Research On Fault Diagnosis Method Of Body-in-white Welding Robot Based On Semi-quantitative Information

In recent years,the stability and safety of complex electromechanical systems such as welding production lines,CNC machine tools,and rail passenger cars have become the core of concern of the whole society.The welding process of body-in-white is one of the four molding processes in automobile industry,and its normal operation has an extremely important impact on production costs and vehicle quality.Body-in-white welding robot is a key part of the welding process,precise and efficacious fault diagnosis is a key technical method to guarantee the security and reliability service of body-in-white welding manufacturing process.It can effectively debase the input of factory cost,enhance the quality of body-in-white,and make full use of limited resources.With the development of the economy and the progress of society,welding robots are also developing in the direction of integration,precision and high-end,and timely maintenance to ensure normal operation has become a huge challenge.Because the mechanical structure and electrical principle of the welding robot are extremely complex,once fault occurs,it will cause great losses.Therefore,it is urgent to improve the reliability and safety of welding robot and develop the fault diagnosis technology for body-in-white welding robot.With the support of the National Natural Science Foundation of China Emergency Management Project,the Key R&D Project of the Jilin Provincial Department of Science and Technology,and the Young and Middle-aged Scientific and Technological Innovation and Leading Talent Team Project of the Jilin Provincial Department of Science and Technology,this paper mainly analyzed and studied the existing problems in the fault diagnosis and performance degradation prediction of the body-in-white welding robot,and proposed to use the method based on semi-quantitative information to solve the existing problems.The main research contents are as follows:Firstly,the basic structure and working mechanism of body-in-white welding robot are analyzed.The whole welding robot is selected as the research target,and the fault mechanism and influence of the robot body and components of the three subsystems are analyzed in detail,which lays a foundation for the subsequent fault diagnosis of body-in-white welding robot.Secondly,aiming at the fault diagnosis problem of the body-in-white welding robot,due to its complex components and strong correlation,the Fault Tree Analysis(FTA)is used to decompose the welding robot layer by layer to reveal its fault mechanism.Aiming at the problems of low fault probability and difficult data acquisition of welding robot,the fault data set is determined by using the fault log data of a factory and software simulation fault probability,and the method of combining Fault Tree Analysis(FTA)and Belief Rule Base(BRB)is used to transform the analysis results of FTA into the BRB,so as to realize the fault diagnosis of welding robot.Aiming at the problem that the reference value of antecedent attribute is not set accurately in BRB model,the method of determining the reference value according to the fluctuation of input data and the reference level setting is proposed.Finally,for the body-in-white welding robot whose performance state degrades slowly,it is not enough to carry out "post-maintenance" after fault happen.If the "pre-maintenance" can be carried out when the fault does not occur or when the fault is represented,it is helpful to formulate the maintenance strategy and control the maintenance timing.To solve the above problems,a Hidden Belief Rule Base prediction method(HBRB)is proposed.Based on the performance degradation evaluation of the welding robot,this method regards performance as implicit behavior,and uses different operating conditions as the observable indicators of the HBRB model to describe its implicit behavior.This method can solve the performance degradation prediction problem under different time and operating conditions.Performance degradation prediction is a further supplement to fault diagnosis.Before a fault occurs,relying on performance degradation prediction technology to predict the occurrence of faults in advance is a new direction for the development of mechanical and electrical system maintenance and security in the future,and also provides effective means and methods for fault diagnosis of body-in-white welding robots.