Research Journal of Recent Sciences ______ ______________________________ ______ ____ ___ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 36 On the Effect of Imperfection on Buckling load of Perforated Rectangular Steel Plates Ali Dadrasi Department of Mechanics, Shahrood Branch, Islamic Azad University, Shahrood, IRAN Available online at: www.isca.in Received 7 th October 2012 , revised 14 th November 2012 , accepted 6 th December 2012 Abstract Buckling behavior of plates and shells is one of the important characteristics in analysis of any structure. One the most significant parameter that must be considered in buckling phenomenon is imperfection. In this paper the effect of imperfection on buckling load of steel rectangular plates under uni - axial in - plane compressive loading is investigated by numerical and experimental methods. The plates were free on two op posite sides and simply supported at the load side whereas the opposite side is either clamped or simply supported. This means that the plate primarily exhibits a type of column’s buckling. Key w ords : Buckling, p erforated rectangular plates, f inite elemen t, i mperfection. Introduction Using plates and shells in fabrication of many structures is unavoidable. These structures may experience axial compression loading in their longevity and buckle through these loads. Furthermore, these structures usually ha ve discontinuities, such as cutouts, which may have effects on their stability. Therefore we must consider any parameter that affect on buckling load. Buckling analysis of thin - walled structures under axial compression has been investigated by many researc hers 1 - 2 . Also thin plates under the concentrated and the distributed forces on the whole plate edge have been investigated by some other researches 3 - 4 . Of course, the stability of these plates is dependent on many parameters such as type of support, type o f loading, ambient temperature, imperfection and etc. a direct matrix method for the bucking loads on structures was applied to the examination of the stability of flat square plates with central square perforations under various combinations of bi - axial l oading by Yetterman and Brown 5 . Their results were given for a wide range of hole to plate size from 0 to 0.8. Christopher et al. 6 studied the elastic stability of plates containing perforations using the conjugate load - displacement method. The effect of the size of a central square hole in a square plate on the elastic buckling load was investigated. Chang - jun and Rong 7 have been provided and solved a system of new boundary integral equations to address the plane stress, critical loads and post - buckling p roblem for perforated thin plates, based on the general bifurcation theory and mathematical model for the stability analysis of perforated thin plates. Their results show that the boundary element method was efficiently applied to the post - buckling analyse s of perforated thin plates. Also in another study Shariati et al 8 investigate the buckling of tubular steel shells with elliptical cutout subjected to oblique loading. In their study the influence of shell length, shell diameter, shell angle and diameter s of elliptical cutouts on the predicted buckling values has been explored. Shanmugam et al. 9 studied on post - buckling behaviour and the ultimate load capacity of perforated plates with different boundary conditions under the uniaxial and biaxial compressi on. They used the finite element method (FEM) and compared their numerical results with the experimental results. They established a design formula to determine the ultimate load of buckling of perforated plates. Suneel Kumar et al. 10 have been presented t he details of tests carried out on the collapse load of stiffened plates with and without cutout, with reinforced cutout and initial imperfections. Based on strut approach and orthotropic plate approach, a generalized computer program for the semi - analytic al solutions proposed by various investigators was developed. They proposed an approximate method based on strut approach to calculate the collapse load of stiffened plates with cutouts and initial imperfections. Eccher et al. 11 provided the application of the isoparametric spline finite strip method to the elastic buckling analysis of perforated folded plate structures. They introduced the general theory of the isoparametric spline finite strip method. A number of numerical examples of flat and folded perf orated plate structures illustrated the applicability and accuracy of their proposed method. The linear buckling analyses of square and rectangular plates with circular and rectangular holes in various positions subjected to axial compression and bending m oment were developed and has been studied by Maiorana et al. 12 The aim of their paper was to give some practical indications on the best position of the circular hole and the best position and orientation of rectangular holes in steel plates, when axial co mpression and bending moment acted together. As depicted in above review, the cut out existing in the plate, the geometry of cut out, the type of loading, the supports of the plates are all among the factors which can have a great effect in the stability of the plates. In this paper, the numerical and the Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 37 experimental investigation on the buckling behavior of the rectangular plates with circular and square cut outs under uni - axial in - plane compressive loading in elasto - plastic range with various loading ba nds are performed. The plates are free on two opposite sides and simply supported at the load side whereas the opposite side is simply supported. This means that the plate exhibits primarily a column type of buckling. Moreover, the relation between stabili ty of the rectangular plates having square and circular cut out with the same cross section has been studied. Furthermore, the effects of the various loading lines have been studied numerically using ABAQUS. Also, some buckling experimental tests have been performed on the samples. Numerical and the experimental results are in good agreement. Material and Methods The meshed geometry and type of loading are shown in f igure 1. In this figure, l is the loading band which varies in the range of ( 0≤ l ≤a ) . The position l=0 is relates to the concentrated load exerted on the middle of the width of the plate, a , and the position l=a represents to the distributed load exerted on the entire the width of the plate. In this investigation, the structural s teel rectangular plates with 100 x 150 x 2.07 mm dimensions are used. These plates have square and circular cut outs. The side of the square cut out is considered to be e=30 mm and for having the same area of two types of cut out, the diameter of the circl e is considered to be D=33.84mm . The lower edge of the plates has been placed in the clamped support and on the other edge, patch compression with various bands has been exerted through a simply support. The tests have been conducted for the width of loadi ng of l = 15,30,50,75 and 100 mm . Only the loaded section of the boundary is constrained and the rest is free as it is shown in Figure 1. In these tests, the post buckling behavior of the plates has been fully studied, too. The mechanical properties of the tested structural steel plates have been specified through the tensile test in accordance with the ASTM - E8 standard using an INSTRON8802 servohydraulic machine with stress - strain curve as displayed in figure 2 - a. The obtained elasticity modulus from line ar elastic region equal to E=218Gpa as it shown in figure 2 - b. Moreover the Poisson's ratio value is considered to be =0.33 . The data of the plastic region of the stress - strain curve has been used for analysis of the plastic behavior in ABAQUS software. The numerical analysis: In Abaqus after defining the geometry, boundary conditions and the applied loading, we must mesh the structure for analysis. This is achieved through using the S8R5 quadrilateral non - linear elements. This elemen t is an 8 - node element with 5 - degree of freedom including 3 displacements in three directions of the coordinate axes and two rotations for each node. In this element, 8 nodes have been placed in such a way that 4 nodes have been placed on the square corner s and 4 other nodes have been placed in the middle of the sides of the square (figure 3.). The S8R5 element is very suitable for the element arrangement of the thin plates and in this research all the analyses have been performed using this type of element 13 . Loading band Figure - 1 The meshed geometry of the plate and the type of the applied loading (a) The plate with circular cut out (b) The plate with square cut out Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 38 (a) (b ) Figure - 2 (a) The specimen being under the tensile test (b) The stress - strain curve resulted by the experimental test Figure - 3 The S8R5 element After meshing the specimen, a linear buckling analysis for getting eigenvalues is performed by using the “Buckle” solver in ABAQUS and the buckling mode shapes are obtained. Since in eigenvalue linear analysis, the plastic properties of the specimen are not taken into account, a linear finite element analysis (eigenvalues), particularly for the plates made of ductile materials, overestimates the real value for buckl ing load. Since buckling usually occurs in smaller mode shapes, a linear analysis should be performed first for all specimens, to find the mode shapes with smaller eigenvalues. For example, the three buckling mode shapes has been displayed in the f igure 4. The displacements for these mode shapes are saved in a specific file and used in the next analysis (Static Riks), so that the effect of mode shapes are considered in buckling analysis. Otherwise, the software would choose the buckling mode in an arbitrar y manner that resulting unrealistic results. This step is called Buckle in the software. For this step, the Subspace solver method of the software was used and three primary mode shapes were obtained. Also a numerical calculation was made over the loaded b and with a constant displacement. Since the imperfection of the plate considerably affects on ultimate buckling load, the imperfection is an important factor which must be considered in buckling of the plates. Imperfection means the every deviance of idea l manner of each structure. To make this point more transparent and clear, the effect of the imperfection on the ultimate buckling load of a rectangular plate has been compared numerically with two different values of the imperfection. The sample is a plat e with a square cut out which is placed with clamped support in lower side and is applied under a 15 mm loading band by a simply support in upper side. The critical buckling loads of 13.298kN and 11.856kN have been obtained for the values of the maximum im perfection amplitude, 0.1 and 0.48 mm, respectively. The difference between these two loads is 1.442kN and as depicted in figure 5, it is a considerable amount. Therefore, prior doing the non - linear analysis, the amount of the imperfection must be determin ed. As a result in this article, the maximum imperfection amplitude (the preliminary bend) of all plates has been considered to be equivalent to 0.48 mm. It is worth mentioning that, this quantity has been calculated through the quantity of the maximum imp erfection amplitude found in the plates used in experimental tests. Effect of loading band on buckling behavior of plates : In this part the effect of the loading band on buckling behavior of perforated plates has been numerically studied. In this research the loading band is varied between 15 to 100 mm. For sample, the behavior of load versus vertical displacement of plate with circular cutout (load - displacement curves) under various loading has been displayed in figure 6. Also load displacement curves fo r plates with circular and square cutouts and perfect (with no cut out) plate for loading band l =100 mm are compared in f igure 7. Also the critical buckling loads of plates with and without cut out for different loading bands are given in tables (1). As s hown in the curves, after the load of the plate reaches to the critical value, the buckling phenomenon is occurred and the plate is bent by a low force. Moreover it is observed that as the loading band increases, the ultimate value of the buckling load is also increased. Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 39 (a) (b) (c) Figure - 4 The mode shapes of the plate with square cut out under a 15mm loading band and the fixed support. (a) Shape of the first mode (b) Shape of the second mode (c) Shape of the third mode Figure - 5 The effect of initial curvature (imperfection) on the ultimate buckling load Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 40 Figure - 6 Figure - 7 The behavior of load - displacement of the Comparison of load displacement curves for plates with plate with circular cut out circular and square cutouts and perfect plate for loading band l=100 mm Table - 1 Results of buckling load for plates with no cutout (perfect) and plates with circular and square cutouts Critical bu ckling load with square cutout (kN) Critical buckling load with circular cutout(kN) Critical buckling load (kN) Loading band (mm) 11.851 12.004 12.986 15 12.256 12.424 13.353 30 12.486 12.676 13.762 50 12.821 13.048 14.126 75 13.076 13.372 14.664 100 As indicated in tables (1), for one specific loading band and under the same boundary conditions, the buckling load of the specimen with circular cut out is a little higher than the buckling load of the specimen with square cut out. For the less di fficulties of manufacturing and creation of the circular cut out than the square cutout and due to the higher buckling load of the specimen with circular cut out than the specimen with square cut out, it is recommended that for perforated structures which don't have any limitations on type of the cut out’s geometry, circular cut out is so better. Experimental results : Experimental technique for measurement of maximum imperfection of plates : As mentioned above, if we want to make a correct comparison betwe en the numerical and experimental results of the plates buckling, we must use the plates with the same imperfections. Therefore, at first a great number of rectangular plates with the desired dimensions were provided and then their imperfection was experim entally measured. For this purpose, the amount of the preliminary curve obtained through pressing the plate along its thickness and drawing its load - displacement diagram simultaneously. This operation which was performed by an INSTRON servohydraulic machin e has been schematically displayed in the figure 8 and figure 9 shows a sample under imperfection’s measuring test. δ in figure 8, indicates the amount of the imperfection. A sample of load - displacement diagram for measuring of imperfection has been displ ayed in figure 10. As it is evident in this figure, where the diagram assumes an asymptotic state, it indicates the amount of imperfection on the horizontal axis which is almost equal to 0.36 mm. Then for prevention of changing of the preliminary bend the boring operations were performed on the plates which had the same preliminary plates by SPARK device. Experimental measurement of buckling load : The lower edge of the plates was placed in clamped support and their upper edge was placed in a simply suppor t with various loading bands. Also the loading conditions are carried out under displacement control with speed of 0.01 mm/s. for sample, the results of experimental tests which were conducted on the rectangular plates with circular cut outs have been disp layed in figure 11. It’s obvious that the trend of experimental results is so similar to numerical and is in agree with some other reported FEM analysis 14 - 16 . Results and Discussion As shown in figure 11, the trend of the experimental diagrams is wholly similar to the numerical diagrams trend and upon the increase of the band of the loading, the buckling load is also increased. For better understanding of the results, the numerical and experimental quantities of the buckling load are compared in proportio n to the boundary conditions and the various loading bands for the specimens having square and circular cut out in tables 2 and 3. Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 41 Table - 2 The numerical and experimental critical buckling loads for plate with square cut out (average error = 6.55%) Numerica l error (%) Experimental buckling load (kN) Numerical buckling load(kN) Loading band(mm) 12.01 10.584 11.851 15 7.61 11.389 12.256 30 5.84 11.796 12.486 50 0.75 12.918 12.821 75 Table - 3 The numerical and experimental critical buckling loads for pl ate with circular cut out (average error = 9.72%) Numerical error (%) Experimental buckling load (kN) Numerical buckling load(kN) Loading band(mm) 11.99 10.718 12.004 15 8.14 11.488 12.424 30 6.43 11.920 12.687 50 12.32 14.881 13.048 75 We can sup pose that these plates buckle as columns. The critical elastic buckling load equation for columns is: (1) Where E is the modulus of elasticity, I is moment of inertia and L is the length of column fo r simply support boundary condition. And for simply - fixed ends boundary condition the critical elastic buckling load equation is: (2) For example for this problem, we have: I= bt 3 / 12 = [ 100*(2.07) 3 /12 ] = 73.914 mm 4 , L = a = 150 mm, E= 218000 N/mm 2 . Therefore: This result is agree with result of figure 7 for load - displacement curve of perfect specimen (critical buckling load, P cr =15 kN). Therefore, we observed that for calculation of nearly critica l buckling load of plates with no cutout and full loading band with simply - clamped ends boundary conditions, we can use Euler's buckling formula with good approximation. Conclusion We can conclude from the results obtained in this research that: When the buckling phenomenon occurs, the capacity of the load toleration is considerably decreased. The effect of the imperfection in the buckling load is considerable and as comparison between buckling load of two similar plates with different imperfection isn’t correct. The initial curvature (the maximum imperfection amplitude) of plate has no effect on post - buckling behavior of plate. The results show that, as loading band increases, the ultimate buckling load also increases. The buckling load of the specimen with circle cut out is a little more than specimens with the square cut out with the equal surface area. Therefore, it is suggested that, if possible, the specimens with circular cut out are used for structures, because the producing of circular cut outs a re very easier than square cut outs. Numerical and experimental results are close to each other and their trend is completely the same. We can use Euler's critical buckling formula for calculation of nearly critical buckling load of perfect plates and ful l loading band with simply - clamped ends boundary conditions. (a) Imperfection Figure - 8 The calculation’s procedure of the amount of the preliminary bends through experimental test: (a) Before pressing the plate. (b) After pressing the plate Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 42 Figure - 9 A sample under imperfection’s measuring test Figure - 10 Load - displacement diagram for calculation of the imperfection (δ=0.36 mm) Figure - 11 Load - displacement behavior of plates with circular cut out. Research Journal of Recent Sciences ______ _ _ _______________________________ ______________ _ ________ ISSN 2277 - 2502 Vol. 2 ( 3 ), 36 - 43 , March (201 3 ) Res. J. Recent Sci. International Science Congress Association 43 Figure - 12 The numerical and experimental comparison of the behavior of the buckling load vs. the loading ba nd for specimen with a circular and square cut out References 1. Dadrasi A., An Investigation on Crashworthiness Design of Aluminium Columns with Damage Criteria, Res. J. Recent Sci ., 1(7) , 19 - 24 (2012) 2. Shariati M., and Allahbakhsh H., Numerical and experim ental investigations on the buckling of steel semi - spherical shells under various loadings, Thin - Walled Structures 48 , 620 – 628 (2010) 3. Timoshenko S.P. and Gere J.M., Theory of Elastic Stability, 2nd edition, New York: McGraw - Hill Book Company (1961) 4. Mignot F., and Puel J. P., Homogenization and Bifurcation of Perforated Plates, Engineering science , 18 , 409 - 414 (1980) 5. Yetterman A.L. and Brown C.J. The Elastic Stability of Square Perforated Plates, Computer and Structures , 21(6) , 1267 - 1272 (1985) 6. Christopher J., and Yettram L., The Elastic Stability of Square Perforated Plates Under Combinations of Bending, Shear and Direct Load, Thin - Walled Structures , 4 , 239 - 246 (1986) 7. Chang - jun Cheng., and Wang R., Boundary integral equations and the boundary element method for buckling analysis of perforated plates, Engineering Analysis with Boundary Elements, 17 , 57 - 68 (1996) 8. Shariati M., Fereidoon A. and Akbarpour A., Buckling Load Analysis of oblique Loaded Stainless Steel 316ti Cylindrical Shells with Elliptical Cutout , Res. J. Recent Sci ., 1(2) , 85 - 91 (2012) 9. Shanmugam NE., Thevendran, V., and Tan, YH. Design formula for axially compressed perforated plates. Thin - Walled Structures, 34 , 1 – 20, (1999) 10. Suneel Kumar, M., Alagusundaramoorthy, P., and Sundaravadivelu, R., Ultim ate Strength of Ship Plating under Axial Compression, Ocean Engineering, 33 , 1249 – 1259 (2006) 11. Eccher G., Rasmussen K. J. R., and Zandonini, R., Elastic buckling analysis of perforated thin - walled structures by the isoparametric spline finite strip method, Thin - Walled Structures , 46 , 165 – 191 (2008) 12. Maiorana E, Elastic stability of plates with circular and rectangular holes subjected to axial compression and bending moment. Thin Walled Structure, 47(3) , 241 - 255 (2009) 13. ABAQUS 6.7 PR11 user’s manual. 14. Lewlyn B.R .N., Rodrigues L.R. and Anjaiah Devineni., Process Parameters Optimization in GFRP Drilling through Integration of Taguchi and Response Surface Methodology Murthy, Res. J. Recent Sci ., 1(6) , 7 - 15 (2012) 15. Magarajan U., Thundil karuppa Raj R. and Elango T., N umerical Study on Heat Transfer of Internal Combustion Engine Cooling by Extended Fins Using CFD, Res. J. Recent Sci ., 1(6) , 32 - 37 (2012) 16. Purkar T. Sanjay and Pathak Sunil., Aspect of Finite Element Analysis Methods for Prediction of Fatigue Crack Growth R ate, Res. J. Recent Sci ., 1(2) , 85 - 91 (2012)