Numerical Simulation and Experimental of Residual Stress Field of SAE1070 Spring Steel Induced by Laster Shock

Transcription

1 Research Journal of Alied Sciences, Engineering and Technology 5(20): , 203 ISSN: ; e-issn: Maxwell Scientific Organization, 203 Submitted: Setember 27, 202 Acceted: November, 202 Published: May 5, 203 Numerical Simulation and Exerimental of Residual Stress Field of SAE070 Sring Steel Induced by Laster Shock Bing Han, Changliang Xu, Jialian Shi and Hua Song School of Mechanical Engineering and Automation, University of Science and Technology Liaoning, Anshan 405, China Abstract: Laser Shock Processing (LSP) is an innovative surface treatment technology. LSP can induce comress residual stress layer in the surface of metallic materials, which is used for imroving the fatigue life. In this study, the roagation course of tyical laser shock wave in material was analyzed according to the theory of lasticity dynamics and the theory of stress wave. The effects of laser shock rocessing, residual stress of SAE070 steel surface layer were studied. The finite element software ANSYS/LS-DYNA was alied to simulate the residual stress field of SAE070 steel surface layer induced by laser shock rocessing. A bilinear isotroic model in a nonlinear inelastic model is created by using ANSYS/LS-DYNA finite element analysis software. Shock wave ressures under different eak loads are adoted to redict residual stress distribution. The exerimental results for the magnitude and distribution of residual stress are consistent with the numerical simulation. Keywords: Laser shock rocessing, residual stress, numerical simulation, finite element INTRODUCTION Laser Shock Processing (LSP) is a new surface modification technology. LSP is a rocess similar to traditional shot eening, in which favorably oriented comressive residual stresses are imarted at and near the surface of the target material. In the rocess of LSP the laser with high ower density and burst ulse irradiates on metallic material, the coating on surface of material vaorized raidly, forming lasma, the lasma exloded to induce a high ressure shock wave. Which can modify the microstructure of material suerficial layer and results in a high comressive residual stress in material suerficial layer? So that LSP can imrove some mechanical roerties of metallic materials, It has been roosed as a cometitive alternative technology to classical treatments for imroving fatigue, corrosion and wear resistance of metals (See et al., 2002). LSP can effectively strengthen materials without changing chemical comonents of material surface. Esecially can rolong the fatigue life of metallic materials such as aerosace industry, automotive engineering and internal combustion engine manufactory and so on. Studies show that the numerical simulation method realized the forecast of residual stress to the traditional eening technology successfully, such as laser eening, ultrasonic eening, water jet eening and water-jet cavitation eening. Meguid et al. (999) the elastolastic behavior of a single rojectile s interaction with metal materials in the traditional simulation shot eening technology obtain the distribution of residual stress. HU and YE redicted the size and distribution of residual stress in the rocess of the laser shock eening technology by the finite element method (Hu et al., 2006; Ding and Ye, 2006).It offered a new effective aroach to otimize arameters of laser shock. From then on, several researchers have used ansys to analyze the laser generated shock waves roagating into different metal materials in the ast six years (Braisted and Brockman, 999; Ding and Ye, 2003; Peyre et al., 2003; Ocana et al., 2007; Zhang et al., 2004). This research is on the basis of the test conditions of laser shock rocessing to SAE070 steel, the use of ANSYS/LS-DYNA analysis latform, it establishes the aroriate finite element model of nonlinear dynamics in the technological rocess and the grid artition, load size, loading history and other key conditions were considered when to simulate the distribution of comressive residual stress field. And comared with the residual stress distribution which determined by X- ray diffraction method finally. LASER SHOCK PROCESSING TEST OF SAE070 STEEL AND PRINCIPLE OF STRENGTH The SAE070 steel is irradiated by a sot which diameter is 7 mm, the laser ower density is 8 GW/cm 2 and duration is 50 ns. A section of the material to be eened is coated with aint and a thin film of water is run over the surface. A high Corresonding Author: Bing Han, School of Mechanical Engineering and Automation, University of Science and Technology Liaoning, Anshan 405, China 4869

2 Fig. : Schematic of the LSP energy laser is aimed at a articular location on the target. A single ulse sent from the laser asses through the water layer and causes the aint layer to vaorize into lasma. As the lasma exands, high ressure, short duration shock waves roceed into the water and the base metal. The resence of the water film tends to confine the energy and increase the intensity of the ressure ulse within the base. The rocess used to laser shock een a target material is shown in Fig.. The method is similor to Ballard (99a) investigated into LSP effects on the residual stress field and fatigue behavior of steel in his doctoral dissertation.to determine eeling and corrosion in samles, as well as residual stress distribution in terms of deth, the samles are olished with electrolytic olishing devices (Proto Electrolytic olisher- Model 888). Stresses generated on the secimen surface were measured by X-ray diffraction and the lastically affected deths were evaluated using successive electrolytical olishing and X-ray diffraction. FINITE ELEMENT SIMULATION Exlicit finite element algorithm: Much exerimental work has been erformed with regard to LSP, but relatively little effort has been sent in the develoment of analytical techniques to redict the residual stresses from LSP. In order to analyze the resonse of the material to the shock ressure with the exlicit finite element method, the dynamic equation at time t can be exressed as follows, according to the virtual work rincile: Greatly LSP has extensive alied foreground in some modem manufacturing: Ω ρ µ iδµ idω + υρµ iδµ idω Ω where, 4870 f Ω + Γ iδµ id Γδµ id σ ijδdijdω () Ω where, ρ Mass density μμ and μμ Nodal acceleration and velocity, resectively v The daming coefficient δμ The virtual dislacement f i The body force density T Ґ The boundary force alied on the boundary Ґ σ ij The Cauchy stress tensors, DD iiii δδuu 2 iiii + δδuu iiii Deformation rate tensors. Discretization of this roblem is accomlished by means of the standard finite element rocedure. After aggregation, we have the dynamic equilibrium equation as follows: Ω int ext (2) M U + CU + F F where, M is the diagonal mass matrix, UU and UU the nodal acceleration and velocity vector, resectively, C diagonal daming matrix, F int the internal element force vector, F ext accounts for external and body force loads. To advance to time t n+, central difference time integration is usually used as follows: - ext int U M ( F F CU ) (3) n+ n- 2 2 n n (4) U U + U n n+ n + 2 U + U t n 2 t + U (5)

3 t n n+ ( t + t ) n+ 2 2 Then we udate the geometry by adding the dislacement increments to the initial geometry: Res. J. Al. Sci. Eng. Technol., 5(20): , 203 n+ 0 n+ X X U (6) + The central difference scheme is conditionally stable. The stability limit can be defined using the element length Le and the wave seed of the material C d : Table : Performance arameters of SAE070 steel Proerties Value Density, ρ/kgm Poisson's ration, v 0.28 Elastic modulus, E/GPa Dynamic yield strength, σ y /MPa stable L C e t (7) d The material constitutive model: Assuming that the model is an ideal elastolastic material, it obeys the Mises yield criterion following an elastic strain constitutive relation. The highest elastic stress level in the direction of the shock wave roagation is defined as the Hugoniot Elastic Limit (HEL). At ressures greater than the HEL, ermanent deformation occurs. If it is assumed that the yielding occurs when the alied stress is greater than HEL, material yield occurs. The dynamic yield strength of the material is Han and Ju (2009): ( 2v) σ y HEL ( v) (8) where, v is Poisson, s ratio. Since the HEL is not strainrate deendent, it is ossible to determine an uer limit on the dynamic yield strength. HEL data for many materials are ublished in the oen literature or can be obtained from flyer late testing. Ballard develoed a one-dimensional analytical model to redict the surface residual stress induced by LSP in an elastic-lastic half-sace. The model was assumed to be ax symmetric since the ressure imacts on the center of the material surface. According to Hook s law, the surface lastic strain is: 2HEL ε ( 3λ + 2u P HEL ) (9) where, is the imact wave ressure; u and λ are Lame, s constants. The mechanical erformance arameters of the 45 steel used in the ideal elastic-lastic model are shown in Table. 487 Fig. 2: Finite element model boundary conditions schemes Finite element model: A finite element analysis method adoting LS-DYNA and ANSYS is described in detail to attain the simulation of single and multile laser shock rocessing. The simulation is comosed of two distinct arts including dynamic analysis erformed by LS-DYNA and static analysis erformed by ANSYS. The hysical model in the exeriment and influence of the reduced imact wave on the edge are shown in Fig. 2. A 40 mm high and 40 mm wide quarter cylinder model is established. Solid 64 is used to mesh in the model. The 0 mm length of the imact zone is divided along the X and Y directions into an average of 30 ortions, with a size roortion of. The rest of the model is also divided into 50 ortions, with a size roortion of 0.2. A 2 mm dee ortion of the imact zone is divided into an average of 30 comonents, with a size roortion of. The remaining ortion is divided into 60 arts, with a size roortion of 0.2. Thus, the model is divided into elements. The two sections of the model are set with a symmetrical boundary and the side and underside of the model are set with a no reflection boundary (Fig. 2). These conditions enable the transmission of the stress wave in the boundary and subsequent simulation of a 3D semi-infinite entity. Boundary conditions are incororated into the model in accordance with the analysis of the model boundary conditions. The satial and temoral distribution of the ressure ulse on the surface of the target material has significant exerimental efforts. For a constant laser ower density I0, the eak ressure P is given by Fabbro et al. (998).

4 where, σ 0 The initial surface residual stress and is set as zero in this study. Imact wave load, method and loading time during WCP are key factors that determine the accuracy of numerical simulations. The duration of imact waves is 70 ns. The variations in the imact wave ressure as a function of time are shown in Fig. 3. RESULTS OF LASER SHOCK PROCESSING AND DISCUSSION Fig. 3: Imact wave ressure changes as a function of time α ( GPa) 0.0 Z(gcm s ) I0( GWcm ) (0) 2α + 3 Dynamics analysis: In the ost rocessor of ANSYS/LS-DYNA, the imact waves are dynamically roagated in the finite element model. The von Mises stress of imact wave roagation in the finite element model at different eriods is shown in Fig. 4. According to the stress wave theory, the elastic wave velocity and the lastic wave velocity can be calculated as follows Wang (985): where, Z is the reduced acoustic imedance [2/Z /Z +/Z 2 ]of the confining (Z ) and target (Z 2 ) materials and α is the efficiency of the interaction. It was also reasonably assumed that 0% (α 0.) of the incident energy density is used for the ressure rise of the lasma. The eak ressure is indeendent of the laser ulse duration or the laser wavelength but is roortional to the square root of the laser ulse intensity. The surface residual stress can be exressed as Ballard (99b) σ surf π σ 0 - ( + υ) µε ( + υ) ( -υ) + σ L r 2 0 () C e l C l ( v) E ( + v)( 2v) E 3( 2v) ρ ρ where, ee cc ll The elastic wave velocity cc ll The lastic wave velocity ρ The material density E The Elastic modulus v The Poisson's Ratio (2) (3) (a) t 999 ns 4872

5 (b) t 996ns (c) t 2998 ns Fig. 4: The dynamic stress wave roagation rocess (d) t ns 4873

6 Fig. 5: Residual X-stress dynamic change as a function of time at different element It is easy to find that the elastic wave velocity is always faster than the lastic one. Then time needed for the first two tyes of stress wave roagating to the bottom surface of the target material is shown as follows: l t (4) e thick e cl l t (5) thick cl where, l thick is the thick of target material. The imact wave sreads from the surface inward and then to the bottom, with a maximal von Mises stress of 579MPa (Fig. 4a). The maximal von Mises stress of the finite element model decreases to 560 MPa when 996 ns (Fig. 4b). The imact wave sreads to the center of the finite element model, the energy gradually increases with deth and the maximum von Mises stress decreases to 574 MPa (Fig. 4c). The internal imact wave and continuous interaction of the material continuously increases the energy, eventually forming a stable residual stress field on the surface of the model. The maximal von Mises stress is 55 MPa (Fig. 4d). The X-dynamic rocess of different tyical units (element no.: 427, 4307, 4487) on the material surface after five continuous cycles of imact waves is shown in Fig. 5. When the material is subjected to five imact waves, element stress shocks occur between comressive and tensile stresses, which become stable after aroximately ns, with a maximum 4874 amlitude of aroximately 300 MPa. Residual stress continuously increases after each imact wave. Final formation of residual stress: The residual stress distribution formed near the surface is deicted in Fig. 6, x and y comonents of residual stresses are only roduced in the layer about mm from the to surface after laser shock rocessing. The residual stress is symmetrically distributed along the X and Y directions and has a consistent value (Fig. 6b and 6c) and z comonent of stress is nearly zero on the to surface. A tensile residual stress is found in the z comonent of stress of the model as shown in Fig. 6d, which is generated by the reversing yielding effect due to the interaction of the various stress waves in the material.the simulation results are consistent with the exerimentally measured residual stresses. The comressive residual stress formed near the surface is aroximately 389 MPa. The final equivalent stress distribution, in which the von Mises stress reaches 55 Ma, is illustrated in Fig. 6a. Influence of different eak loads on residual stress: To determine the varying atterns of residual stress of different nodes along deths with increasing shock waves, The varying atterns of the elements along the imact direction with shock waves and under continuously varying eak values are shown in Fig. 7. Under the different eak value loads, the varying atterns of the residual stresses of the elements with the number of shock waves are different and the residual stresses of the different elements also different. As indicated by the numerical value, the surface residual stress, after the alication of the shock wave functions and a eak load P 2.8 GPa, yields a maximum

7 (a) Contours of effective stress (b) Contours of residual X-stress (c) Contours of residual Y-stress 4875

8 (d) Contours of residual z-stress Fig. 6: Residual stress field numerical simulation is a little lower than exerimental results, broadly seaking, the simulation results for eak load is 3.0 GPa and the exerimental results of 50ns shot eening show high consistency. CONCLUSION Fig. 7: Comarison with residual stress simulation and test results comressive residual stress of aroximately -359 MPa. At a eak load P 3.4 GPa, the maximum comressive residual stress is aroximately -675 MPa. COMPARISON OF SIMULATION AND EXPERIMENTAL Peak load differs after the alication of the imact waves. The numerical simulation and exerimental results for residual stress distribution are comared in Fig. 7: The simulation results for the residual stress near the surface from 0 to mm are slightly lower than the test results, but at the beginning and the ending, the A finite element analysis method adoting ANSYS and LS-DYNA is described in detail to attain the simulation of single and multile laser shock rocessing and a bilinear isotroic model in a nonlinear inelastic model is created. The residual stress distribution on the surface layer of SAE070 steel during LSP under different eak loads can be obtained. On to surface of the material treated, a homogeneous deression with little roughness modification in the action zone of the shock ressure is induced by single LSP, according to the simulation of surface deformation. The simulation of residual stress field for single laser shock is found to be well correlated with the exerimental data and indicates that comressive residual stress of the material is aroximately uniform on shocked region and the stress gradient in deth is small. The simulation results under eak loads of 3.0GPa, agree well with the exerimental results. Thus, the ANSYS/LS-DYNA analysis latform can be used to redict the residual stress distribution induced by LSP. 4876

9 REFERENCES Ballard, P., 99a. Residual stressed induced by raid imact-alications of laser shocking. MA Thesis, Palaiseau, Ecole Poly Technique, France. Ballard, P., 99b. Residual stresses induced by raid imact-alications of laser shocking. Ph.D. Thesis, Ecole Poly Technique. Braisted, W. and R. Brockman, 999. High satial resolution: High energy synchrotron x-ray diffraction characterization of residual strains and stresses in laser shock eened inconel 78SPF alloy. Int. J. Fatigue, 2: 79. Ding, K. and L. Ye, Three-dimensional dynamic finite element analysis of multile laser shock eening rocesses. Surf. Eng., 9(5): Ding, K. and L. Ye, Simulation of multile laser shock eening of a 35CD4 steel alloy. J. Mater. Proc. Technol., 78(-3): Fabbro, R., P. Peyre, L. Berthe and X. Scherereel, 998. Physics and alications of laser-shock rocessing. J. Laser Al., 0(6): Han, B. and D.Y. Ju, Comressive residual stress induced by water cavitation eening: A finite element analysis. Mater. Design, 30(8): Hu, Y.X., Z.Q. Yao and J. Hu, D FEM simulation of laser shock rocessing. Surf. Coatings Technol., 20(3-4): Meguid, A.S., G. Shagal and J.C. Stranat, 999. Finite element modeling of shot-eening residual stresses. J. Mater. Proc. Technol., 92(93): Ocana, J.L., M. Morales, C. Moleceres, O. García, J.A. Porro and J.J. García-Ballesteros, Short ulse laser microforming of thin metal sheets for MEMS manufacturing. Al. Surf. Sci., 254(4): Peyre, P., A. Sollier, I. Chaieb, L. Berthe, E. Bartnicki, et al., Analysis of laser shock waves and resulting surface deformations in an Al-Cu-Li aluminum alloy. Eur. Phys. J. Al. Phys., 23: See, D.W., J.L. Dulaney, A.H. Clauer, et al., Laser eening without coating: Process, effects and alications. Surf. Eng. Al. Electrochem. Wang, L., 985. Stress Wave Theory. Beijing, Post and National Defense Industry Publisher, China. Zhang, W., Y.L. Yao and I.C. Noyan, Microscale laser shock eening of thin films, Part 2: High satial resolution material characterization. J. Manuf. Mater. Sci. Eng., 26: