基于BP神经网络对复杂机电系统可靠性的评估  

Reliability Evaluation of Complex Electromechanical System Based on BP Neural Network

作　　者：阚延鹏 陈玉[1] 刘永明 韩波 KAN Yanpeng;CHEN Yu;LIU Yongming;HAN Bo(School of Mechanical Engineering,Anhui Polytechnic University,Wuhu 241000,China)

机构地区：[1]安徽工程大学机械工程学院,安徽芜湖241000

出　　处：《安徽工程大学学报》2021年第2期34-40,共7页Journal of Anhui Polytechnic University

基　　金：安徽省自然科学基金资助项目(1808085ME127);安徽工程大学引进人才科研启动基金资助项目(2019YQQ004)。

摘　　要：针对传统的复杂机电系统可靠性的评估模型,基于统计理论建立近似模型,但模型误差较大,采用BP神经网络的方法建立工业机器人的复杂系统的可靠性评估模型。其中,电子元器件和机械件寿命分别服从指数分布和威布尔分布。然后根据每个件的失效率计算各件在不同工作时间下的可靠度和系统整体可靠度。数值仿真结果显示,BP可靠性评估模型与基本可靠性模型在工作时间下得到的可靠度值误差很小,验证了模型的准确性、可行性,同时也预测了该系统的可靠度。研究表明,采用该方法能实现由组成部件的可靠度预测整体的可靠度,提高了可靠性评估的精度,解决了基于统计理论分布模型解析表达式难于获得的问题,降低了模型复杂度。Aiming at the traditional reliability evaluation model of complex electromechanical systems based on statistical theory,the approximate model was established,but the model error was large.The BP neural network method is used to establish the reliability evaluation model of the complex system of industrial robots.Among them,the life of electronic components and mechanical parts are subject to exponential distribution and Weibull distribution,respectively.Secondly,the reliability of each part under different working hours and the overall reliability of the system are calculated according to the failure rate of each part.The results of numerical simulation show that the reliability errors obtained in the working time for the BP reliability assessment model and the basic reliability model are very small.The accuracy and feasibility of the model are verified,and the reliability of the system is predicted.It is shown that this method can achieve the overall reliability based on the reliability of the components,improve the accuracy of reliability assessment,and reduce the model complexity.

关 键 词：工业机器人 可靠性 BP神经网络 机电系统 

分 类 号：N945.17[自然科学总论—系统科学] TP242.2[自动化与计算机技术—检测技术与自动化装置]