearity bounds the motion [1]. The occurrence of shock waves on the aircraft aerodynamic surfaces can cause a drop in the flutter boundary in the range of transonic speed. This drop is called transonic dip shown in Fig. 1. The important feature of the transonic dip is the bottom of the dip, which defines the minimum flutter speed at which flutter can occur across the flight envelope of the aircraft. The analysis of flutter by linear aerodynamic methods typically predicts the flutter boundary 470

adequately at subsonic and supersonic speeds, but in the transonic speed range it predicts a higher flutter speed than the experiment [2]. The flutter boundary could be obtained by an inviscid unsteady aerodynamics analysis, e.g. by solving the unsteady transonic small disturbance potential flow, full potential flow, or Euler equations of motions. Although these methods have a capability of capturing shock waves in the flow and transonic dip, they predict a significantly lower flutter speed at the bottom of the transonic dip because they do not involve viscous effects in the calculations. Viscous effects, which act in the form of significant boundary layer thickening, and shock-induced flow separation are responsible for defining the bottom of the transonic dip better.

Fig. 1. Transonic dip

For the flutter analysis, the arbitrary motion of the airfoil is not so often used but the harmonic motion for a single oscillation frequency is of interest. The

*Corr. Author’s Address: Faculty of Mechanical Engineering and Naval Architecture, Ivana Lučića 5, Zagreb, Croatia, frane.majic@fsb.hr

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

objective of such an analysis is to determine the flight conditions that correspond to the flutter boundary (stability boundary), for which one of the modes of motion has a simple harmonic time dependence [3]. In the linear flutter analysis it is assumed that the solution involves simple harmonic motion and also that the excitation force and moment exhibit harmonic behaviour. With this assumption, the equations of motion are then cast into the eigenvalue problem in the frequency domain and are solved for complex eigenvalues. From these eigenvalues, conclusions about stable or unstable oscillations of the airfoil can be drawn. Such flutter analysis cannot provide any definitive measure of flutter stability other than the location of the stability boundary. Despite this weakness of the method, its primary strength is that it needs only the unsteady airloads for a simple harmonic motion of the airfoil. The doublet-lattice method [4] is still present in the actual design analysis because of low computer time consumption and a simple setting procedure of a computational problem. One of the drawbacks of the method is the inability of capturing strong shocks in transonic flows. The Reynolds Averaged Navier Stokes (RANS) simulation for flutter analysis gives much more accurate results, but it uses a large amount of computational time and hence is not the first choice for preliminary design. In addition, RANS needs large grids with high resolution and the problem setting is much more demanding. RANS is also limited with uncertainties in turbulence modelling, difficulties in high quality grid generation [5] and difficulties with the grid deformation algorithm in unsteady flows [6]. In aeroelastic applications where a large number of parameters such as different natural modes, angles of attack, Mach numbers, frequency, etc. must be investigated, methods that solve the unsteady aerodynamic problem in the frequency domain are introduced. These methods are especially suitable for simulations at low reduced frequencies. Recently, numerical methods based on such an alternative approach, namely on the solution to small disturbance Euler equations (SDE) and the linearization of NavierStokes equations, have been presented [7] and [8]. Some papers that analyze the coupling of RANS equations with the boundary layer have been published in recent years, [9] to [11]. These papers have demonstrated the prediction of transition region with the aim to design laminar airfoils with reduced drag. Viscous-inviscid interaction methods, such as the Euler method with viscous boundary layer correction, are a good compromise between the two

methods mentioned above. Euler methods are capable of resolving strong shocks and with the boundary layer coupling they are a good balance between the flow model and the computational efficiency [12]. The viscous-inviscid interaction methods give results comparable to the RANS solvers, but the computer time is several times shorter and this gives a considerable advantage to the fast flutter analysis in the design process [13]. This study is dedicated to the improvement of the viscous-inviscid interaction method with the unsteady Euler as an inviscid solver and a solver of integral boundary-layer equations for the thin viscous region, with interaction by transpiration velocity concept. 1 NUMERICAL METHOD 1.1 Inviscid Model Inviscid model employs two-dimensional Euler equations for an ideal gas. The equations are transformed to a moving body-fitted coordinate system (ξ, η, τ) and are given in a conservative form by:  ∂F  ∂G  ∂Q = 0, + + ∂τ ∂ξ ∂η

(1)

where:

 = JQ, Q  = (− y x + x y )Q + y F − x G , F η τ η τ η η

(2)

 = (− x y + y x )Q − y F − x G. G ξ τ ξ τ ξ ξ Vectors Q, F and G represent the vector of conservative variables, the fluxes in the Cartesian xand y-coordinates, respectively, as follows: ρ  ρu   ρv   ρu   ρu 2 + p    , G =  ρ vu  , (3) Q =  , F =   ρv   ρ uv   ρ v2 + p         ρe  ρ uh   ρ vh  where ρ is the fluid density, p is the pressure, while u and v are the Cartesian x and y velocity components, respectively. In the vectors defined by expressions in Eq. (3), e is the specific total energy (per unit mass):

1 p 1 2 + (u + v2 ) , γ −1 ρ 2

(4)

and h is the specific total enthalpy (per unit mass):

Boundary Layer Method for Unsteady Transonic Flow

471

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

γ p 1 2 2 + (u + v ). γ −1 ρ 2

(5)

In Eq. (2) and in the equations which follow, the subscripts ξ, η and τ represent the derivatives of the physical coordinates with respect to the body-fitted coordinates. J = xξ · yη – yξ · xη is the Jacobian of the transformation. The inviscid model employs the Van Leer flux vector splitting [14]. Correct splitting of the  and transformed flux vectors is done by rewriting F  G as a product of local rotation matrices (TF and TG) and modified flux vectors F and G , respectively, which have the same form as the Cartesian flux vectors but contain the transformed instead of Cartesian velocities [15] and [16]. Rewritten flux vectors are then expressed as follows:  (Q  ) = x 2 + y 2 T F (Q ), F η η F

 (Q  ) = x 2 + y 2 T G (Q ), G ξ ξ G

(6)

where the local rotation matrices TF and TG are equal to:  1   xτ TF =  y τ   x2 + y 2 τ  τ  2  1   xτ TG =  y τ   x2 + y 2 τ  τ  2

0 y

0 x η

− x η

y η

y x − x η y τ η τ

x η xτ + yη yτ

0 x ξ

0 − y

y ξ

x ξ

x ξ xτ + yξ yτ

x ξ yτ − yξ xτ

0  0  0  , (7)  1  0  0  0  . (8)  1 

The modified flux vectors F and G have the following form according to Eq. (6):

 vG   uF   u v   2   G G  u 2 + a  F = ρ  F γ  , G = ρ  2 a2  . vG +    γ    u F vF   v h   u h  F F  G G  Transformed velocities in F are equal to: uF = yη (u − xτ ) − x η (v − yτ ), vF = x η (u − xτ ) + yη (v − yτ ),

vG = − yξ (u − xτ ) + x ξ (v − yτ ).

(11)

The terms x η , yη , x ξ and yξ are normalized as follows: x η =

x ξ =

xη 2 η

2 η

x +y xξ

xξ2 + yξ2

y = η

y = ξ

xη 2 η

x + yη2 yξ xξ2 + yξ2

(12)

The modified total enthalpies hF and hG have the same form as in the Cartesian components, but now with transformed velocities. Splitting of the modified flux vectors can now be performed in the same fashion as in Cartesian coordinates, but in terms of the Mach numbers Maξ = uF a and Maη = vG a . The flux vectors are split in such a way that the Jacobian matrices ∂F + ∂Q and ∂G + ∂Q have only positive eigenvalues and the Jacobian matrices ∂F − ∂Q and ∂G − ∂Q have only negative ones. The split fluxes have the following form: 2 ρa   f1± = ± 1 ± Maξ ) (   4    f ± = a ( γ − 1) Ma ± 2  f ±  ξ  1   2 γ  F± =   , (13) f 3± = vF f1±     ± 2 2 2 f 2 ) vF ±  (  ± γ f1  +  f4 = ± 2 2 ( γ 2 − 1) f1   2 ρ a  g1± = ± (1 ± Maη )   4   g 2± = uG g1±    ± a  ± ± G =  g3 = ( γ − 1) Maη ± 2  g1  . (14) γ     2 ± 2 g3 ) uG ±  (  ± γ2 + g1   g4 = ± 2 2 ( γ 2 − 1) g1  

(9) 1.2 Solution Method Euler equations in the body-fitted coordinates, with flux vector splitting, are now given as: (10)

and in the G flux to: 472

uG = x ξ (u − xτ ) + yξ (v − yτ ),

Majić, F. – Voss, R. – Virag, Z.

 ∂F  + ∂F  − ∂G  + ∂G − ∂Q + + + + = 0, ∂τ ∂ξ ∂ξ ∂η ∂η

(15)

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

 + and G  – are the split fluxes. Eq.  +, F  –, G where F (15) is explicitly discretized and solved as shown in the following equation:

vector by backward or forward formulae depending on which flux is concerned. A general formula for calculating split fluxes is as follows:

 n +1 (i, j ) = Q  n (i, j ) − ∆τ  F  + (i + 1 2 , j ) − Q   −F

( i − 1 2 , j ) + F ( i + 1 2 , j ) −

−

 −F

−

( i − 1 2 , j ) + G ( i, j + 1 2 ) −

 −G

(16)

−

where superscripts n+1 and n represent the new and old time steps, respectively. Indices i, j represent the control volume centers along ξ and η axes, respectively, while indices i+1/2, i-1/2, j+1/2, j-1/2 represent the control volume faces as depicted in Fig. 2. Δτ represents the time increment obtained from the Courant-Friedrichs-Lewy condition (CFL):

CFL =| λ |max

 ( i, j + 1 2 ) = G  Q∓ , m G  i, j + 1 i, j + 1  ,  2 2  ±

( i, j − 1 2 ) + G ( i, j + 1 2 ) − G ( i, j − 1 2 ) −

 ± (i + 1 2 , j ) = F ±  Q∓ , m F  i+ 1, j i+ 1, j  ,  2 2  ± ±   (i − 1 2 , j ) = F Q∓ , m F  i− 1, j i− 1, j  ,  2 2 

∆τ . CVmin

(17)

In Eq. (17), CVmin represents the minimal control volume linear dimension and λ represents the eigenvalues which for a one-dimensional case are λ1 = u + a, λ2 = u and λ3 = u – a.

(18)

 ± ( i, j − 1 2 ) = G ±  Q∓ , m G  i, j − 1 i, j − 1  .  2 2  The term m represents all geometric terms involved in the transformation to the body-fitted coordinates. The extrapolated values of the solution vector are determined with the second order accuracy formulas: Q − 1 = Qi , j + 0.5 ( Qi , j − Qi −1, j ) ,

i+ , j 2

Q + 1 = Qi +1, j + 0.5 ( Qi +1, j − Qi + 2, j ) .

(19)

i+ , j 2

The same formulas are valid for the faces with indices j – 1 and j + 1. 1.3 Boundary Conditions The boundary condition on the airfoil is imposed by setting the normal relative velocity to zero:     (20) ( v − vb − vt ) ⋅ n = 0,    where v , vb and v t are the fluid velocity, the prescribed velocity of airfoil contour and the transpiration velocity, respectively. The transpiration velocity represents the boundary layer effect of growing displacement thickness [17]. This is the way how the boundary layer model is coupled to the inviscid model. The transpiration velocity is defined as:

Fig. 2. Control volume interfaces

The difference between two neighbouring grid lines in the body-fitted coordinates is taken to be unity. The spatial derivatives are approximated by MUSCL differencing (MUSCL - Monotone Upstream-centered Schemes for Conservation Laws), where fluxes are calculated indirectly by extrapolating the solution

vt =

d ( ue δ ∗ ) , ds

(21)

where δ* is the boundary layer displacement thickness and ue is the velocity at the boundary layer edge. The curvilinear coordinate s goes from the stagnation point over the upper and the lower airfoil contour towards the trailing edge. The pressure is determined from the normal component of momentum equation, which is derived from Eq. (20) in the following form:

Boundary Layer Method for Unsteady Transonic Flow

473

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

    Dv  D ( vb + v t )     Dn ⋅n − ⋅ n + ( v − vb − vt ) ⋅ = 0. (22) Dt Dt Dt Assuming that the airfoil contour coincides with the line η = const. and the lines ξ = const. are perpendicular to the airfoil contour, the transformation of Eq. (22) into the body-fitted coordinates yields Eq. (23):

2   u + ρ  J  2 G 2 ( yξ xξξ − xξ yξξ ) +   xξ + yξ

 (23) 2uG + − + − y x x y x y y x ( ξ τξ ξ τξ ) ττ ξ ττ ξ  + xξ2 + yξ2  +  yξ vtxξ − xξ vtyξ   yη ( u − xτ ) − xη ( v − yτ )  +

(

)}

⋅

The terms vtx and vty in Eq. (23) are transpiration velocitiy components in the x and y directions of nonmoving Cartesian coordinates. Variables ξ, η and τ in subscripts represent derivatives with respect to these coordinates. Double subscripts ξξ, τξ or ττ represent second derivatives with respect to these coordinates. The characteristic boundary condition is used on the far field of computational domain. The problem is locally regarded as one-dimensional, i.e. derivatives along the boundary can be neglected (∂( )/∂ξ → 0) and according to [18] the following characteristic condition can be constructed:

D ± ( R ) = 0, Dt

(24)

where R± are the Riemman invariants:

R ± = vnorm ±

2a , γ −1

(25)

where a is the local speed of sound and vnorm the local velocity perpendicular to the far-field boundary. The characteristic equations are used to update the variables on the outer boundary at a new time level. For the two-dimensional case, four primitive variables are concerned and therefore four independent equations are needed. For the subsonic inlet far-field boundary condition, where vnorm < 0, the following expressions are valid: 474

R + = R + (∞), R − = R − (F), vtang = vtang (∞), pT = pT (∞).

R + = R + (F), R − = R − (∞), vtang = vtang (F), pT = pT (F).

(27)

The symbol F means that variables are extrapolated locally from the interior field values, and the symbol ∞ means that variables are calculated from the far-field representation.

∂p 2 ( xξ + yξ2 ) = ∂∂ξp ( xξ xη + yξ yη ) + ∂η

+ J vtxτ yξ − vtyτ xξ

Velocity vtang is the velocity along the far-field boundary and pT is the total pressure. For the subsonic outflow condition at the far-field boundary, where vnorm > 0, the following expressions are valid:

(26)

1.4 Viscous Model The method decouples the inviscid region surrounding the airfoil from the thin viscous region close to the airfoil. The viscous region is evaluated according to Drela’s method of integral boundary layer equations [19], which are the integral momentum equation: dθ Cf θ due = − ( H + 2 − Mae2 ) , ue ds 2 ds

(28)

and the integral kinetic energy equation, also known as the shape parameter equation:  H ∗ due dH ∗ 2CD H ∗ Cf  2 H ∗∗ = − − +1− H  , (29) ∗ ds θ θ 2  H  ue ds where θ is the momentum thickness, Cf the friction coefficient, H the shape parameter, Mae the Mach number at the boundary layer edge, ue the velocity at the boundary layer edge, s is the coordinate originating from the stagnation point and going over upper and the lower airfoil contour towards the trailing edge, H* the kinetic energy shape parameter, CD the dissipation coefficient, and H** is the density shape parameter. The subscript e in Eqs. (28), (29) and all subsequent equations represents the variables for the boundary layer edge. The momentum and shape parameter equations are valid for both the laminar and turbulent boundary layers. These equations contain more than two dependent variables and hence some assumptions about the additional unknowns will have to be made. If θ and δ* are defined as two dependent variables, then four additional closure equations are needed for additional unknown variables Cf, CD, H* and H**. The additional closure equations for the laminar and the turbulent flow are defined as in [20] to [22]. The boundary layer Eqs. (28) and (29) employed in this paper were solved by the fourth order Runge-

Majić, F. – Voss, R. – Virag, Z.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Kutta method. The input values to the boundary layer equations are the fluid velocity ue(s) and the Mach number Mae(s) distribution at the edge of the boundary layer, which are functions of distance from the stagnation point along the airfoil contour. These distributions are the output of the inviscid part of flow, taken at the position of the airfoil contour. The computational grid for boundary layer calculations was one-dimensional with the same number of main nodes as that of control volumes in the inviscid solver bounding the airfoil contour. Between these main nodes, integration was performed on twenty subintervals. The integration of the boundary layer equations starts from the initial solution for the flat plate in laminar flow. The boundary layer initial solution was obtained from the Blasius solution [23] and [24]. The method for the determination of the onset of transition is derived from the spatial amplification theory based on the Orr-Sommerfeld equation [25]. This method is also known as the en method. The growth of these disturbances is responsible for the onset of transition in the boundary layers. The method determines the amplitude of disturbances by the integration of disturbance growth rate from the point of instability. The transition occurs when the amplitude grows by more than a factor en = e9. The exponent n can be different from 9, actually it can vary between 7 and 11 depending mainly on free stream turbulence and surface roughness [21]. In [21], the equation for the amplification ratio n is derived as follows:

m( H ) +1 dn dn 1 l ( H ) , (30) ( H ,θ ) = (H ) θ ds dReθ 2

where: dn = 0.01× dReθ ×

{2.4H − 3.7 + 2.5 tanh 1.5 ( H − 3.1)}

+ 0.25 , (31)

2  1 ( H − 4) s du e  m(H ) = = 0.058 , (32) − 0.068 ue ds  H −1  l ( H )

l (H ) =

ρe ueθ 2 6.45 H − 14.07 = . µe s H2

(33)

The amplification ratio n is a function of coordinate s, and Eq. (30) is integrated downstream from the point of instability scr:

n(s) =

∫ ds ds.

(34)

scr

At the position of instability scr, the Reynolds number referenced by the momentum thickness Reθ is equal to its critical value Reθ = Reθ0 . This critical value can be calculated from the following expression:  1.415   20  log10 Reθ0 =  − 0.489  tanh  − 12.9  + − − 1 H H 1     (35) 3.295 + + 0.440. H −1 The integration of Eq. (30) is completed when the amplification ratio reaches the value of n = 9, and then the turbulent formulation of boundary layer equations is active. The changeover to the turbulent flow is made suddenly without gradual transition. The way of changeover from the laminar to the turbulent flow has a minimal effect on the development of the boundary layer [21]. 1.5 Viscous-Inviscid Coupling The viscous-inviscid coupling between the boundary layer and the Euler equations is made by the transpiration velocity concept or equivalent sources concept, as proposed by Lighthill [26]. The transpiration velocity changes the slope of the net velocity at the boundary and in that way it represents the displacement thickness of the boundary layer and the influence of the boundary layer on the inviscid flow outside the boundary layer. In this study, the viscous-inviscid coupling is made in the direct mode. A scheme of direct mode coupling is shown in Fig. 3. In such an approach the output from the inviscid solver, which is the velocity or pressure at the boundary layer edge, is used as the input in the viscous solver of boundary layer equations. The output from the viscous solver is the displacement thickness, or the transpiration velocity derived from the displacement thickness, which is then used as the input in the inviscid solver to update the boundary condition at the airfoil contour. One of the advantages of such coupling method is its speed and simplicity in application. The disadvantage of the direct coupling is the inability to simulate separated flows because of the appearance of a singularity in the boundary layer equations called Goldstein’s singularity [27].

Boundary Layer Method for Unsteady Transonic Flow

475

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Fig. 3. The direct method scheme of viscous-inviscid coupling

The coupling method used in this study resulted in strong solution oscillation in the near separation test cases and at the position of a sudden increase in the boundary layer thickness. To reduce such oscillatory behavior of the solution and to reach a monotone converged solution, the under-relaxation method was employed. Under-relaxation is performed on the transpiration velocity by the following expression:

vt = vto + β ( vtn − vto ) .

(36)

The superscripts o and n represent the old and the new solution to the transpiration velocity magnitude in two successive iterations of viscous-inviscid coupling, respectively. β represents the under-relaxation factor and its value is smaller than one. In the test cases of near separation, which are the most difficult cases for such methods, the under-relaxation factor adopts very small values around 0.001. The left-hand side of Eq. (36) is the resulting transpiration velocity magnitude and it serves as the old solution in the subsequent iteration. 2 RESULTS AND DISCUSSION Results for the unsteady test case with the appearance of a shock wave are presented. The test case is calculated with a NACA64A010 airfoil. In this case, the airfoil performs harmonic pitch motion. The results are compared with the experimental results from AGARD reports [28], [29] and with the computed RANS results of DLR in-house Tau code (DLR - Deutsches Zentrum für Luft- und Raumfahrt). The RANS results are obtained by an unstructured grid with 50,000 control volumes. A linearised explicit algebraic (LEA) k - ω turbulence model was applied in the RANS calculations. A second order central difference scheme with scalar dissipation as spatial discretization was used. A dual time-stepping with the Runge-Kutta method was employed. Unsteady RANS calculations were performed with a full turbulent 476

flow field without the limitation of turbulence in the laminar part of boundary layer. The initial value of turbulent-to-laminar viscosity-ratio is prescribed for the whole flow region and also for the free stream. The turbulent viscosity ratio was equal to the value of much less than unity (ν t / ν  1) . In order to show the grid independence, the steady Euler results for the normal force coefficient on the NACA0012 airfoil are shown in Fig. 4. The results are calculated for the Mach number Ma = 0.77, and the angle of attack α = 5°, and are presented for three computational grids with a different number of control volumes. The coarsest grid has 4,800 volumes, followed by a finer grid with 9,600 volumes, which is twice as many as the coarsest grid. The finest grid has 19,200 volumes, four times as many as the coarsest grid. The final results for the normal force coefficient are shown in Table 1. Table 1. Normal force coefficient for different grids Grid 4,800 volumes 9,600 volumes 19,200 volumes

cn 1.02849 1.03389 1.04113

Compared with the finest grid (19,200 volumes), the normal force coefficient for the coarsest grid (4,800 volumes) has a difference of 1.2% and medium fine grid (9,600 volumes) has a difference of 0.7%. According to these results and to the fact that such a computational method is aimed for the preliminary phase of aircraft design where many configurations have to be tested, the medium fine grid with 9,600 volumes is selected for further calculations. The CPU time required for the RANS and viscousinvisid coupling calculations differed considerably. In order to show this, steady calculations were performed for the NACA0012 airfoil test case at the Mach number Ma = 0.756, the Reynolds number Re = 4·106 and the angle of attack α = 0°. The RANS calculation was performed by a grid with 50,000 control volumes and the viscous-inviscid calculation by a grid with 9,600 control volumes. Both were performed on a computer with a processor clock speed of 2.4 GHz and 4 GB of random access memory (RAM). The viscousinviscid calculation required 47 seconds while the RANS calculation required 12 minutes and 12 seconds of CPU time. In the preliminary design process where many configurations have to be investigated, this advantage becomes obvious.

Majić, F. – Voss, R. – Virag, Z.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Fig. 4. Solution convergence for different grid size

Fig. 5. NACA64A010 pressure coefficient distribution at phase angle ϕ = 45˚

2.1 NACA64A010 Airfoil Results Table 2 shows the airfoil motion and free stream characteristics for the NACA64A010 test case. The airfoil performed harmonic pitch motion about the axis at a distance of xα / c = 0.239 from the leading edge, where xα is the distance of rotational axis from the airfoil leading edge along the airfoil chord length c. The reduced frequency in Table 2 is defined as follows: ωc (37) ω∗ = . U∞ where ω is the pitch frequency. The boundary layer grid contained 90 main nodes on the airfoil contour. All unsteady calculations were started with the initial condition as a free stream condition. The calculations were performed in five periods, in which the results showed converged periodic solutions. Figs. 5 to 12 present the unsteady pressure coefficient results for the pitching motion according to Table 2.

Fig. 6. NACA64A010 pressure coefficient distribution at phase angle ϕ = 90˚

Table 2. NACA64A010 unsteady test case Mach number (Ma) Reynolds number (Re)

0.797 12.4 ×106

Mean angle of attack (αm) Pitch amplitude (αo)

Reduced frequency (ω

–0.08˚ 2.00˚

Rotational axis position (xα / c)

0.202 0.239

The calculated Euler+BL results are represented by full and dashed lines for the lower and the upper airfoil side respectively. The numerical results for the unsteady RANS (URANS) are represented by triangles pointing up and down for the upper and

Fig. 7. NACA64A010 pressure coefficient distribution at phase angle ϕ = 135˚

the lower airfoil side respectively. The experimental data are represented by full and empty circles for the lower and the upper side respectively. The results are

Boundary Layer Method for Unsteady Transonic Flow

477

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Fig. 8. NACA64A010 pressure coefficient distribution at phase angle ϕ = 180˚

Fig. 11. NACA64A010 pressure coefficient distribution at phase angle ϕ = 315˚

Fig. 9. NACA64A010 pressure coefficient distribution at phase angle ϕ = 225˚

Fig. 12. NACA64A010 pressure coefficient distribution at phase angle ϕ =360˚

Fig. 10. NACA64A010 pressure coefficient distribution at phase angle ϕ = 270˚

presented for phase angles 45, 90, 135, 180, 225, 270, 315 and 360° in the last (fifth) period of simulation where the solution showed periodicity. 478

The Euler+BL method results were in moderate agreement with experimental data at the majority of phase angles, while with the numerical URANS results they showed good agreement at all phase angles. Compared with the experimental data, the Euler+BL results showed slight underprediction of pressure coefficient on the front part of airfoil in front of the shock wave. This shift appears equally on both airfoil sides and therefore should have no influence on the lift force magnitude. At the rear part of airfoil behind the shock wave, the calculated pressure coefficient is in good agreement with the experimental data. The Shock wave position is mostly well predicted. The strength of a shock wave is better predicted at greater angles of attack, where the pressure drop is bigger, than at smaller angles of attack. The peak and the slope of pressure coefficient at the shock wave position are underpredicted at several phase angles.

Majić, F. – Voss, R. – Virag, Z.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Such underprediction happens at the airfoil side with smaller suction than that on the upper side of airfoil in Fig. 7. In such a case, the intensity, position and slope between the calculated and the measured results are not in full agreement. The calculated pressure jump across the shock wave is rather smeared compared with the experimental data. This can indicate that the boundary layer has an impact on the shock wave intensity that is too strong. Due to the presence of boundary layer thickening in the foot of the shock wave, a lambda shaped compression shock appears. As a consequence, the shock wave intensity is reduced and this can be seen in the pressure coefficient distribution on the airfoil contour [30] and [31]. Nearly the same shift in pressure coefficient on the upper and the lower surface can be observed. This could be a consequence of unequal parameters of experimental and numerical calculations, i.e. noncorrected experimental data for the influence of the wind tunnel walls, which is commented on in the AGARD report [29]. Figs. 5 to 12 also show the results from the URANS calculations. The calculated position and intensity of the shock wave for the Euler+BL and the URANS results are in very good agreement at the majority of phase angles. At some phase angles, the shock wave intensity for the Euler+BL results is slightly smeared in comparison with the URANS results. 3 CONCLUSION In this study, a simple and accurate method for unsteady aerodynamic load prediction is developed. The method has produced results which are generally in good agreement with experimental and RANS results. The method has resulted in decreased shock intensity and smeared pressure jump across the shock wave at smaller angles of attack in comparison with the experimental data. The calculated pressure coefficient on the part of the airfoil in front of the shock wave showed a slight shift on both sides with respect to the experimental data. However, it should have no effect on the normal force magnitude since the shift is nearly equal on both sides. The method has showed the oscillatory behavior of pressure coefficient in the vicinity of a strong shock and the trailing edge and such oscillations may cause solution divergence. Therefore, under-relaxation is used, which, on the other hand, can require a greater number of iterations.

The boundary layer inclusion in the unsteady Euler method resulted in a more accurate method for the determination of unsteady aerodynamic loads. The method calculated results with nearly the same accuracy as a higher mathematical model like RANS, while the computational time is shorter and hardware requirements are substantially less demanding. 4 NOMENCLATURE a Speed of sound [ms-1] CD Dissipation coefficient [-] Cf Friction coefficient [-] c Airfoil chord length [m] e Total energy per unit mass [Jkg-1] F, G Flux vectors of conservative variables f, g Members of flux vectors F, G H Shape parameter [-] H* Kinetic energy shape parameter [-] H** Density shape parameter [-] Hk Kinetic shape parameter [-] h Total enthalpy per unit mass [Jkg-1] i, j Control volume node indices J Determinant of the Jacobian matrix of coordinate transformation [-] Ma Mach number [-] Mae Mach number on the outer boundary layer edge [-] m Geometry terms involved in transformation  n Normal vector to the airfoil contour [-] p Pressure [Pa] Q Vector of conservative variables R± Riemann invariants [ms-1] Re Reynolds number [-] TF Transformation matrix for flux vector F TG Transformation matrix for flux vector G u x-component of fluid velocity [ms-1] ue Velocity magnitude on the outer boundary layer edge [ms-1] U∞ Free stream velocity magnitude [ms-1] v y-component of fluid velocity [ms-1]  v Fluid velocity vector [ms-1]  vb Airfoil contour velocity vector [ms-1]  v t Transpiration velocity vector [ms-1] vto Transpiration velocity magnitude at old time step [ms-1] n vt Transpiration velocity magnitude at new time step [ms-1] x, y Spatial coordinates in physical domain [m] xα Rotational axis distance from the airfoil leading edge [m] α Airfoil angle of attack [°] v Kinematic viscosity [m2s-1]

Boundary Layer Method for Unsteady Transonic Flow

479

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

αm Mean angle of attack [°] αo Pitch amplitude [°] β Under-relaxation factor [-]

γ Ratio of specific heats [-] δ Displacement thickness [m] θ Momentum thickness [m] ξ, η Curvilinear coordinates [m] ρ Density [kgm-3] pT Total pressure [Pa] vt Turbulent kinematic viscosity [m2s-1] t Time in computational domain [s] ϕ Phase angle of airfoil harmonic motion [°] ω Angular frequency [rad-1] ω* Reduced frequency [-] Subscripts/Indices e Outer boundary layer edge designation F Domain near the outer boundary designation ∞ Free stream designation 5 REFERENCES [1] Wright, J.R. (2007). Introduction to Aircraft Aeroelasticity and Loads. John Wiley & Sons Ltd., Chichester. [2] Voss, R. (1988). Über die Ausbreitung akustischer Störungen in transonischen Strömungsfeldern von Tragflügeln. Rept. DFVLR-FB 88-13, Institut für Aeroelastik, Göttingen. [3] Hodges, D.H., Pierce, G.A. (2002). Introduction to Structural Dynamics and Aeroelasticity. Cambridge University Press, Cambridge. [4] Albano, E., Rodden, W.P. (1969). A doublet-lattice method for calculating the lift distributions on oscillating surfaces in subsonic flows. American Institute of Aeronautics and Astronautics Journal, vol. 7, no. 2, p. 279-285, DOI:10.2514/3.5086. [5] Džijan, I., Virag, Z., Kozmar, H. (2007). The influence of grid orthogonality on the convergence of the SIMPLE algorithm for solving Navier-Stokes equations. Strojniški vestnik - Journal of Mechanical Engineering, vol. 53, no. 2, p. 105-113. [6] Schuster, D.M., Liu, D.D., Huttsell, L.J. (2003). Computational aeroelasticity: success, progress, challenge. Journal of Aircraft, vol. 40, no. 5, p. 843856, DOI:10.2514/2.6875. [7] Kreiselmaier, E., Laschka, B. (2000). Small disturbance Euler equations: efficient and accurate tool for unsteady load prediction. Journal of Aircraft, vol. 37, no. 5, p. 770-778, DOI:10.2514/2.2699. [8] Pechloff, A., Laschka, B. (2006). Small disturbance Navier-Stokes method: efficient tool for predicting unsteady air loads. Journal of Aircraft, vol. 43, no. 1, p. 17-29, DOI:10.2514/1.14350. [9] Filippone, A., Sorensen, J.N. (1997). Viscous-inviscid interaction using the Navier-Stokes equations. American

480

Institute of Aeronautics and Astronautics Journal, vol. 35, no. 9, p. 1464-1471, DOI:10.2514/2.269. [10] Stock, H.W., Haase, W. (2000). Navier-Stokes airfoil computations with transition prediction including transitional flow regions. American Institute of Aeronautics and Astronautics Journal, vol. 38, no. 11, p. 2059-2066, DOI:10.2514/2.893. [11] Stock, H.W. (2005). Infinite swept-wing Navier-Stokes computations with eN transition prediction. American Institute of Aeronautics and Astronautics Journal, vol. 43, no. 6, p. 1221-1229, DOI:10.2514/1.12487. [12] Whitfield, D.L., Swafford, T.W., Jacocks, J.L. (1981). Calculation of turbulent boundary layers with separation and viscous-inviscid interaction. American Institute of Aeronautics and Astronautics Journal, vol. 19, no. 10, p. 1315-1322, DOI:10.2514/3.60066. [13] Zhang, Z., Liu, F., Schuster, D.M. (2006). An Efficient Euler Method on Non-Moving Cartesian Grids With Boundary-Layer Correction for Wing Flutter Simulations. 44th AIAA Aerospace Sciences Meeting and Exhibit. American Institute of Aeronautics and Astronautics paper no. 2006-884. [14] Van Leer, B. (1997). Towards the ultimate conservative difference scheme V. A second-order sequel to Godunov’s method. Journal of Computational Physics, vol. 135, no. 2, p. 229-248, DOI:10.1006/ jcph.1997.5704. [15] Anderson, W.K., Thomas, J.L., Van Leer, B. (1986). Comparison of finite volume flux vector splittings for the Euler equations. American Institute of Aeronautics and Astronautics Journal, vol. 24, no. 9, p. 1453-1460, DOI:10.2514/3.9465. [16] Carstens, V. (1991). Computation of the Unsteady Transonic 2D Cascade Flow by an Euler Algorithm with Interactive Grid Generation. AGARD-specialists meeting on transonic unsteady aerodynamics and aeroelasticity, San Diego. [17] Yiu, K.F.C., Giles, M.B. (1995). Simultaneous viscous-inviscid coupling via transpiration. Journal of Computational Physics, vol. 120, no. 2, p. 157-170, DOI:10.1006/jcph.1995.1156. [18] Thomas, J.L., Salas, M.D. (1986). Far-field boundary condition for transonic lifting solutions to the Euler equations. American Institute of Aeronautics and Astronautics Journal, vol. 24, no. 7, p. 1074-1080, DOI:10.2514/3.9394. [19] Drela, M., Giles, M.B. (1987). Viscous-inviscid analysis of transonic and low Reynolds number airfoils. American Institute of Aeronautics and Astronautics Journal, vol. 25, no. 10, p. 1347-1355, DOI:10.2514/3.9789. [20] Whitfield, D.L. (1978). Analytical Description of the Complete Turbulent Boundary Layer Velocity Profile. American Institute of Aeronautics and Astronautics, paper no. 78-1158. [21] Drela, M. (1986). Two-Dimensional Transonic Aerodynamic Design and Analysis Using the Euler

Majić, F. – Voss, R. – Virag, Z.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 470-481

Equations. PhD thesis, Massachusetts Institute of Technology, Massachusetts. [22] Majić, F. (2010). Boundary Layer Method for Unsteady Aerodynamic Loads Determination. PhD thesis, University of Zagreb, Zagreb. [23] Blasius, H. (1908). Grenzschichten in Flüssigkeiten mit kleiner Reibung. Zeitschrift für Angewandte Mathematik Physik, no. 56, p. 1-37. [24] Schlichting, H., Gersten, K. (2006). Grenzschicht Theorie. Springer, Berlin. [25] Obremski, H.J., Morovkin, M.V., Landahl, M.T. (1969). A Portfolio of Stability Characteristics of Incompressible Boundary Layer. AGARD-ograph 134, NATO, Neuilly Sur Seine. [26] Lighthill, M.J. (1958). On displacement thickness. Journal of Fluid Mechanics, vol. 4, no. 4, p. 383-392, DOI:10.1017/S0022112058000525.

[27] Goldstein, S. (1948). On laminar boundary-layer flow near a position of separation. Quarterly Journal of Mechanics and Applied Mathematics, vol. 1, no. 1, p. 43-69, DOI:10.1093/qjmam/1.1.43. [28] Thibert, J.J., Grandjacques, M., Zwaaneveld, J. (1979). Experimental Data Base for Computer Program Assessment. Report of the Fluid Dynamics Panel Working Group 04. AGARD-AR 138. [29] Landon, R.H., Davis, S.S. (1982). Compendium of Unsteady Aerodynamic Measurement. AGARD-R 702. [30] Magnus, R., Yoshihara, H. (1970). Inviscid transonic flow over airfoils. American Institute of Aeronautics and Astronautics Journal, vol. 8, no. 12, p. 2157-2161, DOI:10.2514/3.6080. [31] Anderson, J.D.Jr. (1990). Modern Compressible Flow. McGraw-Hill, Singapore.

Boundary Layer Method for Unsteady Transonic Flow

481

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491 DOI:10.5545/sv-jme.2010.023

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P. Jure Marn – Jurij Iljaž* – Zoran Žunič – Primož Ternik

University of Maribor, Faculty of Mechanical Engineering, Slovenia Most numerical simulations about aortic valve are discussing about the blood flow through prosthetic valves, therefore this study covers a biological aortic valve. In this paper the simulation of blood flow around the regurgitated valve as well as around the health valve have been investigated and is showing how the regurgitation affects the heart function and coronary blood flow. Furthermore it also covers the coronary blood flow and its affect on the blood motion around the valve and consequently on the possible calcification. The physical problem has been treated for the period of closed valve, because of the regurgitation, which occurs at that time. For this restriction the assumption of non-elastic geometry has been used. Also the non-Newtonian Power Law model for the blood rheology has been used; because of the regions of slowly moving blood. Results are showing that the coronary blood flow does not have a high impact on the blood motion and thus the calcification is more probable as well that the regurgitation has a strong affect on the heart function, like aortic pressure drop, increased left ventricle volume, etc. The calcification reason has also been stated, but needs some further investigation. This paper covers in detail the geometry of computational domain and physics, so the results can be easily repeated. Keywords: aortic valve, regurgitation, computational fluid dynamics, non-Newtonian

0 INTRODUCTION An aortic valve is lying between aorta and left ventricle (LV) that pumps the blood through the valve to the aorta and consequently to the whole body. Therefore during the cardiac cycle this valve is heavily stressed and can be marked as one of the most important valves in the heart. For this reason the valve failure is more probable and it will have a significant impact on the cardiac cycle, heart and human health itself. This is the reason why it is so important to understand and study the causes for the valve failure and how they affect the cardiac cycle and therefore human health. Many numerical models have been developed through time for the purpose to describe the blood flow around prosthetic aortic valve and its effect on it as presented in the study of Lai et al. [1] and Krafczyk et al. [2]. The reason was to investigate the behaviour of the valve and what was the cause of its failure. At first the fluid-structure interaction was ignored because of the numerical complexity, which has been overcome with the work of De Hart et al. [3] to [6] and Carmody et al. [7]. These numerical models have been developed to describe the blood flow around biological and prosthetic aortic valve with major focus to describe the leaflet motion and induced stresses. They do not describe the blood flow around aortic valve when it fails and how this failure affect the cardiac cycle as well they do not include the coronary arteries and the effects on the coronary blood flow, which is the main objective of this study. The blood flow through coronary arteries has already been 482

covered in study of Johnston et al. [8] and [9] and Boutsianis et al. [10] that used only the part of main coronary artery, left or right, without its origin and its aortic pressure dependent blood flow. We noted the work of De Hart and his coworkers’. They started with two dimensional aspect of the aortic valve involving fluid-structure interaction using Newtonian fluid model, presented in the work [3]. They continue their work by 3D computational analysis of fluid-structure interaction, still using Newtonian fluid model [4]. The 3D geometry of aortic valve has also been used by Carmody et al. [7]. They take the research of blood flow through aortic valve to the next level, with including the LV and blood flow through it. The geometric parameters of biological aortic valve treated in this study have been adopted from Labrosse et al. [11] except for sinus depth and diameter of coronary arteries adopted from De Hart et al. [4] and Johnston et al. [8]. The latter work should also be noted for the use of non-Newtonian blood models. Far more work has been carried out in the area of mechanical heart valves; however mechanical aortic valves were not the subject of our interest. This paper presents a transient simulation of blood flow around healthy and chronically regurgitated aortic valve including the coronary arteries. Regurgitation of aortic valve is the disease or valve failure when it cannot close completely during its closure and consequently enables the blood to leak back into the left ventricle, while the ejection or valve opening is not problematic. For this reason the

*Corr. Author’s Address: University of Maribor, Faculty for Mechanical Engineering, Smetanova ulica 17, 2000 Maribor, Slovenia, jurij.iljaz@uni-mb.si

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

problem has been treated only for the period of closed valve. The second reason for the chosen time interval is also in the coronary blood flow, which is maximal and has the biggest effect on the blood motion around aortic valve and possible calcification. During the closed valve the valve or its cups movement can be neglected and for this reason the assumption of using non-elastic geometry has been made. Also one important comment has to be made here, that the blood behaves as a non-Newtonian pseudo plastic fluid, especially when the shear rate is small, that is less than 100 s-1. Therefore it is necessary to take this into account, because the blood movement around aortic valve during the closed valve is slow. In this place the Power Law model has been used to describe the properties and behaviour of the blood. The problem was solved with Computational Fluid Dynamics (CFD) using commercial software package. The aim of this study is to analysis the blood flow around healthy and regurgitated aortic valve during closed valve as well to analysis the effect of the coronary blood flow on the blood around cusps. This work also presents the base for studying the effects of regurgitated aortic valve on the cardiac cycle (aortic pressure drop, regurgitated blood volume, etc.) and on the coronary blood flow. The remainder of this paper is organized as follows: Section 1 discusses the governing equation for the physical problem; Section 2, the geometry and physics of the problem; Section 3, the results and discussion of this study; and Section 4, conclusions which can be drawn from this work. 1 GOVERNING EQUATIONS Physical problem of blood flow around aortic valve and through coronary arteries can be described mathematically with momentum equation:

∂vi ∂v ∂p ∂τ ji , + ρ v j i = ρ f mi − + ∂t ∂x j ∂xi ∂x j

(1)

where η represents the dynamic viscosity, εij velocity deformation and δij the Knocker delta function. Rheological model Eq. (3) is valid only for Newtonian fluid; however it is also used to model the nonNewtonian fluid flow. Non-Newtonian behaviour is taken into account with additional model, which connects the dynamic viscosity of the fluid η with the shear rate γ and it will be presented in the next chapter, together with the problem physic and boundary conditions. Presented system of governing equation, Eqs. (1) to (3) is highly non-linear and therefore solved numerically using the techniques of CFD. As mentioned, the problem has been solved numerically using commercial software package based on the finite volume method (FVM). 2 NUMERICAL MODEL 2.1 Geometry Geometry of a computational domain is shown on Fig. 1 and it consists of aortic valve, part of the ascending aorta, part of left ventricle (LV) and origin of both coronary arteries. It can also be seen that small geometric simplifications have been made to simplify the modelling. The reason lies in geometry of the real biological aortic valve, which is not symmetrical as well the ascending aorta does not have a perfect tubular shape. Despite this small simplification the presented geometry in this chapter still represent a normal shape of biological aortic valve and its region. Aortic valve consists of three major components; annulus, cusp and commissures. The three half-moon shaped (semilunar) cusps are attached to the aortic root and annulus. Between the free and closing edge of each cusp are two areas known as the lunulas. This way the shape of the cusps and the lunulas gives the valve a good sealant property during diastole of LV.

and continuity equation:

∂ρ ∂ ( ρ vi ) + = 0, ∂t ∂xi

(2)

where νi represents the velocity vector, p static pressure, ρ density, τij the shear stress term and fmi the density of volumetric or mass force. The shear stress term in Eq. (1) is defined with rheological model:

 2 τ ij = 2η εij − ηεkk δ ij , 3

(3)

Fig. 1. Geometry of computational domain; a) healthy aortic valve, b) regurgitation

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved

483

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

Ascending aorta arises behind aortic valve and it consists of sinus and tubular portion. Behind the valve cusps there are three sinuses and their main task is to prevent cusps to stick on the aortic wall. From right and left aortic sinus arise the right and left coronary artery as shown on Fig. 1. They supply the blood to the muscular tissue of the heart called myocardium. The right coronary (RC) artery normally supplies the inferior surface of the LV, the right ventricle (RV) and right atrium (RA) in contrast to the left coronary (LC) artery, which supplies the rest of the LV and left atrium (LA). The important geometric parameters for average biological aortic valve, LV, ascending aorta and coronary arteries are shown on Fig. 2 and gathered in Table 1, together with their values. Introduced geometric parameters have been taken from Labrosse et al. [11] and De Hart et al. [4]. At creating the geometry of the computational domain some assumptions have been made, like equal cusps size, equal diameter of the valve base and sinus tubular junction (STJ) and slightly smaller height of the commissures compared to the sinuses. All these assumptions are not so crucial, because the aortic valve geometry depends on the individual therefore this aortic valve is still representing the normal biological valve. Coronary arteries parameters, like height of origin and position, have been taken from anatomy [12]. Normally both coronary arteries arise at the two thirds of the sinus height from the aortic annulus to the STJ and near the middle between commissures. RC artery arise nearly perpendicular from the aorta, where LC artery arises at an acute angel. Only parameter that has been taken from Johnston et al. [8] has been the diameter of coronary arteries, which is assumed to be equal. The tubular portion of ascending aorta was modelled as a slightly curved tube in a direction of the left cusp, with the same diameter as the STJ. Also the part of LV was designed as the tube with base valve diameter, because there has not been enough data to construct the ventricle around aortic valve. In this manner the computational domain for healthy aortic valve has been created. Computational domain for regurgitated aortic valve has been created in the same way and by the same assumptions. To generate the central hole, which allows the blood to leak back into the LV during the diastole, the base, STJ and aortic diameter have been enlarged from 25 to 27 mm. All other parameters stayed the same. This enlargement created the central hole area of 44.2 mm2, which is 7.7 % of the total 484

valve area. The created geometry of computational domain is shown on Fig. 1. Table 1. Geometric and Power Law parameters Geometric parameters d0 [mm] dSTJ [mm] dC [mm] dA [mm] 25.0 25.0 4.0 25.0 Lf [mm] Lh [mm] LLC [mm] LRC [mm] 28.0 15.0 13.7 12.4 HC [mm] HC0 [mm] HLV [mm] HS [mm] 14.0 12.0 10.0 21.0 Power Law K [Pa·sn] n 0.027844 0.6062

t [mm] 0.3 L A [mm] 21.0 αLC [°] 35.0

RA [mm] 60.0 LS [mm] 6.0 αRC [°] 90.0

Fig. 2. Important geometric parameters of computational domain

2.2 Physics As mentioned before, the numerical simulation of blood flow around healthy and regurgitated aortic valve, during its closed position, has been done. The deformation of the closed valve and its surrounding region, because of the changing aortic and LV pressure as well as the heart movement is assumed to be small enough; therefore the error with assumption of using non-elastic geometry is small. Although the non-elastic geometry was used, the problem has been still treated as time dependent. The blood behaves like a pseudo plastic nonNewtonian fluid for which the Power Law model has been used. It correlates the dynamic viscosity of the fluid η with theshear rate γ by equation:

η = K ⋅ γ n −1 ,

(4)

where and are the Power Law parameters. The values for these parameters have been taken from Ternik et al. [13] and are written in Table 1. When the shear rate is small the dynamic viscosity is high and vice versa. Also the blood has been treated as incompressible fluid, with constant density of 1058 kg/m3, because of the small expected velocities. To determine the boundary conditions, necessary to simulate the treated problem, the physiology of

Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

the heart has to be known. As it is known, during the healthy closed valve the blood flows from aorta to left and right coronary artery and also to smaller arteries to the whole body, but this is not the case of this study. On the other hand, when regurgitation occurs, the blood is also flowing from aorta to the LV, because of the valve leak. Therefore the inlet boundary condition (BC) has been placed on the aortic plane and the outlet BC on the ventricle and both coronary planes. Except for the healthy aortic valve simulation, where the exclusion of the outlet on the ventricle plane has been done, because there is no blood flowing into the LV. Position of the planes is shown on Fig. 3. To solve the presented governing equations for this problem Eqs. (1) to (3) the value of the BC are required, which have been determent on the base of average physiological values for human at rest [14] to [16] and are presented in Table 2.

Table 2. Parameters for definition of boundary conditions: nS the heart rate, pAS the maximal aortic pressure at systole, pAD the minimal aortic pressure at diastole, QLC the blood flow through LC artery, QRC the blood flow through RC artery and tD the diastole time Boundary condition parameters

pAS

pAD

QLC

[1/min]

[kPa]

[ml/min]

QRC

tD [s]

75.0

15.93

10.62

190.0

40.0

0.5

For the healthy aortic valve simulation the inlet BC with time-dependent relative aortic pressure has been prescribed to the aortic plane and its distribution is shown on Fig. 4. On the other hand the outlet BC placed on the RC and LC artery plane has been treated with prescribed mass flow rate. Coronary blood flow, which flows through coronary circulation (CC) is very complex, dependent on the pressure difference between RA and aorta, myocardium contraction,

m i = ρ

∆pi , Ri

(5)

where Δpi is a perfusion pressure and the resistance of CC. The mass flow rate has been evaluated at the end of each time step for the next time step iteration. It is well known that the blood flow rate through coronary arteries is maximal during the diastole, because of the minimal CC resistance. The reason is because myocardium relaxes and the vessels dilatate, especially in endocardium. Therefore the CC resistance is time-dependent and it is shown on Fig. 5. Its distribution has been chosen, based on the normal coronary blood flow at rest, which is known. Some assumptions have been made here in setting the timedependent distribution of resistance, like linear change during the myocardium relaxation and contraction as well as constant value during ventricular filling. To go step further the CC resistance is also controlled with myocardium metabolism, which is higher in LV and consequently causes a higher blood flow rate in LC than RC artery. For this reason the resistance of right CC has been chosen to be smaller than resistance of right CC. Other mechanisms for resistance regulation are myogenic and neural, which has not been implemented in this study, because it exceeds the scope of this work. The perfusion pressure that occurs in Eq. (5) has been defined with

Fig. 3. Boundary condition planes position

etc. Therefore the coronary mass flow rate has been described with linear equation:

∆pi = pic − pRA ,

(6)

where pic represents a pressure in the coronary plane and equal to the pressure in the ascending aorta and pRA represents the pressure in right atrium (RA). The reason is because the blood flows from aorta, through coronary arteries and CC to RA. The prescribed timedependent RA pressure in Eq. (6) is also shown on Fig. 4. The BC for the regurgitated aortic valve simulation has been changed. The outlet BC at the LC and RC artery plane has been treated in the same manner as before with Eqs. (5) and (6). However because the mass flow rate is dependent from an aortic pressure, the mass flow drop has been expected owing to the decreasing pressure. Aortic pressure decreases because of the leaking blood back to the LV. This way the time-dependent relative pressure at the aortic plane has been defined and been calculated at each time step with equation:

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved

485

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

p A ( t + ∆t ) = p A ( t ) − ∆pP − ∆pV ,

(7)

where pA(t+Δt) is the aortic pressure in the next time step, pA(t) is the pressure in the present time, ΔpP is pressure drop because of the blood flow from aorta through systemic circulation (whole body) and ΔpV isthe pressure drop because of the back flowing blood, that is from aorta to LV. Pressure drop ΔpP has been defined from the aortic pressure distribution at healthy aortic valve (Fig. 4). On the other hand the pressure drop ΔpV is not known in advance; therefore it has been described with equation:

∆pV =

m ∆t , ρC

(8)

where m represent a mass flow into the LV (regurgitated blood flow), Δt is a time step and C aortic stiffness. The mass flow has been evaluated from the results at the end of each time interval. Therefore the pressure drop ΔpV has been calculated by Eq. (8) at the end of each time step to update the relative aortic pressure in the next time step defined by Eq. (7). The value for aortic stiffness has taken to be 8.023 m3/Pa and can vary greatly, depends on the person age. To close the system of equations, also the outflow BC on the LV plane has to be defined. For this the time-dependent LV relative pressure has been prescribed and is also shown on Fig. 4. Because the chronic regurgitation of aortic valve has been taken into account, the increase in LV end diastolic pressure was neglected. Finally also the non-slip boundary condition has been prescribed for all the walls.

Fig. 4. Boundary conditions: aortic, LV and RA pressure

The problem has been treated as time-depended; therefore the initial value for the flow field needs to be defined. Therefore the initial values were defined with steady state simulation for initial boundary conditions. 486

For consistence of several figures and results the beginning of valve closure was set to 0 s. This way the BC at regurgitated aortic valve depends on the problem itself as well as the results of the simulation. In this manner it is possible to obtain the effect of regurgitation on the cardiac cycle as well as on the heart, like aortic pressure drop, increased volume of LV, the change of coronary blood flow, etc.

Fig. 5. Boundary conditions: resistance of left and right CC

2.3 Properties Geometry of computational domain has been discretizated with tetrahedral and prism elements to create computational mesh. At the wall the inflation layer has been used for better description of velocity gradients. For regurgitated aortic valve simulation, the computational mesh has also been refined in the region of the central hole also to better describe the high velocity gradients. Representative mesh for healthy and regurgitated valve is shown on Fig. 6. The computational mesh has also been analysed to estimate the numerical error and to prevent inaccurate numerical results. The final mesh consists of 80,376 nodes and 275,046 elements for healthy aortic valve and of 83,180 nodes and 301,820 elements for regurgitation. Numerical model also needs be time discretizated, because of the time-dependent problem. The proper time step has been set by comparing the results of two different time steps of the test case. Therefor the time step at healthy aortic valve simulation has been set to 0.025 s and for regurgitation to 0.001 s. The blood velocity during the diastole is assumed to be low enough to have laminar flow field and therefore the turbulence models have not been included in numerical models. The convergence criteria for solving the problem have been set to 10-4.

Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

Fig. 6. Mesh for computational domain: a) healthy valve, b) detailed view of inflated layer and c) aortic valve regurgitation

3 RESULTS AND DISCUSSION Transient simulation of blood flow through coronary arteries and around aortic valve, during the closed valve, for healthy and diseased (regurgitated) aortic valve, has been done. The duration of closed valve for healthy aortic valve has been set to 0.55 s as it can be seen on Fig. 4. Fig. 7 shows the blood velocity field around the healthy aortic valve for the time 0.15 s, when the coronary blood flow is the highest. The velocity vectors have been normalized to clearly show the direction of the flow. As discussed earlier the blood flow rate through LC artery is substantially higher than through the RC artery, what can also be seen from Fig. 7. The average flow rate through the RC artery during the closed valve has been 36.48 ml/ min and through the LC artery 187.6 ml/min. Even the maximal flow rate through LC artery is pointing to a low Reynolds number defined by:

Re =

vd ρ , η

(9)

where v represents average velocity, η average dynamic viscosity and the characteristic length which is in the diameter of the artery in this case. Maximal value of Reynolds number for LC artery has been 426.7 and for RC artery 42.3, which indicates the laminar flow regime in the arteries and also in the region of aortic valve and support the assumption of excluding the turbulence models in the numerical models. As can be seen from Fig. 7, the blood movement around valve, this is cusps and lunulas, is also very small. First reason for that can be in low coronary blood flow and the other in high dynamic blood viscosity due to the non-Newtonian behaviour. Higher dynamic viscosity increases the resistance of moving and therefore small velocity field. Fig. 8 shows the dynamic viscosity of the blood through left

and right cusp, which supports the above statement. As it can be seen the dynamic viscosity around aortic valve is higher, especially around right cusp, than from the origin of the coronary arteries.

Fig. 7. Blood velocity field for healthy aortic valve (0.15 s)

At aortic valve regurgitation simulation, the blood not only that flows from aorta to the left and right coronary artery but also from aorta through aortic valve hole to LV (regurgitated blood flow). The blood flow field can be seen on Fig. 9, which shows the flow filed at two different time 0.15 and 0.35 s. As it is known this regurgitated blood flow decreases the aortic pressure by Eq. (7), which also gave the pressure drop as a result of simulation. This pressure drop in aorta is shown on Fig. 10, together with normal pressure drop at healthy aortic valve. If a comparison of the aortic pressure for regurgitated and healthy valve simulation is made, it can be seen that the drop is quite large. It can be also noticed that at regurgitation the valve stayed closed for only 0.52 s, which is less than at the healthy valve. The reason lies in the difference between aortic and LV pressure, which is zero or even negative and therefore an opening of the aortic valve, begins.

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved

487

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

Fig. 8. Dynamic viscosity field for healthy aortic valve (0.15 s)

High aortic pressure drop is also causing the drop in the coronary blood flow. Fig. 11 shows the calculated time-dependent coronary blood flow for healthy and regurgitated aortic valve. For regurgitation, the average blood flow rate through RC artery has been 24.47 ml/min, meanwhile through LC artery 132.55 ml/min, which means a 31% drop. With presented numerical models, the volume of regurgitated blood flow can also be calculated. For the threated regurgitation the volume of blood was 56.1 ml. That means that the LV is highly loaded with blood and that the regurgitation in those conditions can be severe. As it can be seen from the Fig. 9 also the strong jet occurs at the central hole. The maximum average velocity of the blood jet has been 3.7 m/s, which gave the Reynolds number of 6084. Reynolds number has been evaluated by Eq. (9), where the hydraulic diameter of 3.8 mm has been taken as a characteristic length. The hydraulic diameter has been defined by equation:

4A , u

where A represent a cross sectional area and u the wetted perimeter. High value of Reynolds number indicates the tendency of turbulent flow, but because of the small time of this phenomena and short distances, the question if the flow is really turbulent, needs some further research. However on the other hand at sever valve regurgitation the regurgitated blood flow is supposed to be turbulent; because it creates a murmur. The strong blood jet also produces the slow recirculation in LV as can be seen on Fig. 9. This recirculation is driven by the velocity of the jet and not by the continuum equation, as it would be if the computational domain would include the whole LV geometry. Therefore this recirculation, because of small computational domain and prescribed boundary conditions, is not physically correct. The strong central jet is also stressing the cusps around the central hole, because of the low pressure field under the cusps. This can also be shown with the shear stress on the surfaces, like it is shown on Fig. 12. Great shear stress on the tissue can cause the lesion or improper function of the tissue. As seen on the Fig. 12, the high shear stress in the region of the central hole tends to move the cusps together to close the central hole. Therefore, some error has been made using non-elastic geometry. For the same reasons as before, also in the regurgitated valve simulation it can be noticed, that the blood movement is slow around the valve cusps. Fig. 13 is showing the dynamic viscosity of the blood, which can be associated with the region of slowly moving blood. Also because the back flow occurs at the beginning of the valve closure,

Fig. 9. Blood velocity field for aortic valve regurgitation; a) 0.15 s, b) 0.35 s

488

(10)

Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

the error of using poor initials values for the flow field is assumed to be minimal. The question arises, if the results from presented numerical models are precise enough, because some assumptions have been made. Like nonelastic geometry, this may be the most important one. To answer that question further investigation or improvement of numerical model should be done. The improvement to use elastic geometry will lead to the whole cardiac cyclic simulation and the error of using poor initial values for flow filed would be excluded. One of the promising methods to incorporate fluidstructure interaction and elastic geometry with high distortion could be meshless particle method described by Petkovšek et al. [17]. Second most important improvement would be to investigate the behaviour and the impact of different non-Newtonian blood models. Also more detailed investigation of important factors that affects the coronary blood flow should be done. This would lead to the improved mathematical model Eq. (5), because the current model has a disadvantage in the CC resistance. This has to be evaluated in advanced, based on the known coronary blood flow. Physically more accurate model would be, if the resistance would be a function of the metabolic, myogenic and neural regulation. As well the model for evaluation of aortic pressure drop because of the regurgitated blood flow Eq. (8) could be improved, to compensate the LV parameters like, LV volume and its pressure rising. These few improvements are stated here just to remind the reader what assumptions have been made and how the evaluation of the results in the further research should be done.

or calcification and with that correlated aortic valve stenosis. The other factor that helps calcium deposition is the deposition time or time the valve being closed, which is longer for resting position than for exercise. Reason for calcification that is stated here is just an assumption, which should need some further investigation.

Fig. 11. Coronary blood flow during closed valve for healthy and regurgitated aortic valve simulation

Fig. 12. Wall shear stress in aortic valve regurgitation

Fig. 10. Aortic pressure for healthy and regurgitated aortic valve

Simulation of blood flow around healthy as well as regurgitated aortic valve shows that the blood is moving real slowly around the leaflet and sinuses. This slow motion can cause a deposition of calcium

Fig. 13. Dynamic viscosity field for regurgitated aortic valve (0.35 s)

De Heart et al. [8] have also done the simulation of the blood wash-out, which indicates that the blood

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved

489

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

particles around the leaflets does not stays in that position longer than two cardiac cycles. The most important assumption in De Heart simulation has been the use of blood as Newtonian fluid, which may lead to the mislead results. As well the coronary blood flow has been neglected. As it can been seen in this study the effect of coronary blood flow on the blood around aortic valve and sinuses was minimal, so the assumption of excluding the coronary arteries may be reasonable. However the effect of coronary blood flow can be substantial during the exercise, where the coronary blood flow can be up to four times larger. Simulation of chronically regurgitated aortic valve has shown that the LV is highly loaded, because of the large regurgitated blood volume. Therefore the effective volume of blood, which had to be pumped from LV each cardiac cycle, has increase severely. For this reason the work done by myocardium has also been increased as well as the demand of the oxygen. However on the other hand the supply of the oxygen has been reduced as shown on Fig. 11, which is because of the pressure drop in ascending aorta. Of course this is not happening in a real life, because other mechanisms, like vasodilatation, etc. came in place to reduce the resistance of CC and to help maintain the demand and supply in balance. But in long term, these mechanisms cannot cope with oxygen demand and myocardial ischemia occurs. In this study only the drop of the coronary blood flow rate at the same CC resistance has been investigated. Nevertheless for these conditions the coronary blood flow drop at 7.7 % regurgitation is quite severe. As stated before, for more precise calculation of coronary blood flow drop, the new mathematical model would be needed. 4 CONCLUSION This study shows the difference between the healthy and regurgitated biological aortic valve, based on presented numerical model. It covers how the regurgitation influences the coronary blood flow, aortic pressure drop, volume load of the LV, cardiac cycle, etc. The computational domain has been narrowed to the aortic valve region with origin of the coronary arteries and its geometry has been covered in detail. However some geometric assumptions have been made, the created geometry still represents normal biological valve and its region. The reason lies in fact that the geometry of the aortic valve distinguishes from person to person. 490

The physical problem has been treated as timedependent and solved only for the period of closed valve. The reason for that lies in the nature of regurgitation and coronary blood flow. Therefore the assumption of using non-elastic geometry has been made. In addition the blood was treated as an incompressible, pseudo plastic non-Newtonian fluid with the use of Power Law model. To solve the physical problem or to close the system of governing equations, the boundary conditions have been applied to the boundary of computational domain, like aortic pressure, LV pressure and mass flow through LC and RC artery. The domain has also been geometric and time discretizated and to evaluate the accuracy of discretization the nodalization analysis have been done. Mass flow through coronary arteries has been described with mathematical model, Eq. (5) as well as aortic pressure drop at regurgitation, Eq. (7). This way the coronary blood flow and also aortic pressure drop have not been prescribed in advance, but are the results of numerical solution. Therefore the comparison between healthy and regurgitated aortic valve can be done. This study shows that regurgitation can highly load the LV; also the oxygen delivery can be greatly decreased, which can lead to myocardium ischemia. Therefore the aortic valve dysfunction, like regurgitation, can have a serious impact on heart function as well as on the human health. Furthermore the coronary blood flow shows little or no affect on the blood movement around the valve and in the ascending aorta. However the question arise what happens during the exercise, when the coronary blood flow increases. At that point the possible reason for calcification has also been stated, but needs some further investigation. The main reason for the slow blood movement around the valve cusps and correlated calcification is in the nature of the blood itself or its dynamic viscosity, which indicates that using the blood as a Newtonian fluid will lead to misleading results. Discussion also covers the further improvements of numerical model, like using elastic geometry and evaluation of the results using different nonNewtonian blood models. With the help of diseased aortic valve simulation, the disease impact on the cardiac cycle as well on the heart function can be evaluated. As well as the evaluation, when the disease progresses too much, that the heart cannot cope with it anymore and urgent surgery is needed. However for this situation, the numerical model would have to be improved, what represents the future work on this study.

Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 482-491

5 REFERENCES [1] Lai, Y.G., Chandran, K.B., Lemmon, J. (2002). A numerical simulation of mechanical heart valve closure fluid dynamics. Journal of Biomechanics, vol. 35, no. 7, p. 881-892, DOI:10.1016/S0021-9290(02)00056-8. [2] Krafczyk, M., Cerrolaza, M., Schulz, M., Rank, E. (1998). Analysis of 3D transient blood flow passing through an artificial aortic valve by Lattice-Boltzmann methods. Journal of Biomechanics, vol. 31, no. 5, p. 453-462, DOI:10.1016/S0021-9290(98)00036-0. [3] De Hart, J., Peters, G.W.M., Schreurs, P.J.G., Baaijens, F.P.T. (2000). A two-dimensional fluidstructure interaction model of the aortic valve. Journal of Biomechanics, vol. 33, no. 2, p. 1079-1088, DOI:10.1016/S0021-9290(00)00068-3. [4] De Hart, J., Peters, G.W.M., Schreurs, P.J.G., Baaijens, F.P.T. (2003). A three-dimensional computational analysis of fluid-structure interaction in the aortic valve. Journal of Biomechanics, vol. 33, no. 1, p. 103112, DOI:10.1016/S0021-9290(02)00244-0. [5] De Hart, J., Baaijens, F.P.T., Peters, G.W.M., Schreurs, P.J.G. (2003). A computational fluid-structure interaction analysis of a fiber-reinforced stentless aortic valve. Journal of Biomechanics, vol. 36, no. 5, p. 699712, DOI:10.1016/S0021-9290(02)00448-7. [6] De Hart, J., Peters, G.W.M., Schreurs, P.J.G., Baaijens, F.P.T. (2004). Collagen fibre reduce stresses and stabilize motion of aortic valve leaflets during systole. Journal of Biomechanics, vol. 37, no. 3, p. 303-311, DOI:10.1016/S0021-9290(03)00293-8. [7] Carmody, C.J., Burriesci, G., Howard, I.C., Patterson, E.A. (2006). An approach to the simulation of fluidstructure interaction in the aortic valve. Journal of Biomechanics, vol. 39, no. 1, p. 158-169, DOI:10.1016/j.jbiomech.2004.10.038. [8] Johnston, B.M., Johnston, P.R., Corney, S., Kilpatrick, D. (2004). Non-Newtonian blood flow in human

right coronary arteries: steady state simulations. Journal of Biomechanics, vol. 37, no. 5, p. 709-720, DOI:10.1016/j.jbiomech.2003.09.016. [9] Johnston, B.M., Johnston, P.R., Corney, S., Kilpatrick, D. (2006). Non-Newtonian blood flow in human right coronary arteries: Transient simulations. Journal of Biomechanics, vol. 39, no. 6, p. 1116-1128, DOI:10.1016/j.jbiomech.2005.01.034. [10] Boutsianis, E., Dave, H., Frauenfelder, T., Poulikakos, D., Wildermuth, S., Turina, M., Ventikos, Y., Zund, G. (2004). Computational simulation of intracoronary flow based on real coronary geometry. European journal of Cardio-thoracic Surgery, vol. 26, p. 248256, DOI:10.1016/j.ejcts.2004.02.041. [11] Labrosse, M.R., Carsten, J.B., Robicsek, F., Thubrikar, M.J. (2006). Geometric modeling of functional trileaflet aortic valves: Development and clinical applications. Journal of Biomechanics, vol. 39, no. 14, p. 2665-2672, DOI:10.1016/j.jbiomech.2005.08.012. [12] Fuster, V., Aleksander, R.W., O’Rourke, R. (eds.). (2001). Hurst´s The Heart, tenth edition. McGraw-Hill, New York. [13] Ternik, P., Žunič, Z., Marn, J. (2007). Blood hemodynamics in carotid bifurcation: influence of rheological models. IWNMNNF: Book of abstracts, Rhodes. [14] Berne, R.M., Levy, M.N. (2000). Principles of physiology, third edition. Mosby, London. [15] Bronzino, J.D. (ed.). (2006). Biomedical engineering fundamentals, third edition. Taylor & Francis Group, New York. [16] Skalak, R., Chien, S. (eds.). (1987). Handbook of bioengineering. McGraw-Hill, New York. [17] Petkovšek, G., Džebo, E., Četina, M., Žagar, D. (2010). Application of Non-Discrete Boundaries with Friction to Smoothed Partile Hydrodynamics. Strojniški vestnik - Journal of Mechanical Engineering, vol. 56, no. 5, p. 307-315.

Non-Newtonian Blood Flow around Healthy and Regurgitated Aortic Valve with Coronary Blood Flow Involved

491

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498 DOI:10.5545/sv-jme.2011.297

Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis Çiçek, A. − Kıvak, T. − Samtaş, G. − Çay, Y. Adem Çiçek1 − Turgay Kıvak2 − Gürcan Samtaş2,* − Yusuf Çay3

2 Düzce

1 Düzce

University, Faculty of Technology, Turkey University, Cumayeri Vocational School of Higher Education, Turkey 3 Karabük University, Faculty of Engineering, Turkey

In this study, the effects of cutting parameters (i.e., cutting speed, feed rate) and deep cryogenic treatment on thrust force (Ff) have been investigated in the drilling of AISI 316 stainless steel. To observe the effects of deep cryogenic treatment on thrust forces, M35 HSS twist drills were cryogenically treated at –196 °C for 24 h and tempered at 200 °C for 2 h after conventional heat treatment. The experimental results showed that the lowest thrust forces were measured with the cryogenically treated and tempered drills. In addition, artificial neural networks (ANNs) and multiple regression analysis were used to model the thrust force. The scaled conjugate gradient (SCG) learning algorithm with the logistic sigmoid transfer function was used to train and test the ANNs. The ANN results showed that the SCG learning algorithm with five neurons in the hidden layer produced the coefficient of determinations (R2) of 0.999907 and 0.999871 for the training and testing data, respectively. In addition, the root mean square error (RMSE) was 0.00769 and 0.009066, and the mean error percentage (MEP) was 0.725947 and 0.930127 for the training and testing data, respectively. Keywords: artificial neural networks, regression analysis, cryogenic treatment, machining, thrust force, predictive modelling

0 INTRODUCTION In recent years, applying different heat treatments to tool steels has been widely executed in order to increase productivity [1]. Specifically, cryogenic treatment has proved for many years to be an effective method to improve tool life and to eliminate residual stress [2] and [3]. Cryogenic treatment is a supplementary process to conventional heat treatments and is a permanent procedure that affects the entire material, unlike hard coatings [4] and [5]. In addition, cryogenic treatment increases the wear resistance of materials by providing a more intensive and homogeneous distribution of carbides due to transformation of the retained austenite into martensite [6]. The cryogenic treatment significantly increases the lifetime of the HSS cutting tools. The improvements in tool life vary from 65 to 343% depending on the cutting conditions used [7]. One of the most important factors in manufacturing costs is energy consumption. Specifically, the power consumed during machining determines the energy consumption. Depending on the specific cutting resistance of the material, the thrust force which determines the consumed power and the energy cost are the most important parameters in the cutting process. A large number of parameters influence the cutting forces, including cutting speed, feed rate, depth of cut, rake angle, nose radius, cutting edge inclination angle, physical and chemical characteristics of the machined part, and chip breaker geometry. Therefore, developing a proper cutting force analytical model is very difficult [8]. Several 492

methods such as artificial neural networks, multiple regression, and finite element analysis have been developed for modelling the cutting force. Hao et al. [9] introduced a predictive cutting force model that used artificial neural networks (ANNs). The inputs to the ANNs consisted of cutting velocity, feed rate, depth of cut, and tool inclination angle, while the outputs were thrust force, radial force, and main cutting force. The performance of the developed cutting force model was fairly satisfactory. Özel and Karpat [10] modelled surface roughness and flank wear in hard turning using ANNs and regression models. Predictive neural network models were found to be better at predicting surface roughness and tool flank wear. Panda et al. [11] developed a fuzzy backpropagation neural network scheme for the prediction of drill wear. Results show a very good prediction of drill wear from the fuzzy back-propagation neural network model. Asiltürk and Çunkaş [12] used ANNs and multiple regression method to model the surface roughness of AISI 1040 steel. The ANNs estimated the surface roughness with high accuracy when compared with the multiple regression method. Nalbant et al. [13] investigated the effects of the coating method, coating material, cutting speed, and feed rate on the surface roughness. For this purpose, a number of experimental studies were performed. Experimental values and ANN predictions were compared using statistical error analysis methods. Here, the surface roughness values were predicted by the ANNs within an acceptable accuracy. Lin [14] adopted an abductive network to construct a prediction model for surface roughness and cutting force. To verify the accuracy of

*Corr. Author’s Address: Düzce University, Cumayeri Vocational School of Higher Education, 81700, Cumayeri, Düzce, Turkey, gurcansamtas@duzce.edu.tr

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

the abductive network, regression analysis was used to develop a second prediction model. The comparison of the two models indicated that the abductive network prediction model was more accurate than the one developed using regression analysis. The objectives of this study are to determine the effects of three different heat treatments and cutting parameters on the thrust force in drilling of AISI 316 stainless steel, and to predict the thrust force using multiple regression analysis and ANNs, without the need for complex and time-consuming experimental studies. 1 EXPERIMENTAL DETAIL In this study, AISI 316 austenitic stainless steel blocks were used as the workpiece material. The dimensions of this workpiece material were 100×170×15 mm. The chemical composition of AISI 316 austenitic stainless steel is given in Table 1. Table 1. The chemical composition of AISI 316 stainless steel C 0.05

Si Mn P S Cr 0.380 0.971 0.039 0.006 16.58

Ni 9.94

Mo Cu 2.156 0.321

The drilling tests were performed using a Johnford VMC 850 model CNC vertical machine centre at four cutting speeds (V, 12, 14, 16, and 18 m/min) and three feed rates (f, 0.08, 0.1, and 0.12 mm/rev). Blind holes were drilled at a constant depth of the cut (13 mm). In the experiments, uncoated M35 HSS (DIN 1897) twist drills were used. Three holes were drilled to confirm the results in each machining condition. A number of drills (Guhring) with a diameter of 6 mm were cryogenically treated in order to observe the effects of deep cryogenic treatment. Three types of uncoated drills (Ct) were used: conventionally heattreated drills (CHT), cryogenically treated drills (CT), and cryogenically treated and 2h tempered drills (CTT) at 200 °C. The cryogenic treatment for the M35 HSS drills was performed by gradually lowering the temperature from room temperature to ‒196 °C at approximately 1.5 °C/min. This minimum cryogenic temperature was held for 24 h, and then raised back to room temperature at the same rate. Before the experiments, the stainless steel blocks were ground to eliminate the adverse effects of any surface defect on the workpiece. During the drilling tests, thrust force values were measured by the Kistler 9257B type dynamometer. 36 experiments were performed at all combinations of cutting speeds, feed rates, and heat treatments applied to the cutting tools.

1.1 Artificial Neural Networks ANNs can be successfully applied to many industrial situations. ANNs are suitable for modelling various manufacturing functions because of their ability to easily learn complex non-linear and multivariable relationships between process parameters [15] to [17]. An ANN consists of three main layers, namely input, hidden, and output layers. The neurons in the input layer transfer data from the external world into the hidden layer. In the hidden layer, outputs are produced using data from the input layer, using bias, summation, and activation functions. The summation function calculates the net input of the cell, as shown in Eq. 1:

NETi = ∑ wij x j + wbi ,

(1)

j =1

where NETi is the weighted sum of the input to the ith processing element, wij is the weight of the connections between the ith and jth processing elements, xj is the output of the jth processing element, and wbi is the weight of the biases between layers. The activation function provides a curvilinear match between input and output layers. In addition, it determines the output of the cell by processing the net input to the cell. The selection of an appropriate activation function significantly affects network performance. Recently, the logistic sigmoid transfer function has been commonly used as an activation function in multilayer perception models, because it is a differentiable, continuous, and non-linear function. For this reason, the logistic sigmoid transfer function was used as the activation function in this study. This function produces a value between 0 and 1 for each value of the net input. The logistic sigmoid function in this study is shown in Eq. 2.

f ( NETi ) =

1 . 1 + e − NETi

(2)

The training and testing data for the ANNs were prepared using 36 experimental measurements from the dynamometer. From within these 36 measurements, 9 were randomly selected as testing data, while the remaining 27 were used as training data. The thrust force was selected as the output data, while cutting speed, feed rate, and cutting tools were used as input. In order to acquire the closest output values to experimental results, the best learning algorithm and optimum number of neurons in the hidden layer was determined. For this reason, both SCG and LM learning algorithms and different

Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis

493

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

numbers of neurons in the hidden layer were used in the built network structure for thrust force. In consequence of trials, the best learning algorithm and network architecture for the prediction of thrust force became SCG: 3-5-1, respectively (Fig. 1).

predicted after ANN training were compared with the values obtained from experimental study. RMSE (root mean square error), R2 (determination coefficient), and MEP (mean error percentage) values were used for comparison. These values were calculated with the following equations:

Fig. 1. ANNs architecture with a single hidden layer

As mentioned above, the logistic sigmoid transfer function was used in this study. One of the characteristics of this function is that only a value between 0 and 1 can be produced. Thus, the input and output data sets were normalised before the training and testing processes. In this study, the input and output values were normalised between 0.1 and 0.9 to obtain optimal predictions. The cutting tool, cutting speed, and thrust force data were normalised by dividing by 6, 60, and 950, respectively. Feed rate was also normalised by multiplying by 4. The cutting tool digits for the ANN were determined to be CHT = 1, CT = 2, and CTT= 3 since they do not have numerical values. In order to understand whether an ANN is making good predictions, thrust force values

 1  RMSE =    ∑ t j − o j  p j  

(

 ∑ j t j − oj R = 1−  2  j oj  ∑ 2

MEP ( % ) =

( )

)

 ,  

 t j − oj   × 100  tj  , p

∑ j 

(3)

(4)

(5)

where t is the goal value, o is the output value, and p is the number of samples. 2 DISCUSSION AND RESULTS 2.1 The Effects of Heat Treatments and Cutting Parameters on the Thrust Forces The variations in thrust force depending on the heat treatment, feed rate and cutting speed in are shown in Fig. 2. As shown in Fig. 2, the thrust force substantially increases with increasing feed rate because the shear area is elevated. In contrast, the thrust force was not considerably affected by the cutting speed within the

Fig. 2. The variations in thrust force depending on the heat treatment, feed rate and cutting speed

494

1/ 2

  

Çiçek, A. − Kıvak, T. − Samtaş, G. − Çay, Y.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

tested cutting range. With increasing cutting speed, the thrust forces for the three different tools decreased until 16 m/min, but slightly increased after this cutting speed. This decrease is due to the softening effect of the work piece at higher cutting speeds. The thrust forces for the CT and CTT drills were less than the CHT drill. This can be attributed to the lower wear rate, lower temperature of the tool tip, and less distortion of the cutting edge of the CT and CTT drills in comparison with the CHT drill during the drilling process [18]. Table 2. Experimental results Cutting speed

Feed rate

12 14 16 18 12 14 16 18 12 14 16 18

0.08 0.08 0.08 0.08 0.1 0.1 0.1 0.1 0.12 0.12 0.12 0.12

CHT 1 Thrust force 680 662 660 668 770 765 756 767 894 871 867 870

Cutting tools CT 2 Thrust force 662 645 640 635 765 757 752 761 855 847 844 848

CTT 3 Thrust force 641 634 628 632 757 742 737 749 846 836 839 842

The uniform distribution of carbide particles, the refinement of small particles, the precipitation of secondary carbides, and the improved toughness led to the CTT drill having the lowest thrust force [2]. The thrust force was also predicted by ANNs and multiple regression analysis, and their affects and reliability were evaluated using statistical methods. The thrust forces measured in the experiments are given in Table 2. The effects of three different cutting tools, cutting speed, and feed rate on the thrust force measured during the drilling experiments were analysed by ANOVA (Analysis of Variance).

2.2 Prediction of Thrust Forces with Multiple Regression Analysis Regression analysis is a statistical tool for evaluating the relationship of one or more independent variables (x1, x2 ,…, xk) to a single, continuous dependent variable (y). This analysis is most often used when the independent variables cannot be controlled, for example, when they are collected in a sample survey or other observational study. As with straightline regression, an ANOVA table can be used to provide an overall summary of a multiple regression analysis. The particular form of an ANOVA table may vary, depending on how the contributions of the independent variables are considered. A simple form reflects a contribution in which all independent variables are considered collectively in order to make the prediction [19]. The analysis of the variance shows that the cutting tools, cutting speed, and feed rate affect the thrust forces within the experimental limits. The obtained ANOVA results are given in Table 3. As shown in Table 3, the feed rate is the most significant factor affecting the thrust force (97.08%). This factor is followed by the cutting tools (1.91%) and the cutting speed (0.23%). The experimental results are evaluated by regression analysis for a significance level of 5% (or 95% confidence level), and regression analysis results are shown in Fig. 3. The mathematical statements established with multiple regression analysis for the thrust force are given in Eq. (6). The coefficient of determination (R2) of the equation was calculated as 0.992:

Ff = 294.861 − 14.458Ct − 1.833V + 5150 f , (6)

where Ff is the thrust force [N], Ct is the cutting tool, V is cutting speed [m/min], and f is the feed rate [mm/ rev]. According to the obtained mathematical model, it was observed that there was a linear relationship between cutting tools, cutting speed, and feed rate with the thrust force.

Table 3. Analysis of variance for thrust force Source Model Cutting Tools Cutting Speed Feed rate Error Total

Degrees of freedom 3 1 1 1 32 35

Sum of square 260238.04 5017.04 605 254616 2010.85 262248.89

Mean square 86746 5.017 605 254616 62.8 -

F value 1380.449 79.8396 9.6278 4051.880 <.0001 -

Pr > F <.0001 <.0001 0.0040 <.0001 -

Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis

Contribution % 1.91 0.23 97.08 0.78 -

495

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

Fig. 3. Comparison of experimental and residual thrust forces with thrust forces predicted by regression analysis

A comparison of the experimental and thrust forces predicted by the multiple regression method is shown in Fig. 4. It was found that the values predicted by the multiple regression method at the feed rates of 0.08 and 0.12 mm/rev were more convergent than the predicted values at 0.1 mm/rev as shown in Fig. 4. This indicates that the effect of the 0.1 mm/rev feed rate in the regression model is less than the other feed rates. 2.3 Prediction of Thrust Forces with ANNs

The training and testing data sets were prepared using the experimental data. In this study, 36 patterns were obtained from the drilling experiments. Among them, 9 patterns were randomly selected and used as the testing data. Cutting speed, feed rate, and cutting tools were used in the input layer, while the thrust force was used in the output layer of the ANN. To obtain accurate results, a single hidden layer with five neurons was used. Table 4 shows the statistical evaluation of results of the ANN with five neurons. The equation of the thrust force is given in Eq. (7). Also, the thrust force can be accurately calculated by the following formula:

1   Ff =  × − ( 3.2694× F1 + 3.0517× F2 + 3.7990× F3 − 7.7062× F4 + 0.0511× F5 + 0.9165 )   1+ e  ×950, (7)

where Fi (i = 1, 2, ..., 4 or 5) can be calculated according to Eq. (8): Fig. 4. Comparison of the actual and predicted thrust forces; a) 0.08 mm/rev, b) 0.1 mm/rev, c) 0.12 mm/rev

496

Çiçek, A. − Kıvak, T. − Samtaş, G. − Çay, Y.

Fi =

1 , 1 + e − Ei

(8)

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

Table 4. Statistical evaluation for the thrust force Goal

Learning algorithm

Number of neurons

SCG

RMSE 0.00769

Training data R2 0.999907

where Ei is the weighted sum of the inputs, and is calculated using the equation in Table 3. The data flow is completed with the weights among the layers. The weight values input and hidden layer are given in Table 5.

MEP 0.725947

RMSE 0.009066

Testing data R2 0.999871

MEP 0.930127

the thrust force values were accurately predicted by the ANN. Fig. 5 shows a comparison of the actual and the predicted thrust force values by the ANN. Table 5. The weights between the input and hidden layers i 1 2 3 4 5

Ei = w1x(Ct/6) + w2x(V/60) + w3x(fx4) + θi w1 w2 w3 -6.7500 -4.4407 -6.8930 -2.3814 -6.5059 -4.7796 -9.1133 -4.9393 6.0572 0.0480 0.1832 -6.4262 3.3620 -8.5526 -2.2916

θi 11.7137 11.8331 -2.7930 3.7340 8.9342

3 CONCLUSIONS In this study, multiple regression and ANN approaches were used to predict the thrust forces in the drilling of AISI 316 stainless steel blocks. The best results were obtained using the SCG algorithm and a network with five hidden neurons. The developed models were evaluated in terms of their prediction capability. In general, the predicted values were found to be close to the experimental values. The proposed models can thus be used effectively to predict the thrust force in such drilling process. The R2 was found to be 0.999 for both the training and testing data in the ANN, while it was 0.992 for the multiple regression analysis. Both methods are suitable for the prediction of the thrust forces within acceptable error limits. In addition, the relationship between cutting parameters and thrust force can be determined by using the ANN. Therefore, the usage of ANNs is highly recommended for the prediction of the thrust force instead of complex and time-consuming experimental studies.

4 REFERENCES c)

Fig. 5. Comparison of the actual and predicted thrust forces; a) 0.08 mm/rev, b) 0.1 mm/rev, c) 0.12 mm/rev

After the training and testing stages, the R2 value for the training data was 0.999907, while the R2 value for the testing data was 0.999871. The RMSE values for the training and testing data were respectively 0.00769 and 0.009066, and the mean error values were 0.725947 and 0.930127, respectively. Therefore,

[1] Simranpreet, G.S., Rupinder, S., Harpreet, S., Jagdev, S. (2009). Wear behaviour of cryogenically treated tungsten carbide inserts under dry and wet turning conditions. International Journal of Machine Tools & Manufacture, vol. 49, no. 3-4, p. 256-260, DOI:10.1016/j.ijmachtools.2008.11.001. [2] Firouzdor, V., Nejati, F., Khomamizadeh, E. (2008). Effect of deep cryogenic treatment on wear resistance and tool life of M2 HSS drill. Journal of Materials Processing Technology, vol. 206, no. 1-3, p. 467-472, DOI:10.1016/j.jmatprotec.2007.12.072.

Modelling of Thrust Forces in Drilling of AISI 316 Stainless Steel Using Artificial Neural Network and Multiple Regression Analysis

497

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, 492-498

[3] Sweeney Jr., T.P. (1986). Deep cryogenics: the great cold debate. Heat Treating, p. 28-32. [4] Smolnikov, E.A., Kossovich, G.A. (1980). Cold treatment of cutting tools., All-Union ScientificResearch Tool Institute, vol. 10, p. 5-7. (in English) [5] Wilkins, C. (1999). Cryogenic processing; the big chill. Electrical Discharge Machining Today, p. 36-44. [6] Huang, J.Y., Zhu, Y.T., Liao, X.Z., Beyerlein, I.J., Bourke, M.A., Mitchell, T.E. (2003). Microstructure of cryogenic treated M2 tool steel. Materials Science and Engineering A, vol. 339, No. 1-2, p. 241-244, DOI:10.1016/S0921-5093(02)00165-X. [7] Da Silva, F.J., Franco, D.D., Machado, A.R., Ezugwu, A.R., Souza Jr.A.M. (2006). Performance of cryogenically treated HSS tools. Wear, vol. 261, p. 674685, DOI:10.1016/j.wear.2006.01.017. [8] Tamas, S. (1999). Cutting force modeling using artificial neural networks. Journal of Materials Processing Technology, vol. 92-93, p. 344-349, DOI:10.1016/S0924-0136(99)00183-1. [9] Hao, W., Zhu, X., Li, X., Turyagyenda, G. (2006). Prediction of cutting force for self-propelled rotary tool using artificial neural networks. Journal of Materials Processing Technology, vol. 180, p. 23-29, DOI:10.1016/j.jmatprotec.2006.04.123. [10] Özel, T., Karpat, Y. (2005). Predictive modeling of surface roughness and tool wearing hard turning using regression and neural networks. International Journal of Machine Tools and Manufacture, vol. 45, no. 4-5, p. 467-479, DOI:10.1016/j.ijmachtools.2004.09.007. [11] Panda, S.S., Chakraborty, D., Pal, S.K. (2007). Monitoring of drill flank wear using fuzzy backpropagation neural network. International Journal Advanced Man ufacturing Technology, vol. 34, no. 3-4, p. 227-235, DOI:10.1007/s00170-006-0589-0. [12] Asiltürk, I., Çunkas, M. (2010). Modeling and prediction of surface roughness in turning operations using artificial neural network and multiple regression

498

method. Expert Systems with Applications, vol. 38, no. 5, p. 5826-5832. [13] Nalbant, M., Gökkaya, H., Toktaş, İ., Sur, G. (2009). The experimental investigation of the effects of uncoated, PVD- and CVD-coated cemented carbide inserts and cutting parameters on surface roughness in CNC turning and its prediction using artificial neural networks. Robotics and ComputerIntegrated Manufacturing, vol. 25, no. 1, p. 211-223, DOI:10.1016/j.rcim.2007.11.004. [14] Lin, W.S., Lee, B.Y. (2001). Modeling the surface roughness and cutting forces during turning. Journal of Materials Processing Technology, vol. 108, no. 3, p. 286-93, DOI:10.1016/S0924-0136(00)00835-9. [15] Karayel, D. (2009). Prediction and control of surface roughness in CNC lathe using artificial neural network. Journal of Materials Processing Technology, vol. 209, no. 7, p. 3125-3137, DOI:10.1016/j. jmatprotec.2008.07.023. [16] Ibrikçi, T., Saçma S., Yıldırım, V., Koca, T. (2010). Application of artificial neural networks in the prediction of critical buckling loads of helical compression springs. Strojniški vestnik - Journal of Mechanical Engineering, vol. 56, no. 6, p. 409-417. [17] Gonçalves, D.V., De Almeida, L.F., Mathias, M.H. (2010). Wear particle classifier system based on an artificial neural network. Strojniški vestnik - Journal of Mechanical Engineering, vol. 56, no. 4 p. 284-288. [18] Sreerama Reddy, T.V., Sornakumar, T., VenkataramaReddy, M., Venkatram, R. (2009). Machinability of C45 steel with deep cryogenic treated tungsten carbide cutting tool inserts. International Journal of Refractory Metals & Hard Materials, vol. 27, p. 181-185. [19] Kleinbaum, D.G., Kupper L.L., Muller, K.E, Nizam, A. (2007). Applied regression analysis and other multivariable methods. Thomson Higher Education, Belmont.

Çiçek, A. − Kıvak, T. − Samtaş, G. − Çay, Y.

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8 Vsebina

Vsebina Strojniški vestnik - Journal of Mechanical Engineering letnik 58, (2012), številka 7-8 Ljubljana, julij-avgust 2012 ISSN 0039-2480 Izhaja mesečno

Razširjeni povzetki člankov Javier López, Jose Breñosa, Ignacio Galiana, Manuel Ferre, Antonio Giménez, Jorge Barrio: Optimizacija mehanske konstrukcije večprstnih haptičnih naprav za aplikacije virtualne manipulacije Andrej Lebar, Luka Selak, Rok Vrabič, Peter Butala: Spremljanje, analiza in brezžično zapisovanje parametrov varilnega procesa v varilski dnevnik Mateja Dovjak, Masanori Shukuya, Aleš Krainer: Eksergijska analiza konvencionalnega in nizkoeksergijskega sistema za ogrevanje in hlajenje skoraj nič-energijskih stavb Mohammad Rabiey, Christian Walter, Friedrich Kuster, Josef Stirnimann, Frank Pude, Konrad Wegener: Poravnavanje CBN-brusnih kolutov s hibridnim vezivom s pomočjo kratkoimpulznega vlakenskega laserja Frane Majić, Ralph Voss, Zdravko Virag: Metoda mejne plasti za nestacionaren transonični tok Jure Marn, Jurij Iljaž, Zoran Žunič, Primož Ternik: Analiza tokovnih razmer pri zdravi aortni zaklopki in njeni regurgitaciji ob upoštevanju koronarnih arterij in nenewtonskega modela krvi Adem Çiçek, Turgay Kıvak, Gürcan Samtaş, Yusuf Çay: Modeliranje podajalnih sil pri vrtanju nerjavnega jekla AISI 316 z uporabo umetnih nevronskih mrež in multiple regresijske analize Osebne vesti Doktorske disertacije, diplomske naloge

SI 89 SI 90 SI 91 SI 92 SI 93 SI 94 SI 95 SI 96

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 89

Optimizacija mehanske konstrukcije večprstnih haptičnih naprav za aplikacije virtualne manipulacije López, J. – Breñosa, J. – Galiana, I. – Ferre, M. – Giménez, A. – Barrio, J. Javier López1,* – Jose Breñosa2 – Ignacio Galiana2 – Manuel Ferre2 – Antonio Giménez1 – Jorge Barrio2 1 Univerza v Almerii, Oddelek za strojništvo, Španija 2 Politehnika v Madridu, Center za avtomatizacijo in robotiko, Španija

Haptične naprave so mehatronski sistemi, ki omogočajo interakcijo z virtualnimi ali oddaljenimi okolji, pri čemer uporabnik občuti sile na svojih prstih. Enoprstne naprave so primerne za enostavne haptične aplikacije, zasnovane za tipanje ali preučevanje obrisov predmetov; za naprednejše naloge navidezne manipulacije pa je potrebnih več prstov. Ena smer razvoja večprstnih haptičnih naprav gre v uporabo večjega števila enoprstnih haptičnih naprav. Glavna slabost takšne konfiguracije so trki med členi, ki občutno zmanjšujejo delovni prostor. V članku je opisana zasnova in optimizacija modularne večprstne haptične naprave za virtualno rokovanje s predmeti, ki zagotavlja učinkovito haptično interakcijo ter hkrati ponuja dober kompromis med stroški in zahtevnostjo naprave. Snovanje takšnih naprav je zahtevna naloga, saj zahteva iskanje kompromisov med več zahtevami: velik delovni prostor, majhna masa in vztrajnost, velika togost, velika nosilnost, izotropne sile in gibanja, odsotnost zračnosti pri spremembi smeri gibanja, majhno trenje, odsotnost singularnosti v delovnem prostoru ter velika pasovna širina. Poveča se tudi zahtevnost zasnove v primerjavi z enoprstnimi haptičnimi napravami, zlasti v zvezi z doseganjem dovolj velikega delovnega prostora, v katerem ne sme prihajati do trkov. Predlagana konstrukcija je zasnovana na modularni in skalabilni konfiguraciji, kjer vsak prst predstavlja ena enota. Vsak modul ima enostavno in kompaktno konstrukcijo, ki se enostavno povečuje in znotraj delovnega prostora omogoča ustvarjanje sile v katerikoli smeri. Vsi moduli so bili optimizirani s kinematičnimi ukrepi kot so delovni prostor naprave, izotropija in nosilnost. Med naštetimi kriteriji je treba poiskati kompromis, zato so bile uporabljene večkriterijske metode optimizacije. Mehanske konfiguracije z dvema ali tremi prsti so zasnovane na eni ali dveh redundantnih oseh. Razmestitev redundantnih osi je bila določena tako, da je dosežen optimiziran delovni prostor brez trkov med deli konstrukcije. Predstavljena je metodologija snovanja večprstnih haptičnih naprav. Osnovo mehanske konstrukcije, ki ima po en modul na vsak prst, je mogoče razširiti v večprstne naprave. Vsak modul je bil optimiziran po večkriterijskih indeksih zmogljivosti. Z ustreznim dodatkom redundantnih prostostnih stopenj je mogoče doseči razširjen delovni prostor brez trkov. V članku sta predstavljeni dve napravi, ki sta zasnovani po opisani metodologiji: dvoprstna in triprstna haptična naprava. Dvoprstna haptična naprava potrebuje eno redundatno prostostno stopnjo, triprstna haptična naprava pa dve redundantni prostostni stopnji. Razmestitev redundantnih osi in optimizirane dimenzije členov zagotavljajo ustrezen delovni prostor, sposobnost manipulacije, sposobnost ustvarjanja sile ter vztrajnost naprave. Glavni prispevek tega članka je v prikazu snovanja večprstnih haptičnih naprav, ki imajo ustrezno zmogljivost, so enostavne in poceni. Določena je optimizirana mehanska konstrukcija enega prsta, ki jo je mogoče razširiti z dodajanjem redundantnih osi za delovni prostor brez trkov. Rešitev ponuja občutno večji delovni prostor v primerjavi z uporabo več povezanih enoprstnih haptičnih vmesnikov, ob primerljivi zahtevnosti rešitve. Sicer obstajajo tudi večprstne robotske naprave, ki nudijo večji delovni prostor, vendar na račun večje zahtevnosti in stroškov. Ključne besede: haptičen, večprstni, virtualna manipulacija, mehanska zasnova, indeksi zmogljivosti, večkriterijska optimizacija

*Naslov avtorje za dopisovanje: Univerza v Almerii, Oddelek za strojništvo, Ctra. Sacramento s/n 04120, La Cañada de San Urbano (Almería), Španija, javier.lopez@ual.es

SI 89

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 90

Spremljanje, analiza in brezžično zapisovanje parametrov varilnega procesa v varilski dnevnik Lebar, A. – Selak, L. – Vrabič, R. – Butala, P. Andrej Lebar* – Luka Selak – Rok Vrabič – Peter Butala Univerza v Ljubljani, Fakulteta za strojništvo, Slovenija

V določenih domenah proizvodnje se soočamo z izredno majhnimi serijami ali celo s posamično izdelavo. Izdelava opreme za hidroelektrarne je značilen primer takšne vrste. V tej domeni je ročno obločno varjenje ena od najbolj časovno potratnih operacij. Zaradi kombinacije izdelave v majhnih serijah in različnih velikosti obdelovancev, je potrebno proizvodnjo organizirati fleksibilno, saj v nasprotnem primeru okoliščine neizogibno vodijo k manjši kakovosti in daljšim čakalnim časom med operacijami. V prispevku je predstavljen nov koncept zajema in shranjevanja podatkov med procesom varjenja, ki omogoča samodejno tvorjenje varilskega dnevnika. Varilski dnevnik je dokument, ki izhaja iz standarda ISO 15609-1 in je namenjen sledenju in zagotavljanju kakovosti zvara. Izpolnjevanje varilskega dnevnika je za varilca zamudno. Poleg tega je bilo ugotovljeno, da je do 15% informacij v varilskih dnevnikih izpolnjenih napačno. Z namenom izboljšanja preglednosti procesa, je razvita metoda za samodejno tvorbo varilskega dnevnika. Avtomatizacija zajema parametrov varilskega procesa je dosežena z vgradnim sistemom, zgrajenim na osnovi mikrokrmilniške platforme in senzorske mreže. Podatki so v realnem času zajeti, obdelani in prenešeni na strežnik. V prispevku je podan pregled komponent vgradnega sistema in razvita metoda za sprotno obdelavo zajetih podatkov na mikrokrmilniku. Napetost med elektrodo in obdelovancem je merjena z uporabo napetostnega delilnika in napetostno izoliranega operacijskega ojačevalnika. Električni tok je merjen s Hallovim polprevodniškim senzorjem, katerega izhodna napetost je proporcionalna merjenem električnem toku. Za sprotno obdelavo signalov sta uporabljena algoritma za sprotni izračun povprečne vrednosti in standardne deviacije. Agregirani podatki so prenešeni na strežnik z uporabo brezžičnega vmesnika, ki temelji na ZigBee (IEEE 802.15.4) protokolu. Prednosti slednjega napram WiFi standardu (IEEE 802.11) so večja zanesljivost prenosa, manjša raba energije in večji doseg signala. Omejitev uporabljenega vmesnika predstavlja nizka hitrost prenosa podatkov, ki je 250 kbit/s, kar pa je za potrebe prenosa agregiranih parametrov dovolj. Predlagani koncept je v okviru študije podprt z dvema zgledoma. V laboratorijskem okolju je preverjeno ujemanje podatkov, zajetih z razvitim vgradnim sistemom, in podatkov zajetih z napravo NI USB 6221. Za eksperimente je uporabljen inverterski varilni stroj. Rezultati potrjujejo primernost razvitega sistema za merjenje električnega toka in napetosti med procesom varjenja. V drugem zgledu je predstavljena industrijska implementacija koncepta avtomatizacije varilskega dnevnika. Pri tem so avtomatsko zaznavani časi začetka in konca varjenja, ter povprečne vrednosti električnega toka in napetosti med varjenjem. Izmerjeni podatki so združeni s podatki sistema vodenja proizvodnje, s čimer je dosežen popoln pregled nad varilskimi operacijami. Ključne besede: obločno varjenje, spremljanje, brezžično senzorsko omrežje, ZigBee komunikacija

SI 90

*Naslov avtorje za dopisovanje: Univerza v Ljubljani, Fakulteta za strojništvo, Aškerčeva 6, 1000 Ljubljana, Slovenija, andrej.lebar@fs.uni-lj.si

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 91

Eksergijska analiza konvencionalnega in nizkoeksergijskega sistema za ogrevanje in hlajenje skoraj nič-energijskih stavb Dovjak, M. – Shukuya, M. – Krainer, A. Mateja Dovjak1,* – Masanori Shukuya2 – Aleš Krainer1

1 Univerza

v Ljubljani, Fakulteta za gradbeništvo in geodezijo, Slovenija za stavbno okolje, Univerza v Tokiu, Japonska

2 Laboratorij

Za izpolnitev zahtev Direktive 2010/31/EU in povečanje števila skoraj nič-energijskih stavb je potreben nov način reševanja problemov v zvezi z visoko rabo energije. Najpomembnejši je celostni pristop, ki vključuje ukrepe na ravni stavbnega ovoja ter ukrepe na ravni učinkovitosti ogrevalnih in hladilnih sistemov. Pri doseganju energijske učinkovitosti stavb ne smemo pozabiti na toplotno ugodje uporabnikov. Doseganje toplotnega ugodja in produktivnosti je pomembnejše od znižane rabe energije. Eksergijske analize obravnavajo procese v človeškem telesu istočasno s procesi, ki potekajo v stavbi. Tako lahko načrtujemo ogrevalne in hladilne sisteme, ki omogočajo zagotavljanje udobnih razmer za individualnega uporabnika ob racionalni rabi energije. Namen študije je primerjava dveh ogrevalnih in hladilnih sistemov (konvencionalni in nizkoeksergijski sistem) z eksergijskega vidika. Vključuje simulacijo toplotnega ugodja za individualne uporabnike in meritve rabe energije. Realni testni prostor je bil opremljen z nizkoeksergijskim sistemom (LowEx-sistem, ogrevalno-hladilni stropni paneli) in nato s konvencionalnim sistemom (električni radiatorji in hladilni sistem z notranjo enoto). Dodatni primer predstavlja toplotno neizolirana testna soba, opremljena s konvencionalnim sistemom. Analiza individualnega toplotnega udobja je bila izvedena s simulacijo eksergijske bilance v človeškem telesu (hbExB), izračunom stopnje rabe eksergije v človeškem telesu (hbExC) in izračunom indeksa PMV. Meritve rabe energije in kontrola mikroklimatskih razmer so bile izvedene s pomočjo integralnega regulacijskega sistema notranjega okolja na osnovi mehke logike (ICsIE sistem). Simulacija je bila izvedena na dveh virtualnih uporabnikih, ki sta bila izpostavljena eksperimentalnim razmeram na osnovi realnih testnih meritev v stavbi. Rezultati simulacije individualnega toplotnega ugodja kažejo, da se stopnja hbExC in indeks PMV spreminjata glede na eksperimentalne razmere, ustvarjene z LowEx in konvencionalnim sistemom. Indeks PMV ne daje dovolj informacij o tem, kateri sistem ustvari toplotno ugodnejše razmere. Rezultati simulacije hbExB podrobno prikazujejo vpliv značilnosti individualnega uporabnika in eksperimentalnih razmer na posamezne dele bilance. Pretekle študije na povprečnih uporabnikih so pokazale, da je stopnja hbExC nižja v toplotno ugodnih razmerah, rezultati študije na individualnih uporabnikih pa kažejo nasprotno. Stopnja hbExC je namreč odvisna od individualnih značilnosti virtualnih uporabnikov. Oba sistema ustvarita toplotno ugodne razmere, vendar mora biti ovoj stavbe dobro toplotno izoliran. V primeru prostora brez toplotne izolacije se pojavi tok hladne sevalne eksergije, ki vodi do toplotnega neudobja. LowEx-sistem, ki je povezan s sistemom ICsIE, omogoča usklajevanje temperature zraka in srednje sevalne temperature za individualnega uporabnika prostora. Na ta način so dosežene toplotno ugodne razmere z regulacijo posameznih delov hbExB. Izmerjena raba energije za ogrevanje testnega prostora je bila pri LowEx-sistemu od 11 do 27% nižja kot pri konvencionalnem sistemu. Izmerjena raba energije za hlajenje je bila pri LowEx sistemu od 41 do 62% nižja. Razlog za manjšo učinkovitost sistema v času ogrevanja je bil v tem, da je bilo za ogrevanje aktiviranih samo 25% površine stropnega elementa. Izračunano zmanjšanje rabe energije v primeru štirikrat večje ogrevalne površine znaša 40%. Pristop istočasne obravnave individualnega toplotnega ugodja in rabe energije, ki je predstavljen v študiji, je pomemben z vidika načrtovanja ogrevalnih in hladilnih sistemov ter za njihovo uporabo v skoraj nič-energijskih stavbah prihodnosti. Ključne besede: eksergija človeškega telesa, ogrevanje in hlajenje stavb, nizkoeksergijski sistem, konvencionalni sistem, raba energije v stavbi, individualno toplotno ugodje

*Naslov avtorja za dopisovanje: Univerza v Ljubljani, Fakulteta za gradbeništvo in geodezijo, Jamova cesta 2, 1000 Ljubljana, Slovenija, Mdovjak@fgg.uni-lj.si

SI 91

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 92

Poravnavanje CBN-brusnih kolutov s hibridnim vezivom s pomočjo kratkoimpulznega vlakenskega laserja

inspire AG, ETH Zurich, Švica 2 IWF, ETH Zurich, Švica

Brusni koluti z novim sistemom hibridnega veziva imajo velik potencial za industrijske aplikacije, saj združujejo prednosti kovinskega veziva (majhna obraba in velika trdnost) in keramičnega veziva (poroznost in dobri žepki za zrna). Problem poravnavanja takšnih visokotrdnih veziv, ki je značilen tudi za kolute s kovinskim vezivom, pa še ni učinkovito rešen. Za kondicioniranje brusov s hibridnim vezivom se trenutno uporabljajo običajni mehanski postopki poravnave s koluti SiC. Kot dobra alternativa se ponuja lasersko poravnavanje. Postopki poravnavanja z laserjem se delijo na dve glavni vrsti. Prvi postopek je lasersko poravnavanje, pri katerem laserski žarek deluje na površino brusnega koluta (tangencialno, radialno ali pod kotom) in ga poravna. Uporaba drugih orodij za poravnavanje ni predvidena. Druga metoda je lasersko podprto poravnavanje, pri katerem laserski žarek deluje na brusni kolut, ga lokalno segreje in tako zmanjša trdnost veziva. Material se odstranjuje sočasno s konvencionalnim orodjem za poravnavanje. V tej študiji je bila uporabljena prva metoda laserskega poravnavanja. Za preučitev učinka parametrov poravnave na rezalne sile, hrapavost obdelovanca in obrabo CBN-brusnega koluta s hibridnim vezivom je bila uporabljena sistematična raziskovalna analiza. V članku je predstavljenih nekaj rezultatov primerjave med konvencionalnim brusnim kolutom, poravnanim s SiC, in brusnim kolutom, poravnanim s kratkimi impulzi vlaknenega laserja. Rezultati kažejo velik tehnološki potencial postopka laserskega poravnavanja v primerjavi s konvencionalno poravnavo. Hibridno vezivo je mogoče poravnati z laserjem brez negativnih učinkov na zmogljivost zrn iz materiala CBN. Največja prednost metode laserske poravnave v primerjavi s konvencionalnim postopkom je v manjših brusilnih silah in specifični energiji, ob enaki površinski hrapavosti ter manjši skupni obrabi brusnega orodja (obraba zaradi poravnave plus obraba zaradi brušenja). Upoštevati pa je treba tudi večjo obrabo brusnega koluta pri brušenju ter večjo površinsko hrapavost brušene površine. Z ustrezno izbiro parametrov procesa je mogoče doseči sprejemljivo zmogljivost ter učinkovit in zanesljiv proces laserske poravnave. Potrebne so dodatne raziskave za preučitev učinka različnih parametrov laserja na učinkovitost in kakovost procesa poravnavanja. Natančno je treba ovrednotiti tudi ekonomski vidik uporabe laserske poravnave z vidika naložb v opremo in časa poravnave. Ti naslednji koraki raziskav se že izvajajo. Prednosti laserskega postopka morajo upravičiti ne le naložbo v laserski sistem, temveč tudi tehnični vidik integracije laserskega sistema z brusilnim strojem. Naslednji izziv je optimizacija postopka laserske poravnave za izboljšanje učinkovitosti procesa ter razširitev možnih aplikacij tako s tehničnega kot s komercialnega vidika. Ključne besede: laser, poravnavanje, brušenje, kondicioniranje, CBN, hibridno vezivo

SI 92

*Naslov avtorja za dopisovanje: inspire AG, ETH Zurich, Tannenstrasse 3, 8092 Zurich, Švica, rabiey@gmail.com

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 93

Metoda mejne plasti za nestacionaren transonični tok

Majić, F. – Voss, R. – Virag, Z. Frane Majić1,* – Ralph Voss2 – Zdravko Virag1 1 Fakulteta za strojništvo in ladjedelništvo, Univerza v Zagrebu, Hrvaška 2 Institut za aeroelastičnost, nemški letalski in vesoljski center, Nemčija Pojav drhtenja (flutter) pri letalih je aeroelastični problem, ki je odvisen od interakcije elastičnih, blažilnih in vztrajnostnih sil konstrukcije ter nestacionarnih aerodinamičnih sil, ki jih ustvarja oscilatorno gibanje same konstrukcije. Najzahtevnejši je izračun nestacionarnih aerodinamičnih obremenitev, zato je bila razvita učinkovita metoda za napovedovanje aerodinamičnih sil. Cilj raziskave je bil predstaviti metodo viskozno-neviskozne interakcije za določevanje nestacionarnih aerodinamičnih obremenitev. Metoda viskozno-neviskozne interakcije se uporablja v preliminarni fazi konstrukcije letala, kjer je treba preizkusiti več konfiguracij. Rezultati so primerjani z eksperimentalnimi podatki in rezultati izračunov po Reynoldsovo povprečenih Navier-Stokesovih enačbah. Predstavljena je dvodimenzionalna metoda nestacionarne interakcije z uporabljenim modelom prehoda mejne plasti in konceptom hitrosti transpiracije kot sredstvom interakcije. Neviskozna okolica aerodinamičnega profila je izračunana z Eulerjevimi enačbami, območje viskozne mejne plasti pa z integralnimi enačbami mejne plasti. Eulerjeve enačbe so bile rešene s pomočjo C-mreže na telesu ter z uporabo Van Leerove delitve vektorja pretoka z natančnostjo drugega reda. Eulerjeve enačbe so bile eksplicitno rešene z natančnostjo prvega reda v času. Integralne enačbe mejne plasti so bile rešene z metodo Runge-Kutta četrtega reda. Hitrost transpiracije je predstavljala vpliv zgoščevanja mejne plasti in je bila uveljavljena na konturi aerodinamičnega profila v Eulerjevem programu za reševanje. Na meji oddaljenega polja je bil uporabljen značilni pogoj meje. Za določitev začetka prehoda v mejni plasti je bila uporabljena metoda en. Viskozno-neviskozna sklopitev je izvedena v neposrednem načinu, kjer se viskozni in neviskozni računi izvajajo zaporedoma. Izračunani viskozno-neviskozni rezultati so bili primerjani z Reynoldsovo povprečenimi Navier-Stokesovimi (RANS) izračuni ter z eksperimentalnimi podatki. Izračunani viskozno-neviskozni rezultati kažejo prednosti v primerjavi z izračuni RANS z ozirom na računski čas, zahtevnost priprave in zahtevano velikost mreže. Študija konvergence je pokazala, da 9.600 kontrolnih volumnov zadostuje za viskozno-neviskozne izračune, medtem ko se izračuni RANS izvajajo na mreži s 50.000 kontrolnimi volumni. Procesorski čas, potreben za izračune RANS, se močno razlikuje od potrebnega časa za viskozno-neviskozne izračune. Prikazani so nestacionarni rezultati za aerodinamični profil NACA64A010 pri transoničnih hitrostih s pojavom udarnega vala. Aerodinamični profil je izvajal harmonično gibanje in udarni val se je premikal v skladu s tem gibanjem. Rezultati tlačnih koeficientov pri viskozno-neviskozni metodi so se delno ujemali z eksperimentalnimi podatki pri večini faznih kotov, medtem ko so se izračunani rezultati RANS dobro ujemali pri vseh faznih kotih. Položaj udarnega vala je večinoma dobro napovedan. Vključitev mejne plasti v nestacionarno Eulerjevo metodo se je izkazala kot natančnejša metoda za določanje nestacionarnih aerodinamičnih obremenitev. Metoda daje rezultate skoraj enake natančnosti kot višji matematični modeli kot je RANS, računski čas je krajši in zahteve glede strojne opreme so manjše. Uporabljena metoda izkazuje oscilatorno vedenje v primerih z majhnim ločilnim mehurčkom za udarnim valom in divergenco v primerih z ločitvijo. Takšno vedenje je pričakovano zaradi neposredne viskozno-neviskozne sklopitve. Boljše vedenje v primerih z majhnim ločilnim mehurčkom bi bilo mogoče doseči z močno viskoznoneviskozno sklopitvijo, kjer bi se enačbe za viskozni tok in enačbe za neviskozni tok reševale sočasno v istem sistemu. Članek prikazuje uspešno uporabo metode prehoda na osnovi Orr-Sommerfeldove enačbe v nestacionarnih aerodinamičnih simulacijah. Prikazana je tudi uporaba koncepta hitrosti transpiracije pri vključitvi mejne plasti v nestacionarne aerodinamične simulacije. Ključne besede: nestacionaren transonični tok, viskozno-neviskozna sklopitev, aerodinamični profil, hitrost transpiracije, napovedovanje prehoda

*Naslov avtorja za dopisovanje: Univerza v Zagrebu, Fakulteta za strojništvo in ladjedelništvo, Ivana Lučića 5, Zagreb, Hrvaška, frane.majic@fsb.hr

SI 93

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 94

Analiza tokovnih razmer pri zdravi aortni zaklopki in njeni regurgitaciji ob upoštevanju koronarnih arterij in nenewtonskega modela krvi Marn, J. – Iljaž, J. – Žunič, Z. – Ternik, P.

Jure Marn – Jurij Iljaž* – Zoran Žunič – Primož Ternik Univerza v Mariboru, Fakulteta za strojništvo, Slovenija

Prispevek obravnava numerično analizo tokovnih razmer pri zdravi aortni zaklopki in pri njeni regurgitaciji, z namenom določitve njenih vplivov na delovanje srca, vplivov koronarnega pretoka krvi na tokovne razmere okoli zaklopke in morebitne kalcinacije, ter makroskopske določitve posameznih vplivov, kot so zmanjšanje koronarnega pretoka, padec tlaka v aorti in povečanje volumna levega ventrikla. Zastavljeni problem določitve tokovnih razmer pri zdravi in oboleli aortni zaklopki z upoštevanjem koronarnih arterij je bil rešen numerično z uporabo programskega paketa za računalniško dinamiko tekočin Ansys CFX, ki temelji na metodi končnih volumnov. Problem je bil obravnavan časovno odvisno za čas zaprtja zaklopke oziroma diastole, kjer je fizikalni pojav regurgitacije oziroma insuficience in pretoka krvi skozi koronarne arterije najbolj izrazit. Geometrija, ki popisuje okolico aortne zaklopke (aortna zaklopka, del levega ventrikla, aorte ter leve in desne koronarne arterije), je bila zato privzeta in obravnavana kot neelastična. Geometrija zaklopke in njene okolice je bila določena na osnovi veljavnih povprečnih vrednosti, pridobljenih iz člankov in druge literature. Reološke lastnosti krvi so bile popisane z nenewtonskim potenčnim modelom, ki upošteva povečanje dinamične viskoznosti pri zmanjšanju strižnih napetosti in tako realneje opiše obnašanje toka krvi. Okolica območja reševanja je bila nadomeščena s primernimi robnimi pogoji, pridobljenimi na osnovi veljavnih povprečnih tokovno-tlačnih vrednosti zdravega srca kot so potek aortnega tlaka, število utripov, povprečna vrednost koronarnega pretoka itd. V ta namen so bile na novo določene nekatere odvisnosti oziroma robni pogoji, s katerimi je mogoče določiti vplive obolele zaklopke na delovanje srca. Prispevek podaja postavitev numeričnega modela za določitev stopnje regurgitacije ter izračun koronarnega pretoka, volumna zatekajoče krvi in padca aortnega tlaka. Rezultati numeričnih simulacij v primeru zdrave aortne zaklopke kažejo na laminarno tokovno polje, tako v koronarnih arterijah, kot tudi v okolici zaklopke, in s tem na mirujoče področje okoli lističev zaklopke. Razlog za mirujoče področje je velika dinamična viskoznost krvi, ki je posledica reoloških značilnosti in premajhnega vpliva pretoka koronarnih arterij. Rezultati v primeru regurgitacije zaklopke, kjer nastane zaradi povečanja premera aorte odprtina velikosti 3,8 mm in kri med diastolo zateka v levi ventrikel, kažejo na močan upad tlaka v aorti, kot tudi koronarnega pretoka. Pri istem uporu miokardne mišice se koronarni pretok skozi levo in desno koronarno arterijo zmanjša za 31% glede na normalni pretok, kar vodi v zmanjšanje dobave kisika in hranilnih snovi. Prav tako pa se mora ventrikel zaradi zatekajoče krvi močno povečati, za 56,1 ml, kar kaže na močno obremenitev ventrikla in kritično regurgitacijo. Zaradi odprtine in zatekanja krvi v ventrikel nastane močan curek, ki še dodatno obremenjuje zaklopko ter lahko zaradi velikih strižnih napetosti povzroči nepravilno delovanje zaklopke oziroma njene poškodbe. Prav tako, kot v primeru zdrave zaklopke, se v zgornjem območju zaradi prej omenjenih razlogov pojavlja območje mirovanja krvi, kar lahko vodi v kalcinacijo zaklopke. Regurgitacija ima tako v danih primerih velik vpliv na delovanje srca, posledično pa tudi na zdravje človeka. Rezultati numeričnega modela veljajo le za privzete omejitve, predpostavke in postavljene modele. Za realnejši opis stanja oziroma dogajanja bi bilo potrebno v nadaljevanju nadgraditi modele uporabljenih robnih pogojev, preizkusiti različne nenewtonske modele krvi, upoštevati elastičnost zaklopke itd. S tako postavljenim numeričnim modelom je mogoče določiti vpliv obolele zaklopke na delovanje srca in zdravje človeka ter vnaprej napovedati kritično stanje, ko srce ne zmore več kompenzirati obolenja in je zato potrebna operacija. Izvirnost prispevka se kaže tako v numerični določitvi posledic obolele zaklopke, kot so na primer sprememba aortnega tlaka, padec koronarnega pretoka in povečanje levega ventrikla, kakor tudi v postavitvi matematičnega modela za določitev koronarnega pretoka, padca aortnega tlaka in izračuna zatekajočega volumna krvi. Predstavljeno delo predstavlja pomemben vir podatkov tako medicini kot inženirjem na področju biomedicine. Avtorji niso našli podobnih raziskav, kar povečuje pomembnost prispevka. Ključne besede: aortna zaklopka, regurgitacija, insuficienca, koronarna arterija, računalniška dinamika tekočin, metoda končnih volumnov, nenewtonska tekočina SI 94

*Naslov avtorja za dopisovanje: Univerza v Mariboru, Fakulteta za strojništvo, Smetanova ulica 17, 2000 Maribor, Slovenija, jurij.iljaz@uni-mb.si

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 95

Modeliranje podajalnih sil pri vrtanju nerjavnega jekla AISI 316 z uporabo umetnih nevronskih mrež in multiple regresijske analize Çiçek, A. − Kıvak, T. − Samtaş, G. − Çay, Y. Adem Çiçek1 − Turgay Kıvak2 − Gürcan Samtaş2,* − Yusuf Çay3 1Univerza

Düzce, Tehniška fakulteta, Turčija Düzce, Visoka poklicna šola Cumayeri, Turčija 3Univerza Karabük, Fakulteta za strojništvo, Turčija

2 Univerza

Cilj predstavljene študije je raziskava vpliva rezalnih parametrov (t. j. rezalne hitrosti, podajanja) in globoke kriogene obdelave na podajalno silo (Ff) pri vrtanju v nerjavno jeklo AISI 316. Vijačni svedri iz hitroreznega jekla M35 so bili za preizkus vplivov globoke kriogene obdelave na podajalne sile podvrženi 24-urni kriogeni obdelavi pri temperaturi –196 °C, po običajni toplotni obdelavi pa so bili še 2 h popuščani pri temperaturi 200 °C. Članek obravnava napovedovanje podajalne sile pri vrtanju v bloke iz nerjavnega jekla AISI 316 z umetnimi nevronskimi mrežami (UNM) in multiplo regresijsko analizo. Za učenje in preizkušanje UNM sta bila uporabljena algoritem učenja s skaliranim konjugiranim gradientom (SCG) in logistična sigmoidna prenosna funkcija, za regresijsko analizo pa je bil uporabljen linearni model. Rezultati eksperimentov kažejo, da so podajalne sile najmanjše pri kriogeno obdelanih in popuščanih svedrih. Rezultati UNM kažejo, da daje učni algoritem SCG s petimi nevroni v skriti plasti za podatke učenja oz. preizkušanja vrednosti determinacijskega koeficienta (R2) 0,999907 oz. 0,999871. Srednja kvadratna napaka (RMSE) podatkov učenja oz. preizkušanja je bila 0,00769 oz. 0,009066, srednja odstotna napaka (MEP) pa 0,725947 oz. 0,930127. Rezultati eksperimentov so potrjeni z regresijsko analizo s stopnjo signifikantnosti 5% (oz. stopnjo zaupanja 95%), izračunani determinacijski koeficient (R2) regresijske analize pa je bil 0,992. Razviti modeli so bili vrednoteni z vidika sposobnosti napovedovanja. Napovedane vrednosti se v splošnem dobro ujemajo z eksperimentalnimi vrednostmi. Predlagane modele je zato mogoče učinkovito uporabiti za napovedovanje podajalne sile pri takšnih procesih vrtanja. Obe metodi sta primerni za napovedovanje podajalnih sil v okviru sprejemljive napake. Analiza variance je pokazala, da je vpliv rezalnih orodij, rezalne hitrosti in podajanja na podajalne sile znotraj eksperimentalno določenih meja. Po rezultatih ANOVA je podajanje najbolj signifikanten faktor vpliva na podajalno silo (97,08%). Sledita mu faktorja rezalnih orodij (1,91%) in rezalne hitrosti (0,23%). Eden najpomembnejših dejavnikov pri proizvodnih stroških je raba energije. Rabo energije določa moč, ki je potrebna za obdelavo. Podajalna sila (Ff) pri vrtanju je odvisna od specifičnega upora materiala pri odrezavanju in je najpomembnejši parameter, ki določa porabo moči in strošek energije. Umetne nevronske mreže (UNM) je mogoče uspešno uporabiti pri različnih industrijskih aplikacijah. UNM so primerne za modeliranje različnih proizvodnih funkcij, saj se lahko enostavno priučijo kompleksnih nelinearnih in multivariabilnih razmerij med procesnimi parametri. Uporaba nerjavnih jekel in kriogenske obdelave v industriji postaja vse pomembnejša. Tipična področja uporabe nerjavnih jekel so v kemični industriji, procesni opremi (npr. črpalke, armature itd.), v prehrambni in farmacevtski industriji, pri dejavnostih na morju in v pohištveni industriji. Za proizvodno industrijo je zelo pomembna tudi določitev optimalne kombinacije parametrov vrtanja, zmanjševanje stroškov in skrajšanje časa obdelave. Ključne besede: umetne nevronske mreže, regresijska analiza, kriogena obdelava, obdelava z odrezavanjem, podajalna sila, prediktivno modeliranje

*Naslov avtorja za dopisovanje: Univerza Düzce, Visoka poklicna šola Cumayeri, 81700, Cumayeri, Düzce, Turčija, gurcansamtas@duzce.edu.tr

SI 95

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 96-100 Osebne objave

Doktorske disertacije, diplomske naloge DOKTORSKE DISERTACIJE Na Fakulteti za strojništvo Univerze v Ljubljani so z uspehom obranili svojo doktorsko disertacijo: dne 1. juniija 2012 Janez KRIŽAN z naslovom: »Laserska fluorescenčna termometrija s sintetiziranimi oksidnimi mono- in nanokristali« (prof. dr. Janez Možina, somentor: izr. prof. dr. Ivan Bajsić); Disertacija prikazuje razvoj merilnih metod in opreme za sintezo in uporabo kristalov v termometriji in optični tehniki. V tem delu razvita oprema za sintezo monokristalov temelji na Verneuilovem postopku, oprema za sintezo nanokristalov pa na razpršilno plamenski pirolizi in zgorevalni sintezi. Kombinacija enih in drugih kristalov za uporabo v zaznavalni in fluorescenčni tehniki je vezana na vrsto aplikacije. Uporaba nano materialov je tudi posledica potrebe, da se nekatere pereče probleme pri uporabi monokristalov za optične namene reši z uporabo sintranih nanomaterialov. S tem se odpira še veliko večje področje možnosti dodajanja primesi in s tem lastnosti, ki jih v monokristalih ni mogoče doseči. Oprema, ki bo omogočala proizvodnjo takih materialov v laboratorijskem pilotnem in industrijskem merilu, je nastala ob razvojnoraziskovalnem delu v pričujoči nalogi. Drugo področje dela se nanša na uporabo sintetiziranih kristalov v optični flourescenčni termometriji. Razvili smo merilno opremo in algoritme za merjenje temperature na fluorescenčnem načelu z lastnimi sintetiziranimi kristali in nano materiali. Ta oprema je primerna za zahtevne aplikacije v različnih tehnologijah in medicini. Eksperimentalno delo je pokazalo, da je mogoče razviti ciljne fluorescenčne materiale, ki so prilagojeni specifični uporabi od inteligentnih ognjeodpornih materialov do termometrije na ravni nanokristalov. Bistvena je tudi prilagoditev zaznaval na različne svetlobne vie, ki so pogojeni z zahtevami uporabe. Zato smo sintetizirali veliko število materialov, ki vsak zase lahko služijo v določenih specifičnih namenih glede na temperaturni obseg, vir osvetljevnja, obstojnost materiala in merilno ločljivost. Pomemben prispevek disertacije je tudi razvoj tehnologije na področju naprav za polindustrijsko proizvodnjo nano prahov in tehnologije sušenja po načelu termoakustičnega reaktorja; dne 19. junija 2012 Jaka TUŠEK z naslovom: »Termomagnetne lastnosti aktivnega magnetnega regeneratorja« (mentor: prof. dr. Alojz Poredoš, somentor: doc. dr. Andrej Kitanovski); V doktorskem delu je predstavljena numerična in eksperimentalna analiza termo-magnetnih lastnosti SI 96

aktivnega magnetnega regeneratorja, kot osnovnega elementa magnetnega hladilnika. V ta namen je bil razvit eno-dimenzionalen, časovno odvisen numerični model, ki temelji na energijski enačbi magnetokaloričnega materiala in fluida za prenos toplote. Na osnovi razvitega numeričnega modela je bila opravljena parametrična analiza različnih geometrijskih lastnosti aktivnega magnetnega regeneratorja, različnih krožnih procesov in različnih magnetokaloričnih materialov. Kot primerjalni kriterij analize sta bila uporabljena specifična hladilna moč in hladilno število (COP) pri določenem temperaturnem razponu. Drugi del doktorske naloge je namenjen eksperimentalni analizi aktivnega magnetnega regeneratorja. Izdelana je bila eksperimentalna proga, ki temelji na recipročnem gibanju strukture iz permanentnih magnetov in sistemu za prečrpavanje delovnega fluida preko aktivnega magnetnega regeneratorja. V okviru doktorskega dela je bilo testiranih šest različnih aktivnih magnetnih regeneratorjev. Ti so kot magnetokalorični material vsebovali različne geometrijske oblike gadolinija. Ugotovljeno je bilo dobro ujemanje numerično izračunanih in izmerjenih trendov odvisnosti temperaturnega razpona in hladilne moči. Geometrija aktivnega magnetnega regeneratorja se je izkazala kot ključen dejavnik za doseganje večjih temperaturnih razponov magnetne hladilne naprave; dne 20. junija 2012 Rok VRABIČ z naslovom: »Krmiljenje avtonomnih obdelovalnih sistemov v proizvodnem okolju« (mentor: izr. prof. dr. Peter Butala); Osnovni problem sodobne proizvodnje je, kako zagotavljati stabilne in predvidljive izhode v vedno bolj dinamičnem in nepredvidljivem okolju. Pri tem se je potrebno soočati s problemi naraščajoče kompleksnosti na vseh nivojih, od izdelkov do procesov. Ključno vprašanje je, kako to kompleksnost obvladovati in kako z boljšim obvladovanjem kompleksnosti zagotavljati trajno konkurenčno prednost. Disertacija gradi na hipotezi, da je boljše obvladovanje kompleksnosti mogoče doseči z uvedbo principov avtonomnega krmiljenja. Z namenom obravnave hipoteze je oblikovan nov model krmiljenja, ki temelji na teoriji proizvodne kibernetike. Obravnavana je kompleksnost v proizvodnih sistemih in vpeljana nova mera za operativno kompleksnost. Razvit je model večagentskega sistema, ki definira osnovni nivo avtonomije v proizvodnih sistemih in služi kot podporna informacijska infrastruktura. Predstavljena je nova metoda strukturiranja

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 96-100

proizvodnega sistema v zaokrožene avtonomne enote – avtonomne delovne sisteme; dne 31. maja 2012 Rok ZUPANČIČ z naslovom: »Strukturiranje avtonomnih delovnih sistemov« (mentor: prof. dr. Alojzij Sluga); V delu je obravnavano strukturiranje proizvodnih sistemov, kar razumemo kot določanje povezav in relacij med funkcionalnostmi. Predlagan je koncept proizvodno orientirane storitvene mreže, v kateri so posamezne proizvodne funkcionalnosti realizirane s storitvami. Storitve se izvajajo s strani storitvenih enot, ki na ta način podpirajo operacije v avtonomnih delovnih sistemih. Razvit je konceptualni okvir za podporo proizvodnim operacijam, ki vključuje potrebne elemente tako za funkcioniranje storitvene mreže kakor tudi za razvoj in implementacijo storitvene podpore. Strukturiranje je osredotočeno na kontrolo kakovosti v proizvodnji, ki predstavlja potencialno področje za uvedbo storitvene podpore. Z uporabo konceptualnega okvirja za podporo proizvodnim operacijam je razvita in prototipno implementirana storitvena podpora kontroli kakovosti na proizvodnem nivoju. * Na Fakulteti za strojništvo Univerze v Mariboru sta z uspehom obranila svojo doktorsko disertacijo: dne 12. junija 2012 Franci GAČNIK z naslovom: »Računalniško modeliranje zobnih vsadkov« (mentor: prof. dr. Zoran Ren); Znano je, da najoptimalnejši naklon zobnega vsadka sovpada z naklonom korenine naravnega zoba, ki je bila nadomeščena z zobnim vsadkom. Pomanjkanje kostne mase (atrofija kosti), zlasti v zadnjih segmentih spodnje in zgornje čeljusti, ter anatomske omejitve, kot so maksilarni sinusi, nosna votlina in mandibularni živec, pa običajno zahtevajo poševno vgradnjo zobnih vsadkov. Vsak takšen naklon zobnega vsadka povzroča nastanek večosnih obremenitev vsadka, kar povzroča resorpcijo marginalne kosti in tako sčasoma izgubo zobnega vsadka. V okviru te doktorske disertacije so bili z namenom določitve mehanskega odziva kostnega tkiva razviti različni natančni numerični modeli vseh sestavnih elementov konstrukcije implantatno nošene zobne prevleke, podprte z zobnim vsadkom, ki je lahko vstavljen v čeljustno kost na mesto spodnjega levega prvega kočnika pod različnimi koti. Začetna geometrija v simulacijah uporabljenih modelov zobnih vsadkov in opornikov zobnih prevlek je bila posneta po dejanskih tovrstnih izdelkih z imenom Ankylos Plus, proizvajalca Dentsply. Pri analizah je

bila upoštevana tudi anizotropnost kostnega tkiva. Kostno tkivo je bilo modelirano kot linearno elastični materialni model, pri čemer so bile elastične lastnosti tkiva odvisne od kostne gostote, ki je izražena v enotah Hounsfield-a. Rezultati parametričnih računalniških simulacij kažejo, da je način modeliranja kostnega tkiva z elastičnim, od kostne gostote odvisnim ortotropnim materialnim modelom najprimernejši, saj je v njem preko kostne gostote neposredno upoštevan vpliv stanja kostnega tkiva. Nadalje se je izkazalo, da je trdnost v simulacijah uporabljenega zobnega vsadka pri večjih vrednostih okluzijskih sil nezadostna, saj v tem primeru pride do njegove plastične deformacije. Najpomembnejša ugotovitev je, da obstaja pri uporabi obravnavanega zobnega vsadka resna nevarnost trajne poškodbe kostnega tkiva pod vplivom vseh treh, v tej nalogi modeliranih vrednosti okluzijskih sil. Tako lahko na osnovi računalniških simulacij sklepamo, da obstoječa geometrija obravnavanega zobnega vsadka pri večjih vrednostih okluzijskih sil ni primerna, zato je v takšnih primerih priporočljivo uporabiti večji dimenzijski razred zobnega vsadka ali celo vsadek z drugačno geometrijo. dne 13. junija 2012 Urška SANCIN z naslovom: »Model inteligentnega sistema za podporo odločanju pri izbiri polimernih materialov v procesu razvoja izdelkov« (mentor: izr. prof. dr. Bojan Dolšak); Proces razvoja izdelka je kompleksen proces, znotraj katerega mora inženir sprejemati mnogo pomembnih odločitev, ki se ne nanašajo le na konstrukcijo, ampak na vse faze dobe trajanja izdelka. Izbira materiala je v sklopu procesa konstruiranja izrednega pomena, saj izbrani material v začetnih fazah razvoja izdelka znatno vpliva na vse prihodnje aktivnosti v procesu, ter na posledice, ki jih ima izdelek na okolico v njegovi celotni dobi trajanja. Pri svojem delu konstrukterji mnogokrat naletijo na dileme v procesu odločanja, pri čemer so mladi, neizkušeni inženirji na začetku kariere, kot tudi mala in srednje velika podjetja, v zapostavljenem položaju, saj pomanjkanje izkušenj posledično pomeni slabši izdelek ali najem strokovnjaka. Vpliv izdelka na okolje je definiran z različnimi parametri ekološkega spektra, ki jih lahko nadzorujemo v procesu izbire materiala. Eden takih je recikliranje, ki je pereča problematika polimernih materialov, saj je delež reciklirane plastike bistveno manjši od deleža drugih materialov. Kvarni vpliv človeka na okolje in izkoriščanje neobnovljivih virov tako kažeta na pomembnost težnje k trajnostnemu razvoju. Z razvojem modela sistema za podporo odločanju pri izbiri polimernih materialov z upoštevanjem okoljskega vidika smo v sklopu doktorske disertacije skušali doprinesti k znanosti SI 97

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 96-100

tega področja ter predvsem razviti računalniško podporo pri konstruiranju okolju prijaznih izdelkov iz polimernih materialov v inženirski praksi. DIPLOMIRALI SO Na Fakulteti za strojništvo Univerze v Ljubljani so pridobili naziv univerzitetni diplomirani inženir strojništva: dne 19. junija 2012: Simon KAPELJ z naslovom: »Razvoj hibridne tehnologije na področju preoblikovanja gradiv« (mentor: doc. dr. Tomaž Pepelnjak, somentor: prof. dr. Karl Kuzman); Uroš PLAZNIK z naslovom: »Eksperimentalna analiza magnetnih termodinamičnih krožnih procesov« (mentor: prof. dr. Alojz Poredoš, somentor: doc. dr. Andrej Kitanovski); Nejc ZOR z naslovom: »Analiza vplivnih parametrov na proces sreženja uparjalnika toplotne črpalke« (mentor: prof. dr. Alojz Poredoš); Drejc ŽNIDARŠIČ z naslovom: »Uvodne raziskave termoformiranja biopolimernih kompozitov« (mentor: prof. dr. Karl Kuzman); Matej MARTELANC z naslovom: »Prototip ergonomsko oblikovane ročke za rehabilitacijo po možganski kapi« (mentor: izr. prof. dr. Slavko Dolinšek, somentor: prof. dr. Janez Kopač); Luka RUS z naslovom: »Razvoj naprave za valjanje navojev« (mentor: prof. dr. Marko Nagode); Blaž SAMBOL z naslovom: »Karakteristični kazalniki toplarniškega postrojenja skladno s standardom SIST EN 16001« (mentor: izr. prof. dr. Mihael Sekavčnik); Miha SPRINČNIK z naslovom: »Modeliranje obremenitvenih stanj z mešanimi gostotami porazdelitve verjetnosti« (mentor: prof. dr. Marko Nagode); dne 22. junija 2012: Barbara BASTAR z naslovom: »Zasnova numeričnega modela umetne žile« (mentor: prof. dr. Boris Štok, somentor: izr. prof. dr. Ivan Bajsić); Mitja DOLENC z naslovom: »Razvoj filtra za uničevanje mikroorganizmov v sistemih vodne hidravlike« (mentor: doc. dr. Jernej Klemenc, somentor: doc. dr. Polona Zalar); Tjaš ŽIGANTE z naslovom: »Hiperelastični konstitutivni modeli za popis obnašanja nelinearno elastičnih materialov« (mentor: prof. dr. Boris Štok, somentor: doc. dr. Nikolaj Mole); Boris KAVDIK z naslovom: »Sistem za avtomatsko karakterizacijo laserskih izvrtin« (mentor: prof. dr. Janez Diaci); SI 98

Blaž PETEK z naslovom: »Spajanje aluminijevih zlitin z jeklom« (mentor: prof. dr. Janez Tušek, somentor: doc. dr. Damjan Klobčar); Boštjan ŠPOLJAR z naslovom: »Eksperimentalno modeliranje tokovnih razmer za prezračevalnimi ventilatorji na modelu garažne hiše« (mentor: prof. dr. Branko Širok, somentor: doc. dr. Tom Bajcar); dne 26. junija 2012: Sebastjan BOŠTJANČIČ z naslovom: »Laserska aktuacija tekočekristalnih elastomerov« (mentor: doc. dr. Matija Jezeršek); Simon GOLOB z naslovom: »Univerzalno avtodvigalo« (mentor: doc. dr. Boris Jerman, somentor: prof. dr. Marko Nagode); Tjaša OŠTIR z naslovom: »Zmanjšanje toplotnih izgub in skrajšanje transportnih časov v vročevodnem distribucijskem omrežju« (mentor: prof. dr. Iztok Žun); Rok VIHAR z naslovom: »Optimizacija energijske učinkovitosti in izpustov onesnažil osebnega vozila« (mentor: izr. prof. dr. Tomaž Katrašnik); dne 27. junija 2012: Gregor CINDRIČ z naslovom: »Deterministično napovedovanje materialnih potreb« (mentor: prof. dr. Marko Starbek, somentor: izr. prof. dr. Janez Kušar); Luka DUHOVNIK z naslovom: »Ekonomska upravičenost projekta« (mentor: prof. dr. Marko Starbek, somentor: izr. prof. dr. Janez Kušar); Jure HROVATIN z naslovom: »Trajnostni test vodnega hidravličnega agregata« (mentor: doc. dr. Jožef Pezdirnik); Jože KAMENŠEK z naslovom: »Zasnova tihega ventilatorja« (mentor: prof. dr. Mirko Čudina, somentor: doc. dr. Jurij Prezelj); Boštjan VELIČEVIČ z naslovom: »Razvoj dvostopenjske ročne vodne črpalke« (mentor: doc. dr. Jožef Pezdirnik). * Na Fakulteti za strojništvo Univerze v Mariboru sta pridobila naziv univerzitetni diplomirani inženir strojništva: dne 7. junija 2012: Niko ŠARUGA TURINEK z naslovom: »Zasnova solarne ulične svetilke« (mentor: izr. prof. Pogačar Vojmir); dne 28. junija 2012: Goran ĐUKIĆ z naslovom: »Inovativen koncept naslona za roke pri zadnjih avtomobilskih sedežih« (mentor: doc. dr. Nataša Vujica Herzog, somentor: izr. prof. Vojmir Pogačar); Branko EKSELENSKI z naslovom: »Obnova kolesnih dvojic tirnih vozil« (mentor: prof. dr. Jože Balič, somentor: prof. dr. Zoran Ren);

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 96-100

Tadej HOLLER z naslovom: »Izguba hladilnega medija v tlačnovodnem jedrskem reaktorju« (mentor: prof. dr. Leopold Škerget, somentor: izr. prof. dr. Jure Marn). Matej KEBRIČ z naslovom: »Pametni pralni stroj« (mentor: izr. prof. Vojmir Pogačar, somentor: izr. prof. dr. Bojan Dolšak); Vid VONČINA z naslovom: »Preračun krmilnega mehanizma vrtavkastega zgrabljalnika SIP Star 85026« (mentor: prof. dr. Iztok POTRČ, somentor: izr. prof. dr. Tone Lerher); * Na Fakulteti za strojništvo Univerze v Mariboru sta pridobila naziv univerzitetni diplomirani gospodarski inženir: dne 26. junija 2012: Matic RAUŠL z naslovom: »Vpliv posodobitve klimatskega sistema na energetsko učinkovitost objektov« (mentor: izr. prof. dr. Igor Drstvenšek, somentor: prof. dr. Duško Uršič); dne 28. junija 2012: Matija MARJETIČ z naslovom: »Analiza zamenjave lesenega vodila lista tračne žage« (mentor: doc. dr. Iztok Palčič, somentor: doc. dr. Karin Širec); * Na Fakulteti za strojništvo Univerze v Ljubljani so pridobili naziv magister inženir strojništva: dne 22. junija 2012: Selami BAYEG z naslovom: »Vrednotenje algoritma časovno temperaturne superpozicije v zaključeni obliki (CFS)« (mentor: prof. dr. Igor Emri); Arman BOROMAND z naslovom: »Neizotermalna kinetika kristalizacijskega procesa PA6 - vpliv biomodalnosti« (mentor: prof. dr. Igor Emri); Huiying JIN z naslovom: »Vpliv mehanskega recikliranja na predelovalnost in trajnost polietilena nizke gostote (LDPE)« (mentor: prof. dr. Igor Emri). * Na Fakulteti za strojništvo Univerze v Mariboru so pridobili naziv magister inženir strojništva: dne 15. junija 2012: Urška KOSTEVŠEK z naslovom: »Modeliranje in izdelava resekcijskih vodil za operacije ramenskega sklepa« (mentor: izr. prof. dr. Igor Drstvenšek); dne 26. junija 2012: Marjana STIPLOVŠEK z naslovom: »Optimiranje distribucije zraka v pnevmatskem omrežju obrata Droge Kolinska« (mentor: doc. dr. Uroš Župerl); Tilen VRANIČAR z naslovom: »Zagotavljanje ustrezne kakovosti meritev pri nadzoru proizvodnega

procesa sestavnega dela kompresorja« (mentor: prof. dr. Bojan Ačko, somentor: doc. dr. Andrej Godina). * Na Fakulteti za strojništvo Univerze v Ljubljani so pridobili naziv diplomirani inženir strojništva: dne 6. junija 2012: Elvis GREGORČIČ z naslovom: »Karakterizacija termo varovalke grelnika vstopnega zraka dizelskega motorja« (mentor: izr. prof. dr. Roman Šturm, somentor: izr. prof. dr. Tomaž Katrašnik); Jan MOŽINA z naslovom: »Obvladovanje procesa rezanja lamel magnetnega jedra transformatorja« (mentor: prof. dr. Karl Kuzman); Tadej ZADEL z naslovom: »Zanesljivost delovanja hidravličnega sistema na letalu CRJ 900« (mentor: prof. dr. Jožef Vižintin); dne 7. junija 2012: Matija LAZNIK STRNAD z naslovom: »Model kavitacijskega mlina za čiščenje pitne vode« (mentor: prof. dr. Branko Širok); Marko PETERNELJ z naslovom: »Priprava tehnologije varjenja opreme iz nerjavnega jekla za živilsko industrijo« (mentor: prof. dr. Janez Tušek); Anton ŠIBANC z naslovom: »Operativni priročnik šolanja pilotov jadralnih letal« (mentor: izr. prof. dr. Tadej Kosel); Urban ULČAR z naslovom: »Primernost uporabe digitalnih elektronskih zaslonov v pilotski kabini športnih letal« (mentor: izr. prof. dr. Tadej Kosel); dne 8. junija 2012: Tadej KOVAČIČ z naslovom: »Analiza preizkušanja senzorja z uporabo orodij Šest sigma« (mentor: prof. dr. Mirko Soković); Matjaž MACEDONI z naslovom: »Izdelava orodij za vakuumiranje« (mentor: prof. dr. Mirko Soković); Nina MOHORIČ z naslovom: »Vplivni parametri membrane na časovni odziv elektromehanskega stikala« (mentor: izr. prof. dr. Ivan Bajsić); Andrej OFENTAVŠEK z naslovom: »Uporaba povratnega materiala v proizvodnji cevi« (mentor: prof. dr. Mirko Soković); Viktor VERDEV z naslovom: »Optimizacija čelne plošče plinskega kuhalnega aparata« (mentor: prof. dr. Iztok Golobič); dne 11. junija 2012: Luka DOLENEC z naslovom: »Optimizacija izdelave plašča koračnega motorja s postopkom upogibanja« (mentor: doc. dr. Tomaž Pepelnjak); Benjamin GRMEK z naslovom: »Standardizacija sistema orodij za CNC obdelovalni center HURCO BMC 4020« (mentor: prof. dr. Janez Kopač); Mitja LIPOVŠEK z naslovom: »Tehnologija izdelave vodilne plošče linearne stiskalnice« (mentor: SI 99

Strojniški vestnik - Journal of Mechanical Engineering 58(2012)7-8, SI 96-100

doc. dr. Davorin Kramar, somentor: prof. dr. Janez Kopač); Sebastjan Janko RAJHER z naslovom: »Izboljšave pri izdelavi orodij za steklene izdelke« (mentor: prof. dr. Janez Kopač); Gorazd URBANIJA z naslovom: »Razvoj, izdelava in ekonomska upravičenost novega izdelka« (mentor: prof. dr. Janez Kopač); dne 20. junija 2012: Gašper BARAGA z naslovom: »Razvoj tehnologije injekcijskega brizganja dveh sestavnih delov črpalke« (mentor: prof. dr. Karl Kuzman); Blaž MOSTAR z naslovom: »Energijska učinkovitost ogrevanja nizkoenergijskih montažnih stavb« (mentor: prof. dr. Vincenc Butala, somentor: doc. dr. Uroš Stritih); Bogdan VELIKANJE z naslovom: »Lasersko označevanje plastičnega ohišja ventilatorja« (mentor: prof. dr. Janez Možina); Gašper KEMPERLE z naslovom: »Uvajanje sistema elektronske pilotske torbe v letalske operacije« (mentor: pred. mag. Borut Horvat, somentor: izr. prof. dr. Tadej Kosel); Dejan MANFREDA z naslovom: »Zasnova in vrednotenje gonila zaganjalnika« (mentor: prof. dr. Marko Nagode); Sebastjan SLUGA z naslovom: »Krmilnik stroja za lasersko spajanje termostatov« (mentor: prof. dr. Janez Diaci); Uroš ŽAGAR z naslovom: »Avtomatizacija laserskega varjenja uporovnega sestava« (mentor: prof. dr. Janez Diaci); dne 22. junija 2012: Peter BLATNIK z naslovom: »Hitra izdelava prototipa mehanizma za nastavljanje višine vzglavnika« (mentor: izr. prof. dr. Slavko Dolinšek, somentor: prof. dr. Janez Kopač); Kolja DEŽMAN z naslovom: »Krajšanje časa v procesu menjave serij« (mentor: izr. prof. dr. Janez Kušar, somentor: prof. dr. Marko Starbek); Stanislav GAZVODA z naslovom: »Uporaba kapljevitega barvnega nanosa na balistični želatini za preučevanje interakcij pri prodiranju izstrelka« (mentor: prof. dr. Iztok Golobič); Metod GRKMAN z naslovom: »Optimizacija razmeščanja delovnih mest« (mentor: izr. prof. dr. Janez Kušar, somentor: prof. dr. Marko Starbek); Urška MOHORIČ z naslovom: »Uvedba vitke proizvodnje z uporabo analize toka vrednosti« (mentor: izr. prof. dr. Janez Kušar, somentor: prof. dr. Marko Starbek); Anton NEMANIČ z naslovom: »Primerjalna analiza procesov razvlaževanja zraka za farmacevtsko industrijo« (mentor: prof. dr. Iztok Golobič); SI 100

Jurij VOLČIČ z naslovom: »Vrtanje kratkih in dolgih izvrtin v trdo« (mentor: prof. dr. Janez Kopač); dne 26. junija 2012: Marko HANČIČ z naslovom: »Določitev mehanskih lastnosti letalskih kompozitov iz steklenih vlaken« (mentor: prof. dr. Franc Kosel, somentor: doc. dr. Viktor Šajn); Tadej HOHNJEC z naslovom: »Pogoni samopostavljivega stolpnega žerjava« (mentor: doc. dr. Boris Jerman); Jože ZORKO z naslovom: »Izračun raketnega motorja in padala reševalne gondole letala« (mentor: prof. dr. Franc Kosel, somentor: doc. dr. Viktor Šajn); Matej ŽITKO z naslovom: »Zagotavljanje kakovosti pri izdelavi zahtevnih varjenih konstrukcij« (mentor: prof. dr. Janez Tušek). * Na Fakulteti za strojništvo Univerze v Mariboru so pridobili naziv diplomirani inženir strojništva: dne 18. junija 2012: Milan SIMETINGER z naslovom: »Načrtovanje kapacitet delovnih mest v podjetju Noži Ravne d.o.o.« (mentor: izr. prof. dr. Borut Buchmeister, somentor: doc. dr. Marjan Leber); dne 20. junija 2012: Saša RAŠEVIĆ z naslovom: »Vzdrževanje transportnega vozička« (mentor: doc. dr. Darko Lovrec, somentor: doc. dr. Samo Ulaga); dne 28. junija 2012: Katja BUH z naslovom: »Analiza procesa proizvodnje zobnikov« (mentor: izr. prof. dr. Borut Buchmeister, somentor: prof. dr. Andrej Polajnar); Žiga CAPL z naslovom: »Linija za transport Alpolizdelkov« (mentor: prof. dr. Iztok Potrč, somentor: izr. prof. dr. Tone Lerher); Danijel HABJAN z naslovom: »Nadgradnja linijskega stroja za brizganje polietilena pe s sistemom za rezanje poliestrskih vlaken« (mentor: viš. pred. dr. Marina Novak, somentor: izr. prof. dr. Bojan Dolšak); Sašo LOVRENČIČ z naslovom: »Uvedba metode SMED za obdelavo družine izdelkov na obdelovalnem centru DOOSAN« (mentor: izr. prof. dr. Borut Buchmeister); Matej SPITAL z naslovom: »Uvedba tehnologije črtne kode v skladišču rezervnih delov v Termoelektrarni Šoštanj« (mentor: prof. dr. Iztok Potrč, somentor: izr. prof. dr. Tone Lerher). * Na Fakulteti za strojništvo Univerze v Mariboru je pridobil naziv diplomirani inženir strojništva (VS): dne 20. junija 2012: Janez ŠPAN z naslovom: »Vzdrževanje stroja za vodni razrez« (mentor: doc. dr. Darko Lovrec, somentor: asist. Vito Tič).

Strojniški vestnik – Journal of Mechanical Engineering (SV-JME) Aim and Scope The international journal publishes original and (mini)review articles covering the concepts of materials science, mechanics, kinematics, thermodynamics, energy and environment, mechatronics and robotics, fluid mechanics, tribology, cybernetics, industrial engineering and structural analysis. The journal follows new trends and progress proven practice in the mechanical engineering and also in the closely related sciences as are electrical, civil and process engineering, medicine, microbiology, ecology, agriculture, transport systems, aviation, and others, thus creating a unique forum for interdisciplinary or multidisciplinary dialogue. The international conferences selected papers are welcome for publishing as a special issue of SV-JME with invited co-editor(s). Editor in Chief Vincenc Butala University of Ljubljana Faculty of Mechanical Engineering, Slovenia Technical Editor Pika Škraba University of Ljubljana Faculty of Mechanical Engineering, Slovenia Editorial Office University of Ljubljana (UL) Faculty of Mechanical Engineering SV-JME Aškerčeva 6, SI-1000 Ljubljana, Slovenia Phone: 386-(0)1-4771 137 Fax: 386-(0)1-2518 567 E-mail: info@sv-jme.eu, http://www.sv-jme.eu Print Tiskarna Knjigoveznica Radovljica, printed in 480 copies Founders and Publishers University of Ljubljana (UL) Faculty of Mechanical Engineering, Slovenia University of Maribor (UM) Faculty of Mechanical Engineering, Slovenia Association of Mechanical Engineers of Slovenia

http://www.sv-jme.eu

58 (2012) 7-8

Chamber of Commerce and Industry of Slovenia Metal Processing Industry Association

Since 1955

Journal of Mechanical Engineering - Strojniški vestnik

Breñosa, Ignacio Galiana, Manuel Ferre, Antonio Giménez,

Strojniški vestnik Journal of Mechanical Engineering

7-8 2012 58

no. year

volume

Strojniški vestnik - Journal of Mechanical Engineering is also available on http://www.sv-jme.eu, where you access also to papers’ supplements, such as simulations, etc.

http://www.sv-jme.eu

58 (2012) 7-8

Since 1955

Contents Papers

431 444 453 462 470 482 492

Journal of Mechanical Engineering - Strojniški vestnik

Strojniški vestnik Journal of Mechanical Engineering

7-8 year 2012 volume 58 no.