Online Quality Monitoring And Diagnosis For Complex Mechanical Product Remanufacturing Assembly Process

The quality of remanufactured product is not only the premise effect for its marketization, but also one of the major restriction factors for the industrialization and large-scale development of remanufacturing. Complex mechanical remanufactured product is a hybrid structure assembly which consists of remanufactured parts, reused parts and new parts that with complicated structure or multiple quality characteristic, according to specific technology and assembly process in order to meet certain technical specifications. Assembly process is the final link during whole remanufacturing process of complex mechanical remanufactured product. Quality control in reassembly (remanufacturing assembly) process is a foundation to guarantee the quality of remanufactured product and enhance the competitiveness of remanufactured enterprises.Take the quality monitoring and diagnosis of assembly process for complex mechanical remanufactured product as an object of study, from system’s perspective, based on the characteristics of reassembly process, adopt theories and technologies such as control chart, Bayesian theory, artificial intelligence technology et al, an on-line quality monitoring and diagnosis method for complex mechanical remanufactured product reassembly process is put forward in this paper.Firstly, through analyzing the characteristics of reassembly process, the limitation of conventional control chart in reassembly process is described in this paper. Using non-informative prior distribution, based on the Bayesian, process capability index and quality control chart model is constructed and verified to realize on-line monitoring for fluctuation of reassembly quality.Secondly, under analyzing the abnormal pattern of reassembly process quality control which based on the Bayesian estimation, the abnormal pattern of control chart is turned into computer recognized form. And an on-line identification model of abnormal pattern for control chart is established. Then, via BP neural network model and algorithm, achieve fast recognition of reassembly process quality abnormal.Thirdly, aim at the uncertainty and fuzzification existing in the complex mechanical remanufactured product reassembly process, this paper establishes mapping relation model between quality abnormal and abnormal control chart patterns based on fuzzy theory. Through using Petri net, we can describe production mechanism and transmission mechanism of abnormal graphically. Then, based on fuzzy production rules, we can achieve on-line diagnostics for reassembly process quality abnormal. Finally, take remanufacturing enterprise as background, combine with related enabling technology such as RFID, machine vision, industrial Ethernet et al, an on-line quality monitoring and diagnosis system for remanufactured engine assembly process is developed to provide technology support in enhancing reassembly quality stability.