Variation propagation network-based modeling and error tracing in mechanical assembling process

To ensure the quality of mechanical products and the assembling process, it is necessary to model the variation propagation flow of the assembly process, identify the key assembly characteristics and control the corresponding error assembling nodes and the error source. A method of modeling and error tracing based on the complex network was proposed. The method was used to construct a self-regulated weighted variation propagation network, taking into account the measured data, the information of characteristic surfaces and the assembly technology in the assembly process. The improved weighted semi-local centrality sorting algorithm was used to identify the key characteristics of the constructed variation propagation network. The backtracking algorithm and the importance rank (IR) index were proposed to identify the error source of the key characteristics in the constructed self-regulated weighted variation propagation network, after which the assembly surfaces which need to be monitored could be distinguished. With the multistage assembly process of a gear shaft as a study case, the proposed method was verified. The method can be used to effectively model the variation flow, as well as identify the key assembly surface and the error source in the multistage assembly process. © 2019, Zhejiang University Press. All right reserved.