Abstraction of Semantic Mid-Surface Based on Rib-Feature Recognition

Abstract

Mid-surface abstraction is an effective simplification method for thin-wall models. The complexity of finite element analysis (FEA) for a mid-surface model can be reduced greatly after abstraction. Although the model decomposition method is adopted for mid-surface extraction, it is hard to obtain the correct mid-surface model for complex models since the existing heuristic rule based methods lack of design intention. In addition, the mid-surface model is not easy to reuse. In this study, a semantic based mid-surface model representation and generation method is proposed. Firstly, a hierarchical semantic mid-surface model based on rib-feature decomposition is proposed. Secondly, based on the reorganization of rib-features and decomposition of the thin-wall model, the rib-features’ semantic information are obtained by the abstraction of the structure and connection in the thin-wall model. Then the hierarchical structure is generated by connection semantics. According to the various structure semantics, different abstraction methods will be employed to get the mid-surface patch for each sub region. Finally, the hierarchical semantic mid-surface model is constructed by the generation of the connection relationship between mid-surface patches based on the connection semantics between the rib-features. This semantic model ensures the high efficiency and accuracy of mid-surface regeneration when local modifications occur to a thin-wall model. A typical example is given to demonstrate the process.