Robustness Evaluation And Tolerance Prediction For A Stamping Process With Springback Calculation By The FEM
- Universita di Cassino, Dip. Ingegneria Industriale (Italy)
The FEM simulation of a sheet forming process can be seen as a method for calculating a deterministic vector set of output responses, starting from a vector set of input variables. In the real world, though, most process parameters should be considered as random variables (material properties, lubrication conditions, etc.), with a given statistical distribution. As a consequence, any FEM simulation which is run with nominal values of the input variables does not provide the 'true' solution, but only a nominal response, related to a distribution around an expected value. The robustness of a given solution can be evaluated by several methods: sensitivity analysis, Monte Carlo method, Response Surface Method (RSM), etc.The sensitivity analysis is usually performed by running one or two simulations for each random variable, setting it at an extreme value. The method does provide a rough measure of the process variability, but it does not usually guarantee a quantitative estimation of the response's confidence interval. Statistical techniques, such as Monte Carlo or RSM are very effective in predicting the true process variation. However, both methods are computationally very expensive, if the problem is complex (e.g., when forming, trimming and springback operations must be simulated for a large part).In the present paper the solution's robustness of a Numisheet '05 benchmark case is faced and solved with an approximate and computationally inexpensive method. The purpose of the study is to determine the tolerance interval of the part surface, induced by the random variation of 11 process parameters. The variability is modeled with a joint multinormal distribution. The analyzed response is the geometrical error of the process with respect to a reference 'deterministic' geometry. The proposed method provides an upper bound estimation of the tolerance interval, by exploring the boundaries of the ellipsoidal probability density function of the multinormal input. Finally, the effectiveness of the method is tested by running a conventional Monte Carlo analysis of the problem.
- OSTI ID:
- 20726135
- Journal Information:
- AIP Conference Proceedings, Vol. 778, Issue 1; Conference: NUMISHEET 2005: 6. international conference and workshop on numerical simulation of 3D sheet metal forming process, Detroit, MI (United States), 15-19 Aug 2005; Other Information: DOI: 10.1063/1.2011231; (c) 2005 American Institute of Physics; Country of input: International Atomic Energy Agency (IAEA); ISSN 0094-243X
- Country of Publication:
- United States
- Language:
- English
Similar Records
Nuclear Data Uncertainty Quantification for Adjoint-Weighted Quantities via XGPT in Monte Carlo Codes
Kinematic Hardening: Characterization, Modeling and Impact on Springback Prediction