You are on page 1of 22

Reprinted from

CMES
Satya N. Atluri

Computer Modeling in Engineering & Sciences
Founder and Editor-in-Chief:

ISSN: 1526-1492 (print) ISSN: 1526-1506 (on-line)

Tech Science Press

Copyright © 2010 Tech Science Press

CMES, vol.69, no.3, pp.261-279, 2010

Analysis and Prediction of Edge Effect on Inherent Deformation of Thick Plates Formed by Line Heating
Adan Vega, Naoki Osawa, Sherif Rashed and Hidekazu Murakawa

Abstract: A three dimensional thermal-elasto-plastic FEA has been performed to predict the heat induced (inherent) deformation produced in thick steel plates formed by line heating. Using this FEA, the edge effect on inherent deformation is clarified. From the results of this study, a method to predict the edge effect is developed. Using this method, the edge effect on inherent deformation, for a wide range of plate thickness and heating condition, can be easily predicted, been this, an important step toward the automation of the process. Keywords: Line heating, inherent deformation, FEM, edge effect, side edge effect, induction heating. 1 Introduction

Line heating using torch, induction or more recently laser heating is one of the most important plate forming processes used in the shipbulding industry (Kyrsanidi, Kermanidis and Pantelakis (1999), Moshaiov and Latorre (1985), Moshaiov and Vorus (1987), Tsuji and Okumura (1989)). However, the line heating process is far to be fully automated causing delays in production, even when some intends have been made not fully automatic machine for plate forming exist (eg. Ueda, Murakawa, Rashwan, Okumoto and Kamichita (1993 and 1994), Jang and Moon (1998)). The main reason of this is the fact that the relation between applied heat and final plate deformation, the key to automate the process, is too complicate to analyze by using simple mechanical models. Aiming to solve this, many researchers has presented theories to explain both the thermal and mechanical problem (eg. Jang, Seo and Ko (1997), Chang, Liu, and Chang (2005), Ling and Atluri (2006), Osawa, Hashimoto, Sawamura, Kikuchi, Deguchi, and Yamaura (2007), Liu (2006), Liu, Liu, and Hong (2007), Terasaki, Kitamura and Nakai (1999), Moshaiov and Shin (1991)). However most of these researches have been focused on the analysis of single heating lines applied over small plates.

On the other hand. is small compared with that at the interior of the plate. Magee. in forming of spherical shapes by line heating it is necessary to apply large amount of plastic (inherent) strain at the plate edges. pp. Further details may be found in Nishikawa. The solution procedure consists of two steps. Second.69. based on the iterative substructure method. 2010 More recently the authors proposed a method which considers all the influential factors affecting the process such as: geometry of the plate. Cheng. Stresses. Pratt and Fan (2005). Rashed. Serizawa and Murakawa (2005). . An iterative approach is used to ensure continuity of tractions between the linear and nonlinear regions. In addition. material properties. In fact the inherent deformation produced at plate edges does not behave in the same way as that at the central region of the plate (Bao and Yao (2001). Ishiyama and Murakawa (2008). Tango. 2 2. the uncoupled formulation considers the contribution of the transient temperature field to stresses through thermal expansion. However. the transient temperature distribution history obtained from the heat transfer analysis is employed as a thermal load in a subsequent mechanical analysis. This variation of plastic strains with the length of heating influences the four components of inherent deformation reducing the accuracy of the existing predictive method. the region close to the plate edges. strains and displacements are then evaluated. which presents complex distribution of inherent deformation. multi-heating lines. This approach aims to reduce the computational time for complex thermal-elastic-plastic analyses by separating the model into regions which are linear or weakly nonlinear and those which are highly nonlinear. heat-induced curvature.261-279. the temperature distribution history is computed using transient heat transfer analysis.262 Copyright © 2010 Tech Science Press CMES. Watkins and Steen (1997)). Liu. Yao. This variation of inherent deformation along the heating line is studied in details in this paper. no. location of the heating. the usage of simple straight heating lines may not be enough because the plastic (inherent) strain produced by straight heating varies along the plate length (been smaller at the plate edges). At the end. as well as temperature-dependent thermo-physical and mechanical properties. First. The thermo-mechanical behavior of plate forming by line heating is analyzed using uncoupled formulation. In those cases. a new method to predict edge effect on inherent deformation is presented. cooling condition. residual stresses. vol. and inter-heating temperature (See Vega (2009) and Vega.1 Formulation of the 3D thermal elastic-plastic FEM Method of analysis Thermal and mechanical analyses were undertaken using a proprietary finite element code.3.

By holding the line heating energy constant. For those cases where the heat input and/or the speed of the heating source were varied. The analysis conditions for most of the studied case are as follows: speed (v) equal to 3 mm/sec. which is defined as the integration of the plastic strain over the cross section of the plate.2 Computed results The deformation of the plate is expressed in terms of the inherent deformation. Cooling is defined corresponding to natural cooling in air. transverse shrinkage (δ yy ). These four fundamental components of inherent deformation can be determined by integrating the inherent strain over the cross section of the plate of thickness h as follows. specific heat.1.1.Analysis and Prediction of Edge Effect 263 2. Heat imput (Q) equal to 3125 J/mm. including thermal conductivity. Generally. inherent deformations are classified into longitudinal shrinkage (δ xx ). longitudinal bending (θ xx ) and transverse bending (θ yy ). Mild steel thermal. 2. The same finite element model used in the thermal analysis is employed in mechanical analysis. it may be observed . δxx = δyy = θxx = θyy = p εxx dydz/h (1) (2) (3) (4) p εyy dydz/h p εxx (z − h/2)/(h3 /12)dydz p εyy (z − h/2)/(h3 /12)dydz 3 Inherent deformations produced at plates edges Figure 3a shows a typical distribution of inherent transverse shrinkage produced by a straight-line heating along the heating line.1 Mesh generation and process parameters All analyses were carried out using rectangular flat plates of 2000x2000xh (mm). Young’s modulus. 30. Poison ratio. In this figure. Values of thickness (h) equal to 20. and yield stress (see Figure 1 and 2). values of heat input from 2000 to 10000 J/mm and values of speed from 1 to 25 mm/sec were studied. Necessary constraints are added to eliminate rigid body motion. 40 and 50 mm were used. The highest temperature on the surface in the heating zone is kept at about 850˚C. physical and mechanical properties with temperature dependency is considered. the input energy per unit length along the heating line is kept unchanged even though power and velocity change.

the center and the exit edge of the plate.69. 3b.261-279. respectively. 3 η(J/m2/s/°C) 0. no. bars (a). 2010 1.1 Transverse components of deformation First. close to the edges transverse shrinkage decreases and exhibits a complex distribution.6 ρ (g/mm3)x10-2 λ (J/mm/s/°C)x10-1 0. The heating and cooling process of each bar is given by state (1) to (2). However. This is an undesirable effect in plate forming because it causes deviations from the original design.0 0 0 250 500 750 1000 1250 1500 Τ [°C] Figure 1: Thermal and physical properties of mild steel that if the heating line is long enough.2 Thermal Properties 0.3. 3c and 3d). 3. 6 0. Where. 9 0. To understand the causes of the variation of inherent deformation along the plate length let us examine the process of forming inherent deformation. 2 c (J/g/°C) 0. The process of forming inherent transverse deformation in each state is described as follows: • In the initial state (state (1)). heating is applied to the bar (a). the distribution of transverse shrinkage is almost uniform in the interior of the plate. In this state. (b) and (c) represent the entrance edge. Similar tendency can be observed in the other three components of inherent deformation (see Fig. let us study the process of forming transverse components of inherent deformation using the three bars model show in Figure 4. vol.3 0. pp.9 Thermal Conductivity[λ ] Heat Transfer [η] Density [ρ] Specific Heat [c] 1. .264 Copyright © 2010 Tech Science Press CMES.

• When the moving heating source reaches bar (c). the large restraint has similar influence on the thermal expansion. • In the last state (state (4)). When bar (a) is at high temperature. bars (a) and (b) are cooling down and therefore they are shrinking. The thermal contraction along the heating line during the cooling of bars (a) and (b). large amount of heating is quickly dissipated and therefore the maximum temperature is lower. the restrain in the transverse direction of the plate . it tries to expand however it is pushed back due to the large restraint. thus shrinkage increases (state (2)). allows the area of the plate under high temperature (bar (c)) to freely expand in the transverse direction (state (3)). and thermal expansion of the bar (a) is fully transformed into transverse deformation. bar (a) is cooling down. while the heating source is moving toward the plate center. However. Then. this deformation is small due to the lower temperature caused by the heat lost at the plate edge. however in this case the heat input has reached the steady state and therefore the temperature is larger.Analysis and Prediction of Edge Effect 265 300 0 2 50 5 00 750 1000 1 25 0 1500 3 00 Mechanical Properties 250 200 150 100 50 0 0 σy (Mpa) E (Gpa) H (MPa) Poisson's Ratio [ σy ] Yield Stress Young Modulus [ E ] Thermal Expansion [ α ] Strain Hardening [ H ] [γ] 2 50 2 00 α (1/°C)x10 -7 1 50 1 00 γ ( x10-2) 250 500 750 1000 1250 50 0 1500 Τ [°C] Figure 2: Mechanical properties of mild steel since the heat is applied at the plate edge. • When the moving heat source reaches bar (b).

(b) and (c) is small) and therefore no significant additional plastic deformation in the transverse direction is produced at the plate exit edge.266 0.08 0 Copyright © 2010 Tech Science Press 500 1000 1500 2000 2500 3000 0.1 0 δ sx 0 -0. no.3 0 500 1000 1500 2000 2500 3000 L [mm] L [mm] (a) 0.005 0 θ ex -0.08 δx [mm] δy [mm] 0 -0. bar (b) represents the heating area while bars (a) and (c) represent the area of the plate that is not heated.006 0.2 -0.012 θx [radians] θy [radians] 0.0025 -0.015 0.04 -0. (c) Longitudinal bending and (d) Transverse bending have become small (the gradient of temperature between bars (a).2 Longitudinal components of deformation The process of forming longitudinal components of inherent deformation is described using the mechanical model shown in Figure 5.1 0.003 -0.005 (b) 0. creates negative transverse shrinkage (transverse expansion). the expansion created in state (c).012 0.005 500 1000 1500 2000 2500 3000 0 0 0 500 1000 1500 2000 2500 3000 L [mm] L [mm] (c) (d) Figure 3: Comparison between computed and assumed distribution of inherent deformation (a) Longitudinal shrinkage.006 -0.0025 0 500 1000 1500 2000 2500 3000 0.2 0.009 0.003 θ sy θ iy 0.0025 0 θ sx θ ix 0 0 500 1000 1500 2000 2500 3000 0. In addition. 2010 0. 3. The process is as follows: • At the beginning of heating.1 -0.005 0. pp.04 0 500 1000 1500 2000 2500 3000 -0.2 δ sy δiy 0 0.04 δ ey -0.69.009 0.0025 θ ey 0.12 0. vol.2 0. (b) Transverse shrinkage.12 CMES.04 δ ix δ ex 0.3 -0.015 0.3.1 0 500 1000 1500 2000 2500 3000 0.261-279. the thermal expansion of the plate is smaller due . Where.

the restraint L . • When the moving heat source reaches the exit edge of the plate. the longitudinal deformation at the interior of the plate increases compared with that at the entrance edge. Due to this.Analysis and Prediction of Edge Effect 267 a b c a b c L State (1) a b c a State (2) b c State (3) State (4) Figure 4: Mechanical model used to analyze the transverse component of inherent deformation Heat input u a b c L a b c L a ΔL b c Figure 5: Mechanical model used to analyze the longitudinal components of inherent deformation to the heat lost and the small restraint of the plate. the restraint of the plate decreases meanwhile the area behind the heating. is contracting. similar to the explanation given in previous section. • When the moving heating is approaching the plate center.ΔL .

In order to quantify this variation and consider it in the analysis of plate deformation. edge effect on transverse shrinkage (τ yy ). Therefore. Moreover. it is necessary to separate them into eight (8) components given as follows: s τxx = i s δxx − δxx i δxx i e δxx − δ δxx i δxx i s δyy − δ δyy i δyy i e δyy − δ δyy i δyy i s θxx − θ θxx i θxx i e θxx − θ θxx i θxx (5) e τxx = (6) (7) (8) (9) (10) (11) (12) s τyy = e τyy = s ψxx = ψxxe = s ψyy = i s θyy − θ θyy i θyy i e θyy − θ θyy i θyy e ψyy = . the edge effect has four components: edge effect on longitudinal shrinkage (τ xx ). edge effect on longitudinal bending (ψ xx ) and edge effect on transverse bending (ψ yy ).69. However. no.3. 2010 in the longitudinal direction has decreased. vol. 4 Edge effect on inherent deformation The inherent deformation produced by straight heating lines applied from edge to edge varies along the plate length as is explained in previous sections. we define the ratio between the maximum and the minimum inherent deformation at the entrance and exit edges of the plate as edge effect. pp.268 Copyright © 2010 Tech Science Press CMES. In addition. the edge effect at the entrance edge is different to that at the exit edge of the heating line as explained above.261-279. The reason why the longitudinal bending decreases at the exit edge is because the longitudinal bending is strongly influenced by the combination of other components of inherent deformation and thus it may vary when other components vary. Under this. no significant change in the distribution of longitudinal shrinkage is observed. the longitudinal bending decreases at the exit edge.

4) and therefore the edge effect on transverse shrinkage became larger. peak temperature is lower at the lower velocity due to the more time for heat dissipation. transverse shrinkage and transverse bending do not significantly changes with heat input (slightly decrease) while the edge effect on longitudinal bending slightly increases. 4. At the lower velocity. longitudinal shrinkage and transverse bending may cause increase of the edge effect on longitudinal bending as shown in the figure. middle and finishing part of the heating line. the restrain in the transverse direction of the plate become small (state (d) in Fig. average of inherent deformation along the interior of the plate is used in these relations. The reason of this small variation is the fact that when the heat input is changed. the edge effect on transverse bending decreases with heat input. the difference between the temperature at the top and bottom surface is also smaller than that at higher velocity for the same reason. positive values of edge effect mean decrease while negative values mean increase of inherent deformation.1 Variation of edge effect with heat input Figures 6 (a) shows the variation of edge effect with heat input at the entrance edge of the plate. The phenomenon changes with large values of heat input as seen in the figure. the edge effect does not significantly change (slightly decreases). In case of edge effect on longitudinal shrinkage. The increase of edge effect on transverse shrinkage with heat input is associated with a reduction of peak temperature at the heat surface. The variation on edge effect on transverse shrinkage. the edge effect on longitudinal shrinkage and transverse bending decreases while that on transverse shrinkage and longitudinal bending increases.5 kJ/mm when the plate thickness is 40 mm. The main reason of that is found in the relation between speed of the heating source and heat input as shown in Figure 7. As mentioned. i. The increase of edge effect on longitudinal bending is due to the same reason explained in previous section. the peak temperature of the heated surface varies and therefore the inherent deformation (see Vega (2009)). Because of this. In addition. and e are the starting. the relation between speed and heat input for large values of heat input is close to constant (contrary to the case of small heat input in which . Henceforth.Analysis and Prediction of Edge Effect 269 Where the superscripts s. large heat input is obtained at low speed. Figure 6 (b) shows the variation of edge effect with heat input at the exit edge of the plate. It is important to note that the inherent deformation at the interior of the plate is not exactly uniform therefore. As seen in this figure. It is noted that the edge effect on longitudinal shrinkage. it decreases with heat input because the decrease of restraint in the longitudinal direction caused by the lower speed. Because of this. It is also noted that for larger values of heat input. It is noted that up to Q equal to approximately 3. respectively.

.2 T = 0 80 3 0 °C °C 0. Consequently. 2010 Entrance edge effect [mm/mm] 2. vol.5 T = 800 ° C v [mm/sec] Thickness [h] 20 mm 30 mm 40 mm 50 mm T = 800 ° C 0.1 0 0 4 8 12 16 20 Q [KJ/mm] Figure 7: Relationship between neat heat input and speed of the heating source any small change on speed produce large variation of heat input).261-279.4 0.5 2 1. pp.69.5 0 1 2 3 4 5 6 2 3 1 2 3 4 5 6 3 Exit edge effect [mm/mm] ψ sx τ sx τ sy ψ sy 2 τ ey 2 ψ ex 2 1 1 1 τ ex ψ ey 1 0 0 1 2 3 4 5 6 0 0 1 2 3 4 5 6 Q [KJ/mm] Q [KJ/mm] (a) (b) Figure 6: Variation of edge effect with heat input (a) entrance plate edge (b) exit plate edge 12 9 0 2 4 6 8 10 0. no. the peak temperature at the heated surface does not significantly changes when the heat input increases.3.3 6 T= 800 0.270 Copyright © 2010 Tech Science Press CMES. the peak temperature at the heated surface does not changes and therefore no visible change in edge effect is observed.5 1 0. Because of this.

008 0. It therefore seems reasonable to conclude that the edge effect is independent of the plate size.012 0 500 1000 1500 2000 2500 3000 0.003 0. Similar conclusion is obtained after comparing the variation of inherent deformation with plate width (See Vega (2009)).Analysis and Prediction of Edge Effect 271 4. in the analysis of plates .1 -0. can be obtained by superposing the edge effect obtained from the analysis of a larger plate.04 0 Plate length [L] 3000 mm 1400 mm 800 mm 400 mm 0.2 0. and (d) Transverse bending Figure 9 shows the variation of edge effect with plate thickness at the entrance and exit edge of the plate.1 -0. (a) Longitudinal shrinkage.2 Variation of edge effect with plate geometry Figure 8 shows the variation of inherent deformation with plate length.08 0 500 1000 1500 2000 2500 3000 0.006 0.003 0 500 1000 1500 2000 2500 3000 0. Hence. the edge effect changes with plate thickness.006 0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000 L [mm] L [mm] (c) (d) Figure 8: Variation of inherent deformation with plate length. the distribution of edge effect for any plate size (same thickness). As seen in the figure. respectively.006 (b) 0.2 0. considering the similarity of inherent deformation at the entrance and at the exit edge of plates with different lengths larger than about 800 mm. first.012 θx [radians] 0 θy [radians] -0.08 δx [mm] δy [mm] 0.004 0 -0.2 -0.003 -0.004 -0.12 0.003 Plate length [L] 3000 mm 1400 mm 800 mm 400 mm 0.12 0.016 0.04 0 -0.2 0 Plate length [L] 3000 mm 1400 mm 800 mm 400 mm 0 -0.016 0.1 0 500 1000 1500 2000 2500 3000 0. (b) Transverse shrinkage.008 0.04 0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000 L [mm] L [mm] (a) 0. It is clearly seen that the inherent deformation produced at both plate edges does not significantly vary with plate length.1 0. (c) Longitudinal bending.04 -0. 0.006 0 Plate length [L] 3000 mm 1400 mm 800 mm 400 mm 0 0. There are two reasons for this.

200. this conclusion is based on analysis of small plates. 2010 Entrance edge effect [mm/mm] 3 15 25 35 45 55 3 4 15 25 35 45 55 4 Exit edge effect [mm/mm] 2 ψ sx 2 3 2 1 0 15 ψ ex 3 τ sx τ ey 2 1 τ sy ψ sy 1 1 ψ ey τ ex 0 0 15 0 25 35 45 55 25 35 45 55 h [mm] h [mm] (a) (b) Figure 9: Variation of edge effect with plate thickness (a) entrance plate edge. pp. 600. It is important to note that the variation of edge effect with plate thickness is large and therefore it may be considered in the analysis. Second. In this section. it is a fact that thicker plates have large difference of temperature between top and bottom surface (restraint through the thickness increases with plate thickness).272 Copyright © 2010 Tech Science Press CMES. which do not represent the actual size of plates used in shipbuilding. the inherent deformation is obtained by averaging the value at the central region of the plate. For those cases that are spaced more than 300 mm (in the case of longitudinal . a numerical study is performed to clarify the influence of side edge on inherent deformation of large plates. Thus. 400. transverse and longitudinal component of inherent deformation decrease as the heating line becomes closer to the free edge. especially angular distortion. no.3. 800 and 1000 mm) is computed. It may be clearly seen that both. inherent deformation of single heating lines applied at different distances from the free edge (0. as is explained in previous section. the heat input has been changed in order to attain same temperature at the heated surface (speed is kept constant). the edge effect may decrease.261-279. Figure 10 shows the resulting inherent deformation. Here. (b) exit plate edge) with different thickness. Plate models (2000 x 2000 x 40 mm) as well as heating and cooling condition are the same in all cases. Therefore. the edge effect changes. They concluded that plate deformation.3 location of heating lines Cheng et al (2005) performed numerical investigations aimed to clarify the side edge effect on plate deformation. greatly decreases when the heating line becomes closer to a free edge. vol. At first. 4. However.69.

Figure 11 shows the variation of edge effect on transverse shrinkage with thermal material properties.008 0.4 Variation of edge effect with material properties Magee et al.02 0.016 0.012 0.2 d W h θy δy δx δ [mm] 0.Analysis and Prediction of Edge Effect 273 components) and 400 mm (in case of the transverse components) from the free edge. θ [radians] .25 0 200 400 600 800 1000 0.004 0. In both cases. Despite the fact that the reduction of inherent deformation due to side edge effect is only found in a small area of the plate.05 0 0 θx 200 400 600 800 0 1000 Figure 10: Variation of inherent deformation with the distance from the plate edge 4. thermal properties such as specific heat. As it is clearly seen. these influences are small and can be neglected. 0.1 0.15 0. In order to evaluate this effect on inherent deformation the influence of material properties on inherent deformation produced at plate edges is considered. Almost similar conclusion is obtained after examining the variation of edge effect with mechanical material properties as is shown in Figure 12. the maximum variation of edge effect is of about 10% and it happens when the material properties have been increase (compared with the original value obtained from experiment) in about 20%. Therefore. However. density and thermal conductivity influences the edge effect. the influence is large. the influence of plate side edge on inherent deformation must be considered. the inherent deformation does not significantly change. in accurate prediction of inherent deformation due to line heating. (1997) showed that the extent of the edge effects on angular distortion is also related to the material properties.

the speed of the heating source.69. It has been also demonstrated that these variation can be quantified through the edge effects given in Equations 5 to12. (b) exit plate edge 5 Method to predict edge effect on Inherent deformation As demonstrated in previous sections.6 0.5 1 0.6 0.8 CMES. 2010 0.8 0. the edge effects can be related to the heat input parameter Q/h2 as shown in Figures 13 and 14. the edge effect on longitudinal shrinkage generally increases with the heat input parameter (Figure 13 (a)).274 2.5 0 -20 Young's modulus Yield stress Strain hardening coefficient Poisson's ratio 1.2 0 -20 Density Specific heat Heat transfer coefficient Thermal conductivity 15 25 35 45 55 0. no.2 0 -10 0 10 20 -10 0 10 20 Variation of thermal properties [%] Variation of thermal properties [%] (a) (b) Figure 11: Variation of edge effect on transverse shrinkage with thermal properties (a) entrance plate edge.5 0 -20 0 Density Specific heat Heat transfer coefficient Thermal conductivity 0.2 0.6 τ ey [mm/mm] 1.5 0 -10 0 10 20 0 -20 0 -10 0 10 20 Variation of mechanical properties [%] Variation of mechanical properties [%] (a) (b) Figure 12: Variation of edge effect on transverse shrinkage with mechanical properties (a) entrance plate edge.2 0. In order to create a method of prediction that considers all these variations.8 0.4 0.8 2 τ ey [mm/mm] 0.5 τ sy [mm/mm] 1 Young's modulus Yield stress Strain hardening coefficient Poisson's ratio 0. vol. the inherent deformation at the entrance and exit edge of the plate varies with the heat input.5 0.2 15 25 35 45 55 0.4 τ sy [mm/mm] 0.5 2 15 25 35 45 55 2.5 2 15 Copyright © 2010 Tech Science Press 25 35 45 55 0. Similar behavior is observed in Figure 13 (c) where the edge effect on . pp.3. the plate thickness and with materials properties. As it is seen in these figures.4 0.5 1 0.6 1.261-279. at the starting edge.4 0.8 0. (b) exit plate edge 2.4 0.6 0.

8 0 2 4 6 8 10 2 1.25 0 2 4 6 8 10 3 (c) 1 0. However.8 0.5 0. (c) Longitudinal bending and (d) Transverse bending longitudinal bending is shown.4 0 Thickness [h] 20 mm 30 mm 40 mm 50 mm τsy [mm/mm] Thickness [h] 20 mm 30 mm 40 mm 50 mm 1.25 0 1. the edge effect on transverse components of inherent deformation increases with the heat input parameter up to approximately 2.75 0.6 0. up to Q/h2 equal to approximately 10. up to Q/h2 equal to approximately 8. the edge effect on longitudinal shrinkage shows a complex distribution.2 0.2 0 1. Similar behavior is observed in Figure 14 (c) where the edge effect on longitudinal bending is shown.6 0 2 4 6 8 10 2 275 1 0. after that it slightly decreases as shown in Figures 13 (b) and (d).4 0 0 0 4 8 12 16 20 0 4 8 12 16 20 3 Q / h2 [J/mm] 3 Q / h2[J/mm ] (a) 3 2. it has been . (b) Transverse shrinkage.8 0. Despite the fact that edge effect shows a complex distribution.6 1. On the other hand.2 0. On the other hand.75 0 Thickness [h] 20 mm 30 mm 40 mm 50 mm 1.25 1.8 0.75 0 2 4 6 8 10 3 2. At the exit edge.75 0 0 0 4 8 12 3 h2[J/mm ] 16 20 0 4 8 12 3 h2[J/mm ] 16 20 Q/ Q/ (b) (d) Figure 13: Relation between edge effect and heat input parameter Q/h2 at the entrance plate edge (a) Longitudinal shrinkage.4 0.Analysis and Prediction of Edge Effect 2 1. However.6 τsx [mm/mm] 1.5 0.4 0.2 0.5 0. the edge effect on longitudinal bending does not significantly changes with the heat input parameter Q/h2 .25 ψ sx [rad/rad] ψ sy [rad/rad] Thickness [h] 20 mm 30 mm 40 mm 50 mm 2. edge effect on transverse components of inherent deformation decreases with heat input parameter as shown in Figures 14 (b) and (d).5 0. Generally. the edge effect on longitudinal bending starts to decrease with and increasing heat input parameter Q/h2 . it increases with the heat input parameter as seen in Figure 14 (a).

3. (c) Longitudinal bending and (d) Transverse bending demonstrated that it is possible to relate it to the heat input parameter Q/h2 .8 0.2 1. If this distribution is approximated as uniform distribution. To overcome this.261-279.5 -0.8 0. (b) Transverse shrinkage.2 0.5 Thickness [h] 20 mm 30 mm 40 mm 50 mm 0.4 0 0 0 4 8 12 16 20 0 4 8 12 16 20 3 Q / h2[J/mm ] 3 Q / h2[J/mm ] (a) 20 15 0 2 4 6 8 10 3 (c) 1 0 2 4 6 8 10 1 2.4 0 2 4 6 8 10 2 τex [mm/mm] 1.25 ψ ex [rad/rad] ψ ey [rad/rad] 0. no.8 0.2 0.276 2 1.6 1. vol. the edge effect varies along the heating line and it decreases toward the plate center.6 0 Thickness [h] 20 mm 30 mm 40 mm 50 mm 1.5 0 4 8 12 16 20 3 Q / h2[J/mm ] 3 Q / h2[J/mm ] (b) (d) Figure 14: Relation between edge effect and heat input parameter Q/h2 at the exit plate edge (a) Longitudinal shrinkage.5 0 0 0.69.4 0. the edge effect can be considered as a linear function of the extension of the edge effect.2 0. .6 Copyright © 2010 Tech Science Press 0 2 4 6 8 10 2 CMES. pp. However. 2010 3 2. the accuracy of the inherent deformation is significantly reduced.5 10 5 0 Thickness [h] 20 mm 30 mm 40 mm 50 mm 1. Examining the distribution of computed inherent deformations for cases with different conditions it was found that the extension of the entrance edge effect and the exit edge effect are approximately equal to 300 and 400 mm respectively. Using the relation given in Figures 13 and 14 the edge effect for a wide range of heating condition and plate thickness can be obtained.4 0 τey [mm/mm] Thickness [h] 20 mm 30 mm 40 mm 50 mm 1.8 1.6 1.75 0 0 4 8 12 16 20 -0. Note that values given in these figures correspond to the value at the plate edges.

. (2005): A Group Preserving Scheme for Inverse Heat Conduction Problems. S. D. Pratt. S. C.. Yao.13-38. Seo. C. 123:53-61. Journal of Manufacturing Process. E. C.2. CMES: Computer Modeling in Engineering& Sciences. These two types of edge effect are undesirable since they adversely affect the accuracy of the plate forming. G. 28-40. ASME Vol. (2006): Stability Analysis for Inverse Heat Conduction . K. Jang. C. Journal of Materials Processing Technology.. 13(1):22-27. Based on the analysis performed here. J. Pantelakis S. SME Vol. Journal of Manufacturing Science and Engineering. and Atluri. Moon. Yao. T.. Cheng. (2005): Analysis and Prediction of Size Effect on Laser Forming of Sheet Metal. 87:281-290. D. Y. X. Journal of Ship Production. Ko. (1997): A study on the prediction of deformations of plates due to line heating using a simplified thermal elasto-plastic analysis. 10. Ling. a database of edge effect is presented in Section 7. (2001): Analysis and Prediction of Edge Effects in Laser Bending. pp. B. 1. Chang. References Bao. Acknowledgement: Financial support from The Secretary of Science and Technology of The Government of The Republic of Panama is gratefully acknowledged..Analysis and Prediction of Edge Effect 277 6 Conclusions It has been demonstrated that the variation of inherent deformation along the plate length (known as edge effect) need to be separated into two: that at the entrance edge and that at the exit edge of the plate. A. C. P. L. L.S... Liu.R.1 of this thesis..N.. Jang. Journal of Ship Production.. Y. Y. Chang. D. (1998): An Algorithm to Determine Heating Lines for Plate Forming by Line Heating Method. S. Fan. J. 7/No. Kermanidis. C. Kyrsanidi. It is explained that the first is produced by the heat lost at the plate edge while the second is mainly produced by the reduction of restraint in the transverse direction of the plate during the time the plate is being heated toward the exit edge of the plate.1. D.. (1999): Numerical and experimental investigation of the laser forming process. Through the study of the influence of some influential factors on inherent deformation produced at plate edges it has been concluded that it is necessary to consider the variation of edge effect with plate thickness and heat input in order to improve the accuracy of the predictive model. 14(4):238245.W.. Liu.

A.. I. (1987): The Mechanics of the Flame Bending Process: Theory and Applications. 13.278 Copyright © 2010 Tech Science Press CMES. pp. J. A. (1991): Modified strip model for analyzing the line heating method-part 2: Thermo-elastic-plastic plates. Y. pp. pp. W.3. Vorus. pp.. Kamichika. Sidhu J. G. Calder.. Rashwan. Terasaki.. Y.149-160 (in Japanese).1-18.. CMES: Computer Modeling in Engineering & Sciences. Ueda.... (2007): Development of Heat Input Estimation Technique for Simulation of Shell Forming by Line-Heating. S. Journal of Ship Research. Kirby. pp.S. Latorre.399-408. K. (2007): Highly Accurate Computation of SpatialDependent Heat Conductivity and Heat Capacity in Inverse Thermal Problem.. 76. vol. C. J. 1. (2006): An Efficient Simultaneous Estimation of Temperature-Dependent Thermophysical Properties. 35(3):266-275. Okumoto. R. H.K.. M. Kikuchi.. H. Trans. M. Steen. CMES: Computer Modeling in Engineering & Sciences. Kamichika. Journal of Ship Production 10(4):248-257. Liu. CMES: Computer Modeling in Engineering & Sciences.. Okumoto. CMES: Computer Modeling in Engineering & Sciences. 14. Hong. Naval Architects.. pp... The West-Japan Soc. H. T. pp. Y. (1985): Temperature Distribution during Plate Bending by Torch Flame Heating. Nishikawa. Naval Architects. pp. (Report 3) – relation between heating condition and deformation. The West-Japan Soc. no.. Serizawa.379.69..77-90. K. R. J.. Okumura Y. (1993): Development of Computer-Aided Process Planning System for Plate Bend- .. Deguchi. 1. Proceeding of the LANE’97. Watkins. N. Y. N. H. J. 3. (1999): Predictive Equation for Thermal Cycle Generated by Line Heating Method. pp. 99. J. T.W. W. A. 2010 Problems. 2.321-329 (in Japanese). pp. Moshaiov... C. Journal of the Japan Society of Naval Architects. Murakawa. Osawa.. (1994): Development of Computer-Aided Process Planning System for Plate Bending by Line Heating.” Laser assisted net shape engineering 2.261-279. 20. Liu...S. H. Sawamura. (1997):“Edge effect in laser forming.45-53. Meisenbach Bamberg. Y. Moshaiov A. Ueda. 17. Murakawa. N.. (2005): Development of a Largescale FEM for Analysis Mechanical Problems in Welding. Yamaura. 2. G.. Hashimoto. Moshaiov A. M. Journal of Ship Research. M. Rashwan. R.219228. Murakawa. Magee. H. 31(4):269-281. Tsuji. 29.. Trans.. (1988): A Study on Line Heating Process for Plate Bending of Ship Steels. Nakai. Kitamura. Journal of Ship Research.1-11. 1. Liu L. Shin.

(Report 4) Transactions of Japan Welding Research Institute. S. . 1993. Y. CMES: Computer Modeling in Engineering & Sciences.. Japan. 1. Vega. M. Osaka University. Rashed. Doctoral Thesis.. Vol. Murakawa. A. Tango.. pp. A. Ishiyama. 1-14. No. 22(2):305-313.. (2008): Analysis and prediction of multi-heating lines effect on plate forming by line heating. 28. H.Analysis and Prediction of Edge Effect 279 ing by Line Heating. (2009): Development of Inherent Deformation Database for Automatic Forming of Thick Steel Plates by Line Heating Considering Complex Heating Patterns. Vega.

.

. GA 30097-8220. MathSci Net.techscience.com Website: http://www.techscience.com/cmes/ Manuscript submission http://submission. Zentralblatt fur Mathematik. INSPEC Databases. Suite G108 Duluth. Science Alert.com CMES is Indexed & Abstracted in Applied Mechanics Reviews. Computing & Technology. Science Navigator. USA Phone (+1) 678-392-3292 Fax (+1) 678-922-2259 Email: sale@techscience.com Subscription: http://order. Engineering Index (Compendex). Current Contents: Engineering. Mathematical Reviews. Materials Sciences & Engineering.CMES: Computer Modeling in Engineering & Sciences ISSN : 1526-1492 (Print).com Published by Tech Science Press 5805 State Bridge Rd. and Computer & Information Systems Abstracts Database).techscience. Mechanics.techscience. CompuMath Citation Index. Cambridge Scientific Abstracts (Aerospace and High Technology. 1526-1506 (Online) Journal website: http://www. Science Citation Index.