翻邊接頭沖壓成形工藝與模具設(shè)計,翻邊接頭沖壓成形工藝與模具設(shè)計,接頭,沖壓,成形,工藝,模具設(shè)計 PROCESS ANALYSIS AND DESIGN IN STAMPING AND SHEET HYDROFORMING DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University By Ajay D. Yadav, M.S. * * * * * The Ohio State University 2008 Dissertation Committee: Professor Taylan Altan, Adviser Associate Professor Jerald Brevick Professor Gary L. Kinzel Approved by: - Adviser Industrial and Systems Engineering Graduate Program ii ABSTRACT This thesis presents initial attempts to simulate the sheet hydroforming process using Finite Element (FE) methods. Sheet hydroforming with punch (SHF-P) process offers great potential for low and medium volume production, especially for forming (a) lightweight materials such as Al- and Mg- alloys and (b) thin gage high strength steels (HSS). Sheet hydroforming has found limited applications and is thus still a relatively new forming process. Therefore, there is very little experience-based knowledge of process parameters (namely forming pressure, blank holder tonnage) and tool design in sheet hydroforming. For wide application of this technology, a design methodology to implement a robust SHF-P process needs to be developed. There is a need for a fundamental understanding of the influence of process and tool design variables on hydroformed part quality. This thesis addresses issues unique to sheet hydroforming technology, namely, (a) selection of forming (pot) pressure, (b) excessive sheet bulging and tearing at large forming pressures, and (c) methods to avoid leaking of pressurizing medium during forming. Through process simulation and collaborative efforts with an industrial sponsor, the influence of process and tool design variables on part quality in SHF-P of axisymmetric punch shapes (cylindrical and conical punch) is investigated. In stamping and sheet hydroforming, variation in incoming sheet coil properties is a common problem for stamping plants, especially with (a) newer light weight materials for automotive applications (aluminum-, magnesium- alloys) and (b) thin gage high strength steels. Even though incoming sheet coil may meet tensile test specifications, high scrap rate is often observed in production due to inconsistent material behavior. Thus, tensile test specifications may not be adequate to characterize sheet material behavior in production stamping/hydroforming iii operations. There is a strong need for a discriminating method for testing incoming sheet material formability. The sheet bulge test emulates biaxial deformation conditions commonly seen in production operations. This test is increasingly being applied by the European automotive industry, especially for obtaining reliable sheet material flow stress data that is essential for accurate process simulation. This thesis presents a new inverse-analysis methodology for calculating flow stress curves at room temperature, using the biaxial sheet bulge test. This approach overcomes limitations of previously used closed-form membrane theory equations and exhibits great potential for elevated temperature bulge test application. To verify the developed methodologies presented in this thesis, selected case studies are presented, to (a) demonstrate the successful application of finite element (FE) simulation in tool design, process sequence design and springback reduction in stamping and sheet hydroforming and, (b) validate the developed methodology for automation/standardization of tool and process sequence design procedure and recording of existing design guidelines in transfer die stamping. iv ACKNOWLEDGMENTS I am immensely grateful to my graduate adviser, Dr. Taylan Altan, for his guidance and encouragement during my doctoral studies at the Engineering Research Center for Net Shape Manufacturing (ERC/NSM). This dissertation work would not have been possible if not for his continuous support and enthusiasm for applied research. I sincerely appreciate the suggestions and comments of my candidacy and dissertation committee: Dr. Jerald Brevick, Dr. Gary L. Kinzel and Dr. Jose Castro. I am grateful to sponsors of the Center for Precision Forming (CPF) for partially supporting this research. I am also thankful to ERC/NSMs industrial sponsors for supplying us with an adequate number of metal forming problems. I acknowledge all students and visiting scholars of ERC/NSM and CPF (years 2001 - 2008) for their contributions. Special thanks to Dr. Hariharasudhan Palaniswamy, Mr. Lars Penter (Dresden), Mr. Parth Pathak, Mr. Dario Braga (Brescia) and Mr. Gianvito Gulisano for their co- operative efforts. Thanks are also due to Dr. Manas Shirgaokar, Dr. Ibrahim Al-Zkeri, Mr. Thomas Yelich, Mr. Giovanni Spampinato, Dr. Yingyot Aue-u-lan, Dr. Hyunjoong Cho, Dr. Hyunok Kim, Mr. Shrinidhi Chandrasekharan and Dr. Gracious Ngaile (during my Masters) for moral and intellectual support. I wish to thank all my friends for their encouragement, advice and support. Finally, I dedicate my graduate study efforts to my parents for their everlasting support, patience and personal sacrifice. v VITA Oct 7, 1979 Born Pune, India. Jul 1997 Jun 2001 B.E. Mechanical Engineering, All India Shri Shivaji Memorial Societys College of Engineering (AISSMS COE), University of Pune, Maharashtra, India. Aug 2001 May 2003 M.S. Industrial and Systems Engineering, The Ohio State University, Columbus, Ohio, USA. Aug 2001 Jun 2008 Graduate Research Associate, NSF I/UCRCs Center for Precision Forming (CPF) and Engineering Research Center for Net Shape Manufacturing (ERC/NSM). vi PUBLICATIONS Yadav A. D., Palaniswamy H., and Altan, T., (2008), Sheet Hydroforming: Room and Elevated Temperature Chapter for ASM Sheet Metal Forming, Editor: T. Altan (in progress). Yadav, A.D., Pathak P., Altan, T., (2008), Progressive and Transfer Die Stamping Chapter for ASM Sheet Metal Forming, Editor: T. Altan (in progress). Yadav, A. D., Kaya S., Altan, T., (2008), Servo Drive Presses Stamping Technology Chapter for ASM Sheet Metal Forming book, Editor: T. Altan (in progress). Yadav A. D., Penter, L., Pathak, P., Altan, T., (2008/2009), Flow stress determination with biaxial sheet bulge test using inverse analysis approach (Manuscript in preparation) for peer review journal. Yadav A. D., Palaniswamy H., and Altan, T., (February, March, April 2006), Three-part series on “Sheet Hydroforming”, Stamping Journal R Initial sample thickness for ERCs VPB bulge test = 1.3 mm Mirtsch 2006 . 65 xvi Figure 4.9: Chart showing the comparison of the thickness measured at dome apex with predictions from Excel macro. Calculated values of thickness are 4% lower than measured thickness values. . 66 Figure 4.10: Schematic shows springback error (exaggerated) in dome height measurement . 67 Figure 4.11: Comparison of flow stress curves, with and without springback correction . 69 Figure 5.1: Inverse analysis methodology to determine flow stress curve (for materials following the power law fit) using the dome geometry evolution from the biaxial bulge test. . 75 Figure 5.2: Two dimensional schematic of FE model in LS-DYNA v9.71 . 76 Figure 5.3: Schematic illustration of the objective function. (a) Changing dome height for five time increments (t = 1 through 5). Comparison of measured and simulated dome height at final time t = 5. . 78 Figure 5.4: (a) 2D view of design space describing the two variables in flow stress determination problem Design variable 1: K (in MPa), design variable 2: n- value (b) 3D view of design space, showing objective function (response) obtained for each K and n-value from simulation. The objective of this study is to find the minimum on the design surface created from response values. . 79 Figure 5.5: Schematic of a full factorial set of FE simulations for a given input range for K and n-value. For K = 400 to 500, a set of 11 values is chosen (namely, 400, 410, 420 500). Similarly, for n-value = 0.1 to 0.5, a set of 11 points is chosen. The number of points is user defined. Thus, for a full factorial set of simulations, 11 x xvii 11 = 121 simulations are planned by LS-OPT. This is computationally expensive. . 80 Figure 5.6: (a) Selective FE simulations using a D-optimal design (b) An example response polynomial surface (second order) obtained from the first 10 FE simulations conducted using the D-optimal design. . 82 Figure 5.7: Minimization of objective function E using RSM. In each iteration (set of 10 FE simulations), the margins for K and n-value shrink to a smaller design space, until objective function is minimized. . 83 Figure 5.8: Results for bulge tests for AA 5754-O (t = 1.01 mm) conducted at industrial sponsor: Convergence history of strength coefficient K at end of eight iterations, K converged to 432 MPa (b) Convergence history of the strain hardening exponent n-value. At the end of eight iterations, n-value converged to 0.3. . 84 Figure 5.9: Flow stress results using inverse analysis methodology at room temperature: Aluminum alloy AA5754-O, sheet thickness = 1.01 mm. Biaxial hydraulic bulge tests were conducted at industrial sponsor (seen in Figure 4.3). . 85 Figure 5.10: Flow stress results using inverse analysis methodology at room temperature: Aluminum alloy AA5754-O, sheet thickness = 1.3 mm. Biaxial bulge experiments were conducted at ERC/NSM using the VPB test tool geometry. . 86 Figure 5.11: Flow stress results using inverse analysis methodology at room temperature: Stainless steel SS 304, sheet thickness = 0.86 mm. Biaxial bulge experiments were conducted at ERC/NSM using the VPB test tool geometry. . 87 xviii Figure 5.12: (a) Elevated temperature hydraulic bulge test set up Erlangen, Germany (b) 3D and 2D view of a deformed Mg-alloy AZ31-O dome (measured by CCD camera) Hecht et al. 2005. . 88 Figure 5.13: Elevated temperature hydraulic bulge tests at Virginia Commonwealth University, USA Koc et al. 2007 . 90 Figure 5.14: (a) Schematic of the elevated temperature bulge test tools at University of Darmstadt, Germany (die opening = 115 mm, die corner radius = 4 mm) (b) Sample Mg-alloy specimen at 25C and 225C Kaya et al. 2008. . 91 Figure 5.15: Deviation of the dome surface from spherical shape (at elevated temperature) Kaya et al. 2008. . 92 Figure 5.16: (a) Definition of objective function (b) Proposed inverse analysis methodology for elevated temperature bulge test, to determine flow stress curve for materials following the power law fit = K () n () m using dome geometry evolution. . 96 Figure 5.17: Two dimensional schematic of the axisymmetric of the biaxial sheet bulge test for elevated temperature flow stress determination. Die diameter = 4 inches (101.8 mm), die corner radius = 0.25 inch (6.35 mm) . 98 Figure 5.18: Methodology for selecting lockbead dimensions for elevated temperature bulge test tools. . 99 Figure 5.19: (a) 2D schematic of the designed axisymmetric elevated temperature bulge test tool. (b) Effective strains in the clamped sheet (within lockbeads) are lesser than uniform elongation value for AA5754-O material. Thus, design is safe. . 101 xix Figure 5.20: Locating the three LVDTs for measuring the dome curvature. One LVDT will measure the dome apex, while the other two are designed for flexibility in measuring location. . 102 Figure 6.1: Schematic shows top view of example non-axisymmetric (rectangular) part. Due to asymmetry, the state of stress in the deforming sheet metal is different along the perimeter. . 107 Figure 6.2: Pot pressure P a is not sufficient to lift the sheet metal off the die corner at Section 1. Thus, insufficient pot pressure causes rubbing of sheet with the die corner leading to excessive thinning. . 107 Figure 6.3: Sheet metal flow in sections 1 and 2 (from Figure 1) for the pot pressure of P b (Note: P b P a ). An increased pot pressure P b ensures that sheet metal does not rub at the die corner at Section 1. However, this increased pot pressure P b causes the sheet metal to “bulge against drawing direction” at Section 2. . 108 Figure 6.4: Process limits on pot pressure are limited by sheet bulging in straight sections (for high pot pressure) and sheet rubbing at die corner (for low pot pressure). 109 Figure 6.5: Schematic of the 90 mm cylindrical cup tooling at Schnupp Hydraulik. 110 Figure 6.6: Photograph of the 800 kN (80 ton) hydraulic press used for experiments at Schnupp Hydraulik, Germany . 111 Figure 6.7: (a) Pot pressure and blank holder force (BHF) curve estimated by Schnupp Hydraulik through trial and error experiments. (b) Picture of 90 mm cup hydroformed cup using the estimated process parameters. . 113 Figure 6.8: (a) The optimum BHF curve was estimated using numerical techniques coupled with FE simulation. The “ERC/NSM pot pressure curve” was estimated by xx Contri, et al. 2004 through trial-and-error FE simulation (b) Part formed successfully, using ERC/NSM pot pressure curve and optimum BHF curve estimated by Braedel, et al. 2005 . 114 Figure 6.9: FE-Model (quarter geometry) in PAMSTAMP 2000 (AQUADRAW) for the 90 mm cylindrical cup SHF-P process. . 116 Figure 6.10: Flow stress of St1403 sheet material of thickness 1 mm estimated by Viscous Pressure Bulge (VPB) test at ERC/NSM. Uniaxial tensile tests were conducted to obtain the anisotropy coefficients (r 0 , r 45 and r 90 ). . 117 Figure 6.11: In investigating influence of punch-die clearance in SHF-P of cylindrical cups, the previously estimated optimum pot pressure curve (Max. value 400 bar) was too high. Thus, a reduced pot pressure curve (Max. value 180 bar) was used. . 119 Figure 6.12: Optimum blank holder force (BHF) curves for experiments 1 and 2 for pot pressure 180 bar (see Figure 6.11), for punch corner radius = 5 mm. . 120 Figure 6.13: Optimum blank holder force (BHF) curves for experiments 3 and 4 for pot pressure 180 bar (see Figure 6.11), for punch corner radius = 10 mm. . 120 Figure 6.14: Formed cylindrical cups for the four experimental cases. . 121 Figure 6.15: Comparison of (a) material draw-in and (b) thinning distribution (along rolling direction of sheet) between FE simulation and experiment, for Experiment 1. . 122 Figure 6.16: Comparison of (a) material draw-in and (b) thinning distribution (along rolling direction of sheet) between FE simulation and experiment, for Experiment no. 2. . 123 Figure 6.17: Comparison of (a) material draw-in and (b) thinning distribution (along sheet rolling direction) between FE simulation and experiment, for Experiment 3. . 124 xxi Figure 6.18: Comparison of (a) material draw-in and (b) thinning distribution (along rolling direction) between FE simulation and experiment, for Experiment 4. . 125 Figure 6.19: Sidewall wrinkles in the part were observed for punch-die clearance 5.5 mm (a) Formed part from Experiment no. 2 (b) Formed part from Experiment no. 4. . 128 Figure 6.20: Pot pressure curve (180 bar) was insufficient to form the cup completely against the punch. Therefore, pot pressure curve was modified by Schnupp Hydraulik and experiments were repeated. . 129 Figure 6.21: Blank holder force (BHF) curves by trial-and-error, for experiments 1 and 2 (punch corner radius = 5 mm) using Schnupp pot pressure curve (Figure 6.20). . 129 Figure 6.22: BHF curves obtained by trial-and-error, for experiment nos. 3 and 4 (punch corner radius = 10 mm) using Schnupp pot pressure curve (Figure 6.20). . 130 Figure 6.23: Four cups were formed with the modified pot pressure curve (one cup for each experimental case). . 130 Figure 6.24: Thinning distribution comparison (along sheet rolling direction) for Experiments. 3 and 4. . 131 Figure 6.25: Schematic shows three punch stroke locations using conical punch. Punch-die clearance changes with punch stroke. . 134 Figure 6.26: (a) Conical punch geometry selected for the study (b) Tool and blank dimensions for conical punch simulations/experiments . 135 Figure 6.27: FE model (quarter geometry) and input parameters used in FE simulations in PAMSTAMP 2000 (AQUADRAW) for conical cup SHF-P process. . 137 Figure 6.28: Pot pressure curves for conical punch FE simulations. . 138 xxii Figure 6.29: Sheet bulging observed for the selected conical punch geometry, at punch stroke 3.5 mm for Simulation no. 1. This bulge was observed for pot pressure = 20 bar and punch-die clearance 28 mm. . 139 Figure 6.30: Data points indicating bulging/no bulging for the planned FE simulation matrix. With increasing punch stroke (reducing punch-die clearance) higher values of pot pressure are needed to bulge the sheet against drawing direction. . 141 Figure 6.31: State of stress in example rectangular (non-symmetric) part geometry. . 143 Figure 6.32: Example non-symmetric punch geometry. This punch geometry will be used for (a) prediction of optimum pot pressure and BHF curve, (b) an investigation on the effect of punch-die clearance on sheet deformation. . 144 Figure 7.1: (a) Large reflector (b) Schematic of the reflector geometry (2D section view) . 146 Figure 7.2: 2D schematic of sequence of SHF-D forming operation for large reflectors. . 147 Figure 7.3: 3D finite element model (quarter geometry) in PAMSTAMP. . 149 Figure 7.4: Plastic strain ratio along rolling direction (0), diagonal direction (45) and transverse direction (90) used in FE simulation . 150 Figure 7.5: Thinning distribution predicted by FE simulation for anisotropic material (Inset: Schematic of the curvilinear length of deformed 50 inch reflector) . 152 Figure 7.6: Springback in the formed 50 inch reflector predicted by FE simulation for anisotropic sheet material (r 0 = 1.4, r 45 = 1.6 and r 90 =1.2) . 152 Figure 7.7: Schematic of the assumed tooling for FE simulations . 153 Figure 7.8: Comparison of Z displacement due to springback in the hydroformed part for the die geometry with different flange angle (for initial sheet thickness = 0.25 in) . 156 xxiii Figure 7.9: Comparison of radial displacement due to springback in the hydroformed part for the die geometry with different flange angle (for initial sheet thickness = 0.25 in) . 156 Figure 7.10: Comparison of springback predictions in 12 m reflector for different interface friction conditions, for flange angle 60 (for initial sheet thickness = 0.25 in). . 157 Figure 7.11: Cross sectional view of a typical formed axisymmetric part. . 162 Figure 7.12: Methodology for automating the process sequence and tool design. . 164 Figure 7.13: Steps followed in tool and process sequence design . 165 Figure 7.14: User interface to accept final part dimensions (in inches). . 167 Figure 7.15: Surface area calculation of the final part. Source: Progressive Dies Certificate program, Society of Manufacturing Engineers, at Dearborn, MI 2004. . 168 Figure 7.16: Deliverable of developed methodology: Process sequence for an axisymmetric part (with 4 drawing stages). . 170 Figure 7.17: Deliverable of developed methodology: Tool design and drawings, generated using the developed Visual Basic methodology. . 172 Figure 7.18: (a) Stainless steel case for a flow measuring device. Sketches of conceptualized process sequences to form possible final part geometries, namely (b) donut- shaped part (c) half donut-shaped part. The donut-shaped part was selected for process/tool design using commercial finite element (FE) code. . 174 Figure 7.19: Material properties (flow stress data) for SS304 and SS304L Source: Kalpakjian, S., Manufacturing Processes for Engineering Materials, Addison Wesley 1991 . 175 xxiv Figure 7.20: Finite element (FE) model. (a) Two dimensional view of the 1st stage forming process, inner curvature forming. (b) Two dimensional view of the 2nd stage forming process outer curvature forming. . 176 Figure 7.21: (a) Quarter geometry view showing thinning distribution in formed part, maximum value 30% (b) Thinning distribution along curvilinear length of the part. . 178 Figure 7.22: Quarter geometry view showing “springback” in the formed part. Maximum springback predicted is 2.5 mm (0.1 inches) (a) Back view (b) Front view . 178 Figure