You are on page 1of 140

First Edition 2008 MOHD. HANIM OSMAN 2008 All rights reserved.

. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopy, recording, or any information storage and retrieval system, without permission in writing from Universiti Teknologi Malaysia, Skudai, 81310 Johor Darul Tak'zim, Malaysia. Perpustakaan Negara Malaysia Cataloguing-in-Publication Data

Analysis, design and performance of steel section with trapezoid web / editor Mohd. Hanim Osman. Includes index ISBN 978-983-52-0575-0 1. Structural engineering. 2. Structural design. I. Mohd. Hanim Osman. II. Universiti Teknologi Malaysia. Fakulti Kejuruteraan Sivil. 624.1 Pereka Kulit: MOHD. NAZIR MD. BASRI Diatur huruf oleh / Typeset by
MOHD. HANIM OSMAN & RAKAN-RAKAN

Fakulti Kejuruteraan Awam Universiti Teknologi Malaysia 81310 Skudai Johor Darul Ta'zim, MALAYSIA Diterbitkan di Malaysia oleh / Published in Malaysia by
PENERBIT UNIVERSITI TEKNOLOGI MALAYSIA

34 38, Jalan Kebudayaan 1, Taman Universiti, 81300 Skudai, Johor Darul Ta'zim, MALAYSIA. (PENERBIT UTM anggota PERSATUAN PENERBIT BUKU MALAYSIA/ MALAYSIAN BOOK PUBLISHERS ASSOCIATION dengan no. keahlian 9101) Dicetak di Malaysia oleh / Printed in Malaysia by
UNIVISION PRESS

Lot 47 & 48, Jalan SR 1/9, Seksyen 9 Jln. Serdang Raya, Tmn Serdang Raya 43300 Seri Kembangan, Selangor Darul Ehsan MALAYSIA

Contents

CONTENTS
Preface Chapter 1 Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading Izni Syahrizal Ibrahim, Mohd Hanim Osman and Fathoni Usman Lateral Torsional Buckling of Beam with Trapezoid Web Mohd. Hanim Osman, Sarifuddin Saad, Fatimah Denan and Abdul Karim Mirasa Analytical Study on Secondary Bending moment in Trapezoid Web Beam Mohd Hanim Osman, Abd. Latif Saleh, A. Aziz Saim and Fong Shiaw Ween Buckling Analysis of Column with Trapezoid Web Mohd Hanim Osman, Tan Cher Siang and Abd. Latif Salleh Performance Test on Simply Supported Composite Beam with TWP Steel Section Goh Kee Keong and Mohd Hanim Osman Experimental Test on Steel Beam with Partial Strength Connections using Trapezoid Web Profiled Steel Sections Arizu Sulaiman, Mahmood Md Tahir and Anis Saggaff
vi 1

Chapter 2

12

Chapter 3

26

Chapter 4

43

Chapter 5

56

Chapter 6

68

Contents

Chapter 7

Field Capacity of Steel Pile with Corrugated Web Surface Mohd Hanim Osman, Fauziah Kasim and A. Aziz Saim Second Moment of Area about the Minor Axis of Trapezoid Web Section Mohd.Hanim Osman, Sarifuddin Saad, Fatimah Denan and Tan Cher Siang Local Flange Buckling of Trapezoid Web Profile Ng Zee Leong and Mohd Hanim Osman

82

Chapter 8

99

Chapter 9

121

Preface

PREFACE

Built-up steel sections with trapezoid web have been used for years in some countries, particularly to increase the out of plane stiffness and shear buckling strength without the use of vertical stiffeners. Steel plate with corrugated web is not a new idea and has for so many years been used in aircraft design, shipbuilding for containers, as cold-formed webs for long span roof beam and later for civil engineering applications in buildings and bridges. The use of corrugated webs allows for the use of thin plates without the need of stiffeners, therefore it considerably reduces the cost of the fabrication and improves their fatigue life. The sections are fabricated by welding the trapezoid web steel plate to two flanges and proved to be an alternative to the conventional hot rolled and welded sections in respect to its strength/weight ratio. Theoretical and experimental studies have been conducted to study the bending capacity, shear, local flange and lateral beam buckling and column buckling. Other structural behaviours studied were secondary bending moment, torsional behaviour and flexural stiffness about major and minor axis. Finite element analysis is used to study the critical buckling by using the Eigen value analysis, and also to study the stress distribution in the web and around the web boundary. Experimental work to study the performance of composite beam using the trapezoid section as the steel component was also carried out. The purpose of the study is to utilise the section in the

Preface

design of long span girders such bridges. Another scope of study was the soil skin friction resistance of the section when used as foundation pile. The results of the studies are used to develop a new design guide for the section. This is to promote the use of the section in construction industries in this region.

Mohd. Hanim Osman Faculty of Civil Engineering Universiti Teknologi Malaysia 2008

1
BUCKLING ANALYSIS OF PLATE GIRDER WITH TRAPEZOID WEB SUBJECTED TO SHEAR LOADING
Izni Syahrizal Ibrahim, Mohd Hanim Osman and Fathoni Usman
1.1 INTRODUCTION

Plate girder structure, which is constructed from welded steel plate, is well recognized. Economical design of girders and beams normally requires thin webs. But if the web is extremely thin the problem of plate buckling may arise. Possible ways to reduce this risk consist of using thicker plates, web stiffeners or strengthening the web by making it corrugation. The conventional provision of stiffeners to allow the use of thin webs has two disadvantages, i.e. high fabrication cost and a possible reduced life due to fatigue cracking that may initiate at them stiffener weld. The use of corrugated plates (Fig. 1.1) to replace the flat stiffened plates as the web can eliminate both disadvantages. The uses of corrugated webs have been increasingly used in many diverse engineering structures from structural engineering to aerospace and marine engineering. The corrugation of the webs acts as transverse stiffeners that allows for the use of thin plates. Because of the high slenderness ratio, web stability against shear buckling is of prime importance. Various formulas for shear buckling of corrugated section have been proposed by previous researchers. However, they were applicable only to certain configuration for which they were derived. Experimental and analytical studies are being carried out

Analysis, Design and Performance of Steel Section with Trapezoid Web

to develop formula that can be used for all types of web configuration. Part of the analytical study is presented herein. Eigen-value buckling analysis in the finite element method was employed in this study to determine the elastic shear buckling of various corrugation types. Analysis on normal web panel, in which the formula of shear buckling capacity has been well established, was also carried out as a comparison.

Figure 1.1 Trapezoidally Web Plate Girder

1.2 THEORETICAL CAPACITY

FORMULAS

FOR

SHEAR

It has been reported that corrugated panels loaded in shear exhibit both local and global buckling depending on the geometry of the corrugations. Hamilton reported that differences in panel thickness, aspect ratio, and configuration of corrugation influence the behavior of the panel [1]. Turner et.al. investigated that from numerical analysis results corrugation angle influenced the buckling behavior of thin corrugated panel [2]. Elgaaly determined that in the local buckling mode, the corrugated webs acts as a series of flat subpanels that mutually support each other along their vertical (longer) edges and are supported by the flanges at their horizontal (shorter) edge [4]. These flat plate subpanels are subjected to shear, and the elastic buckling stress is

Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading

given by:

l = k
(1.1)

2E 12( 1 2 )( b / t w ) 2

where k is a buckling coefficient, which is a function of the panel aspect ratio (b/hw) and the boundary support conditions. For the longer edges simply supported and the shorter edges clamped, k = 5.34 + 2.31(b/hw) - 3.44(b/hw)2 + 8.39(b/hw)3, and in the case where all edges are clamped, k = 8.98 + 5.6(b/hw)2. The geometrical nominations of the trapezoid web section are shown in Fig. 1.2. The designed shear force: Vd = l hw tw
q s

(1.2)

tw

hr

Figure 1.2 Geometrical nomination of trapezoid web section

1.3 SHEAR BUCKLING ELEMENT METHOD

ANALYSIS

BY

FINITE

The objective of the finite element analysis is to study the critical shear by using Eigenvalue-buckling analysis. Eigenvalue buckling analysis is a linear analysis, which may be applied to relatively stiff structures to estimate the maximum load that can be supported to structural instability or collapse [9]. Five series of model have

Analysis, Design and Performance of Steel Section with Trapezoid Web

been developed and the thickness of each model increased from 1 to 5 mm. Semiloof thin shell element (QSL8) in LUSAS structural analysis software was chosen for meshing the surface element of the models. QSL8 is grouped as shell element [10]. Elastic-Isotropic material is applied to each element. Other element attributes used on the model include geometry of the model, support conditions and loading. A series of finite element model of corrugated web were carried out. The geometrical corrugation is shown in Table 1.1. For shear model support condition, AD is restrained in each translation and rotational direction. BC is restrained in x and z directions translation. Shear load is applied only on BC, as shown in Fig. 1.3.
D C V A B
Figure 1.3 Finite element models Table 1.1 Dimension for finite element models
FEM E fy hw q s b d hr Model kN/mm2 N/mm2 mm mm mm mm mm mm Index TWP 400 205 355 400 400 457 170 80 80 TWP 600 205 355 600 400 457 170 80 80 TWP 800 205 355 800 400 457 170 80 80 TWP 1200 205 355 1200 400 457 170 80 80 TWP 1600 205 355 1600 400 457 170 80 80

45 45 45 45 45

Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading

The result of the Eigenvalue analysis as shown in Table 1.2 is used to build the empirical equation. To obtain the contribution of the corrugation to the critical shear capacity of the trapezoid web panel, the equation proposed is based on three key parameter; hw/tw, b/tw, and b/hw. It is assumed that the web panels of the models are simply supported at both top and bottom flanges. From the analysis, a new equation for critical shear stress is proposed as follows:

cre = k

1.6 2 E 12(1 2 )(hw / t w ) 2 1.8/(b/hw)2-(b/hw)3+ 8*(b/hw)+9

(1.3)

where, k =

The dimensions of the experimental models and the comparison between the value of Vcr derived from the proposed equation and the experimental results that previously held in UTM [7, 8] are shown in Table 1.3, where depth hw and thickness tw of the web is subjected to a vertical shear force V, constant in the x-direction. The Von Mises criterion is used, so the yield stress in shear is:

y =
(1.4)

fy 3

0 .6 f y

In the case where cre > 0.8 y , inelastic buckling will occur and the inelastic buckling stress cri can be calculated by [4]:

cri = 0.8 cre y y

(1.5)

The results of critical shear capacity compared with experimental results shows that the proposed equation is sufficient and the following equation could be used to calculate the critical shear

Analysis, Design and Performance of Steel Section with Trapezoid Web

capacity:

Vcr = cr * hw * tw
Table 1.2 FEA results as raw data
FEM Model Index TWP 400-1 TWP 400-2 TWP 400-3 TWP 400-4 TWP 400-5 TWP 600-1 TWP 600-2 TWP 600-3 TWP 600-4 TWP 600-5 TWP 800-1 TWP 800-2 TWP 800-3 TWP 800-4 TWP 800-5 TWP 1200-1 TWP 1200-2 TWP 1200-3 TWP 1200-4 TWP 1200-5 TWP 1600-1 TWP 1600-2 TWP 1600-3 TWP 1600-4 tw mm 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 hw mm 400 400 400 400 400 600 600 600 600 600 800 800 800 800 800 1200 1200 1200 1200 1200 1600 1600 1600 1600 b mm 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 170 Vcr kN 9.56 75.54 249.11 571.66 1070.97 12.94 102.46 337.90 767.79 1409.00 16.50 130.73 433.95 1003.25 1789.34 22.69 180.20 554.79 1295.26 2626.91 30.44 241.60 804.35 1869.85

(1.6)

N/mm2 23.90 94.42 207.59 357.28 535.48 21.57 85.39 187.72 319.91 469.67 20.62 81.71 180.81 313.52 447.33 18.91 75.08 154.11 269.85 437.82 19.02 75.50 167.57 292.16

Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading

Table 1.3 Dimensions of experimental models and comparison between proposed equation and experiment results
Specimen Mark CWB 1 CWB 2 UCW 1 UCW 2 UCW 3 UCW 5 UCW 6 TS 600-3 TS 800-2 TS 1300-2 TS 1300-3 fy tw mm 4.0 4.0 2.4 2.1 1.3 2.1 1.9 3.0 2.0 2.0 3.0 hw mm 351 351 380 380 380 380 380 600 800 1300 1300 b mm 170 170 130 130 130 130 130 170 170 170 170 b/tw hw/tw b/hw k

cre

cri

355 355 275 275 275 275 275 275 275 275 275

42.5 87.8 42.5 87.8 55.1 61.9 99.2 63.1 67.4 56.7 85.0 85.0 56.7 161.0 181.0 290.1 184.5 196.9 200.0 400.0 650.0 433.3

0.48 0.48 0.34 0.34 0.34 0.34 0.34 0.28 0.21 0.13 0.13

20.4 20.4 27.1 27.1 27.1 27.1 27.1

786.7 786.7 309.6 245.1 95.4 235.9 207.1

158.8 158.8 158.8 158.8 158.8 158.8 158.8 152.1

Vcr kN Eq. (1) 287.8 287.8 142.4 126.7 47.5 124.3 116.4 285.8 149.9 210.3 593.0

Vexp kN 265.3 247.4 150.0 132.5 122.5 117.5 127.5 270.0 169.8 275.0 650.0

Vcr/ Vexp

1.08 1.16 0.95 0.96 0.39 1.06 0.91 1.06 0.88 0.76 0.91

33.7 249.5 50.6 93.7 115.3 80.9 115.3 182.0

Note: unit for fy, cre and cri is N/mm2

1.4

DISCUSSION

The buckling pattern as shown in Fig. 1.4 indicates that the inclined parts provide a resistant to shear buckling from crossing from one flat sub-panel to the adjacent one. As a comparison to normal web, the buckling pattern is always in the diagonal direction of the panel so the trapezoid web will have higher ultimate shear strength [6].

Figure 1.4 Buckling pattern of the model

Analysis, Design and Performance of Steel Section with Trapezoid Web

The buckling formula proposed in Eq. (1.3) is given in the function of the web slenderness parameter; hw/tw, and b/hw. From the experimental results, the thickness of the web gave significant contribution to the shear capacity of the model. Fig. 1.5 and Fig. 1.6 show that straight line parallel with x-axis is the yield stress of the model and the curve was derived from Eq. (1.3). The curve will change to straight line when inelastic buckling stress is greater than the yield stress. Therefore, it can be proposed that the buckling capacity curve is of straight line and follow the curve from Eq. (1.3). The proposed curve is indicated by the dotted line.

500 450 400 350 300 cr 250 200 150 100 50 0 0 100 200 300 400 500 600 700 UCW2, 166.04 UCW1, 167.26 UCW5, 150.10 UCW6, 173.85

UCW3, 246.08

hw/t w

Figure 1.5 Graph derived form proposed formula compared with experiment result as on Table 3

Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading

500 450 400 350 300 cr 250 200 TS1300-3, 166.67 150 100 50 0 100 TS1300-2, 105.77

200

300

400

500

600

700

800

900

hw/t w

Figure 1.6 Graph derived form proposed formula compared with experiment result as on Table 3

1.5

CONCLUSION

Parametrical study with Eigenvalue buckling analysis in the finite element method have been successfully employed to built an equation that were to be used to calculate the elastic shear buckling capacity of the trapezoid web subjected to shear loading. From the experimental results and the analytical study, it can be concluded that: 1. The shape of the buckling pattern shows diagonally occurs along the web panel. The assumption that the corrugated webs acts as a series of flat sub-panels that mutually support each other along their vertical (longer) edges and are supported by

10

Analysis, Design and Performance of Steel Section with Trapezoid Web

the flanges at their horizontal (shorter) edge had been proven. 2. The finite element analysis results show that the proposed formula of shear capacity as in Eq. (1.3) is sufficient when cre < 0.8 y . Factor k in the equation is derived with the assumption that the web is simply supported along the flanges. NOTATION

b d E fy hr hw k tw Vcr

y cre cri

horizontal width of sub panels of a corrugation fold horizontal width of the inclined fold Youngs modulus yield stress thickness of corrugation depth of corrugation web buckling coefficient thickness of web critical shear force angle of corrugation profile yield shear stress elastic critical shear stress inelastic critical shear stress Poissons ratio

Buckling Analysis of Plate Girder with Trapezoid Web Subjected to Shear Loading

11

REFERENCES 1. Hamilton R W, Behavior of Welded Girders with Corrugated Webs, A Ph.D. Thesis, University of Maine, August 1993. 2. Turner, S J, Van Erp, G M and Yuen S W, The Influence of Boundary Conditions on the Shear Buckling Behavior of Thin Corrugated Plates, Faculty of Engineering and Surveying, University of Southern Queensland, Toowoomba, Queensland, Australia. 3. Johnson R P and Cafolla J, Corrugated Webs in Plate Girders for Bridges, Proceeding Instn Civil Engineering Structures and Buildings, May1997. 4. Elgaaly, M, and Seshadri, A., Steel Built-up Girder with Trapezoidally Corrugated Web, Engineering Journal, Fisrt Quarter, 1998. 5. Elgaaly, M, and Hamilton, R.W, Shear Strength of Beams with Corrugated Webs, Journal of Structural Engineering, April 1996. 6. Scheer, J, Beritch Nr. 6203, Institute fur Stahlbau der Technichen Universitat Braunschweigh, Germany, 1996. 7. Osman, M H, Ibrahim, I S and Tahir, M M, Shear Strength of Trapezoid Web Plate Girder, Faculty of Civil Engineering, University Teknologi Malaysia, Malaysia, 1999. 8. Osman, M H, Ibrahim, I S and Tahir, M M, Strength Behaviour of Trapezoid Web Plate Girder, Faculty of Civil Engineering, University Teknologi Malaysia, Malaysia, 1999. 9. LUSAS Modeller User Manual, Version 13, FEA Ltd., United Kingdom, 1999. 10. LUSAS Element Reference Manual, Version 13, FEA Ltd., United Kingdom, 1999. 11. Narayanan and Rockey, Ultimate Load Capacity of Plate Girder, Proc. Instn Civ. Engrs, Part 2, 1981.

2
LATERAL TORSIONAL BUCKLING OF BEAM WITH TRAPEZOID WEB
Mohd. Hanim Osman, Sarifuddin Saad, Fatimah Denan and Abdul Karim Mirasa

2.1 INTRODUCTION 2.1.1 Trapezoid Web Beam Economical design of structural steel sections normally requires thin webs to increase the shear buckling strength. The conventional method which uses intermediate stiffeners welded to web to allow the use of thin webs has two disadvantages i.e. high cost of fabrication and reduced service life of the element. The use of corrugated sheets (Fig. 2.1) to replace flat sheets as webs of a girder eliminate both disadvantages [1,2,3]. In addition, it reduces the total weight of the structure, thus allowing longer spans and savings in foundation design. Previous researches have been carried out to study the performance of trapezoid web section in shear in web, secondary bending moment in flange, bending, and axial buckling [4,5,6]. 2.1.2 Lateral Torsional Buckling

When a beam is loaded, it will deflect vertically. If the beam does not have sufficient lateral stiffness or lateral support along its length, the beam will also deflect out of the plane of loading. The load at which this buckling occurs may be substantially less than the beams in plane load carrying capacity. For an idealized perfectly straight elastic beam, there will be no out-of-plane

Analysis, Design and Performance of Steel Section with Trapezoid Web

13

deformations until the applied moment reaches the critical value Mb, when the beam buckles by deflecting laterally and twisting. These two deformations are interdependent: when the beam deflects laterally, the applied moment exerts a component torque about the deflected longitudinal axis which causes the beam to twist. This behaviour, which is important for long unrestrained Ibeams whose resistances to lateral bending and torsion are low, is called elastic lateral torsional buckling. Experimental and numerical study on lateral torsional buckling of steel section with trapezoid web is presented in this paper. The objectives of the study is to determine the lateral torsional buckling capacity of trapezoid web profile in comparison with normal flat web beams using experimental and finite element method.

Figure 2.1 Typical beam sections with trapezoid web and flat web

2.2 EXPERIMENTAL STUDY 2.2.1 Test Procedure Lateral torsional buckling tests were conducted on three sets of beams, each set consists of two specimens i.e. sections with trapezoid web numbered as TWP3 (TWP3A and TWP3B), trapezoid web profile TWP4 (TWP4A and TWP4B) and flat web FW3 (FW3A and FW3B). The difference between TWP3 and TWP4 is in their web corrugation thickness, where TWP3 has full

14

Lateral Torsional Buckling of Beam with Trapezoid Web

corrugation thickness (hr = B) while TWP4 has only half corrugation (hr = 0.5B). The test was designed based on the test method by Dirk [7] and Salina [8]. Fig. 2.2(a) and 2.2(b) show the diagrammatic view of the test set-up. The photographs are shown in Fig. 2.3(a) and 2.3(b). A point load was loaded at mid span of the specimen through a specially designed loading device. The L-shape roller bearing guide was used to ensure that the jack always seated on a horizontal surface so that the direction of the loading was kept vertical under increasing loading. The roller on the top flange under the loading was used to ensure that there was no horizontal restraint that might inhabit lateral buckling during loading. Two types of lateral restraint at the support were used, i.e. type A and type B. For type A, the bottom flange of the beam at both ends were fixed, whilst for type B, the bottom flange and web which were both restrained from deflect laterally. It is shown in Fig. 2.4. For each type of beam, two different spans were used i.e. 4000 mm and 5000 mm. Displacement transducers were placed at five different locations to measure the lateral deflection of the beams. Loads were applied gradually, with the increments of 1.0 kN. The displacement was recorded at each increment. The lateral displacements of beam specimens were measured at 400 mm (left and right) from mid span of the beam (for top and bottom flange) and also at the centre of the web.

Analysis, Design and Performance of Steel Section with Trapezoid Web

15

A Beam specimen

Load actuator Load cell Roller

fixed at flange

Lateral deflection transducer

fixed at flange

Figure 2.2(a) The overall view of the test set up

Test frame

Load actuator Load cell L-Shape roller bearing guide

Beam specimen

Lateral deflection transducer

Figure 2.2(b) Details of the loading device

16

Lateral Torsional Buckling of Beam with Trapezoid Web

Figure 2.3(a) The overall view of the test set up

Figure 2.3(b) Details of the loading device

Fixed at flange only (Type A)

Fixed at flange and web (Type B)

Figure 2.4 Lateral restraint Type A and B at the beam end support

Analysis, Design and Performance of Steel Section with Trapezoid Web

17

2.2.2

Test Results

The test was stopped when buckling occurred, as determined in the graph of moment versus lateral displacement. In the test, all beam specimens were found to be still in elastic state after the tests. Relationships between bending moment and lateral deflection were plotted. In general, the lateral deflection increases linearly with the vertical bending moment. Then, the increase becomes non-linear, followed by a stage when the deflection increases monotonically. The value of lateral torsional buckling moment resistance was determined from the intersection of tangent of the first and second curve. The intersection method was known as knee joint which has been used by many researchers [6,7,8,9] to determine the moment resistance of connection. The values of lateral torsional buckling resistance, Mb for all specimens were determined when a knee shape was observed. In each graph, two tangent lines were drawn and the intersection of these two lines gives the Mb value.

TWP3A(A)5m 10.0 8.0 Moment, M (kNm) 6.0 4.0 2.0 0.0 0.0 1.0 2.0 Lateral displacement (mm) 3.0
12.00 10.00 M o m en t, M (k N m ) 8.00 6.00 4.00 2.00 0.00 0.00

TWP3A(B)5m

Mb = 8.80 kNm

Mb =10.30 kNm
0.20 0.40 0.60 Lateral dispalcement (mm) 0.80

Figure 2.5 Typical graphs for the determination of Mb

18

Lateral Torsional Buckling of Beam with Trapezoid Web

Table 2.1 Test results of Mb for beams with normal flat web and beams with trapezoid web profile
Span of beam (mm) Beam mark Support type A Mb kNm 8.80 9.20 8.00 8.40 7.10 7.00 9.50 9.80 8.90 9.30 8.10 8.50 Mb kNm 10.30 11.30 9.00 9.50 8.20 8.90 11.00 11.40 9.20 10.00 8.50 8.80 8.65 8.30 9.25 8.55 11.20 9.60 8.20 7.05 9.65 9.10 Support type B Average Mb kNm 10.80 Average Mb kNm 9.00

5000

TWP3A TWP3B TWP4A TWP4B FW3A FW3B TWP3A TWP3B TWP4A TWP4B FW3A FW3B

4000

The deflection from the midspan was used for the determination of Mb. The Mb value for each beam was indicated in each graph (Fig. 2.5). The result of the experiment is presented in Table 1. From the table, it is observed that: (i) Beams with flat webs and 5 m have the average of lateral torsional buckling moment, Mb lower than that of beam 4 m, for each of the Type A and B support. The same finding for TWP3 and TWP4 was obtained. As expected, the beam with Type B support has higher Mb value that those with Type A support for both spans. TWP section performs better than that of flat web in terms of lateral torsional buckling moment resistance. The beam with trapezoid web profile section with full

(ii) (iii) (iv)

Analysis, Design and Performance of Steel Section with Trapezoid Web

19

shape corrugation (hr = B) are better than the beam with trapezoid web profile section with half shape (hr = 0.5B) and flat web in their lateral torsional buckling moment resistance. This is because the Iy value for a TWP beam is higher than Iy of flat web of similar dimensions. 2.3 FINITE ELEMENT STUDY ON LATERAL TORSIONAL BUCKLING BY FINITE ELEMENT In this study, all models were assumed to buckle under perfect conditions, where there is no initial imperfectness and eccentric load. The buckling moments were then compared with result obtained from testing. Eigenvalue analysis of LUSAS Modeller [6] was used to determine the buckling loads. A linear buckling analysis is a useful technique that can be applied to relatively stiff structures to estimate the maximum load that can be supported prior to structural instability or collapse. The assumptions used in linear buckling analysis are that the linear stiffness matrix does not change prior to buckling and that the stress stiffness matrix is simply a multiple of its initial value. 2.3.1 Modelling

LUSAS models are defined in terms of geometric features that must be subdivided into finite elements for solution. This process of sub division is called meshing. Mesh datasets contain information about element types, element discretisation and mesh type. The I-beam models were assigned ungraded mild steel for its material property with Youngs modulus, Es = 209 103 N/mm2, shear modulus, G = 79 103 N/mm2 and Poisson ratio of 0.3. The beams are simply supported and unrestrained laterally.

20

Lateral Torsional Buckling of Beam with Trapezoid Web

The convergence of the mesh was established by independently increasing the mesh density in each part of the model beam section. The model was also analysis with increased mesh density in all parts of the section simultaneously, and with higher-order elements (QSL8). 2.3.2 Eigenvalue Buckling Analysis

The main objective in the Eigenvalue buckling analysis is to obtain the critical buckling load, by solving the associated Eigenvalue problem. In LUSAS, there are two methods to obtain information regarding buckling loads and their respective deformation mode i.e. The linear Eigenvalue buckling analysis and the full geometrically non-linear analysis. Fig. 2.6 shows a typical buckling shape in Mode 1. Modes 1, 2 and 3 represent the buckling shape of the element. In this study the result of mode one would be considered; because it was found that all the beam specimen failed in the tests due to this mode. This is also because mode one is the least value. It will be unrealistic to choose the higher modes 2 and 3 to get the critical buckling load. The resulted Eigenvalues are actually the load factors to be multiplied to the applied loading, to obtain critical buckling load. The Eigenvalue buckling analysis in LUSAS Modeller will provide both local and global buckling modes. Engineering judgment is necessary to determine which buckling mode is the most critical in order to select the appropriate buckling load factor. It is, of course possible to visually examine the resultant modes in LUSAS Modeller.

Analysis, Design and Performance of Steel Section with Trapezoid Web

21

Figure 2.6 The buckling shape mode 1

2.4

RESULT AND DISCUSSION

The results of Mb are summarized in Table 2.2. In addition, the Mb value derived from the design calculation is given for each beam. The method of calculating the design Mb value is also given in BS 5950: Part1: 2000, by neglecting the contribution of web. It can be summarized as follow: The critical buckling loads and the lateral torsional buckling moments results of Eigenvalue analysis theory calculation for trapezoid web profile and flat web are presented in Table 4, for both type A and type B support. It is shown that, as expected, as the lengths of the two beams with trapezoid web profile increase, the lateral torsional buckling moment decreases. It is found that, trapezoid web profile sections need higher load to buckle compared to flat web. This is because of the higher value of Iy for the TWP section compared to the flat web section [ ]. In terms of the effect of corrugation shape, the results show that

22

Lateral Torsional Buckling of Beam with Trapezoid Web

web with full corrugation (TWP3; hr = B) have a higher resistance to lateral torsional buckling compared to that of the half corrugation shape (TWP4; hr = 0.5B). As a conclusion, trapezoid web profile section have higher resistance in lateral torsional buckling behaviour and hence suitable for structural applications. Fig. 2.7 shows the comparison between the TWP3, TWP4 and FW in their lateral torsional buckling resistance for Type A support. Both set of results show a similar trend i.e. as the beam length increases, the buckling moment decreases. In all beam cases, the finite element prediction is more than that of the test. For support Type B, the beam length increases, the buckling moment decreases. For all specimens, the buckling moment results from the finite element predictions are bigger than that of the test results for all lengths of beams.
Table 2.2 Percentage difference of Mb for beams with normal flat web and beams with trapezoid web profile (finite element analysis)
Support Span (mm) 6000 5000 4000 3500 6000 5000 4000 3500 Mb TWP3 5.84 7.05 8.30 9.35 11.62 14.35 15.41 15.66 TWP4 5.54 6.45 7.81 8.75 9.75 11.28 13.79 15.05 (FW) 3.70 4.09 4.39 4.40 7.23 8.87 11.50 13.43 Mb (design) 3.15 3.85 4.73 5.51 3.32 8.60 10.96 13.08

Support Type A

Support Type B

Analysis, Design and Performance of Steel Section with Trapezoid Web

23

Type A 10
Mb (kNm)

8 6 4 2 0 3000 4000 Span 5000 6000

TWP3 TWP4 (FW) Design

Type B 20
TWP3 TWP4 (FW)

Mb (kNm)

15 10 5 0 3000 4000 5000 6000

Design

Span (m)

Figure 2.7 Buckling moment resistance for different span

Fig. 2.7 shows that as the length increases the value of Mb decrease. In both figures, comparison between different corrugation ratio shows that the value of Mb for TWP3 (halfcorrugation) is higher than TWP4 (full-corrugation). The overall observation shows that the result from the finite element analysis is higher than the test for support Type B but not for the Type A support. In comparison, in all cases, the value of Mb for finite element analysis and testing are more than the value Mb from the design formula. In finite element analysis, the value of Mb at Type A support was less than Mb value from design formula. From both figures, comparison between different types of restraint shows that Type B gives extra value than Type A support. This is in

24

Lateral Torsional Buckling of Beam with Trapezoid Web

accordance to the theory, i.e. Type B support is supposed to get higher value of Mb than Type A. This is because of the specimen was more difficult to move in Type B support. Therefore, the value of Mb will be higher for Type B. 2.5 CONCLUSION

From the experimental and analytical studies on the lateral torsional buckling on trapezoid web section, it can be concluded that : (1) Steel beam with trapezoidally corrugated web section has higher resistance to lateral torsional buckling compared to that of section with flat web. (2) The result shows that corrugation thickness influences the resistance to lateral torsional buckling. Sections with thicker corrugation have higher resistance to lateral torsional buckling. (3) Higher value of moment of inertia about minor axis for the section with thicker corrugation contributes to the higher resistance to lateral torsional buckling. (4) Finite element method can be used to determine the elastic lateral torsional buckling moment of the section. REFERENCES 1. Elgaaly M, Seshadri A, Steel built-up girders with trapezoidally corrugated webs, 1988, Engineering journal, 1st.Quarter, London:AISC. Elgaaly,M, Seshadri,A, Hamilton, R.W, Bending strength of steel beams with trapezoid corrugated webs, ASCE Journal of Structural Engineering, 1997, Vol.123. Johnson,R.P and Caffola, J, Local flange buckling in plate girders with corrugated webs, Proceeding Instn Civil Engineering Structures and Buildings, 1997. Osman, M.H, Shear buckling of trapezoid web profile section, International Conference on Numerical Method in

2.

3.

4.

Analysis, Design and Performance of Steel Section with Trapezoid Web

25

5.

6.

7.

8.

9.

Structural Engineering, Indonesia, 2001. Osman, M.H, Secondary bending moment in trapezoid web section, Civil Engineering Research Seminar, Universiti Teknologi Malaysia, Johor, Bahru. Osman, M.H, C,H, Tan, Axial buckling of column with trapezoid web, Asia Pacific Structural Engineering Conference, 2003, Johor Bahru, Malaysia. Dirk, P.P, Lateral torsional buckling of of end notched steel beams, International Colloquium on Stability of Structures under Static & Dynamic Loads, ACSE, 1977. Salina, J, Lateral torsional buckling of beam with trapezoid web profile, Master thesis, Universiti Teknologi Malaysia, 2003. Anis, Behaviour of connection of composite beam with trapezoid web profile PhD. Thesis, Universiti Teknologi Malaysia, 2007.

3
ANALYTICAL STUDY ON SECONDARY BENDING MOMENT IN TRAPEZOID WEB BEAM
Mohd Hanim Osman, Abd. Latif Saleh, A. Aziz Saim and Fong Shiaw Ween
3.1 INTRODUCTION

Trapezoid web beam is a type of steel I-section in which the web is made corrugated in a trapezoidal profile form. The corrugated thin web is continuously welded to the flanges along the top and bottom edges. Trapezoid web beam is a built up section that able to support vertical loads over long spans. The higher bending capacity is achieved by increasing the depth of the section. Increasing the depth will increase the slenderness of web and hence reducing the shear buckling capacity. Ordinarily, the economic design of steel web I-beam requires thin web. To avoid shear buckling, intermediate stiffener has to be used, or alternatively, the web can be made corrugated in trapezoidal profile. When beams with corrugated webs are compared with those with stiffened flat webs, it can be found that trapezoidal corrugation in the web enables the use of thinner webs and trapezoidal web beams eliminate costly web stiffeners. The flange of trapezoid web beam carries the bending moment and the trapezoidal web carries the shear force. Due to the shear force subjected on the trapezoidal web, a lateral bending moment is induced in the flange, which is known as secondary bending moment, Myf. It will cause a minor reduction in the bending moment capacity of the web, as given in the second term of the

Analysis, Design and Performance of Steel Section with Trapezoid Web

27

interactive equation:
M yf Mx + <1 M cx M cyf

(3.1)

where Mx is the applied bending moment, Mcx is the section bending capacity, V is the applied shear force on the section, Mcyf is the secondary bending moment capacity of each flange about the minor axis Mcyf = py . Zyf where py is the design strength of the flange and Zyf is the elastic modulus of each flange in y-axis. There is limited information and research on the secondary bending moment phenomena. A manufacturer of the trapezoid section in Germany has produced a table of properties for its standard section, which include the secondary bending moment coefficient [1]. However, there is no explanation and formula on how it was derived. Hence, this study is carried out to develop a formula, which can be used for any size of section. In this study, the finite element analysis was used to determine the lateral reactions at the flange of the trapezoidal web beam. 3.2 OBJECTIVE AND SCOPE OF STUDY

The main objectives of this study is to determine the secondary bending moment, Myf in the flange, and then using the results to derive a formula of secondary bending moment coefficient, Co which is applicable to any size of section. Finite element analysis was used to study the behaviour of the secondary bending moment for various sections. To achieve the objective, lateral reaction in the flanges of the trapezoidal web beams subjected to shear loading were determined. Then the secondary bending moment, Myf induced in the flanges and the

28

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

value of Co due to the lateral forces were calculated. Subsequently, from the parametric study of various section sizes, and web depth and thickness, the general formula of Co was developed. 3.3 SECONDARY BENDING MOMENT IN TRAPEZOID WEB BEAM Fig. 3.1 shows a short segment of a web with a trapezoid profile, subjected to a vertical force at its end. Due to the corrugation of the web, it tends to bend laterally, which then exerted lateral force to the flange which are weld connected to the top and bottom edges of the web. These reactions become imposed lateral force to the flange, resulted in the bending moment about the minor axis of the flanges.
Top edge Shear force Top view Bottom edge

Lateral force to the flanges

Figure 3.1 The lateral reaction created by a shear force

Analysis, Design and Performance of Steel Section with Trapezoid Web

29

The shear force subjected to a corrugated web section is resisted by its web, through the development of a shear stress flow as shown in Fig. 3.2. The oblique orientation of alternate web sub panels induces component forces, Q from the shear stress flow, in lateral direction (z direction) of the section in each of the sub panel. These component forces act in opposite direction. This will result into couples and induces secondary bending moment, Myf about the yaxis of the section. Myf = Qa (3.2) The secondary bending moment, Myf in each flange change linearly with the applied shear force and can be written by: Myf = VCo (3.3) where V is the direct shear loading applied to the web, Co is the geometric constant called secondary bending moment coefficient, in unit mm. From Eq. (3.2) and Eq.(3.3), Co becomes:
Co = Q.a V

(3.4)

The lateral reaction depends on the section properties and increase linearly with the applied shear force. Thus, secondary bending moment coefficient, Co is induced. In a flat web plate, there is no inclined web, thus no transverse component of shear stress flow. Nina [2] had made some analysis on Co and Co/Mcyf values for various web depths. Graphs of Co and Co/ Mcyf versus web depth from the analysis and the existing manufacturers table of properties were plotted. She concluded that the values of Mcyf and Co from analysis give reasonable comparison with the existing design table.

30

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

V V V V V V V V V V
y x Shear stress flow

V V

Shear force

Myf = Q.a Q

Q a

Figure 3.2 The shear stress flow and secondary bending moment

3.4

FINITE ELEMENT ANALYSIS

A finite element model which was made up of two cycles web panel was studied. The number of cycles in trapezoidal web was found to be of no effect to the result. The model consists of top and bottom flanges, trapezoidal web and right and left side stiffeners plates as shown in Fig. 3.3.

Analysis, Design and Performance of Steel Section with Trapezoid Web

31

Each surface is formed from 4 nodal lines. Semiloof curved thin shell element (QSL8) which is a family of shell element in 3D dimension was chosen to represent the element type of model in this study. Thin shell is selected as the generic element type, Quadrilateral as the element shape with quadratic as the interpolation order. Regular mesh was assigned by allowed transition pattern with 1 local x and y division. Surface geometric are assigned to the web with the thickness of 2 mm. Flange and side plates were of different thicknesses. Isotopic material properties are assigned to the model with Youngs modulus of 205 106 kN/m2 and Poissons ratio of 0.3. The nodes along the supported side were restrained in translation in x, y, z direction. The nodes along the loaded side are restrained in x and z direction only. Nodes connecting web and flanges are restrained in z direction. The total concentrated load in vertical direction, 20 kN is assigned to the points along the loaded side of the plate which allowing deformation in vertical direction (y direction only). Fig. 3.3 shows the loading arrangement assigned to the model in the analysis. The finite element analysis was carried out to determine the reactions at lateral direction at the nodes along the top and bottom edges of the web. A parametric study was carried out on more than forty finite element models with the web depth ranging from 300 to 1300 mm, flange width ranging from 120 to 300 mm, flange thickness from 10 to 30 mm, and web thickness from 2 to 8 mm.

32

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

Y X Z

Figure 3.3 Support and loading condition assigned to the model

3.5

RESULTS OF THE FINITE ELEMENT ANALYSIS

The lateral reactions of each section size, in each oblique part for the trapezoidal web are summarized in Table 3.1 for five typical sizes. Fig. 3.4(a) illustrates the phenomena of the lateral reactions on the nodes along the oblique web of a section with 300 mm depth web. The sum of the forces, Q on each sub-panel is shown in Fig. 3.4(b), which shows that Q is equal for all sub-panels. The sum of lateral forces on one oblique sub-panel at the upper flange and bottom flange is equal in magnitude but is opposite in direction. At the oblique part of the web, the total forces at the top flange are balanced to the total forces at the oblique part of bottom flange. These opposite forces forms a couple, which is known as secondary bending moment, Myf due to the lever arm between them, a = 250 mm for each couple along the section. Myf = Q 250 (3.5)

Analysis, Design and Performance of Steel Section with Trapezoid Web

33

Table 3.1 Lateral reaction on nodes of inclined web for some typical section sizes
Lateral reaction each node top flange (kN) Web 1 Web 2 Web 3 Lateral reaction each node bottom flange (kN) Web 1 Web 2 Web 3

Section size

300 120 10

-2.357 -0.004 -3.752 -0.003 -1.079

1.727 0.004 3.748 0.004 1.727 7.210

-1.080 -0.003 -3.752 -0.004 -2.358 -7.197

2.357 0.004 3.752 0.003 1.079 7.195

-1.727 -0.004 -3.748 -0.004 -1.727 -7.210

1.080 0.003 3.752 0.004 2.358 7.197 7.201

Total, Q

-7.195

350 120 10

-2.022 -0.003 -3.218 -0.003 -0.926

1.481 0.003 3.213 0.003 1.481 6.182

-0.927 -0.003 -3.218 -0.003 -2.023 -6.173

2.022 0.003 3.218 0.003 0.926 6.172

-1.481 -0.003 -3.213 -0.003 -1.481 -6.182

0.927 0.003 3.218 0.003 2.023 6.173 6.176

Total, Q

-6.172

400 120 10

-1.771 -0.003 -2.817 -0.002 -0.811

1.297 0.003 2.812 0.003 1.297

-0.811 -0.002 -2.817 -0.003 -1.771

1.771 0.003 2.817 0.002 0.811

-1.297 -0.003 -2.812 -0.003 -1.297

0.811 0.002 2.817 0.003 1.771

Q (Average)

34

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

Table 3.1 (continued) Lateral reaction on nodes of inclined web for some typical section sizes
Lateral reaction each node top flange (kN) Web 1 Web 2 Web 3 Lateral reaction each node bottom flange (kN) Web 1 Web 2 Web 3

Section size

450 120 10

-1.575 -0.003 -2.504 -0.002 -0.721

1.153 0.003 2.499 0.002 1.153 4.812

-0.722 -0.002 -2.504 -0.003 -1.575 -4.806

1.575 0.003 2.504 0.002 0.721 4.806

-1.153 -0.003 -2.499 -0.002 -1.153 -4.812

0.722 0.002 2.504 0.003 1.575 4.806 4.808

Total, Q

-4.806

500 120 10

-1.418 -0.002 -2.254 -0.002 -0.650

1.039 0.002 2.250 0.002 1.038 4.331

-0.650 -0.002 -2.254 -0.002 -1.419 -4.327

1.418 0.002 2.254 0.002 0.650 4.326

-1.039 -0.002 -2.250 -0.002 -1.038 -4.331

0.650 0.002 2.254 0.002 1.419 4.327 4.328

Total, Q

-4.326

Q (Average)

Analysis, Design and Performance of Steel Section with Trapezoid Web

35

1.727 0.004 3.748 0.004 1.727

2.357 0.004 3.752 0.003 1.079

1.080 0.003 3.752 0.004 2.358

(a)

Force on each node at the top flange due to the reaction to the oblique web
7.195kN 7.197kN 7.210kN

7.210kN 7.197kN (b)

7.195kN

Total lateral forces at each oblique sub panel

Figure 3.4 Lateral forces induced in the flanges at oblique sub panel for the web depth 300 mm

Q varies linearly with V. Therefore Myf varies linearly with V and giving, Myf = V Co (3.6)

where Co is a coefficient called the coefficient of secondary

36

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

bending moment, which has the unit in mm. Co is obtained by calibrating between the applied shear loading and the secondary bending moment resulted from the couples. Co = Q 250 / V (3.7) The values of Co from finite element analysis are calculated based on the equation above. The values are compared to Co values from German existing table capacities. Table 3.2 shows the ratio of Co/Mcyf from calculation compared to Co/Msec,o ratio from the existing capacities table. The relations of Co/Msec,o versus web depth are illustrated in Fig. 3.5.
Table 3.2 The comparison of Co between the FEM analysis and the existing manufacturers standard table
D (mm) 300 350 400 450 500 550 600 650 700 750 Qavr (kN) 7.201 6.176 5.406 4.808 4.328 3.935 3.608 3.310 3.094 2.888 V (kN) 20 20 20 20 20 20 20 20 20 20 Co (mm) FEA 90.01 77.20 67.58 60.10 54.10 49.19 45.10 41.38 38.68 36.10 Existing 98.60 89.00 80.30 73.00 66.90 61.10 56.80 52.60 49.60 46.50 Mcyf kNm 8.52 8.52 8.52 8.52 8.52 8.52 8.52 8.52 8.52 8.52 Msec,o kNm 11.64 12.38 12.87 13.23 13.53 13.66 13.88 13.98 14.16 14.32 Co/Mcyf FEA 10.58 9.06 7.93 7.05 6.35 5.77 5.29 4.86 4.54 4.24 Co/Msec,o Existing 8.47 7.19 6.24 5.52 4.94 4.47 4.09 3.76 3.50 3.25

Note: D = overall depth, Flange width = 120 mm, Flange thickness = 10 mm, web thickness = 2 mm

The results from the finite element analysis show that the values of Co and ratio Co/Mcyf decrease with the increase of the web depth. From the Table 3.2, values of Co from finite element analysis are

Analysis, Design and Performance of Steel Section with Trapezoid Web

37

lower than Co given in the existing manufacturers standard table of properties. However, values of ratio Co/Mcyf from the analysis are higher than the existing Co/Msec,o , as shown in Fig. 3.5. This is because Mcyf is lower than Msec,o. From Fig. 3.5, the pattern of graphs plotted from finite element analysis is similar with the existing standard table of capacities. Hence, the derivation of secondary bending moment, Myf and secondary bending moment, Co from the finite element analysis give reliable comparison with the existing table capacities.

3166.3x-1.000

B = 120mm T = 10mm

3252.4x-1.044

Figure 3.5 Co/Msec,o vs. depth of web, from the finite element analysis and the existing value

3.6 DERIVATION OF THE SECONDARY BENDING MOMENT

COEFFICIENT

OF

In order to develop a general formulation for the value of Co, a parametric study was carried out, by using sections with various dimensions, i.e. depth of section ranging from 300 to 1300 mm,

38

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

flange width from 120 to 350 mm, flange thickness from 10 to 30 mm and web thickness form 2 to 8 mm. It was noted that the web thickness do not affect much on the secondary bending moment. From the data of the analysis, a best fit equation has been developed for the value of Co/Mcyf, i.e.
Co 2 x 1010 = M cyf B 3.2235 x D (10.0008T )

(3.8)

where B is the flange width (mm), D is the web depth (mm), T is the flange thickness (mm), Co is the coefficient of secondary bending moment (mm) and Mcyf is the minor axis moment capacity of each flange (kNm). By substituting various dimension data, a typical graph relating the secondary bending, section depth and flange width was plotted, as shown in Fig. 3.6. The formula developed in Eq. (3.8) is compared with the finite element results and the values in the existing table, as shown in Fig. 3.7. They have a similar pattern of curve. It shows that the formula for the secondary bending moment developed in this study can be reliably used for design. Another significant note is that the curve generated from the formula is higher than the curve for the existing table. This means that the formula will give a more conservative design because it over estimate the values of the secondary bending moment.

Analysis, Design and Performance of Steel Section with Trapezoid Web

39

Figure 3.6 Co/Mcyf vs. web depth from formula result for various section

40

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

(a)

(b) Figure 3.7 Graph Co/Mcyf vs. section depth and flange width

3.7

CONCLUSION

Secondary bending moment may reduce the bending moment capacity of a web with trapezoid web. A series of finite element analysis on various size and properties of trapezoid web beam

Analysis, Design and Performance of Steel Section with Trapezoid Web

41

under shear loading was carried out by using LUSAS finite element analysis. The lateral reactions at the top and bottom flange were obtained for different geometric parameters. The data were used to calculate the secondary bending moment, Myf induced in each flange and the coefficient of secondary bending moment, Co. From the parametric studies using a wide range of geometrical section properties, curves were plotted relating the values of Co, the moment capacity about minor axis of each flange, Mcyf, and various section properties, a general formula for Co/Mcyf has been developed. The formula was verified by comparing with the existing manufacturers standard table of properties and found to be more conservative to be used in design. By referring to the formula, it will be become easier for local engineers to perform their design work, choose and determine a safer and more suitable size of trapezoid web beam. Various size of trapezoid web beam can be chosen and manufactured. It is also suggested that experimental work should be carried out to obtain the value of lateral reactions in future research. From the experiment, the results can be compared with the formula of Co that derived in this study. By comparing the results from both the experimental and analytical studies, the formula of Co can be improved and verified. This will result in more reliable value of Co which could be considered in the future study and developed into trapezoid web beam design work. REFERENCES
1.

Hussein, W.Q. (1997), Design Concept of Trapezoid Web Profile, Lecture notes and technical report, Johor: Antara Steel Mill Sdn. Bhd. 2. Nina Imelda (2003), The Effect of Opening on the Strength of Corrugated Web Plate Girder Subjected to Shear, Master Thesis, Universiti Teknologi Malaysia, Skudai. 3. R. C. Hibbeler (2003), Mechanics of Materials fifth edition, Upper Saddle River, New Jersey: Pearson Education, Inc.

42

Analytical Study on Secondary Bending Moment in Trapezoid Web Beam

4. Ferdinand P. Beer, E. Russell Johnston, Jr. (1981), Mechanics of Materials, United States of America: McGraw-Hill, Inc. 5. Hussein, W.Q. (1998), Design Concept for Steel Plate Girder with Corrugated Webs (TWP), Pasir Gudang, Johor: Trapezoid Web Profile (TWP) Sdn. Bhd. 6. LUSAS Modeller User Manual, Version 13.5 (2002), United Kingdom: FEA Ltd. 7. Fong, S,W,Secondary Bending Moment in Trapezoid Web Section subjected to shear force, Master Thesis (2006), Universiti Teknologi Malaysia.

4
BUCKLING ANALYSIS OF COLUMN WITH TRAPEZOID WEB
Mohd Hanim Osman, Tan Cher Siang and Abd. Latif Salleh
4.1 INTRODUCTION

Trapezoidal web section is a type of flanged steel section with the web part made corrugated in trapezoidal form (Fig. 4.1). Plate girders fabricated from TWP section offer an alternative to conventional hot rolled and welded sections of uniform web profile. They have been proven to be more economical in terms of materials used and the time of fabrication [1]. Study on trapezoid web girder sections has been conducted in the Steel Technology Centre, Universiti Teknologi Malaysia to develop a design guide for beam with trapezoid web profile. For compression member, a comprehensive analytical and experimental study is required to complete the design guide. There is no evidence of research effort in this subject. The objective of this research is to develop the formulation for the buckling capacity of column with trapezoid web. The scope of the study is focused on the comparison of lateral bending behaviour between flat web (FW) and TWP section to study the effect of corrugated profile on the second moment of area (I) of the section, by experimental, analytical and numerical approaches. It was followed by the study on buckling load using eigenvalue analysis in the LUSAS finite element program. The deflection and buckling study was for the minor axis only. For the comparison between theoretical and experimental results,

Analysis, Design and Performance of Steel Section with Trapezoid Web

44

Euler buckling load is assumed in the theoretical elastic critical buckling load. Based on the results of the finite element analysis, and in comparison with the experimental results, a design guide for the design of column with trapezoid web profile is proposed. Experimental study is the best approach to determine the actual behaviour of TWP column. K.J.R. Rasmussen [2], Saleh H. AlSayed [3,4] and E.M. Basista [5] had done tests on columns fabricated from slender plates, single angle columns and channelcolumns. Their studies show that a direct column testing requires sophisticate and well-planned set-up. They also involved a large amount of testing specimens (about 10 ~ 26 specimens for each testing), which may be not practical and fund-effective in this research. Therefore, computer modelling is recommended.

hr

x D y y

Figure 4.1 Trapezoid web profile

45

Buckling Analysis of Column with Trapezoid Web

4.2 BUCKLING CURVE IN COLUMN DESIGN 4.2.1 Buckling Curve for Columns with Flat Web Section
The basic formula for the buckling of pin ended strut based on Euler is given by:
PE =

2 EI
L2

(4.1)

In BS 5950: Part 1: 2000 [6], the basic buckling formula is modified empirically according to the types of section, thickness of steel and methods of fabrication. For the welded I section, the buckling curve is given in Table 24(b), which is based on PerryRobertson formula given in the standard. 4.2.2 Proposed Design Formula for Column with Trapezoid Web Section In the design of the compression member with trapezoid web, the proposed design guide is based on the same standard and neglecting web in the calculation of the moment of inertia of the section. The design buckling strength is reduced by 10% in order to obtain a conservative but safer design load, after considering that the deflection of trapezoid web about its major axis is 10% higher than flat web [7]. However, no evidence of experimental and analytical study have proved this proposed method so far, thus it is unclear of the conservativeness level. 4.3 SECOND MOMENT OF INERTIA OF TRAPEZOID WEB From the formula, the second moment of area (I) stands as an important factor in the column strength. The column usually starts

Analysis, Design and Performance of Steel Section with Trapezoid Web

46

buckle at weakest axis (minor axis or y-y axis), so Iy will be the main concern in this research. For TWP, Iy cannot be calculated mathematically because the corrugated web gives a non-uniform profile. The first approach is to neglect the web, made the Iy value smaller but safer in use. There are few alternative proposed formula for the moment of inertia: (i) Neglect the contribution of web, a simplified formula proposed by Wail [1] and Osman [8]. (ii) Include the contribution of web by assuming similar to symmetrical section as is studied herein. (iii) Similar to (ii) but reducing 10% to consider the finding by Ihsan [7] that the deflection of trapezoid web is greater than the flat web. (iv) Include the web as proposed by unquoted reference [1]. In the proposed Eq. (4.1) the formula is given as:
I y = I yf = t f x B3 12

(4.2)

In the proposed Eq. (4.2) the formula is given as:


t f x B3 d x tw3 + I y = I yf + I yw = 12 12

(4.3)

In the proposed Eq. (4.3) the contribution of web is given by:


2 3 h t x hr I yw = 2 x d x t w r + w 2 6 x sin

(4.4)

Although Eq. (4.4) tries to take into account the contribution of web in more detail, it results in an unreliable value of moment of

47

Buckling Analysis of Column with Trapezoid Web

inertia, thus is not considered further here. For the range of standard sections of TWP, it has been found that the contribution of web to the moment of inertia about the minor axis of the section is less than 0.5% which is not significant. This is because the web is very thin compared to other element. 4.4 STUDY OF LATERAL BENDING

The deflection of simply supported beam, length L under a midpoint load P is given by the simple formula as:
max =
PL3 48EI

(4.5)

The value of the second moment of area I is the only sectional parameter influencing the deflection. By knowing the P and value, I can be obtained as :
P I = max L3 48 E

(4.6)

For two beams of the same material and length , the ratio of their moment of inertia can be written as:
I1 I2 P L3 48 E 2 1 = = 3 1 P L 48 E 2

(4.7)

i.e. I is inversely proportional to deflection . For the conventional section with uniform web profile, I is simply calculated. For nonuniform section such as the trapezoid web, there are a number of proposed formulas which none of them have been verified. A load-deflection test was carried out for both beam specimens with trapezoid web and flat web, 300 120 10 2 (D B T t) sections. The span of beams was 5 m, 4 m and 3 m, with a load

Analysis, Design and Performance of Steel Section with Trapezoid Web

48

at mid-span. Load actuator 100 kN capacity was used to apply point load on the beam and the mid-span deflection was measured using displacement transducers and recorded in the data acquisition logger. The result is presented in Fig. 4.2 which shows lower deflection of trapezoid web section (range from 2 6%). It means that the moment of inertia of section with trapezoid web is higher than that of flat web accordingly.

12.0 10.0 Loadi ng, P (kN) 8.0 6.0 4.0 2.0 0.0 0.00 10.00 20.00 30.00 40.00 50.00 TWP Test Res ult FW Tes t Result Theoretic al Value TWP FEM FW FEM

Deflec tion (mm)

Figure 4.2 Load deflection relationship of 5 m length beam section

Numerical study was carried out on the same model using LUSAS finite element method. The finite element model of trapezoid web beam is shown in Fig. 4.3, which also shows the deflection in minor axis. The deflection of TWP beam was 3% lower than flat web. The experimental and numerical results, together with the analytical calculation using Eq. (4.6) were shown in Fig. 4.3 for a typical length of 5 m. It is shown that the deflection of trapezoid web beam is slightly lower than the flat web beam of about 3%. It is shown that the result of finite element analysis is equal to the theoretical formula as in Eq. (5). The experimental results are found to be less than

49

Buckling Analysis of Column with Trapezoid Web

the theoretical results. However, this does not affect the conclusion of the study because this is a comparative study which is based on the flat web section as control.

Figure 4.3 The finite element model of beam and the deflection in minor axis

4.5 CRITICAL BUCKLING ELEMENT METHOD

LOAD

BY

FINITE

The Eigenvalue buckling analysis in the finite element method was used to determine the elastic critical buckling load of column. A total of 23 trapezoid web column models were analysed, ranging from depth 300 to 1500 mm, covering the full range of standard sections produced by fabricators. The lengths of 4, 6 and 8 meter were used for each section, giving the slenderness of column ranging from 50 to 256, and to the total of 253. Almost all models ran successfully, few were failed in local buckling before achieving elastic member buckling. Fig. 4.4 shows a typical buckling mode of column.

Analysis, Design and Performance of Steel Section with Trapezoid Web

50

Figure 4.4 A typical buckling mode of column

The results for the length of 4 meter were presented in Table 4.1. It is noted that the buckling load increased linearly with the increase in the second moment of area of the sections, as shown in Fig. 4.5. This is as expected by Eq. (4.1) i.e. the buckling load for a certain length of column increase linearly with the second moment of inertia of the section. It also verifies the reliability of the modelling technique and the use of Eigenvalue buckling analysis in the finite element program. The theoretical buckling load (assuming flat web section) is also shown in the figure. The buckling load of trapezoid web column is found to be higher (maximum 2.71%) than the flat web in the range of smaller

51

Buckling Analysis of Column with Trapezoid Web

sections, but becomes lower (minimum 1.49%) than the flat web for larger sections. The phenomena that the buckling load of trapezoid web decreases for larger section is clearly shown in Fig. 4.6. The reason is due to the ratio between the corrugation thicknesses, hr in relation to the width of section, B. For the larger sections, the ratio, hr/B is low. This is in agreement with the report by Siang [9], that the deflection in minor axis direction increases when the corrugation thickness decrease. Since I is inversely proportional to deflection, it follows that the critical buckling load decrease accordingly. The variation of (Pcr-Pe)/Pe for different hr/B is shown in Fig. 4.7. It seems that there exists a consistent relationship between the two parameters. Taking the lowest envelop as the conservative values, the point of hr/B = 0.45 is the reference for which the buckling of trapezoid web and flat web of the same web and flange thickness are equal. Therefore, a formula can be derived for the buckling capacity of trapezoid web column, Pcr, as follows:
Pcr = h Pe 1 + 10 r 4.5 B

(4.8)

where Pe is the buckling capacity of the section assuming as flat web, which can be obtained from the buckling curve in BS 5950: 2000. In the calculation of Pe, it does not make any difference of whether the web is included or not.

Analysis, Design and Performance of Steel Section with Trapezoid Web

52

Table 4.1 Buckling load finite element method and Euler formula
Web included No Geometrical dimension (mm) D 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 300 450 450 550 550 650 650 650 650 750 750 750 900 900 900 1000 1000 1000 1000 1200 1500 1500 1500 B 120 140 180 120 180 140 140 160 180 120 180 250 140 180 250 180 200 240 270 200 200 250 300 T 10 10 12 10 12 10 10 12 12 10 12 20 10 12 20 12 16 16 16 16 16 18 20 t 2 2 4 2 4 2 2 2 4 2 4 4 2 4 4 2 2 2 4 2 3 3 3 Iy (mm4) 2880187 4573620 11666272 2880353 11666805 4573753 4573753 8192417 11667339 2880487 11667872 52087120 4573920 11668672 52087920 11664651 21333979 36864645 52493163 21334112 21336636 46878294 90003285 Pe (kN) 364 578 1475 364 1475 578 578 1036 1475 364 1475 6587 578 1476 6587 1475 2698 4662 6638 2698 2698 5928 11381 Web neglected Iy (mm4) 2880000 4573333 11664000 2880000 11664000 4573333 4573333 8192000 11664000 2880000 11664000 52083333 4573333 11664000 52083333 11664000 21333333 36864000 52488000 21333333 21333333 46875000 90000000 PE (kN) 364 578 1475 364 1475 578 578 1036 1475 364 1475 6586 578 1475 6586 1475 2698 4662 6637 2698 2698 5928 11381 Pcr TWP (kN) 374 587 1487 374 1487 587 587 1040 1487 374 1487 6521 587 1487 6521 1476 2687 Fail! 6551 2687 2693 5864 11211 Diff. % 2.71 1.5 0.78 2.71 0.79 1.52 1.52 0.42 0.81 2.72 0.81 -0.99 1.51 0.82 -0.98 0.08 -0.4 0 -1.3 -0.39 -0.16 -1.07 -1.49

53

Buckling Analysis of Column with Trapezoid Web

12000 10000
Buckling load (kN)

(a)
Flat web

8000 6000 4000 2000 0 0 2000 4000 6000 8000 10000


Moment of inertia (cm4)

Figure 4.5 Buckling load versus the second moment of area

3 2
(Pcr-Pe)/Pe (%)

1 0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

-1 -2
Moment of inertia (cm4)

Figure 4.6 Variation of Pcr/Pe with the second moment of inertia

Analysis, Design and Performance of Steel Section with Trapezoid Web

54

3 2 (Pcr-Pe)/Pe 1 0 0.000 -1 -2 hr/B

0.100

0.200

0.300

0.400

0.500

0.600

0.700

Figure 4.7 The influence of corrugation thickness ratio to the buckling load

4.6

CONCLUSION

From the study on the derivation of buckling capacity of column sections with trapezoid web profile, it can be concluded that: (a) The contribution of web to the second moment of area of trapezoid web of standard sections is negligible. (b) Experimental results showed that the deflection of trapezoid web beam is less than the flat web section. The deflection of beam with trapezoid of lower hr/B is than the flat web, but is greater for low hr/b. (c) The Eigenvalue buckling analysis has been used successfully to estimate the critical buckling load of compression member. (d) The buckling of column with trapezoid web has been determined. It is higher than flat web for higher hr/B and lower for lower hr/B. The most important conclusion is that the buckling formula for trapezoid web column has been proposed.

55

Buckling Analysis of Column with Trapezoid Web

REFERENCES 1. Wail Q. Hussein (2001). Design Guide for Steel Plate Girders With Corrugated Webs (TWP). Johor: Trapezoid Web Profile Sdn. Bhd. Nethercot, D.A. & Lawson, R.M. (1992). Lateral Stability of Steel Beams and Columns Common Cases of Restraint. Berkshire: The Steel Construction Institute. K.J.R. Rasmussen, al et (1990). Design of Columns Fabricated from Slender Plates. Journal of Construction Steel Research Vol. 17. Great Britain: Elsevier Science Publishers Ltd. Saleh H. Al-Sayed & Reidar Bjorhovde (1989). Experimental Study of Single Angle Columns. Journal of Construction Steel Research Vol.12. Great Britain: Elsevier Science Publishers Ltd. E.M. Batista & E.C. Rodrigues (1994). Buckling Curve for Cold-Formed Compressed Members. Journal of Construction Steel Research Vol. 28. Great Britain: Elsevier Science Publishers Ltd. British Standards Institution (2000). BS5950: The Structural Use of Steelwork in Building, Part 1. London: BSI. Hasni@Ihsan b. Atan (2000). Flexural Behavior of Trapezoid Web Plate Girder. UTM: Master Thesis. Hanim Osman & Salina Jamali (2001). Report of Design Guide Using Trapezoid Web Profile (Calculation). UTM Steel Technology Center: Technical Report (Sept). Tan Cher Siang, The second moment of area of trapezoid web section, Technical Report submitted to Faculty of Civil Engineering, UTM, 2003.

2.

3.

4.

5.

6. 7. 8.

9.

5
PERFORMANCE TEST ON SIMPLY SUPPORTED COMPOSITE BEAM WITH TWP STEEL SECTION
Goh Kee Keong and Mohd Hanim Osman

5.1

INTRODUCTION

Steel beams with trapezoidal web profile (TWP) have been widely and increasingly used in recent years mainly due to corrugated webs have allowed the use of thin plates without the need for stiffeners. On the other hand, beams with steel-concrete composite action are one of the most commonly used structural elements because they considerably increase flexural strength and stiffness of steel beams. However, there are few or no experimental test data available which to check the performance of TWP steel section acting compositely with concrete. Consequently, the search for this experimental data has been the main concern in this project and the experiment is briefly described and discussed in this paper. The objectives of this project are: (a) to compare the ultimate performance of composite beam with TWP steel section to composite beam with I-plate girder; (b) to study the suitability of adopting composite design method by BS5950: Part 3: Section 3.1 [1] for designing composite beam with TWP steel section; (c) to obtain the stress and strain distribution and the position of neutral axis of the composite beam with TWP steel section. These will also allow better understanding of the true behaviour of composite beam with TWP steel section.

Analysis, Design and Performance of Steel Section with Trapezoid Web

57

5.2 5.2.1

METHODOLOGY Test Specimens

A steel-concrete composite beam, with trapezoidal web steel section, is designed base on BS5950: Part 3: Section 3.1. In which, the design assumed that the ultimate bending strength may be taken as the plastic moment capacity of the steel and concrete acting compositely, given as the function of their plastic section modulus and their material yield strength. From the design, the proposed test specimen with steel cross section of 300 120 mm and concrete flanges of 1000 mm breadth and 110 mm thickness is shown in Fig. 5.1. As control, a similar set of composite beam with I-plate girder will also be tested. 5.2.2 Test Setup

The apparatus for the testing programme as shown in Fig. 5.2 and 5.3 was arranged in such a way that a plastic failure mechanism could be developed in the specimens without incurring significant frictional forces [2]. Frictional forces developed at the loading and reaction points were reduced through the use of rollers at the reaction and loading points. Torsion restraint was provided at the loading points by adjustable torsion restraints. At each section labeled A, B, and C, five strain gauges were attached; one at concrete and the remaining gauges at the steel section. Concrete and steel surfaces where gauges will be attached were grinded to remove paint and rust as well as to provide a smooth surface for effective bonding.

58 Performance Test on Simply Supported Composite Beam with TWP Steel Section

1000 110 10

300 2 120

280

10

(a)

Composite beam with TWP section


1000 110

10

300 2 120

28

10

(b)

Composite beam with I-plate girder (Control specimen) Figure 5.1 Test specimens

Analysis, Design and Performance of Steel Section with Trapezoid Web

59

LOAD

Free Rollers

Load Cell Transfer Beam

B 2m

A 1m 5m

C Load S preader

Figure 5.2 Elevation of the test arrangement, with strain gauges attached to provide sections information at A, B, and C. (Torsion restraints are not shown)
Line Loads

Rollers

1.0 m Torsion Restraints

Figure 5.3 Plan view of the testing arrangement

60 Performance Test on Simply Supported Composite Beam with TWP Steel Section

5.3

TEST RESULTS Beam with I-Plate Girder (Control

5.3.1 Composite Specimen)

The bottom steel flange of composite beam starts to yield when the load is about 175 kN at mid span, which is at the point of maximum bending moment. Following that, at the distance 1.0 m from mid span at B as shown in Fig. 5.2, the bottom flange started to yield at load 180 kN. On the other hand, the yielding effects continue to happen and spread towards the neutral axis at the mid span. However, premature web buckling then occurs at loading 198 kN at section B, which prevents the development of the full bending strength. Elastic neutral axis and plastic axis is different. Neutral axis gradually shifted upwards from elastic condition to plastic condition. From the experiment, the elastic neutral axis (E.N.A.) is 88.50 mm from concrete top surface whereas plastic neutral axis (P.N.A.) is 67.24 mm with reference to the same base line. The concrete never reached its maximum compression stress throughout the experiment. As concrete is known for its weak tension resistance, hairline cracks are observed at mid span of the beam at the bottom edge of concrete and along the longitudinal direction at middle of the beam, which is believed to be resulted from longitudinal splitting forces. It is concluded that a combined failure of bending and shear happens to the beam.

5.3.2

Composite Beam with TWP Steel Section

At the load of 150 kN, the bottom steel flange of the specimen reaches its yield strength at the mid span, it follows that the yield of bottom flange starts to happen at C after 160 kN. Excessive

Analysis, Design and Performance of Steel Section with Trapezoid Web

61

deflection occurs when the load reaches 200 kN, which gave a significant change of the section from elastic condition to lower bound of plastic condition. First observed hairline crack appears at the load of 180 kN, and there is minor vertical buckling of flat web near mid span at this stage. Starting at 210 kN and onwards, more minor cracks can be easily observed along both of the concrete edges at the mid span region. These cracks include the longitudinal splitting cracks on top of the concrete surface. From the experimental data, both steel flanges and concrete material reached their tensile and compressive strength respectively at 230 kN, meaning that full plastic section is developed. Trapezoidal web buckling also initiated at B at this particular loading. The elastic neutral axis (E.N.A.) is 81.35 mm whereas elasticplastic neutral axis is 69.71 mm. During the end of experiment, concrete reaches its maximum compression stress at mid span with the corresponding load of 230 kN.

5.3.3

Data Analysis and Comparison

Considering the deflection of the two specimens in Fig. 5.4, within the elastic behaviour of the beams, the composite effect with the concrete material has brought to the common deflection for both types of steel section. Previous test on the deflection of only the steel sections has shown that TWP section generally deflect more than I-plate girder under the same loading. And, some other researchers of corrugated web beam have agreed such result. However, within the elastic-plastic region, the composite beam of TWP section deflects less than the control specimen of I-plate girder.

62 Performance Test on Simply Supported Composite Beam with TWP Steel Section

250

200

Loadings (kN)

150

100 I-girder TWP

50

0 0 5 10 15 20 25 30 35 40 45

Midspan Deflection (mm)

Figure 5.4 Plot of loading (kN) against mid span deflection (mm)

0 50

Experimental E.N.A. = 88.5mm


100 Section Depth (mm) 150 200 250 300 350 400 -1000 -500 0 500 Strains ( ) 1000 1500 2000 2500 50 100 150 180 195 198

Figure 5.5 The distribution of mid span strain at loadings 50, 100, 150, 180, 195, and 198 kN for control specimen. (198 kN is the failure load)

Analysis, Design and Performance of Steel Section with Trapezoid Web

63

0 50 100 Section Depth (mm) 150 200 250 300 350 400 -1000 -500 0 500 Strains ( ) 1000 1500 2000 2500 50 100 150 180 200

Experimental E.N.A. = 81.35mm

Figure 5.6 The distribution of mid span strain at loadings 50, 100, 150, 180, and 200 kN for specimen with TWP. (200 kN is the failure load)

The control specimen gave a set of satisfying plots of strain distribution along the section depth, see Fig. 5.5, in which the value of positive strains (showing tensile stresses) increase proportionally to the distance from neutral axis. Such strain distribution is often been idealized into linear strain distribution. When composite beam with TWP section is tested, the experimental results show that the web contributes much lesser to the tensile strength of the steel section. This effect is particularly obvious at the diagonal web as shown in Fig. 5.6. For this reason, the flanges of the TWP specimen will have to resist the additional tension force that is not taken by the trapezoidal web. This explained a slightly higher elastic neutral axis.

64 Performance Test on Simply Supported Composite Beam with TWP Steel Section

5.4

DISCUSSION

The objective of stressing the composite beam with I-plate girder to ultimate failure is not reached due earlier occurrence of shear failure before the whole steel section at mid span yield. No stiffeners have been provided to stiffen the web at the shear area is the main reason to this. However, stiffeners are not proposed in this experiment because that will eliminate the advantage of using TWP. Thus, the results obtained from this experiment as sgiven in Table 5.1 correspond to the lower bound condition of the ultimate failure. Considering the control specimen of I-girder, the obtained position of elastic neutral axis is rather accurate with only 0.2% and 2.23% differences to the theoretical and design value respectively. However, the plastic neutral axis is much greater than the expected value. This can be explained through several reasons. Before that, it should be noted that the difference of theoretical and design value is due to the introduced partial safety factor of concrete material, m. The first reason for such difference is that the experimental value is obtained when the steel section is not fully yielded. In the rigid plastic theory of composite beam, in order to maintain equilibrium, the whole steel element must yield and only part of the concrete element will be fully yielded [4]. The second reason is the possible slip strain, which is hardly avoided though full interaction shear connectors have been provided. Moreover, the strain gauge has not been prepared for the bottom surface of the concrete. When comparing the elastic neutral axis of composite beam with TWP, higher position of neutral axis can be explained through the tension strain that is concentrated at the flanges of TWP steel section, and not because of a stronger steel section is being used. A slightly rise in concrete strength might as well be the reason because the concrete ages are different in the two experiment, but its significance is yet to be determined.

Analysis, Design and Performance of Steel Section with Trapezoid Web

65

In the elastic-plastic stage, particularly after the yielding of bottom flange, composite beam of TWP steel section is able to enhance greater force and deflects less before the web buckles when compared to control specimen of I-girder.
Table 5.1 Comparison of the Position of Elastic Neutral Axis and Plastic Neutral Axis (all values are in mm, calculated from the concrete top surface)
Design [3] (BS5950: Part 3: Section 3.1)

Composite Elastic 88.66 86.57 88.50 Beam with (E.N.A.) I-plate girder Plastic 35.10 56.31 67.24 * (Control (P.N.A.) specimen) Composite Elastic 83.16 79.52 81.35 beam with (E.N.A.) TWP steel Plastic 29.15 46.81 69.71 * section (P.N.A.) * Position of initial elastic-plastic neutral axis

2.5

CONCLUSION

The performance of composite beam with TWP steel section, before the first yielding occurs, needs to be monitored carefully. TWP section does not show any advantage compared to I-plate girder. However, within the elastic-plastic region, especially after the bottom flange reached its yield strength, TWP section shows a better performance with less deflection and stiffer web from

Experimental

Neutral Axis

Theoretical

Specimen

66 Performance Test on Simply Supported Composite Beam with TWP Steel Section

buckling. It was also found that elastic neutral axis slightly rise when I-plate girder is replaced with TWP steel section in composite beam. The performance of TWP-composite beam, before the first yielding at bottom flange occurs, does not show any excel advantage compared to I-girder-composite beam. However, within the elasticplastic region, especially after the bottom flange reached its yield strength, TWP section shows a better performance with less deflection as well as stiffer web from buckling. Finally it is concluded that from the experiment carried out in this project, BS5950: Part 3: Section 3.1 is less conservative to be used in designing TWP-composite beam. However, this remark is limited to the service limit state of the beam. The performance of TWP-composite beam in ultimate state is yet to be determined. As an overview to the project, it was found that the flexural behaviour of TWP-composite beam is as expected by the BS5950: Part 3 because the design method neglects the strength of steel web in calculating composite moment capacity. Such behaviour is found also in the TWP-composite beam from the experiment. But when compared to a conventional I-girder-composite beam, it seems that the TWP-composite is less conservative when BS5950: Part 3 design method is adopted. Hence, it is concluded that by introducing suitable coefficient into the design method, particularly for the serviceability limit state design, BS5950: Part 3 might still be a practical design method for designing TWP-composite beam. REFERENCES 1. BS5950: Part 3: Section 3.1: 1990. 2. Byfield M.P., Nethercot D.A., An Analysis of the True Bending Strength of Steel Beams, Institution of Civil Engineers Structures & Buildings, 1998.

Analysis, Design and Performance of Steel Section with Trapezoid Web

67

3.

Commentary on BS5950: Part 3: Section 3.1-Composite Beams, The Steel Construction Institute, 1990. 4. Oehlers D.J., Bradford M.A., Composite Steel And Concrete Structural Members Fundamental Behaviour, Pergamon, 1995. 5. Chapman J.C., Balakrishnan S., Experiments On Composite Beams, The Structural Engineer, Vol. 42, Nov. 1964.

6
EXPERIMENTAL TEST ON STEEL BEAM WITH PARTIAL STRENGTH CONNECTIONS USING TRAPEZOID WEB PROFILED STEEL SECTIONS
Arizu Sulaiman, Mahmood Md Tahir and Anis Saggaff

6.1

INTRODUCTION

For a typical steel building frame, the connection between the beam and column is either assumed as pinned, where only nominal moment from the beam is transferred to the column, or rigid or full strength, where full continuity of moment transfer exists. Alternatively, EC 3 (ENV 1993-1-1: 2002) [1] allows building frames to be designed as semi-rigid using the partial strength connection, provided that the moment resistance of the connection can be quantified. When incorporated into the construction of a whole frame, the type of construction that uses the partial strength connection is referred to as a semi-continuous construction, due to the partial continuity that exists between the beam and column. Unlike the conventional design approaches (simple and rigid), semi-rigid design requires the moment-rotation relationships of partial strength connection, which includes the moment resistance and rotational stiffness (rigidity), to be established prior to its usage in design. In this research, the behaviour of partial strength connections with TWP sections as beams has been studied. The purpose of using TWP sections is to take advantage of the benefits offered by the sections in general.

Experimental Test on Steel Beam with Partial Strength Connection

69

6.2 6.2.1

BACKGROUND Partial Strength Connections

By definition, if the moment resistance or capacity of a connection is lower than the moment capacity of the connected beam, the connection is referred to as the partial strength connection. Two types of partial strength connections that are commonly used are the flush endplate connection and extended endplate connection. These connections consist of a plate, which is welded to the beams end in the workshop, and then bolted to the column on site. In the case of extended endplate, the plate is extended above the flange of a beam and with one row of bolt in case of extended endplate connection). In the design of braced multi-storey steel frames, the steel weight of the connections may account for less than 5% of the frame weight; however, the cost of the fabrication is in the range of 30% to 50% of the total cost [4]. The fabrication of partial strength connections may be marginally more expensive since some degree of rigidity has to be provided. However, by using partial strength connections instead of simple connections, beam sizes could be reduced and significant overall savings of frame weight could be acquired [3], [4]. In Malaysia where the cost of labour is relatively low compare with the Europe, the use of the proposed connections will be an added advantage. It has been reported that the savings in steel weight of using partial strength connection alone (noncomposite) in multi-storey braced steel frames using British hotrolled section was up to 12% [5]. The overall cost saving was up to 10% of the construction cost, which is quite significant according to the cost of labour in the United Kingdom [6].

70

Analysis, Design and Performance of Steel Section with Trapezoid Web

6.2.2

Trapezoid Web Profiled (TWP) Steel Sections

A trapezoid web profile plate girder is a built-up section made up of two flanges connected together by a thin corrugated web usually in the range of 2 to 8 mm. The web is corrugated at an angle of 45 degree and welded to the two flanges by using automated machine. Since the web and flanges may comprise of different steel grades, TWP section is also classified as a hybrid steel section. The steel grade of the flanges is usually designed for S355, so that the flexural capacity of the beam can be increased, whilst the steel grade of the web is usually designed for S275, so that the cost of steel material can be reduced since the shear capacity is usually not critical [6]. The use of different steel grades in the fabrication of TWP section leads to further economic contribution in addition to the contribution from using partial strength connections. The thick flanges, thin web and deeper beam of a TWP section in comparison to a hot-rolled section of the same weight lead to larger load carrying capacity and greater beam span. 6.3 EXPERIMENTAL PROCEDURE PROGRAMME AND

The aim of the experiment is to study the effects of partial restraint provided by the partial strength connections on the ultimate and serviceability of the TWP beam. A series of two isolated bare steel beam-to-column joints and two bare steel sub-assemblage beamto-column joints were tested on a full-scale basis. A purpose-built test rig was designed and erected to accommodate a column height of 3 m and a beam span of up to 6 m. The rig consists of channel sections pre-drilled with 22 mm holes for bolting purposes. The sections were fastened and bolted to form loading frames, which were subsequently secured to the laboratory strong floor as shown

Experimental Test on Steel Beam with Partial Strength Connection

71

in Fig. 6.1(a) for the isolated tests and Fig. 6.1(b) for the subassemblage tests. The height of the column was kept at 3 m to represent the height of a sub-frame column of multi-storey steel frame. The column was restrained from rotation at both ends whilst the beam was restrained from lateral movement as shown in the figures. In the isolated tests, the load was applied at a distance of 1.3 m from the face of the column. This distance was deemed adequate to cover the distance of the contra flexural point between the negative end moment of the joint and the positive moment of the beam. As for the sub-assemblage tests, the load was applied on the 6 m beam using a hydraulic jack at the mid-span, and was converted into a two-point load using a spreader of 1.8 m spreading distance. This distance was still within the standard distance of one third of the length of the beam so that a bending situation was assured. After the instrumentation system had been set-up and the specimen had been securely located in the rig, the data collection software in the computer was checked to make sure that all channels connecting to the instruments on the specimen indicated a properly working condition. Correction factors from calibration and gauge factors from manufacturer were input into the software prior to each test. A 5 kN increment was adopted in order to have a gradually applied loading condition. Each specimen was then loaded until a substantial deflection of the beam was observed. The load application was continued at this point but adopting a 2 mm increment in the deflection instead of the load as before. This procedure was continued until the specimen had reached its failure condition. The failure condition was considered to have been reached when an abrupt or significantly large reduction in the applied load or when a large increment in the deflection of the beam has been attained.

72

Analysis, Design and Performance of Steel Section with Trapezoid Web

Roller

Connection Bracing Column Beam

Load Cell Hydraulic Jack Roller

Base Plate Base Beam

Strong Floor

Figure 6.1(a) Arrangement for isolated tests

Roller Load Cell

Connection Beam Bracing Column

Hydraulic Jack Spreader

LTB Bracing Base Plate Base Beam

Strong Floor

Figure 6.1(b) Arrangement for sub-assemblage tests

Experimental Test on Steel Beam with Partial Strength Connection

73

6.4 6.4.1

DESCRIPTION OF SPECIMENS Isolated Tests

In the isolated joint tests, two sets of beam-to column joints, which consisted of a 1.5 m beam, were tested. Each of the beam specimens was connected to the column flange using a different type of partial strength connections. The first arrangement was a flush endplate connection (FEP) as shown in Fig. 6.2(a), whilst the other arrangement was an extended endplate connection (EEP) as shown in Fig. 6.2(b). The geometry of the two connections was identical except that there were two additional bolts on the extended part of the extended endplate connection. Details of the specimens for the two arrangements in the isolated joint tests are as shown in Table 6.1. 6.4.2 Sub-Assemblage Tests

For the sub-assemblage tests, two new arrangements, which consisted of a beam with a typical length of 6 m, were tested. The beam in each arrangement was connected to the column flange at both ends using the identical FEP connection and EEP connection as in the isolated joint tests. The full-scale tests were conducted using a 6 m length beam so as to observe the influences of the mid-span deflection on the behaviour of the connections. Those influences can be listed as followed: i) On the moment resistance ii) On the rotational rigidity iii) On the ductility

74

Analysis, Design and Performance of Steel Section with Trapezoid Web

Table 6.1 Details of specimens


Model Name Isolated Test F2R20P1 (N9) Beam Size TWP Column Size UC Connection Type Bolt Row (Top-Bot) End Plate / Bolt Beam Length (m) 1.5 1.5

E3R20P1 (N7) Sub-assemblage Test FS-FSF2R20 (N10) 450 160 50.2/12/4 FS-E3R20P1 (N12)

450 160 50.2/12/4

305 305 118

FEP EEP

2(4-4) 3(6-4)

200 12 / M20 200 12 / M20 200 12 / M20 200 12 / M20

305 305 118

FEP EEP

2(4-4) 3(6-4)

6.0 6.0

200 55 90 50 40 60 90 150 90 60 25

200
55

55 90 55 25 60 90 150 90 60 25

(a) FEP

(b) EEP

Figure 6.2 Geometry configuration of partial strength connections

6.5

DISCUSSION OF RESULTS

The results of the experiments were focused on the behavioural

Experimental Test on Steel Beam with Partial Strength Connection

75

characteristics of the flush endplate and extended endplate connections in isolated tests, and the deflection at mid-span and the applied loads of the beam in sub-assemblage tests. The deflection at mid-span was compared with the deflection limit suggested by BS5950: 2000, Part 1 as L/360 for brittle material and L/200 for other than brittle. On the other hand, the moment resistance and rotation of the connections that associated with the mid-span deflection and load were also observed. The behaviour in term of the moment-rotation relationship of both types of connections was obtained from the isolated tests results. Results of the all tests are shown in Table 6.2. Fig. 6.3 shows the plots of moment versus rotation for the FEP and EEP, whereas Fig. 6.4 and Fig. 6.5 show the plots of load versus mid-span deflection and load versus rotation for the two sub-assemblage tests. It was noticed that although both specimens of the sub-assemblage tests failed due to the buckling of the top flange at the centre of the beam, the connections possessed a ductility characteristic with a rotation capacity of 33.5 mRad for FEP and 32.8 mRad for EEP. The moment capacities, on the other hand, are 95 kNm for FEP, which are 0.28Mp and 225 kNm for EEP, which is 0.67Mp. These values are between the 25% and 100% of the capacity of the connected beam, which categorise both connections as partial strength connections. Theoretically, the plastic moment, Mp or capacity of the beam is calculated to be 335.8 kNm.
Table 6.2 Test results

Reference Moment Capacity (kNm) Rotation Capacity (mRad) Initial Stiffness (kNm) Max Load (kN) Mid-Span Deflection at Max Load (mm) Rotation at Max Load (mRad) Moment Resistance at Max Load (kNm) Failure Mode of Isolated Specimens Failure Mode of Sub-Assemblage Specimens

F2R20P1 95 33.5 30770 235 23 5.20 85 Endplate yielding Buckling of top flange at mid-span

E3R20P1 225 32.8 30770 380 34 6.10 150 Endplate yielding plus bolt slipping Buckling o f top flange at mid-span followed by end plate yielding

76

Analysis, Design and Performance of Steel Section with Trapezoid Web

(a)

(b)

Plate 1 Failure mode of a) FEP and b) EEP

Plate 2 Failure mode of beam on sub-assemblage tests

Experimental Test on Steel Beam with Partial Strength Connection

77

Moment-Rotation Curve for N9


160.0 140.0

MU=138 MR=95

Moment (kNm)

120.0 100.0 80.0 60.0 40.0 20.0 0.0


0.0

=33.5
5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0

Rotation (mRad)

Moment-Rotation Curve for N7


300.0

MU=266 MR=225

Moment (kNm)

250.0 200.0 150.0 100.0 50.0 0.0 0.0

=32.8
5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0

Rotation (mRad)

Figure 6.3 Moment versus rotation for FEP and EEP

78

Analysis, Design and Performance of Steel Section with Trapezoid Web

P vs Rotation - FEP (N10)


250.0 200.0

Pu=235 P1=185

Load (kN)

150.0 100.0 50.0 0.0 0.0

1=2.8
1.0 2.0 3.0

u=5.2
4.0 5.0 6.0

Rotation (mRad)

P vs Rotation - EEP (N12)


400.0 350.0 300.0

Pu=380 P2=375

Load(kN)

250.0 200.0 150.0 100.0 50.0 0.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0

P1=240

1=3.0

2=5.2

u=6.1

Rotation(mRad)

Figure 6.4 Load versus rotation for FEP and EEP

Experimental Test on Steel Beam with Partial Strength Connection

79

P vs Mid-Span Deflection - FEP (N10)


250.0 200.0
Pu=235

Load (kN)

150.0 100.0 50.0

P1=185

1=17.5

u=23.0

0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0

Deflection (mm)

P vs Mid-Span Deflection - EEP(N12)


400.0 350.0 300.0 250.0 200.0 150.0 100.0 50.0 0.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0

Pu=380 P2=375

Load (kN)

P1=240

1=17.5

2=31.5

u=34.0

Deflection (mm)

Figure 6.5 Load versus mid-span deflection for FEP and EEP

6.6

CONCLUSION

Based on the results obtained, several observations have been made which lead to the following conclusion: i) Both types of connections have the same value of initial stiffness, however the moment capacity of the extended endplate connection is more than double the moment

80

Analysis, Design and Performance of Steel Section with Trapezoid Web

capacity of the flush endplate connection. ii) The failure mode for the flush endplate in the isolated tests is endplate yielding, whilst for the extended endplate, the failure mode is endplate yielding and bolt slipping. The mid-span deflection of the sub-assemblage specimens has reached its limit before the moment capacity of the endplate connections has been reached. The failure modes for both specimens are due to the buckling of the top flange at the mid-span of the beam. This shows that the compression force induced in that area has caused the buckling of the flange to occur before any typical mode of failure of the connection occurs. Partial strength connections with TWP sections can provide sufficient moment capacity and rotational capacity as outlined by the EC3 for semi-continuous construction.

iii)

iv)

v)

REFERENCES 1. Eurocode 3. (2002). Design of Steel Structures: ENV 1993-11: Part 1.1: General Rules and Rules for Buildings. CEN, Brussels. British Standards Institute BS 5950-1. (2000). Structural Use of Steelwork in Building Part 1: Code of Practice for Design Rolled and Welded Sections. British Standards Institution, London. Md Tahir, M (1995). Structural and Economic Aspects of The Use of Semi-Rigid Joints in Steel Frames. PhD Thesis. University of Warwick, United Kingdom. Steel Construction Institute and British Constructional Steelwork Association. (1996). Joints in Steel Construction. Volume 1: Moment Connections. London. Weynand, K., Jaspart, J. P., Steenhuis, M. (1998). Economy Studies of Steel Frames with Semi-Rigid Joints. Journal of

2.

3.

4.

5.

Experimental Test on Steel Beam with Partial Strength Connection

81

6.

Constructional Steel Research. Vol 1. No. 1-3. Paper No. 63. Couchman, G. H. (1997). Design of Semi-Continuous Braced Frames, Steel Construction Institute Publication 183, Silwood Park, Ascot, Berkshire SL5 7QN, U.K

7
FIELD CAPACITY OF STEEL PILE WITH CORRUGATED WEB SURFACE
Mohd Hanim Osman, Fauziah Kasim and A. Aziz Saim

7.1

INTRODUCTION

Trapezoid web steel section is built up by a welded flange and web of trapezoidally corrugated profile (Fig. 7.1). Originally, trapezoid web profile sections (TWP) were used for years in the design of beams and columns. The difference with the conventional section is that the web in the conventional section is flat. It is used particularly to increase the shear strength of beam. In this research, the trapezoid web section is used for foundation pile system. By using the new section, the inclined web is expected to provide extra skin friction resistance due to the induced normal component of the friction force. This normal component force increases with increasing inclination of the web. The purpose of the study is to investigate the performance of pile with trapezoid web in resisting loading from the building foundation. The specific objectives of this research project are to compare the ultimate load bearing capacity of different H-pile section with different angles of corrugated and number of corrugated including the flat web H-pile and to study on the behavior of the corrugated web H-pile with respect to performance in skin friction resistance under axial load, when act as pile foundation system. A number of series of tests had been carried out, i.e., tests on both full scale pile and small scale pile specimens, and consolidation

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

83

test of soil surrounding the pile. In the small scale test series, the load test on corrugated web pile and flat web H shape piles with the sizes of 75 mm 75 mm cross section and 450 mm length had been carried out. The settlement versus load increment was measured. The small scale pile specimens had also used to study the consolidation of soil surrounding the surface of the pile. The behaviour of the consolidation process is important because the contribution of the web to the skin friction resistance is achieved by assuming that the soil around the pile surface is fully consolidated. The problem to the new system is the longer time taken for the consolidation of soil around the pile. The tests were carried out to determine the advantages of corrugated shape web in term of settlement when subject to vertical load due, to its larger contact surface.

Flange Web

Figure 7.1 Typical pile section with trapezoid web and flat web

7.2

BEARING CAPACITY OF PILE

A pile subject to load parallel to its axis, will carry the load partly by shear generated along the shaft, and partly by bearing stresses generated at the base of the pile. The ultimate capacity, Qu of the

84

Analysis, Design and Performance of Steel Section with Trapezoid Web

pile under axial load is equal to the sum of the bearing capacity, Qb and the skin friction resistance, Qs. The relative magnitude of the skin friction and the bearing capacity depend on the geometry of the pile and the soil profile. The skin friction capacity of a pile is mobilized at much smaller displacements of the pile (typically 0.5% - 2% of the pile diameter) than is the bearing capacity. The bearing capacity of pile is derived from the formula; Qu = Qb + Qs (7.1)

where Qb is the bearing resistance given by the cross section of the pile and Qs is the skin friction resistance given by the frictional force between the surface of the pile with the soil surrounding the pile. Qb = Ab qb (7.2)

where Ab is the pile base area and qb is the bearing strength coefficient of the soil. Qs = As qs (7.3)

where As is the pile skin area and qs is the friction resistance coefficient. For the section with trapezoid web profile, additional force mechanism is available, i.e., the normal component of friction force at the inclined part of the web (Fig. 7.2). The extra resistance force given by the inclined web depends on some geometrical factors. In this research, the factors to be studied are the angle of the inclination of the web, the number of cycles of the trapezoid and the pitch thickness of the corrugation (Fig. 7.2). The steel-soil friction coefficient may also influence the frictional

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

85

resistance at the inclined web. The extra resistance at the inclined web can be obtained when complete consolidation of soil has been achieved, at some period after the driving of the pile. The consolidation is due to the surcharge pressure from the soil above and also loadings from the buildings on the ground surface.

Skin friction resistance, Qs

Skin friction resistance, Qs

Extra normal force at the inclined web

End bearing resistance, Qb

End bearing resistance, Qb

(a) Pile with flat web

(b) Pile with trapezoid web

Figure 7.2 The bearing capacity of pile with flat web and trapezoid web

7.3 7.3.1

PILE LOADING TEST Pile Test Sample

A full scale test was carried out in a construction site. The test was based on the British Code of Practice BS8004: 1986 requirements

86

Analysis, Design and Performance of Steel Section with Trapezoid Web

and design criteria. The samples were tested with Static Maintained Load Test method in order to determine their respective ultimate bearing capacity on shaft friction. Data were collected and transformed into loading versus settlement graphical curve for the comparison study. Soil investigation was carried out for the selected site area, to determine in detail the geological structure of the site, including the thickness, sequence and extent of the strata, ground water conditions and mechanical properties of soil. Based on the soil investigation report, the value of qs is equal to 12 N/mm2, and qb is equal to 1200 N/mm2. The bearing capacities of the piles are therefore; Q = Ab qb + Af qs = 32 kN. The sizes of the specimens used were 200 mm 80 mm and 5 m long, with 5 mm thickness flanges and 2 mm thick web. Six piles of different web corrugations were used in the test. The profiles of the web are shown in Fig. 7.3. The samples are non typical sections and thus unable to estimate the loading capacity using standard chart or table for the test load. Basically, the evaluation of the pile capacity is divided into two parts, i.e., estimation by numerical calculation, and quick load test on site for the verification of the estimated value. Fig. 7.4 shows a pile being
driven into soil. This is the initial for the entire field test with the purpose of the determination and the confirmation of the pile capacity obtained from the empirical formula analysis. To verify this, the first pile was loaded with hydraulic jack using constant rate of penetration method. The load test confirmed that the estimated pile capacity obtained by empirical analysis was correct and the pile achievable to maximum loading capacity of 3.2 tons for corrugated web Hpile section. Therefore the design load for the pile shall be 1.6 tons

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

87

after taken into consideration safety factor of 2. But, as predicted the loading capacity for the flat web H-pile is lower than the corrugated web H-pile. As such, in order to make the test easier and the data can be collected within the range of test load, the test load was fixed at 10 kN as design load and 20 kN as working load for all pile tests. The dynamic load tests were carried out prior to the actual test (Fig. 7.5). The purpose was to estimate the bearing capacity of the pile test more reliably. The results of the dynamic test are shown in Table 7.1. The dynamic load test was conducted just after the driving-in of piles completed.
80 mm

FW
80 mm 40 mm

TWP 1
80 mm

TWP 2 Figure 7.3 The profiles of the piles

88

Analysis, Design and Performance of Steel Section with Trapezoid Web

Figure 7.4 Driving of one of six full-scale pile specimens

Figure 7.5 Dynamic load test on site

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

89

Table 7.1 Results of Dynamic Load Test


Pile No. and Mark 1-FW-A 2-TWP1-A 3-TWP2-A 4-TWP2-B 5-TWP2-B 6-TWP1-B Load Capacity from Dynamic Test (kN) 20 43 37 37 37 20

7.3.2

Method of Loading

The load test assembly was set up using the jacking frame with Kentledge reaction system as shown in Figs. 7.6, 7.7 and 7.8. The Kentkedge was provided by a heavy mass of dead load, exceeding the maximum test load required, supported centrally over the test pile on suitable constructed temporary staging. A mild steel plate of sufficient thickness was mounted on the top of the pile head to accommodate the loading and settlement measuring equipment, such as hydraulic jack, and to prevent damage due to the concentrated application of load from the loading equipment. The dial gauges and the leveling survey instrument scales were used to measure the settlement of the pile top. Two settlement dial gauges were placed equally at the sides of the test pile to measure the settlement of the pile and firmly held by magnetic base as shown in Fig. 7.7. The average of the two readings was taken as the pile settlement. A temporary bench mark was established on a well-founded structure, served as a second check for the pile settlement. After plotting the load against settlement curve, the load test was considered as pass or successful if the settlement is less than 5 %

90

Analysis, Design and Performance of Steel Section with Trapezoid Web

of the pile width. The residual settlement after unloading shall be less than 6.5 mm at working load or less than of total settlement at twice working load.
Concrete blocks

Reaction beam Loading jack 1m

Load cell

Bearing plate

Displacement transducer

Pile under test

Figure 7.6 The test set-up of the pile load test.

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

91

Figure 7.7 The Test Set-up of Pile Load Test on Site

Figure 7.8 The on Site Pile Test

92

Analysis, Design and Performance of Steel Section with Trapezoid Web

7.3.3

Loading Procedure

The field test is sub divided into two stages, i.e., the maintained load and the constant rate of penetration tests. For constant rate of penetration test, the load was applied in increment of 1.0 kN up to failure. The jack was operated to cause the pile to penetrate at a uniform speed of 0.75 mm/min. The readings of penetration and jacking were made at appropriate interval. In the maintained static load test, the piles were loaded in two cycles. In the first cycle, the load was applied in the increment of 25% of design load and maintain for 20 minutes at each load. The load was maintained for 2 hours at design load. The load was decreased at 25% each, and maintained for 20 minutes at each load. The level readings of the support of the reference beams and the test pile were recorded and verified with the level of the benchmark. After loading, the piles were unloaded to obtain the shape of the rebound curve. In the second cycle, the load was applied in increment of 25% of design load and maintain for 20 minutes at each load. The load was increased for the same increment up to twice the design load, at which was maintained for 12 hours. The load was decreased at 50% each decrement to zero loads. The level readings of the support of the reference beams and the test pile were also recorded. 7.4 RESULTS AND ANALYSIS

The performance of the shaft friction capability of a foundation element in cohesive soil, instrumented load testing is compulsory to justify the design method and criteria. Backward analysis determines the skin friction and end bearing for both types of piles for comparison. The tests were carried out to compare the friction capability of the pile. Therefore, both types of piles were tested on similar ground conditions and method of installation under the

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

93

control environment, so that the results of the tests are reliable for the comparison study. The ultimate bearing capacities of all piles are shown in Table 7.2. Upon collection and compilation of the pile load tests data, the load versus settlement curves were plotted to present the actual behaviour of the pile that subjected to loading, and the comparison study was made based on the analysis on the individual curve. Fig. 7.9 shows the load versus settlement curves for all pile samples. The results of the constant rate of penetration and maintained load test are plotted in the same graphs. Fig. 7.9 illustrates how the ultimate pile load carrying capacity is interpreted from the load versus settlement curve using Van Weele method. This method is selected for the analysis because it is the primary load test interpretation method generally used. In Van Weele method, the graph is interpreted based on the concept that the load is carried mostly by skin resistance until the shaft slip is sufficient to mobilize the limiting value. When the limiting skin resistance is mobilized, the point load increases nearly linearly until the ultimate point capacity is reached. At this point, further applied load results in direct settlement (load curve becomes vertical). These settlements are translated as follows: From the point of origin to point a the capacity is based on the skin resistance plus any small bearing capacity contribution. The skin resistance capacity is the principal load-carrying mechanism in this region. Point a usually requires some visual interpretation. From point a to b the load capacity is the sum of the limiting skin resistance (now a constant) plus the bearing capacity. From bearing capacity b the curve becomes vertical as the ultimate point capacity is reached. Often the vertical asymptote is anticipated (or the load to some value is adequate) and the test terminated before a vertical curve branch is established. According to Van Weele, if the dashed line c tod through the

94

Analysis, Design and Performance of Steel Section with Trapezoid Web

origin and parallel to the point capacity region from a to b is drawn, the load-carrying components of the pile can be shown in the figures. These figures indicated at settlement = 4 mm (5% of pile width) for the load carried as shown in Table 7.3. Based on the graphical analysis illustrated in Fig. 7.9, piles with corrugated web perform better in skin friction resistance. The estimation using Van Weele Method also shows that the corrugated web piles have higher friction capacity than that of the flat webs pile. It is also shown that the pile sections with full corrugation have higher friction resistance compared with the section of flat web. The maintained static load tests were conducted in two groups having the results plotted in load versus settlement curves in two loading cycles. The curves of load versus settlement are combined in single graphs as shown in Fig. 7.9. From the graphs, there is no clear comparison between the various sections in their settlements at design load. However, it is shown that piles with corrugated web show good resistance when loaded to twice the design load in second cycle, as shown by pile No. 3 and 4 (full corrugation). They also show good rebound effect during releasing of the test load. Pile No. 6, i.e., pile with flat web totally failed in second cycle when it exceeded the allowable settlement limit of 4 mm. As predicted the corrugated web piles have greater capability in friction mainly resulted from its larger contact surface with soil due to normal force executed on the inclined web surface couple with interlocking effects. It is also generally predicted that the fiction capability generates by the flat web pile mainly relies on its surface texture, shaft area and the degree of cohesiveness of the soil. Therefore, its friction capability is relatively weak compared to the corrugated web section which in line with the general practice that H-pile is not suitable to be used as fiction pile in foundation application.

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

95

Table 7.2 Ultimate Bearing Capacity of Piles Pile No. 1-FW-A 2-TWP1-A 3-TWP2-A 4-TWP2-B 5-TWP1-B 6-FW-B Load 16 kN 24 kN 27 kN 29 kN 27 kN 17 kN

Table 7.3 The Results of the Analyses of the Load Settlement Curves
Settlement at design load 0.7 mm 0.9 mm 0.9 mm 0.9 mm 0.6 mm 0.4 mm Settlement at twice design load 2.3 mm 2.2 mm 1.5 mm 1. 5mm 2.0 mm Fail

Bearing resistance (kN) 10.0 4.5 5.0 6.0 1.5 -

1-FW-A 2-TWP1-A 3-TWP2-A 4-TWP2-B 5-TWP1-B 6-FW-B

FW-A FW-B TWP1-A TWP1-B TWP2-A TWP2-B

Skin resistance (kN) 14.0 21.0 24.0 28.0 26.0 -

Code No.

Pile No.

96

Analysis, Design and Performance of Steel Section with Trapezoid Web

Load vs Settlement : Pile No.1 - FW 1

Pile No.2 - TWP 1A 50 40 30 20 10 0

35 Load (kN) 25 15 5
c a d b

Load (kN)

b
Skin Resistance Bearing Resistance

Skin Resistance Bearing Resistanc

d c

0 1 2 3 4 5 6 7 8 Settlement (mm)

-5

0 1

2 3 4 5 6 7 8 Settlement (mm)

Load vs Settlement - Pile No.3 - TWP 2A 40 30 Load (kN) 20 10 0


d c
Bearing Resistanc

Pile No.4 - TWP 2B 40


b a

b a
Skin Resistanc

Load (kN)

30 20 10
c

Skin Resistan ce d Bearing Resistance

2 3 4 5 6 7 -10 0 1 Settlement (mm) 8

0 1Settlement 5 6 7 8 2 3 4 (mm)

Pile No.5 - TWP 2B 40


Load (kN) 40
a b

Pile No.5 - TWP 2B 30 20 10 0


c d
Skin Resistanc Bearing Resistance

Load (kN)

30 20 10 0
c

Skin Resistance

Bearing Resistance

0 1 2 3 4 5 6 7 8 Settlement (mm)

0 1 2 3 4 5 6 7 8 Settlement (mm)

Figure 7.9 Load versus Settlement Graphs for all Test Piles

7.5

CONCLUSION

From the results of the experimental works, it indicates that the total bearing capacities of pile with trapezoid web is higher than

Performance of Loading Bearing Capacity of Steel Section with Corrugated Web

97

that of the flat web section of the same cross section area. The analytical method shows that the higher capacity is due to the increased portion of pile skin friction, as is expected in the hypothesis. The bearing capacities of the piles with full corrugations are in the range of 30 kN while the piles with flat web are only 20 kN. It is also observed that the capacity of piles with half corrugation are in-between the two lower and the upper values. Thus the new section can be used as a better alternative to the conventional section in term of the load bearing capacity. REFERENCES Abdul Karim Mirasa, Noor Adli Sapari, Mohd. Zain Yusuf and Ramli Nazir. (2001). Design Guide for Piles Using Locally Produced Steel H-Section. Skudai: Penerbit UTM. Page 9 23. ASCE Deep Foundation Committee. (1984). Practical Guidelines for the Selection, Design and Installation of Piles. U.S.A: ASCE Deep Foundation Committee, 5 7. Bowles, J.E. (1968). Foundation Analysis and Design, International Student Edition. Tokyo: McGraw-hill Book Company. 454. Mohd. Nordin Bin Mustapha. (2002). The Performance of Trapezoid Web Profile (TWP) Pile In Loose Sand. University Teknologi Malaysia: Thesis B.Eng. Ong Ka Thiam. (2004). Friction Capability of Steel H-Pile with Corrugated Web Section. University Teknologi Malaysia. Thesis B.Eng. Goh Wei Liong. (2005). Friction Capacity of Steel H-Pile with Corrugated Web Section. University Teknologi Malaysia. Thesis B.Eng Brahma, S.P. 91685). Foundation Engineering. New Delhi: Tata McGraw-Hill Publishing Company Limited. 232 - 290. Briton Carson, A. (1965). Foundation Construction. New York: McGraw-hill inc. 148 213.

98

Analysis, Design and Performance of Steel Section with Trapezoid Web

Dawson, W.A. (1981). Pile Driving. Ice Works Construction Guides. London: Thomas Telford Ltd. 1 26. Fleming, W.G.K., Weltman, A.J., Randoplh, M.F. and Elson, W.K. (1985). Piling Engineering. London: Surrey University Press. 254 365. Henry, F.D.C. (1986). The Design and Construction of Engineering Foundation. 2nd ed. New York: Chapman and Hall ltd. 652 738. Highway Research Board, Special Report 36. (1958). A Study of the Comparative Behavior of Friction Piles. Washington D.C.: National Academy of Sciences National Research Council. 1 23. Smith, G.N. and Pole, E.L. (1980). Elements of Foundation Design. London: Granada Publishing Limited-Technical Books Division. 55 98. British Code of Practice BS8004:1986

8
SECOND MOMENT OF AREA ABOUT THE MINOR AXIS OF TRAPEZOID WEB SECTION
Mohd.Hanim Osman, Sarifuddin Saad, Fatimah Denan and Tan Cher Siang

8.1

INTRODUCTION

In this chapter, a parametrical study on the trapezoid web profile (TWP) to develop a formula for the second moment of area about the minor axis is discussed. Finite element analysis (FEM) was carried out on locally fabricated trapezoid web profile models. Several factors that influence the second moment of area about the y-y (minor) axis, Iy value of a trapezoid web profile were being studied. Based on the results of the parametrical study using the finite element method, an equation for Iy of the trapezoid web profile is proposed. To validate the equation, the value of Iy computed from the proposed equation is compared with the experimental results. At the end of this chapter, design table which include calculation of section properties, capacities and design of lateral torsional buckling moment resistance for TWP section are proposed and developed. Fig. 8.1 show the shape and dimension which will be used in this derivation.

Analysis, Design and Performance of Steel Section with Trapezoid Web

100

(a) Trapezoid Web Profile (TWP) Section

(b) Front view

one cycle = L (c) One cycle trapezoidal web

Figure 8.1 Shape and dimension of TWP section

101 Second Moment of Area About the Minor Axis of Trapezoid Web Section

one cycle = L (d) Plan view T is the flange thickness, B is the flange width, tw is the web thickness, d is the depth of web, hr is the depth of corrugation, D is the overall depth, S is the length for each eccentric flat plates, L is the length of one cycle of corrugation shape and = 45

Figure 8.1 (continued) Shape and dimension of TWP section

8.2 PARAMETRIC STUDY FOR FORMULATION OF SECOND MOMENT OF THE AREA FOR TRAPEZOID WEB PROFILE The modelling using the LUSAS software of load-deflection analysis gives good prediction of the second moment of area for both flat web (FW) and TWP sections. This fact proves that the modelling technique used is accurate and acceptable. Thus, various TWP sections can be modelled to give further understanding on their behaviour, which cannot be done via experimental test. In this investigation on the sensitivity of the Iy value with different geometrical properties of the TWP section, Tan [45] suggested that the Iy value depends very much on the corrugation depth to width of section ratio (hr/B) of the TWP section. Other than that, Tan also found that the S/L also had an effect on the Iy formulation, where S was taken as the length of each eccentric flat plates and L as the length of one cycle of corrugation shape as shown in Fig. 8.1. From the findings, the effects of corrugation depth to width of

Analysis, Design and Performance of Steel Section with Trapezoid Web

102

section (hr/B) and S/L of the Iy value of the TWP section should be investigated in more detail for the derivation purpose. Theoretical prediction and FEM analysis has shown that a beam with a trapezoid web profile has a higher stiffness than of that the I-beam. To conclude this finding, it was targeted to derive a formula for the Iy of the TWP section. For this purpose, the two factor i.e. hr/B and S/L that were found to influence the value of Iy formulation were further investigated. The objective of this study is to find the sensitivity of the minor axis stiffness of beam-section with different hr/B and S/L. FEM analysis was carried out to determine the load-deflection relationships for sections with different web profile. The Iy for the sections were subsequently determined by using Eq. (8.1):
3 P L I = 48 E

(8.1)

where I is the second moment of area, P is the applied point load, L is length of span, E is the elastic modulus of steel and is the vertical deflection at mid span. The simply supported beam will be analyse as shown in Fig. 8.2. Using the above formula and the relationship of Iy TWP = If + Iw. Iw will be defined as shown in Table 1, where If is the second moment of area of flange and Iw is the second moment of area of web.

103 Second Moment of Area About the Minor Axis of Trapezoid Web Section

Web
Y

Flange

Figure 8.2 Typical model with eccentric web only - EW model (side

view) with top and bottom web in diagonal shape

8.3 STUDY ON THE EFFECT OF S/L TO THE SECOND MOMENT OF AREA OF THE WEB, Iw The non-uniform shape of the web makes it impossible to determine its second moment of area via theoretical calculations. However, FEM analysis can give the answer. A few shapes of sections with different web profile were studied using LUSAS software. The size of the sections were kept as 80 mm (flange width, B) 200 mm (depth, D) 5 mm (flange thickness, T). In the process of modeling a problem using LUSAS program, the processing step is almost similar to other FEM analysis. Several

Analysis, Design and Performance of Steel Section with Trapezoid Web

104

essential steps are required in order to achieve the correct result or output. These steps are: Creating the geometry of the model problem. Mesh the model with appropriate type of element (QSL8). Define and assign the geometry and material properties. The whole trapezoid web profile beam is assign ungraded mild steel for its material property with Youngs modulus, E = 209 103 N/mm2, shear modulus, G = 79 103 N/mm2 and Poissons ratio of 0.3. Define and assign the boundary condition. Define and assign the loading. Specify type of analysis and analysis option/control. Run the modeling and view the result/output.

Sections with eccentric web only (EW) were studied in this case. The shapes of EW model are shown in Fig. 8.2. Two examples of modeling involved in this analysis are shown in Fig. 8.3(a) and 8.3(b). Fig. 8.3(a) shows the model with flange only with a transverse plate at loading point (EW 000). Fig. 8.3(b) shows section with eccentric web only with S/L= 0.4 (EW 400). Support and loading condition are shown in Fig. 8.3(c) and 8.3(d). The computer analysis outputs taken were the maximum deflections, which were used later to derive Iy as the parameter of stiffness for the section. All models have been run successfully. Theoretical calculations of the Iy were made for uniform section by using Eq. (8.2) to Eq. (8.4). For a two-flange section (without the web):
TB 3 I yTF = 2 12

(8.2)

whereas for an I-section:

105 Second Moment of Area About the Minor Axis of Trapezoid Web Section

TB 3 dt 3 I y = 2 12 + 12

(8.3)

For a box-section or rectangular hollow section (RHS):


2 TB 3 dt 3 h + 2 + dt r I yRHS = 2 12 2 12

(8.4)

Y X

Figure 8.3(a) Flange only with a transverse plate at loading point (EW 000)

The objective of this study was to find out the effects of the length of the eccentric flat web, S (see Fig. 8.1) on the beam stiffness. Referring to Fig. 8.1, S is the length of each eccentric web for both sides (bottom and top), and L is the length of the section. When the S/L equals to zero, the section has no web but contains the two flanges only. If the ratio equals to 1.0, then the web covers all top and bottom area, or in other words, it becomes a rectangular

Analysis, Design and Performance of Steel Section with Trapezoid Web

106

hollow section. Models EW with different S/L were generated and analyse. The analysis of load deflection results on models EW with different S/L was recorded in Table 1. From the results, the second moment of area for webs were calculated as:

I w = I y I flanges

(8.5)

where the Iflanges equals to the Iy for the model with a S/L equals to zero. Graphs of Iw/IRHS versus S/L are plotted as shown in Fig. 8.4. It shows a smooth increasing relationship when the S/L increases. But for S/L between 0.6 and 1.0, the Iw/IRHS shows a rapid increment. This is due to the continuity of the web which gives higher second moment of area values.

Support condition

Y X

Loading condition
Z

Figure 8.3(b) Section with eccentric web only with S/L=0.4 (EW 400)

107 Second Moment of Area About the Minor Axis of Trapezoid Web Section

Taking EW result but ignore the last data (i.e. when S/L =1.0), a curve was plotted in Fig. 8.4 to give the relationship between Iw/IRHS and S/L. This relationship may be used to derive a formula representing Iy for the trapezoidal web profile at the end of the chapter.

Y X

Figure 8.3(c) Support condition

Figure 8.3(d) Loading condition

I / IRHS

Analysis, Design and Performance of Steel Section with Trapezoid Web

108

1.20 1.00 0.80

y = 1.28x2 - 0.27x

II/I-box / IRHS

0.60 0.40 0.20 0.00 0.00 -0.20

0.20

0.40

0.60

0.80

1.00

1.20

SS/LL /
AS1-proposed AS1-fea Poly. (AS1-proposed)

Figure 8.4 Graph Iw/IRHS versus S/L for 0 S/L < 1.0 Table 8.1 FEM results for EW (mm) Iy (cm4) Iw (cm4)
1.24 1.19 1.19 1.03 0.61 0.43 42.67 81.1 83.91 96.47 163.41 233.99 0.00 38.4 41.24 53.83 120.75 191.33

Model EW 000 EW 200 EW 400 EW 600 EW 800 EW 1000

S/L 0.0 0.2 0.4 0.6 0.8 1.0

Iw/IRHS 0.00 0.20 0.22 0.28 0.63 1.00

8.4 STUDY ON THE EFFECT OF WEBS CORRUGATION THICKNESS (hr) TO WIDTH OF FLANGE (B) RATIO From the observation of the test results, hr/B influence the value of Iy of TWP sections. From tests, it was found that specimen TWP3 (hr/B = 1.0) has a bigger Iy value compared with specimen TWP4 (hr/B = 0.5). This presents a study on that effect. There were three models used in this study with different hr/B as shown at Table 2. The angle of corrugation was fixed at 45o and web thickness was

109 Second Moment of Area About the Minor Axis of Trapezoid Web Section

fixed at 2 mm. The results are recorded in Table 8.2. Fig. 8.5 show the graphs of deflection versus hr/B and this linear relationship will be included in the Iy formula later. Thus it is concluded that the corrugation depth to flange width ratio brings an important effect to the stiffness of the TWP section about the minor axis.
Table 8.2 FEM results for hr/B Model hr/B (mm)
hrB00 hrB05 hrB10 0.0 0.5 1.0 2.33 2.24 2.15

2.35

2.30 y = -0.18x + 2.33 Deflection (mm) 2.25

2.20

2.15

2.10 0 0.2 0.4 0.6 hr/B 0.8 1 1.2

Figure 8.5 Graph of deflection vs. hr/B

8.5 FORMULATION OF SECOND MOEMNT OF AREA FOR TRAPEZOID WEB PROFILE Wail discussed the formula for trapezoid web profile sections

Analysis, Design and Performance of Steel Section with Trapezoid Web

110

properties in Design Guide for Steel Plate Girders with Corrugated Webs (TWP) [50]. Fig. 8.6 shows the notations used in the Wail formula. He gave the formula for the second moment of area about minor axis as For the web of trapezoid web profile section:
2 3 h t h I yw = 2 d t w r + w r 2 6 sin

(8.6)

where tf is the flange thickness, bf is the flange width, tw is the web thickness, d is the depth of web and hr is the depth of corrugation.

tf (web) bf

tw
(flange) (flange)

(a) Flanges dimension labels


d hr

tw

(b) Webs dimension labels Figure 8.6 Flanges and webs dimension labels of Wail formula

The formula for the flanges, Iyf was calculated from the two flanges of the section:

111 Second Moment of Area About the Minor Axis of Trapezoid Web Section

t f bf 3 t f bf 3 = I yf = 2 12 6

(8.7)

For the flanges, the formula for Iyf is similar to that used for a flat web section. However, there were several arguments appeared for the formula for web, Iyw. The equation consists of two portion for the eccentric flat plates:

I yw

2 hr = 2 d t w 2

(8.7a)

and for the transverse (diagonal) plates:

I yw

t w hr3 = 6 sin

(8.7b)

In Eq. (8.7a), the moment of inertia about its own axis is not considered due to its thinness. However, the 2 expression is still questionable, because at any part of the section, the number of eccentric web is only one, not two. In Eq. (8.7b), the axis of stiffness is not clear. It does not seem to be the minor axis (y-y) of the section. Now closer study of Eq. (8.7b) suggests that it is obtained as follows: I=
h bh 3 , where b = tw and h = r 12 sin
3 3 h h tw r tw r 3 sin = sin = t w hr = 2 12 6 6 sin

There are two transverse plates in a corrugation cycle, thus:

I yw

(8.7c)

Analysis, Design and Performance of Steel Section with Trapezoid Web

112

There may be several arguments against the formula. First, the sin term should be in the power of three. Secondly, the axis of consideration for this formula is also not clear. Wail did not explain in detail how he obtained the proposed formula for the second moment of area about minor axis for the TWP section. In his discussion, the web portion was neglected in the calculation of the trapezoid web profile sections properties. Thus, it did not affect other properties for the trapezoid web profile, i.e. radius of gyration, ry, elastic modulus, Zy, compressive strength, pcy etc. Therefore, the proposed formula cannot be applied as explained in above comments. Tan [45] also studied the derivation of the formula for Iy. The dimension that used in the Tan formula is shows in Fig. 1. He proposed a new formula for the Iy of a trapezoidal web profile as below:
2 2 TB3 dt 3 h S S + 2 I yTWP = 2 + dt r 0.4585 0.011 12 L L 2 12

dt 3 + (cos ) 12

hr 1 + 0.1 B 0.5

Simplified as a:

TB 3 8 + 0.0082dt 3 + 78.4dt 0.95 + 10 4 (8.8) I yTWP = 6 B

The formula was based on the result of FEM analysis due to bending.

113 Second Moment of Area About the Minor Axis of Trapezoid Web Section

8.6 PROPOSED SECOND MOMENT OF AREA ABOUT MINOR AXIS FOR TRAPEZOID WEB PROFILE SECTIONS

The starting point of the new formulae proposed in this work is the same as that of Wail, where the second moment of area about minor axis, Iy consists of the contribution of Iy for the two flanges and Iy for the web: I y = I f + Iw (8.9)

In the case of the Iy for the flanges, the formula is the same as that proposed by Wail in Eq. (8.7):
TB 3 I f = 2 12

(8.9a)

For the web portion, it is assumed that Iy is contributed by the eccentric flat plates (S) and the transverse plates (C) (see Fig. 8.7): Iw = Is + Ic (8.10)

(a) Formula for Is: Taking S as the length for each eccentric flat plates as shown in Fig. 1, if the S/L = 0, the sections Iy only contributed by the two flanges. If the S/L = 1.0, there exist a full continuous plates on top and bottom of the section, which formed the section into a box shape, with the second moment of area is contributed by both flanges at both sides and both webs at top and bottom. Thus a formula is proposed for the second moment of area about the minor axis for eccentric web, Is:
2 dt 3 hr S I s = 2 + dt 2 L 12

(8.11)

Analysis, Design and Performance of Steel Section with Trapezoid Web

114

It is believed that the Is for a TWP section would lie between two limits of S/L.
Web
Y

Flange

=
Eccentric flat (S )
Y

Flange

+
Transverse plates (C)
Y

Flange

Figure 8.7 Eccentric flat plates and transverse plates of a TWP

A LUSAS FEM analysis for the TWP section with different S/L has been carried out and discussed in section 8.2. The Is slightly increased as the S/L increase, but a sudden large increment happened when S/L = 1.0. By neglecting the value of S/L =1.0, a curve can be drawn in the graph. Taking this polynomial relationship, the proposed formula for Is would be:
2 2 dt 3 hr S S I s = 2 + dt 1.28 0.27 for L L 2 12

115 Second Moment of Area About the Minor Axis of Trapezoid Web Section

S < 1.0 L

(8.12)

(b) Formula for Ic Observation of FEA analysis showed that 90o transverse plates along a two-flanged section would not influence the Iy. When these transverse plates were made 0o, they formed a continuous flat plate in the middle of the section and the Iw is given as:
dt 3 Iw = 12

It is proposed that a formula to represent the second moment of area for transverse plates, Ic can be written as:
dt 3 Ic = 12 (cos ) ,

0o < < 90o

(8.13)

A FEM analysis to calculate the values of Ic for a TWP section with different angles of slope of transverse plate was done by Tan [45]. The Ic results were found to be higher than those calculated using the proposed equation. It is concluded that the proposed Ic formula (Eq. (8.13)) can be either neglected, or added to the Iy formula of the whole TWP section to take into account of the presence of the transverse plates. From the study, by taking into account Eq. (8.7), Eq. (8.12) and Eq. (8.13), it is proposed that the Iy for the TWP section:

I y = I f + I w ; and
Iw = Is + Ic

I y = I f + Is + Ic

Analysis, Design and Performance of Steel Section with Trapezoid Web

116

2 2 TB 3 dt 3 hr S S + 2 I y = 2 + dt 1.28 0.27 12 L L 2 12

dt 3 (sin ) + 12

(8.14)

8.6.1

Other Factors that Influenced the Iy of TWP Sections

Parametric study on various sizes of TWP sections showed that the hr/B would influence the Iy of the section [45]. For hr/B =1.0, the Iy of the TWP section is higher than a two-flanged section; but when the hr/B decreases, the Iy will proportionally decrease too, even lower than Iy of two-flanged section. The relationship between h deflection and hr/B (see Fig. 8.5) is y = 0.18 r + 2.33 . From B this equation and with value suggested -0.93, thus it is suggested that a correction is made for the proposed Eq. (8.14):
h I ycorrection = I y 0.18 r + 2.33 0.93 B h I ycorrection = I y 0.18 r + 1.40 B
I yTWP = TB 3 2 12
2 2 dt 3 h S S + 2 + dt r 1 . 28 0 . 27 L 12 2 L 3 dt + 12 sin

hr + 1 .4 0 . 18 B

(8.15)

This equation will be used to propose and develop the design guide for TWP section later in this chapter.

117 Second Moment of Area About the Minor Axis of Trapezoid Web Section

8.6.2 Comparison of Iy TWP by using Current Design Method (BS 5950), Wails Formula, Tans Formula and the Proposed Formula

Wail [52] proposed a formula for the second moment of area about the minor axis for TWP sections:
3 2 TB 3 hr t hr Iy = + 2 d t + 6 2 6 sin

10 4 cm4

(8.16)

where = 45o. But for the current application, the web area is considered very small and is being neglected. Thus the current design formula [4] is: Iy = TB 3 10 4 cm4 6 (8.17)

Tan [45] proposed another formula for Iy trapezoid web profile as given in Eq. (8.8). From the study using FEM analysis reported in this chapter, a new formula is proposed as shown in Eq. (8.15), that covers the corrugated webs contribution and the effect of hr/B. By obtaining the Iy using these three formulae, comparisons were made with FEM Iy results for different size of TWP sections, as given in Table 8.3. The Wails formula over estimated the Iy TWP between 5.2 to 326.2 %. Current design equation as given in Eq. (8.17) gave a closer formula to actual Iy TWP, where the percent differences were between -3.4 and 1.4%. The Tans formula also gives a closer formula to actual Iy TWP, where the differences were between -1.0 and 1.9%. The authors proposed formula gave the best comparison, where the difference was between 0.3 and 2.7 %. From the comparison, it is concluded that Wails formula may not provide a good value for the Iy value for TWP sections. The current design formula and Tans formula predicted the Iy TWP very closely, but they sometimes underestimate the Iy TWP. The

Analysis, Design and Performance of Steel Section with Trapezoid Web

118

proposed new formula gave a better comparison with the FEM value. The proposed formula seems to be the best formula to calculate the value of Iy TWP.
8.7 CONCLUSION

After a few studies conducted to investigate about the sensitivity of the Iy value with different dimension properties of the TWP section, it was found that the Iy value depends very much on the S/L ratio and the corrugation depth to width of section ratio (hr/B) of the TWP section. Numerical study on sections with different eccentric web profile was carried out. From this study, a new formula representing the second moment of area about the minor axis for TWP sections, Iy TWP was finally proposed as Eq. (8.15). Comparison of Iy TWP values between the proposed new formula, current design formula, Tans formula and Wails formula finite element Iy TWP result, gives the following differences: Wail Tan Current (BS) Proposed - between 5.2 and 326.2 %; - between -1.0 and 1.9 %; - between -3.4 and 1.4 % and - between 0.3 and 2.7 %.

It is concluded that the authors equation give the best prediction to Iy TWP, where the difference were in the range of only in between 0.3 and 2.7%. In application to the lateral torsional buckling load analysis, it can be forecasted that a TWP members may have greater buckling resistance compared to that of the FW member, but this advantage would decrease when the hr/B of the section is less than 0.5. The compatibility of the proposed new equation, Iy TWP to calculate the buckling load for a TWP section will be discussed in the next chapter.

Table 8.3 Comparison of Iy TWP using Wails Eq (8.16), current design equation, Eq. (8.17), Tans Eq. (8.8) and the proposed new equation, Eq. (8.15)
% Difference Eq. (15)/ Eq. (17)/ Eq. (8)/ Eq. (16)/ Iy FEM Iy FEM Iy FEM Iy FEM 0.7 0.5 0.3 1.6 2.7 827 5205 2135 2.6 1.7 2.5 -3.4 1.4 -2.7 -2.7 -0.7 -0.9 0.1 -0.1 0.9 1.9 0.4 -1.0 0.5 0.4 0.6 0.9 326.2 80.7 255.6 42.2 21.9 15.1 5.2 6.1

Section BS 5950, Eq. (17) 42.7 42.7 150 288.0 1166.4 819.2 5208.3 2133.3 2153.3 5237.4 5474.2 2266.3 830 952.1 1181 1432.3 293.0 420.9 296 1175 154.8 548.3 154.2 42.9 76.08 42.1 44.6 188.4 44.2 Tan, Eq. (8) Wail, Eq. (16) Iy FEM TWP

Second moment area about the minor axis, Iy (cm4)

No

(D B T t)

Proposed equation, Eq. (15) (Fatimah)

200 80 5 2

44.5

200 80 5 2

42.3

170 100 9 4

154.7

300 120 10 2

300.9

450 180 12 4

1207

650 160 12 2

848.7

900 250 20 4

5291.8

119 Second Moment of Area About the Minor Axis of Trapezoid Web Section

1200 200 16 2

2189

9
LOCAL FLANGE BUCKLING OF TRAPEZOIDAL WEB PROFILE
Ng Zee Leong and Mohd Hanim Osman

9.1

INTRODUCTION

Steel girders with a cross section of I shape has been used since the advent of structural steel material into the construction industry. Thus, various sizing of flange, web and stiffeners and welding specifications has been developed for this built-up section. The design specifications of this steel girder are also available in different standards and codes. Nowadays, steel girders with trapezoidal web profile has been produced and used in the steel construction industry. This type of steel girder is rather new in the industry but rapidly gaining popularity because of its increased laterally restraint effect due to the trapezoidal web profile. However, there has not been a standard way of defining the strength properties and design methods for these steel girders. For instance, the effect that the trapezoidal web contributes to the local buckling resistance of the compressive flange is still undefined. The main objective of this paper is to define the effect of the trapezoidal web profile to the flange under compression that is welded to it. This is followed by the sub-objectives like finding out the different effect resulting from different web corrugation angles and the spacing between corrugations. This project covers the flanges subject to compression, e.g. flanges in a column and the flange under compression in a steel beam. The flange is assumed to be simply supported on one side of the

122

Local Flange Buckling of Trapezoidal Web Profile

longitudinal edge and the other is free edge. This project was carried out in two phases. In the first phase, a test rig with customized fittings was set up to facilitate the experimental testing on both the flanges with plane web and corrugated web. This phase included the following stages: (a) Preliminary conceptual design of the customized fittings (b) Finalized the design and detailed drawings (c) Sent for fabrication at the metal workshop After the fabrication of the customized fittings, the fittings were fixed on the test rig. It was then followed by the experimental work, which was carried out in three different sizes of specimens (flanges), i.e. Specimen 1: 100 350 4 mm Specimen 2: 200 350 4 mm Specimen 3: 400 350 4 mm

These three steel plates will be referred as Specimen 1, 2 and 3 throughout this report starting from this point. All these specimens were tested in five different arrangements, i.e., changing the corrugation angles and spacing between corrugations of the trapezoidal web profile as listed below. (i) Testing on flanges with plane web (without trapezoidal web profile) (ii) Testing on flanges with trapezoidal web profile slanted at an angle of 30 and corrugation spacing of 100 mm (iii) Testing on flanges with trapezoidal web profile slanted at an angle of 45 and corrugation spacing of 100 mm (iv) Testing on flanges with trapezoidal web profile slanted at an angle of 30 and corrugation spacing of 150 mm (v) Testing on flanges with trapezoidal web profile slanted at an angle of 45 and corrugation spacing of 150 mm

Analysis, Design and Performance of Steel Section with Trapezoid Web 123

9.2

EXPERIMENTAL SETUP

The flange was simply supported on three edges and one edge free. The supports were built-up by welding right angle sections at the angles to 10 mm thick mild steel plates. The lower bearing plate was supported on a stiffened square hollow section, which is rested on the frame of the test rig. Trapezoidal web restraints were provided by clipping four steel solid rods (with a diameter of 8mm and 113 mm in length) to the side of the specimen. These trapezoidal web restraints were supported on roller in vertical direction and fixed horizontally. This was to ensure that the web restraints could displace along with the flange in compression as in the actual condition. The whole set-up was fitted to a jacking machine with a capacity of 200 kN at the centre of the top bearing plate to create a compression condition to the specimen. Three points of the specimen were monitored for their respective displacements. The vertical displacement was taken at the top edge of the specimen by attaching one LVDT to the top of the test rig. The other two lateral displacements were measured by using two LVDTs supported separately (not attached to the test rig) by builtup stands placed beside the test rig. The set-up of the whole experimental work is shown in Fig. 9.1.

124

Local Flange Buckling of Trapezoidal Web Profile

Upper bearing plate

200 kN pressure jack

Lower bearing plate

(a)
200 kN pressure jack Upper bearing plate

Front view

Trapezoidal web restraint S pecimen

Lower bearing plate

(b)

Side view

Figure 9.1 Experimental set-up of customized fittings and specimen

Analysis, Design and Performance of Steel Section with Trapezoid Web 125

9.3 RESULTS 9.3.1 Theoretical Results The critical buckling stress of the flange, cr, simply supported on 3 edges and 1 longitudinal edge free was calculated from the Euler formula as shown below. The main results calculated from theoretical formula were shown in Table 9.1. This table lists the buckling coefficient, the critical buckling stress and the critical buckling load. Graph used to find the buckling coefficient is shown in Fig. 9.2.
cr = t 2 b 12(1 v) k 2 E
2

(9.1)

where k is the buckling coefficient, E is the Modulus Young, is the Poissons ratio, t is the thickness of the plate and b is the width of the plate.
k values for 1 edge Simply Supported & 1 edge free
7 6 5 4
0.875 1.75 3.5

k curve b=400 b=100

k
3 2 1
0 .87 5

b=200

0
0 0.5 1 1.5

1.75 2 2.5 3 3.5

3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5

a/b

Figure 9.2 Buckling coefficients for simply supported plates in compression with longitudinal edges simply supported and free. (Adopted from Bradford and Azhari, 1995)

126

Local Flange Buckling of Trapezoidal Web Profile

Table 9.1 Critical buckling stress and load obtained from theory calculations
Specimen dimension (mm) 1003504 2003504 4003504 Buckling coefficient, k Critical Buckling Stress, cr (N/mm2) 154.2 58.5 33.7 Critical Buckling Load, Pcr (kN) 61.66 46.84 53.95

0.52 0.79 1.82

9.3.2

Experimental Results

After all the testing of the specimens, 15 sets of test results were obtained from 3 specimens with the aforementioned 5 arrangements for each specimen. The critical buckling stresses were estimated from the Load-deflection curves plotted according to the experimental results. Table 9.2 shows the critical stresses due to buckling of the specimens with various configurations.
Table 9.2 Critical buckling stress obtained from experimental testing
Critical Buckling Stress, cr (N/mm2) Type of flange 100 mm spacing with corrugation angle at 30 100 mm spacing with corrugation angle at 45 150 mm spacing with corrugation angle at 30 53.05 46.88 28.64 Plane web Specimen dimension (mm) 150 mm spacing with corrugation angle at 45 55.55 48.96 30.69

1003504 2003504 4003504

51.13 41.55 25.00

56.25 48.15 26.46

66.81 51.44 26.95

Analysis, Design and Performance of Steel Section with Trapezoid Web 127

9.4

DISCUSSION

As seen from the results listed in Table 9.2, the stress value for specimens with plane web via experiment has a decrease of 25.81% for Specimen 4, 28.97% for Specimen 2, and 66.84% for Specimen 1. Nevertheless, flanges with plane web have the lowest value of critical buckling stress among all of the test variations. This is mainly due to the fact it has no extra restraints to prevent it from buckling when compared to the flange with corrugated web. Corrugated web acts like stiffeners to the flange in compression, thus providing more buckling resistance. The additional resistance is 22.75 - 30.68%, it shows that the steel beam with corrugated web has an added advantage over conventional I beam in flange compressive capacity besides its known advantages like its lightweight due to elimination of stiffeners and thinner web and increased lateral torsional buckling resistance. The test results also shows that the effects of the corrugated web restraint is more significant in specimen with a smaller width, e.g. Specimen 1 buckled at a 30.68% higher critical buckling stress while Specimen 3 achieved an increment of 22.75% for all corrugated web testing arrangements in this project. As expected, each local buckle of the compressed flange occurred in the centre region of the longitudinal free edge. However, there was also difference in terms of the critical buckling stress for different corrugation spacing. It has been found that the specimens of the same corrugation angle but with a different corrugation spacing, i.e. 100 mm spacing has achieved a higher critical buckling stress that those with a corrugation spacing of 150 mm spacing for Specimen 1 and Specimen 2 and vice-versa for Specimen 3. This could be due to effect of a longer widest flange outstand. Besides, specimens that are similar in corrugation spacing but differ in corrugation angles also contribute to different critical buckling strength. Specimen that has a corrugated web restraint of 45 reached a higher critical buckling load than its 30 counterpart in all specimens, with the highest value reaching 20.66%, though it

128

Local Flange Buckling of Trapezoidal Web Profile

was not significant for the Specimen 3, only 1.96 - 7.15% in difference. Observing from the load vs. vertical deflection curve as shown in Fig. 9.3 and 9.5, it follows a linear path until it almost reached the elastic-plastic state. Meanwhile, the load vs. horizontal displacement curve (Fig. 9.4 and 9.6) was not as linear as the load vs. vertical displacement curve. It started to diverge from the linear path at around three quarter of the critical buckling load. This could be caused by the slight geometrical and set-up imperfection. Therefore the Southwell plot was introduced to find out the initial deflection of the specimen. As shown in Fig. 9.7 of the Southwell plot (deflection against deflection/load) of Specimen 1 with plane web, it has a initial deflection of 3 mm and a buckling load of 51.73 kN deriving from the steepness of the linear curve.

Analysis, Design and Performance of Steel Section with Trapezoid Web 129

35

30

25

20 Load (kN) 15 10 5 0 0 0.5 1 1.5 2 2.5 3 Vertical D isplacemen t (mm)

Figure 9.3 Load vs. vertical displacement for specimen 1 with plane web
35

30

25

Load (kN)

20

15

10

0 0 1 2 3 4 5 6 Horizontal Displacement (mm)

Figure 9.4 Load vs. horizontal displacement for specimen 1 with plane web

130

Local Flange Buckling of Trapezoidal Web Profile

40

35

30

25 Load (kN)

20

15

10

0 0 0.5 1 1.5 2 2.5 3 3.5 Deflection (mm)

Figure 9.5 Load vs. vertical displacement for Specimen 1 with trapezoidal web restraint (100 mm spaced; 45 in direction)
40

35

30

25 Load (kN)

20

15

10

0 0 1 2 3 4 5 6 Horizontal Displacement (mm)

Figure 9.6 Load vs. horizontal displacement for Specimen 1 with trapezoidal web restraint (100 mm spaced; 45 in direction)

Analysis, Design and Performance of Steel Section with Trapezoid Web 131

Deflection Vs Deflection/Load 6 y = 51.734x - 2.9987

Deflection

0 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18

-2

-4

-6 Deflection/Load

Figure 9.7 Southwell plot of Specimen 1 with plane web

9.5

CONCLUSION

Experimental testing have been carried out to identify the effects that a corrugated web steel girder on the local buckling of the flange, and it has been found that the main effect was that corrugated web did contribute to higher critical local buckling stress of the corresponding flange. However, it has to be noted that replacing a plane web of a plate girder by a corrugated one may worsen the slenderness class of the compression flange, for local buckling, if the flange width is too large.

132

Local Flange Buckling of Trapezoidal Web Profile

The main conclusions that can be drawn from this experimental testing include: 1. 2. 3. Corrugated web did provide a higher buckling strength for the flange when compared to an ordinary I-beam with plane web. The effect of the corrugated web becomes smaller with the increasing width of the flange, bf. Corrugated web at a direction of 45 is found to provide more additional buckling resistance, to the flange with a width of 400mm when compared to a direction of 30. The corrugation spacing of 100 mm is provides more buckling resistance than the 150 mm spacing for Specimen 1 and 2. However, it was the other way round for Specimen 3.

4.

You might also like