• CN: 11-2187/TH
  • ISSN: 0577-6686

Journal of Mechanical Engineering ›› 2018, Vol. 54 ›› Issue (13): 179-191.doi: 10.3901/JME.2018.13.179

Previous Articles     Next Articles

Modeling and Analysis of Adaptive Weighted Variance Propagation Network in Hybrid Multistage Machining Processes

ZHENG Xiaoyun1, YU Jianbo1, LIU Haiqiang1, CHENG Hui2, SUN Xiwu2, WU Hao3   

  1. 1. School of Mechanical Engineering, Tongji University, Shanghai 201804;
    2. Shanghai Aerospace Equipment Manufacturing Factory, Shanghai 201100;
    3. Shandong Special Equipment Inspection Institute, Shandong 250101
  • Received:2017-08-08 Revised:2017-12-04 Online:2018-07-05 Published:2018-07-05

Abstract: For hybrid multistage machining processes (HMMPs), the final product quality is affected by the propagation and accumulation of variation from all machining stages. To ensure the machining processes stability, a dynamic real-time analysis of variation propagation and error source identification based on adaptive weighted variance propagation network (AWVPN) is proposed. The mechanism of variation propagation is used to construct the Network. Based on the AWVPN model and the PageRank ranking algorithm, an evaluation index named AWVPN-NodeRank for node importance and corresponding evaluation method is proposed to identify the key machining surface. To ensure the priority of the nodes, the method of level structuring is used to recognize the sorted error sources of key machining surface. Typical HMMPs of the main bearing cap is used to verify the effectiveness of the proposed method. The results show that these methods can effectively identify the weak process and the problematic machining equipment.

Key words: adaptive weighted network, error source diagnosis, key process, multiple variance flow, variation propagation

CLC Number: