2002 · 13 citations · 0 references
EngineeringManufacturing EngineeringMechanical EngineeringComputer-aided DesignMechanical DesignSimulation-based DesignDesign For AssemblySheet Metal FabricationSolid Mechanics ApplicationsStamping ProcessesProcess ParametersProcess Parameters OptimizationSheet Metal FormingSheet Metal StampingSpringbackSpringback CompensationStampingTool Geometry
The purpose of this study is to develop a methodology for springback compensation in sheet metal stamping operations. An optimization method is employed to minimize the difference between the simulation results and the intended design. This procedure results in an optimized die shape. LS-DYNA, LS-OPT and TrueGrid are used to input original tool geometry, material, and process parameters, identify design variables, perform springback simulations, and output optimized tool geometry. It is found that springback trends are consistent with changes in the die shape, which provides an effective strategy for springback compensation. The standard NUMISHEET’96 S-Rail is used as a benchmark example in this study. Metal Forming Technology 7 International LS-DYNA Users Conference 15-2 Introduction Springback is an elastic deformation which occurs at the end of a sheet metal stamping process, as the stamped part is removed from the stamping tools. Springback has the effect of changing the part’s finished shape so that it no longer matches the tools. If this shape deviation is large, it can cause difficulty during a subsequent assembly process, or cause twisting in the assembled part. Accordingly, it is important to produce parts whose finished shape closely match the designed surface. Usually corrections to compensate for springback are made by modifying the shape of the stamping tools. The design of these modifications, or die compensation, is a very complex process. Two commonly used methods are the trial-and-error and spring-forward methods. The trial-and-error method predicts die modifications based on engineering experience. Usually many years of dieshop experience are necessary before an engineer can successfully guess how to change the dies. The trial-and-error method is also very time consuming: to make a modified die set usually takes months of time. In addition, several trial-and-error corrections are frequently required before adequately compensated parts are obtained. Accordingly, the trial-and-error process is very expensive, often requiring over one million dollars to make a die that produces “good” parts. When new materials are used or when a new design is adopted, previous experience cannot be applied directly. The difficulties associated with the trial-and-error method can result in costs and lead-times that are out of control. Computer simulation has gained popularity in the stamping industry due to its speed and low cost, and it has been proven to be effective in prediction of formability and springback behavior. However, to date, no effective simulation method has been found to compensate the die based on the predicted springback. The spring-forward method is based on numerical simulation by Finite Element Analysis (FEA). The method begins by performing a stamping simulation, from which information for the deformed part is obtained while it is still positioned in the closed dies. This information includes the geometry, and material stress and strain data. The method then assumes that subsequent springback deformation will be driven by material stress, and that if the stress distribution through the material thickness is reversed, the resulting springback deformation will also be in the reversed direction, as compared to the actual part. Based on this logic, the geometry which is obtained by springback analysis with reversed stress can be used to predict modifications to the dies. This method is very simple to apply, and it is most popular numerical method. However, the method suffers from two major shortcomings that prohibit use in many practical applications. These are under-cut (interference between the parts during the stamping process) and accuracy. The purpose of this study is to investigate an effective algorithm for springback compensation. TrueGrid is used as a pre-processor to parametrically define geometry of the rigid tools, LSDYNA software (LSTC, 2001) is used as a Finite Element based solver and LS-OPT (LSTC, 1999, Stander & Craig, 2001) is used to guide the solution to an optimum. The NUMISHEET’96 S-RAIL example is used as a benchmark. 1 Registered Trademark of XYZ Scientific Applications Inc., Livermore, CA 7 International LS-DYNA Users Conference Metal Forming Technology 15-3 Approach The die compensation analysis presented here uses an optimization method. Many stamping/springback simulations are performed during an iterative process. Using the parametric preprocessor TrueGrid, the user can determine a set of design variables. These may include geometric variables, such as critical locations where tool elevation should be modified, important tool radii, and starting blank geometry. Process variables may also be specified, such as drawbead restraining forces and binder and pad loads. Simulation cost will increase linearly with the number of design variables. A set of constraints may also be specified for each design variable. These are used to limit the range of design variations to reasonable values. For example, tool radii could be limited to be no less than two millimeters, and elevation changes could be limited not to exceed five millimeters. In addition, strength criteria such as maximum thickness reduction of the sheet metal or FLD criteria can be specified. Using this design variable information, LS-OPT will automatically create a family of stamping/springback models (according to an experimental design method), and submit several simultaneous simulation jobs. These jobs may run in parallel or be queued on one or more computers, and since they execute independently, will exhibit near perfect parallel efficiency. LS-OPT will also automatically monitor job progress, interrogate results, and catalog results data in an organized way. After collecting and processing results from the first set of simulations, LS-OPT will predict optimized values for each design variable. This prediction will be made using a response surface method. Using these optimized design variables, the next set of simulations (experimental design) will be automatically created and submitted. This iterative process will continue until each variable has been determined within a specified tolerance, or until a limiting number of iterations have been completed. Various options for the optimization objective can be made available, e.g. RMS or maximum discrepancy with respect to a desired geometry. During the optimization process, interactive software will be available to view all results obtained to date and monitor the evolution of each design variable. Example: Numisheet 96 S-Rail The tools and sheet-metal blank of the Numisheet 96 springback benchmark problem are shown in Figure 1. The punch is controlled at a constant 1m/s while the binder is driven by a piece-wise linear force curve as shown in Fig. 2. Metal Forming Technology 7 International LS-DYNA Users Conference 15-4 Fig. 1: Numisheet 96 benchmark: Punch, die, binders and blank (baseline design) Fig. 2: Binder Force as a function of time 0 10 38 Time (ms) Force (kN)