Reliability Index Prediction Method Of Industrial Robot Under Epistemic Uncertainty

In our country,industrial robot industry is already in a stage of rapid development,and is gradually becoming in line with international standards in the realization of digitalization,intelligence,and intelligent manufacturing industrialization.With the continuous development of the science and technology and the implementation of “Made in China2025”,Chinese industrial robots have a higher market share and have huge market potential.In order to shorten the gap between Chinese industrial robots and foreign countries,and respond to the urgent needs in the market,it is imperative to strengthen the research and innovation of domestic industrial robots in terms of reliability.The reliability index prediction method of the industrial robot is studied in this thesis.(1)This thesis elaborates and analyes the composition and function of the industrial robot system,and then divides the subsystem according to its own characteristics,selects the reliability endexes,and constructs the reliability model of the robot system and draws the reliability block diagram.thus lays foundation for reliability analysis and reliability index prediction.(2)The industrial robot subsystem is used as the target for reliability analysis,that is FMECA(Failure Mode Effects and Criticality Analysis).According to the collected data,found the main failure mode of the system,and determine the cause of the failure mode and the impact of the failure mode.Then,put forward the detection methods and the establish measures for improvement.At the same time,based on the influence of epistemic uncertainty factors,this thesis uses the fuzzy comprehensive evaluation method to analyze the criticality of the system,and focuse on the reliability level of key components,thus provides a reference for improving the reliability of the whole machine.(3)Development of the reliability data analysis of industrial robots.Collecting the reliability data related to industrial robots,and judging the appropriate distribution model according to the distribution characteristics of the data.Then,estimating the parameters of established models and using K-S to test the goodness of fit,determining the distribution model of the industrial robots' fault data,and obtain the reliability index parameters.This method provides an important reference for the following reliability index prediction.(4)Development of predicting the reliability index of the industrial robot subsystem.Different prediction methods are adopted for different subsystems.Because of the characteristics of industrial robot subsystem,it can be used the component counting method to predict the reliability index,and the robot subsystem and the drive subsystem can be used similar product method to predict the reliability index.Aiming at the problems caused by epistemic uncertainty in similar product method,we use the interval analytic hierarchy process to quantify the influencing factors of subsystems,and built the comprehensive evaluation hierarchy model for reliability correction factor.And then,combined with the reliability indicators of similar products,the reliability indicators of the subsystems are predicted.The method can fully consider the fault data information of similar products,and can combine the rich experience accumulated by experts to effectively solve the epistemic uncertainty of the reliability index prediction process.It can determine the weight of each influencing factor,and calculate the reliability correction factor of subsystem.Baesd the reliability index of similar products,the reliability index of the subsystem will be predicted.