Fuzzy Reliability Modeling of Mechanical Meta-action Assembly Accuracy. (August 2020)

• Record Type: Journal Article
• Authors: Ge, Hongyu; Liu, Baoqiang; Zhang, Chuanwei; Yang, Manzhi
• Is Part Of: Journal of physics. Volume 1601:Number 6(2020)
• Publication Date: 2020-08
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1601/6/062017
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: Meta-action assembly units are the basic units to ensure the reliability of mechanical function. For the problem that the reliability of meta-action assembly units are difficult to be finely modeled. Combined with the fuzziness of assembly process, this paper introduces fuzzy theory into the reliability analysis of meta-action assembly accuracy.By analyzing the structure of meta-action assembly units, it is concluded that meta-action assembly units are composed of support member, power source, transmission fastener and actuator.Then, an unified structural model of meta-action assembly units is established; according to the mechanism of assembly accuracy, the accuracy information of a meta-action assembly unit is mapped to each assembly item (assembly item is consisted of individual part feature and corresponding part feature). Combined with the fuzzy feature of assembly information, the concepts of assembly accuracy reliability and assembly accuracy fuzzy reliability are proposed. By analyzing the fuzzy characteristics of assembly accuracy of assembly items, membership function of assembly accuracy is determined. According to probability definition of the fuzzy events, a fuzzy reliability model of the meta-action assembly accuracy is established. In view of whether the assembly project of scientific research enterprise has faults, a large number of historical data are analyzed, and the overall 0-1 distribution of observation data is obtained. Precision test of fuzzy reliability model is got on by maximum likelihood estimation of the 0-1 distribution. The example shows that the reliability value obtained by using the model is different from the test result by 0.0029. The results indicate that the method has greater feasibility, simplicity and flexibility.