Sub-assembly recognition algorithm and performance analysis in assembly sequence planning

Abstract

With the highly intelligent development of the manufacturing industry, assembly problems have become one of the reasons for the high cost of products. One of the main challenges in the automatic planning of assembly sequences is the identification of sub-assembly. The sub-assembly identification can effectively realize the parallel development and planning of the assembly. In the paper, the focus is concentrated on a sub-assembly recognition algorithm. It is used to recognize the sub-assemblies in various assemblies in order to simplify assembly sequence and decrease assembly difficulty and costs. Firstly, the principle of sub-assembly recognition algorithm is clarified, and the structural information model of the assembly is defined by the weighted undirected connection graph. It is then converted into an adjacency incidence matrix to judge the membership relations between parts in the assembly. Second, mathematical modeling is performed according to the process and principle of sub-assembly recognition. Thirdly, the sub-assembly recognition algorithm is programmed in Matlab, and the feasibility of the algorithm is verified by an example. Finally, the sub-assembly center initialization, the quantity of sub-assemblies c, and the fuzzy index m among the important factors in the algorithm are selected to analyze the parameters and optimize the parameters.