Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 51, 1, pp. 195-202, Warsaw 2013 LOAD-DEFLECTION RELATIONS OF T-SECTION RAILS UNDER LATERAL LOADS Youshuo Song, Zhonghua Yu Zhejiang University, Institute of Modern Manufacturing Engineering, Hangzhou, China e-mail: caq 221@zju.edu.cn This paper studies the load-deflection relations of T-section rails under lateral loads based on elastic-plastic theory. A linear-hardening model and an elastic-plastic power-exponent hardeningmodel of thematerial are adopted in this study.The analytical expressions for the load-deflection relations in the loading process are given. Compared with the experimental results, it is found that the load-deflection curves calculated with the elastic-plastic power- exponent hardeningmodel are closer to the experimental results than thosewith the linear- hardening model. Key words: T-section rail, load-deflection relations, lateral loads, linear-hardeningmodel 1. Introduction Straightening is an importantprocess in theproduction route of aT-section rail (Biempica et al., 2009;Volebov et al., 1994). Inpractice, themost commonmethod forT-section rail straightening is the threepoints reversebendingbasedonhandworkandtheworker’s experience.So thequality and efficiency is hard to guarantee. Tomeet higher production requirements, an automatic rail straightening machine is to be developed, and the load-deflection model of the bending process is suitable to be used on the straightening control (Li et al., 2004). Deflections of sandwich beams subject to concentrated or localized loads have been studied, and it is sufficient to accurately predict thevertical displacements in the face sheets of a sandwich beam (Shen et al., 2004; Sokolinsky et al., 2003). Also, a number of studies have been proposed in order to investigate the load-deflection relationships of concrete beams reinforced with FRP bars (Ou et al., 2004; Masmoudi et al., 1998). Tsai and Kan (2008) studied the load-deflection model of the uniformly loaded rectangular cross-section cantilever beam. According to Tsai et al. (2009), various loading scenarios can result in different deflection profiles, albeit with the same tip deflection. Thematerial hardening parameters can be determined by comparing load–displacement cu- rves fromFEsimulationswith those fromtests in thecyclic three-pointbendingtest (Omerspahic et al., 2006; Eggertsen andMattiasson, 1998). Studieswere also conducted on the load-deflection relations of some components that have regular cross sections such as shafts, pipes and the rec- tangular section part (Li and Xiong, 2007; Katoh and Urata, 1993). However, because of its complicated cross section with large size, few investigations have be done on the load-deflection relations for the T-section rail straightening process. In thispaper, the load-deflection relationsofT-section rails under lateral loadsare studiedba- sed on elastic-plastic theory. The linear hardeningmodel and the elastic-plastic power-exponent hardening model are used for the analysis. Two experiments are also carried out to verify the accuracy of themodel. It will be applied on the automatic T-section rail straighteningmachine. 196 Y. Song, Z. Yu 2. Establishment of the load-deflection relations 2.1. Elastic deformation stage The following assumptions are accepted in this study: • The plane section assumption and uniaxial stress assumption are reasonable. • TheBauschinger effect and the reverse plastic deformation duringunloading are neglected. • Compared with the distance between the two support points, the deflection of the rail is small. TheT shaped cross section of the rail is shown in Fig. 1. In this work, analysis is carried out on the lateral bending deformation of the rail. Fig. 1. Plastic region and bending moment distribution At the elastic deformation stage, the relationship between the deflection and loading force can be written according to material mechanics δ= l3 6EI F F ¬ 2Mt l (2.1) where I = (H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4B 3 3 12 is the inertia moment against neutral axis of the cross-section, and 2l is the distance between the two support points, and E is Young’s modulus. And themaximum elastic bendingmoment Mt can be written as Mt = σs 6B1 [(H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4)B 3 3] (2.2) where σs is the yield stress of the material. 2.2. Elastic-plastic deformation stage 2.2.1. Linear-hardening model At this stage, plastic deformation occurs from the outer fiber in thematerial and the dashed area in Fig. 1 is the plastic region (Li et al., 2004). In the linear-hardening case, the stress-strain relation for the rail material is as follows σ= { Eε for ε¬ εs σs+EP(ε−εs) for ε­ εs (2.3) Load-deflection relations of T-section rails... 197 where εs is the strain at the yield point, Ep the linear hardening modulus, σ the true stress and ε the true strain. Figure 2 shows the stress distribution through the rail section after bending, so that (Johnson and Yu, 1981) σ=        σs z zs for 0¬ z¬ zs σs [ 1+λ ( z zs −1 )] for zs ¬ z¬ B1 2 (2.4) where λ=EP/E. Fig. 2. The stress distribution through the rail section after bending: (a) stress distribution after bending, (b) stress-strain curve of linear hardening material Thewidth of the rail foot is much larger than that of the rail web and head, so generally no fibers of the rail web and head are strained beyond the yield limit during deformation. Then, for the force equilibrium across the section Mx =2 zs ∫ 0 z z zs σs(H1−H3) dz+2 B2 2 ∫ 0 z z zs σs(H3+H4) dz+2 B3 2 ∫ 0 z z zs σs(H2−H4) dz +2 B1 2 ∫ zs z [ 1+λ ( z zs −1 )] σs(H1−H3) dz (2.5) where Mx is the bending moment imposed on the rail section and zs is the distance between fibers which are just at yield and the neutral axis. Equations (2.5) and (2.2) reduce to the following equation Mx Mt = 1 ξ [ 1+ 1 2 (λ−1)(H1−H3)B 3 1(ξ 3 −3ξ+2) (H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4)B 3 3 ] (2.6) where ξ= zs/(B1/2) is the relative elastic zone ratio and ξ­B3/B1. And Mx can also be written as Mx = F 2 x (2.7) Let CΣ be the relative change of total curvature of the rail, then based on the plane supposition of elastic-plastic bending and elastic mechanics (Wu et al., 2000) ξ= 1 CΣ = Ct CΣ (2.8) where Ct =Mt/(EI). 198 Y. Song, Z. Yu Then, from equations (2.6)-(2.8), the relation between CΣ and x can be obtained Fx 2Mt = CΣ Ct + 1 2 (λ−1)(H1−H3)B 3 1 [( CΣ Ct )−2 +2CΣ Ct −3 ] (H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4)B 3 3 2Mt F ¬x¬ l (2.9) And the loading stroke at the middle point, i.e. deflection, can be obtained (Chui, 1994) δΣ = Fl3t 6EI + n ∑ i=1 xiCΣxi∆x F ­ 2Mt l (2.10) where lt = 2Mt F ∆x= l− lt n xi = lt+ i∆x So, combining equations (2.9) and (2.10), if the loading force F is given, δΣ can be calculated under the linear-hardening model. 2.2.2. Elastic-plastic power-exponent hardening model The elastic-plastic power-exponent hardeningmodel is as follows (Johnson and Yu, 1981) σ= { Eε for ε¬ εs Eε−K(ε−εs) n for ε­ εs (2.11) where K, and n­ 1 are the hardening coefficient and the hardening exponent, respectively. Figure 3 shows the stress distribution through the rail section after bendingunder the elastic- plastic power-exponent hardeningmodel. Similar to equation (2.4), it can be proven that (Johnson and Yu, 1981) σ=        σs z zs for 0¬ z¬ zs σs z zs −K′ ( z zs −1 )n for zs ¬ z¬ B1 2 (2.12) where K′ =K(σs/E) n. Fig. 3. The stress distribution through the rail section: (a) stress distribution after bending, (b) residual stress distribution after springback, (c) stress-strain curve of elastic-plastic power-exponent hardening material Load-deflection relations of T-section rails... 199 Then, for the force equilibrium across the section Mx =2 zs ∫ 0 z z zs σs(H1−H3) dz+2 B2 2 ∫ 0 z z zs σs(H3+H4) dz +2 B3 2 ∫ 0 z z zs σs(H2−H4) dz+2 B1 2 ∫ zs z [ σs z zs −K′ ( z zs −1 )n] (H1−H3) dz (2.13) Equations (2.13) and (2.2) reduce to the following equations Mx Mt = 1 ξ − 3B31(H1−H3)K ′ξ2 [ ( 1 ξ −1 )n+2 n+2 + ( 1 ξ −1 )n+1 n+1 ] σs[(H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4)B 3 3] (2.14) Then, from equations (2.7), (2.8) and (2.14), the relation between CΣ and x can be obtained (2Mt/F ¬x¬ l) Fx 2Mt = CΣ Ct − 3B31(H1−H3)K ′ ( CΣ Ct )−2 [ ( CΣ Ct −1 )n+2 n+2 + ( CΣ Ct −1 )n+1 n+1 ] σs[(H1−H3)B 3 1 +(H3+H4)B 3 2 +(H2−H4)B 3 3] (2.15) So combining equations (2.10) and (2.15), the load-deflection curves can be plotted under the elastic-plastic power-exponent hardeningmodel. 3. Experiments To verify the load-deflection model, experiments are carried out respectively on the universal testing machine and the self-made hydraulic straightening machine. Take T/89 rail for test and the material considered for the rail is low carbon steel Q235 with following parameters: E =206000MPa, σs =235MPa; K =200000MPa, n=1.02, λ=0.0173. And the rail section has dimensions of H1 =20.9mm, H2 =41.1mm, H3 =11.3mm, H4 =7.7mm, B1 =89mm, B2 =10mm, B3 =15.88mm. Fig. 4. Experimental set-up: (a) on the universal testing machine (2l=500mm), (b) on the self-made hydraulic straighteningmachine (2l=1000mm) Figure 4a shows the experimental set-up on the universal testing machine AG-1 introduced fromShima dzuCorporation.The rail ismoment-free supported at both ends.And the distance 200 Y. Song, Z. Yu between the end supports is 500mm, that is 2l=500mm.During the test, the loading part in the middle is moved at the speed of 0.1mm/s, providing the loading force. And the results of this test are shown in Fig. 5a. Figure 4b shows the experimental set-up on the self-made hydraulic straightening machine. The rail is alsomoment-free supported at both ends. But the distance between the end supports is 1000mm, that is 2l = 1000mm. During the experiment, the loading part is moved at the speed of 0.6mm/s. The results are illustrated in Fig. 5b. Fig. 5. Comparison of the results by experiment andmodel: (a) on the universal testing machine (2l=500mm), (b) on the self-made hydraulic straighteningmachine (2l=1000mm) Figures 5a and 5b show the comparison of results by the experiment and by the model. It can be found that the load-deflection curves calculated with the elastic-plastic power-exponent hardening model are closer to the experimental results than those with the linear-hardening model. As shown in Fig. 5a, the relative error of the results obtained with the elastic-plastic power- exponent hardening model will become larger with the increase of the loading stroke δ. So in the case of large elastic-plastic bending deformation, the real material hardeningmodel is more complicated, and taking into account contact stresses and strains in the vicinity of the loading device is necessary to enhance the accuracy of the analytical results. 4. Conclusions In this paper, the load-deflection relations ofT-section rails under lateral loads are studiedbased on elastic-plastic theory. The linear hardening model and the elastic-plastic power-exponent hardening model are used for the analysis. To verify the model, experiments are carried out respectively on a universal testing machine and a self-made hydraulic straightening machine. From the results presented in this paper, the followings are concluded: • The load-deflection curves calculated with the elastic-plastic power-exponent hardening model are closer to the experimental results than those with the linear-hardeningmodel. • In the case of large elastic-plastic bending deformation, the real material hardeningmodel is more complicated, and taking into account contact stresses and strains in the vicinity of the loading device is necessary to enhance the accuracy of the analytical results. Acknowledgements This research was partially funded by the National Natural Science Foundation of China under GrantNo. 50835008 andNo. 71071138, science and technology project of Zhejiang province under Grant No. 2008C21124. Load-deflection relations of T-section rails... 201 References 1. BiempicaC.B., Diaz J.J.D., NietoP.J.G., Sanchez I.P., 2009,Nonlinear analysis of residual stresses in a rail manufacturing process by FEM,Advanced Mathematical Modeling, 33, 1, 34-53 2. Chui Pu, 1994, Straightening Theory and Parameters Calculation, Mechanical Industry Press, China 3. Eggertsen P.-A., Mattiasson K., 2010, An efficient inverse approach for material hardening parameter identification from a three-point bending test, Engineering with Computers, 26, 2, 159-170 4. Johnson W., Yu T., 1981, On springback after the pure bending of beams and plates of elastic work-hardeningmaterialsIII, International Journal of Mechanical Sciences, 23, 10, 687-695 5. Katoh T., Urata E., 1993, Measurement and control of a straightening process for seamless pipes,ASME J. Eng. Indus., 115, 347-351 6. Li J., Xiong G., 2007, Study on calculation method of press straightening stroke based on stra- ightening process model,Key Engineering Materials, 340-341, 2, 1345-1350 7. Li J., Zou H.J., Xiong G.L., 2004, Establishment and application of load-deflection model of press straightening,Advances in Engineering Plasticity and its Applications, 274, 1/2, 475 8. Masmoudi R., Theriault M., Benmokrane B., 1998, Flexural behavior of concrete beams reinforced with deformed fiber reinforced plastic reinforcing rods, ACI Structural Journal, 95, 6, 665-676 9. Omerspahic E., Mattiasson K., Enqvist B., 2006, Identification of material hardening para- meters by the three-point bending of metal sheets, International Journal of Mechanical Sciences, 48, 1525-1532 10. Ou J.P.,WangB., He Z., ZhangX.Y.,QianM.Z., 2004, Load-deflection response of concrete beams reinforced with FRP bars,Advances in Structural Engineering, 7, 5, 427-436 11. Shen H.B., SokolinskyV.S., Nutt S.R., 2004, Accurate predictions of bending deflections for soft-core sandwich beams subject to concentrated loads,Composite Structures, 64, 1, 115-122 12. Sokolinsky V.S., Shen H., Vaikhanski L., Nutt S.R., 2003, Experimental and analytical study of nonlinear bending response of sandwich beams,Composite Structures, 60, 2, 219-229 13. Tsai Shang-His, Kan Heng-Chuan, 2008, The exact solution of the load-deflectionmodel of a uniformly loaded rectangular cross-section cantilever beam, Journal of Physics D: Applied Physics, 41, 095502 14. Tsai Shang-His, Wang Yeng-Tseng, Kan Heng-Chuan, 2009, Analysis on the uniformly loaded rectangular cross-section cantilever by amodified load-deflectionmodel, Journal of Physics D: Applied Physics, 42, 045505 15. Volebov I.F, Polyakov A.P, Kolmogorov S.V., 1994,Mathematicalmodel of rail straighte- ning and experimental estimation of its adequacy, Journal of Materials Processing Technology, 40, 1/2, 213-218 16. Wu B.J., Chan L.C., Lee T.C., et al., 2000, A study on the precision modeling of the bars produced in two cross-roll straightening, Journal of Materials Processing Technology, 99, 202-206 202 Y. Song, Z. Yu Charakterystyki obciążeniowo-odkształceniowe szyn o przekroju teowym poddanych obciążeniom poprzecznym Streszczenie W pracy zaprezentowano charakterystyki obciążeniowo-odkształceniowe obciążanych poprzecznie szyn o przekroju teowym otrzymane na podstawie teorii lepkosprężystości.W badaniach przyjętomodel materiału o liniowym oraz potęgowym umocnieniu. Podano analityczne wyrażenia dla krzywych obcią- żenia w funkcji odkształcenia. W wyniku weryfikacji eksperymentalnej stwierdzono, że charakterystyki obciążeniowo-odkształcenioweotrzymane przy zastosowaniumodelu z potęgowymumocnieniemmateria- łu są bliższe rezultatom doświadczalnym niż w przypadkumodelu z umocnieniem liniowym. Manuscript received October 28, 2011; accepted for print May 4, 2012