Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 48, 3, pp. 703-714, Warsaw 2010 STATIC AND DYNAMIC INTERACTIVE BUCKLING REGARDING AXIAL EXTENSION MODE OF THIN-WALLED CHANNEL Zbigniew Kołakowski Technical University of Lodz, Department of Strength of Materials, Łódź, Poland e-mail: zbigniew.kolakowski@p.lodz.pl The present paper deals with an influence of the axial extension mode on static and dynamic interactive buckling of a thin-walled channelwith imperfections subjected to uniform compressionwhen the shear lag phe- nomenon and distortional deformations are taken into account. A plate model is adopted for the channel. The structure is assumed to be simply supported at the ends. Equations of motion of component plates were obtained from Hamilton’s principle taking into account all components of inertia forces. In the frame of first order nonlinear approximation, the dynamicproblemofmodal interactivebuckling is solvedbythe transition matrixusing theperturbationmethodandGodunov’s orthogonalization. Key words: thin-walled structures, dynamic interactive buckling, axial mode 1. Introduction Thin-walled structures composed of plate elements have many different buc- kling modes that vary in quantitative and qualitative aspects. In the case of finitedisplacements, differentbucklingmodes are interrelated evenwith the lo- ads close to their critical values. Thepostcritical behavior cannot be described any more by a single generalized displacement. When the postcritical beha- vior of each individual mode is stable, their interaction can lead to unstable behavior, and thus to an increase in the imperfection sensitivity. A nonlinear stability theory shoulddescribe allmodes and interactive buckling, taking into consideration imperfections of the structure. The theory of interactive buckling of thin-walled structures subjected to static and dynamic loading has been already widely developed for over forty years. Although the problem of static coupled buckling can be treated as 704 Z. Kołakowski pretty well recognized, the analysis of dynamic interactive buckling is limited in practice to columns (adopting their beam model), single plates and shells. In theworld literature, a substantial lack of the nonlinear analysis of dynamic stability of thin-walled structures with complex cross-sections can be felt. In this study, a special attention is focused on the influence of the axial extensionmode on the global flexural bucklingmode of a thin-walled channel. 1.1. Static interactive buckling When components of the displacement state for the first nonlinear order approximationare taken intoaccount, it canbe followedbyadecrease invalues of global loads. In the case when the critical values corresponding to global buckling modes are significantly lower than those for local modes, then their interaction can be considered within the first nonlinear approximation. It is possible as the post-buckling coefficient for uncoupledbuckling is equal to zero for the second order global mode in the Euler column model, and in the case of an exact solution, it is very often of little significance. The theoretical static load-carrying capacity obtained within the frame of the asymptotic theory of the nonlinear first order approximation is always lower than the minimum value of critical load for the linear problem, and the imperfection sensitivity can only be obtained. Since the late 1980’s, theGeneralizedBeamTheory (GBT) (Basaglia et al., 2008; Camotim et al., 2008; Silva et al., 2008) has been developed extensively. Recently, a new approach has been proposed, i.e., the constrained Finite Strip Method (cFSM) (Adany and Schafer, 2006a,b; Dinis et al., 2007; Schafer, 2006). These two alternativemodal approaches to analyze the elastic buckling behavior have been compared in papers. In the current decade, in more and more numerous publications (Adany and Schafer, 2006a,b; Basaglia et al., 2008; Camotim et al., 2008; Dinis et al., 2007; Schafer, 2006; Silva et al., 2008), the attention has been paid to the global axial mode, which is considered only in the theoretical aspect in linear issues, that is to say, in critical loads. Adany and Schafer (2006a) said that ”it should be noted that this axial mode is a theoretically possible buckling mode, even though it has little practical importance”. In the axial extension mode, longitudinal displacements of the cross-section dominate and thismode can be referred to as the shortening one (Fig.1). The axial mode is symmetric with respect to the cross-section axis of symmetry and it is symmetric with respect to the axis of overall bending (Kołakowski andKowal-Michalska, 2010). In the present study, a trial to dispute with the statement included in the paper by Adany and Schafer (2006a) has been undertaken and the attention Static and dynamic interactive buckling... 705 Fig. 1. Longitudinal displacement distributions of the axial mode of the channel has been focused on an interaction of the global flexuralmodeof bucklingwith the global axial extension mode in the first order nonlinear approximation of the perturbation method (Kołakowski andKowal-Michalska, 2010). The present study is based on the numerical method of the transition matrix usingGodunov’s orthogonalization. Instead of the finite stripmethod, the exact transition matrix method is used in this case. A plate model of the columnhas been adopted in the study to describe global buckling,which leads to lowering the theoretical value of the limit load. In the solution obtained, the co-operation between all the walls of structures being taken into account, the effects of an interaction of certain modes having the same wavelength, the shear lag phenomenon and also the effect of cross-sectional distortions are included. The distortion instability of the channel is investigated using the nonlinear theory. The solution method was partially presented in paper Kołakowski and Królak (2006). 1.2. Dynamic interactive buckling Dynamic pulse load of thin-walled structures can be divided into three categories, namely: impact with accompanying perturbation propagation (a phenomenon that occurs with the soundwave propagation speed in the struc- ture), dynamic load of amean amplitude and a pulse duration comparable to the fundamental flexural vibration period, and quasi-static load of a low am- plitude and a load pulse duration approximately twice as long as the period of fundamental natural vibrations. As for dynamic load, effects of damping can be neglected in practice. Dynamic buckling of a columncanbe treated as reinforcement of imperfec- tions, initial displacements or stresses in the column through dynamic loading in such amanner that the level of dynamic response becomes very high.When the load is low, the column vibrates around the static equilibrium position. On the other hand, when the load is sufficiently high, then the column can 706 Z. Kołakowski vibrate very strongly or can move divergently, which is caused by dynamic buckling (i.e. dynamic response). In the literature on this problem, various criteria concerning dynamic sta- bility have been adopted. Themost widely used is the Budiansky-Hutchinson criterion (BudianskyandHutchinson, 1966;Hutchinson andBudiansky, 1966), inwhich it is assumed that the loss of dynamic stability occurswhen theveloci- ty with which displacements grow is the highest for a certain force amplitude. Other criteria were discussed in Ari-Gur and Simonetta (1997), Huyan and Simitses (1977), Petry and Fahlbusch (2000), Volmir (1972), for instance. The critical values of dynamic stresses, obtained for the first nonlinear ap- proximation,maybehigher than theminimumvalues of static critical stresses, respectively. The dynamic response to the rectangular pulse load of the duration corre- sponding to the fundamental period of flexural free vibration has been analy- sed. 2. Formulation of the problem The cross-section of the structure composed of a few plates is presented in Fig.2 along with local Cartesian systems of co-ordinates. A long prismatic thin-walled channel built of panels connected along longitudinal edges has been considered. The channel is simply supported at its ends. In order to account for all modes and coupled buckling, a plate model of the thin-walled channel has been assumed. The material the channel is made of is subject to Hooke’s law. For each plate component, precise geometrical relationships are assumed in order to enable the consideration of both out-of-plane and in-plane bending of the i-th plate (Kołakowski and Królak, 2006) εxi =ui,x+ 1 2 (w2i,x+v 2 i,x+u 2 i,x) εyi = vi,y+ 1 2 (w2i,y +u 2 i,y+v 2 i,y) (2.1) 2εxyi = γxyi =ui,y +vi,x+wi,xwi,y+ui,xui,y+vi,xvi,y and κxi =−wi,xx κyi =−wi,yy κxyi =−2wi,xy (2.2) where: ui, vi,wi are components of the displacement vector of the i-th plate in the xi, yi, zi axis direction, respectively, and the plane xiyi overlaps the central plane before its buckling. Static and dynamic interactive buckling... 707 Fig. 2. Prismatic plate structure and the local co-ordinate system When the full tensor of membrane strains (2.1) εxi, εyi, γxyi = 2εxyi is taken into account, then a ”full” analysis of all bucklingmodes, including the axial extension mode, can be carried out. Themain limitation of the assumed theory lies in an assumption of linear relationships between curvatures (2.2) and second derivatives of the displace- ment w. This is themost often applied limitation in the theory of thin-walled structures. The attention has beendrawn to the necessity of considering the full strain tensor and all the components of inertial forces in order to carry out a proper dynamic analysis in the whole range of length of the structures. For thin-walled structures with initial deflections, Lagrange’s equations of motion for the case of an interaction of N eigenmodes can be written as (Kowal-Michalska, 2007; Schokker et al., 1996; Sridharan and Benito, 1984) 1 ω2r ζr,tt+ ( 1− σ σr ) ζr+aijrζiζj−ζ ∗ r σ σr + . . .=0 for r=1, . . . ,N (2.3) where: ζr is the dimensionless amplitude of the r-th buckling mode, σr, ωr, ζ∗r – critical stress, circular frequency of free vibrations and dimensionless amplitude of the initial deflection corresponding to the r-th buckling mode, respectively. Due to the fact that the axial mode is taken into consideration, in the present paper it is assumed that the absolute maximum value of one of the components of the displacement field of the r-th mode is equal to the first plate thickness t1. The expressions for aijr are to be found in Kołakowski 708 Z. Kołakowski andKrólak (2006), Kowal-Michalska (2007). In equations of motion (2.3), the inertia forces of the pre-buckling state and second order approximations have been neglected (Kowal-Michalska, 2007; Schokker et al., 1996). The initial conditions have been assumed in the form ζr(t=0)=0 ζr,t(t=0)=0 (2.4) The static problem of interactive buckling of the thin-walled channel (i.e. for ζr,tt =0 in (2.3)) hasbeen solvedwith themethodpresented inKołakowski and Królak (2006). The frequencies of free vibrations have been determined analogously as in Teter and Kołakowski (2003), whereas the problem of inte- ractive dynamic buckling (2.3) has been solved bymeans of the Runge-Kutta numerical methodmodified by Hairer andWanner. At the point where the load parameter for static problems σ reaches its maximum value σs (the so-called theoretical load carrying capacity) for the imperfect structurewith regard to the imperfection of the bucklingmodewith the amplitude ζ∗r , the Jacobian of the nonlinear system of equations (2.3) is equal to zero. 3. Analysis of the calculation results 3.1. Eigenvalue problems A detailed analysis of the calculations is conducted for the compressed channel with the following dimensions of its cross-section (Fig.3) (Adany et al., 2008) b1 =150mm b2 =60mm b3 =15mm t1 = t2 = t3 =2mm Fig. 3. Geometry of the thin-walled channel Static and dynamic interactive buckling... 709 Each plate is made of steel characterized by the followingmechanical pro- perties: E=210GPa, ν =0.3, ρ=7850kg/m3. The global flexural mode (m = 1) and the axial extension mode (for m = 1) for the assumed length ℓ have been analyzed. The following index symbols have been introduced: 1 – flexural mode for m = 1; 2 – axial mode for m=1. Calculations have been conducted for four lengths of the column: ℓ=10000, 7500, 5000, 2500mm.The lengths ℓ have been selected as tomake the values of global critical stresses (i.e., for m= 1) lower than local critical loads, which enables us to analyze the column buckling within the first order approximation. In Table 1 values of critical loads σr for four lengths ℓ of the channel under investigation are shown. The critical loads σ2 for the axial mode, which is identical for the lengths considered, are collected. For the channel whose length is 2500 ¬ ℓ ¬ 10000mm, the maximum displacements in the cross-section plane (i.e., v(2), w(2)) are equal to approx. 2% of the longitu- dinal displacements (i.e., u(2)) at most (Kołakowski and Kowal-Michalska, 2010). Thedisplacements u(2) are practically constant for the cross-section for x = const. Thus, this mode can be called the ’pure’ axial extension mode. For the axial mode, the displacements are equal to the thickness t1 (that is to say, u (2) max ∼=u(2) = t1). Values of the natural frequencies ωr of free vibrations corresponding to the two modes under analysis for different column lengths ℓ are presented in Table 1, too. Vibration frequencies were determined taking into account all components of the inertia forces (Kowal-Michalska, 2007; Teter andKołakow- ski, 2003) (i.e. in-plane ρu,tt, ρv,tt and out-of-plane ρw,tt). The same index symbols have been adopted as for the static problems. Table 1.Critical stresses σr and natural frequencies ωr of the channel ℓ σ1 σ2 ω1 ω2 [MPa] [MPa] [s−1] [s−1] 10000 10.69 80212 11.595 1624.8 7500 19.00 80212 20.609 2166.4 5000 42.69 80211 46.332 3250.4 2500 167.88 80211 183.73 6498.6 710 Z. Kołakowski 3.2. Interactive buckling in the first order approximation Static coupling buckling Detailed results of the static interactive buckling analysis are presented in Table 2 for the channel. The following two variants of the imperfections are assumed: 1. ζ∗1 = |ℓ/(1000t1)|, ζ ∗ 2 =0, 2. ζ∗1 = |1.0|, ζ ∗ 2 =0. In each case, the sign of the imperfection ζ∗1 has been selected in the most unfavorable way, that is to say, as to obtain the lowest theoretical load- carrying capability σS for the given level of imperfectionwhen the interaction of buckling modes is accounted for. In Table 2, the ratio of the theoretical load carrying capacity to the global flexural critical stress σs/σ1 for the assumedvariants of imperfections is given. A comparison of the results presented in Table 2 allows us to state that when the axial mode is accounted for in the interaction, then the theoretical load carrying-capacity σs is considerably decreased. A decrease in the load- carrying capacity σs/σ1 does not exceed 40%. Table 2.Theoretical static load carrying capacity σs/σ1 Imperfection σs/σ1 variant ℓ=10000 ℓ=7500 ℓ=5000 ℓ=2500 1 0.6369 0.6852 0.7455 0.8238 2 0.8495 0.8495 0.8492 0.8452 It follows from this comparison that the consideration of the axialmode in the interaction is necessary as it results in a visible decrease in the theoretical load-carrying capacity in the first order approximation. The nonlinear coefficient of system of equations (2.3), namely a211, exerts the main influence on the decrease in the load σs. The key role in the inte- raction of the buckling mode is played by the coefficient a211, i.e., the term a211ζ2ζ 2 1 of the third order in the expression for potential energy (Kołakowski andKowal-Michalska, 2010). It is related to the term σ(2)L2(U (1)) (following the notation of papers by Kowal-Michalska (2007), Schokker et al. (1996), Sridharan and Benito (1984)). This term arises by the product stress σ(2) as- sociatedwith the axial modewith the term representing themidsurface strain L2(U (1)) and integrating the same over the structure. The longitudinal displa- cements u(2) of the axialmode are of the sameorder as the displacements w(1) Static and dynamic interactive buckling... 711 for the global flexural mode due to the assumed conditions for the mode nor- malization (i.e., u (2) max =w (1) max). In the theory of thin plates, we always have w(i) ≫ u(j) for i 6= j, apart from the interaction of the mode with the axial mode analyzed in this study. Itmakes the coefficient a211 play such a key role for the first order approximation. Dynamic response Further on, an analysis of dynamic interactive buckling (i.e. dynamic re- sponse) of the channel under considerationwas conducted. Identically as in the static analysis, an interaction of the same modes was considered. A detailed analysis was conducted for: a rectangular pulse load σ(t)=σD for 0¬ t¬T1 and σ(t)= 0 for T < t. This case corresponds to the pulse duration equal to the period of the fundamental flexural free vibration T1 =2π/ω1. In Table 3, values of the critical dynamic load factors DLFcr = σ BH D /σ1 for various column lengths ℓ and for two imperfection variants under analysis are given, where σBHD denotes the critical value of dynamic stress determined from the Budiansky-Hutchinson criterion (Budiansky and Hutchinson, 1966; Hutchinson and Budiansky, 1966). The values of critical dynamic load factors DLFcr =σ BH D /σ1 presented in Table 3 correspondwith some accuracy to themaximum values of deflections ζrmax within the applicability of the assumed theory (i.e. the total maximum deflection of the column is at least sixty times as high as the cross-section wall thickness) (Kowal-Michalska, 2007), andnot to asymptotic values ((Budiansky andHutchinson, 1966; Hutchinson andBudiansky, 1966). Themain limitation that results from the adopted theory is the assumption of a linear dependence between the curvatures and second order derivatives of the displacement. Table 3.Critical dynamic load factors DLFcr =σ BH D /σ1 Imperfection DLFcr =σ BH D /σ1 variant ℓ=10000 ℓ=7500 ℓ=5000 ℓ=2500 1 0.669 0.746 0.850 1.045 2 1.199 1.148 1.138 1.095 For the cases of column dimensions under analysis, the obtained values of DLFcr are higher than the respective values of theoretical static load carrying capacity σs/σ1. For the second variant of imperfections DLFcr > 1, whereas for the first variant DLFcr < 1 for the column length 10000¬ ℓ¬ 5000 and DLFcr > 1 for ℓ=2500, correspondingly. 712 Z. Kołakowski 4. Conclusion In the study, special attention has been focused on the coupled buckling of the Euler global mode of buckling with the axial mode in the first nonlinear ap- proximation of the perturbationmethod. In theworld literature, it is probably the first study, to the author knowledge, devoted to the dynamic interaction of buckling with the axial mode. This problem may be of great significance and it requires further investigations. According to the author’s opinion, a further analysis of the interactive buckling ought to include an interaction of the axial mode with global and local modes. Therefore, the interactive buckling should be further analyzed and comprehensively and thoroughly discussed. Acknowledgement This publication is a result of the research work carried out within the project subsidized over the years 2009-2011 from the state funds designated for scientific research (MNiSW –NN501113636). References 1. Adany S., Schafer B.W., 2006a, Buckling mode decomposition of single- branchedopen cross-sectionmembers via finite stripmethod:Derivation,Thin- Walled Structures, 44, 563-584 2. Adany S., Schafer B.W., 2006b, Buckling mode decomposition of single- branched open cross-sectionmembers via finite strip method: Application and examples,Thin-Walled Structures, 44, 585-600 3. Adany S., SilvestreN., Schafer B.W., Camotim D., 2007,On the iden- tification and characterization of local, distortional and global modes in thin- walledmembers using the cFSM andGBT approaches,Proceeding of the Sixth International Conference on Steel and Aluminium Structures ICSAS’2007, Be- ale R.G. (Ed.), Oxford Brookes University, 760-767 4. Adany S., Silvestre N., Schafer B.W., Camotim D., 2008, Buckling mode identification of thin-walledmembers: A comparison between cFSM and CBT approaches,Proceedings of the fifth International Conference on Coupled Instabilities in Metal Structures CIMS2008, Rasmussen K. and Wilkinson T. (Eds.), University Publishing Service, The University of Sydney, 249-256 5. Ari-Gur J., Simonetta S.R., 1997, Dynamic pulse buckling of rectangular composite plates,Composites Part B, 28B, 301-308 Static and dynamic interactive buckling... 713 6. Basaglia C., Camotim D., Silvestre N., 2008, GBT-based post-buckling analysis of thin-walled lipped I-section steel columns,Proceedings of the fifth In- ternational Conference on Coupled Instabilities inMetal StructuresCIMS2008, Rasmussen K. and Wilkinson T. (Eds.), University Publishing Service, The University of Sydney, 241-248 7. Budiansky B., Hutchinson J.W., 1966, Dynamic buckling of imperfection- sensitive structures,Proceedings of the Eleventh International Congress of Ap- plied Mechanics, Goetler H. (Ed.), Munich, 636-651 8. Camotim D., Basaglia C., Silvestre N., 2008, GBT buckling analysis on thin-walled steel frames, Proceedings of the fifth International Conference on Coupled Instabilities in Metal Structures CIMS2008, Rasmussen K. and Wil- kinsonT. (Eds.), UniversityPublishing Service, TheUniversity of Sydney, 1-18 9. Dinis P.B., Camotim D., Silvestre N., 2007, FEM-based analysis of the local-plate/distortional mode interaction in cold-formed steel lipped channel columns,Computers and Structures, 85, 1461-1474 10. Hutchinson J.W., BudianskyB., 1966,Dynamic buckling estimates,AIAA Journal, 4-3, 525-530 11. Huyan X., Simitses G.J., 1977, Dynamic buckling of imperfect cylindrical shells under axial compression and bending moment, AIAA Journal, 35, 8, 1404-1412 12. Kołakowski Z., Kowal-Michalska K., 2010, Influence of the axial mode on interactive buckling of the thin-walled channels in the first nonlinear ap- proximation,Shell Structures: Theory and Applications, 2, PietraszkiewiczW., Kreja I. (Eds.), Taylor and Francis Group, 125-128 13. Kołakowski Z.,KrólakM., 2006,Modal coupled instabilities of thin-walled composite plate and shell structures,Composite Structures, 76, 303-313 14. Kowal-Michalska K. (Ed.), 2007, Dynamic Stability of Composite Plated Structures,WNT,Warszawa [in Polish] 15. Petry D., Fahlbusch G., 2000, Dynamic buckling of thin isotropic plates subjected to in-plane impact,Thin-Walled Structures, 38, 267-283 16. Schafer B.W., 2006, Review: The direct strength method of cold-formed member design,Proceeding of International Colloquium on Stability and Ducti- lity of Steel Structures SDSS 2006, CamotimD., SilvestreN.,Dinis P.B. (Eds.), Instituto Superior Tecnico, Lisbon, 49-65 17. Schokker A., Sridharan S., Kasagi A., 1996, Dynamic buckling of com- posite shells,Computers and Structures, 59, 1, 43-55 18. Silva N.M.F., Camotim D., Silvestre N., 2008, GBT cross-section ana- lysis of thin-walled members with arbitrary cross-sections: A novel approach, Proceedings of the Fifth International Conference on Thin-Walled Structures 714 Z. Kołakowski 2008, Mahendran M. (Ed.), Queensland University of Technology, Brisbane, Australia, 1161-1171 19. Sridharan S., Benito R., 1984, Columns: static and dynamic interactive buckling, J. Engineering Mechanics, ASCE, 110, 1, 49-65 20. Teter A., Kołakowski Z., 2003,Natural frequencies of thin-walled structu- res with central intermediate stiffeners or/and variable thickness,Thin-Walled Structures, 41, 291-316 21. VolmirS.A., 1972,NonlinearDynamics of Plates and Shells, Science,Moscow [in Russian] Statyczne i dynamiczne interakcyjne wyboczenie cienkościennego ceownika uwzględniające osiową postać wzdłużną Streszczenie Wprezentowanej pracy omówionowpływ osiowej wzdłużnej postaci na statyczne i dynamiczne interakcyjne wyboczenie cienkościennego ceownika z niedokładnościa- mi poddanego równomiernemu ściskaniu przy uwzględnieniu zjawiska shear-lag oraz dystorsyjnej deformacji. Przyjęto płytowy model ceownika. Konstrukcja jest przegu- bowopodpartanaobukońcach.Równania ruchupłyt składowychotrzymanoz zasady Hamiltona, biorąc pod uwagę wszystkie składowe sił bezwładności. Dynamiczne za- gadnieniemodalnego interakcyjnegowyboczenia w ramachpierwszego rzędu nielinio- wej aproksymacji rozwiązanometodąmacierzy przeniesienia i metodą ortogonalizacji Godunova. Manuscript received December 17, 2009; accepted for print January 25, 2010