Improvement And Programming On Contact Searching Algorithm Of Sheet Forming Simulation

With the advent of increasingly stringent requirements on lightweight car body, crash safety and machining accuracy in automobile manufacturing, the complexity of sheet forming process is increasing, meanwhile, reducing the demands of automotive design cycle time and development costs is also becoming more and more intense, which puts forward severe challenges to the simulation of sheet stamping. Different approach models and algorithms of contact problem are the key factors to the sheet forming simulation effects. Developing more reliable, accurate and efficient algorithms is the consistent pursuit for researchers who devote themselves on sheet forming simulation contact problems. Considering the three goals above, this paper carries out study of algorithm optimization, contact model improvement and modularized programming on contact module of KMAS, a sheet stamping simulation software.Firstly, Memory consumption is improved based on the original contact search model, including memory consumption improvement by using2D grid correlation method, memory consumption improvement by using offsetting method. A new extended territory geometry model is proposed with time and memory consumption both improved, and its algebraic formulation is given. Memory consumption is improved by eliminating redundancy for the correlation information list in contact search. Furthermore, we have probed into the extension value ranges of extend territory, and get a local memory consumption optimal value through theoretical derivation and simulation test verification.Secondly, we put forward a new sheet forming simulation contact search model and search algorithm based on the contact domain and algebraic description of its boundary, which effectively overcomes the flaw of nonuniqueness of a contact pair that between a sheet node and a mold element in the original contact judgement algorithm. The algebraic equation of contact domain model is deduced in detail. Global search correlation norm and search process of new algorithm is constructed. Furthermore the new algorithm has been applied to KMAS software, through testing and comparing some typical examples, the improved results from the new algorithm demonstrated its effectiveness, also showed improvement both in solution time and memory consumption. Finally, a framework refactor is launched on KMAS contact searching module based on the modular programming and object-oriented concept. Throughout the refactor, the program is transferred from FORTRAN language to C language, which is with predominant development environment and superior software integration. Besides, we have reshaped the data structure and interface of the contact searching programs. In the refectored software framework, programming achieves the new algorithm and improvement work. Refactored contact search program, is more extendable and maintainable, and through optimization in the new Microsoft Visual Studio platform, based on OpenMp parallel computing method, it greatly reduces contact search time, significantly promotes the efficiency of stamping simulation.