Research On Data Driven Fault Diagnosis And Fault Tolerant Control Integrated Design Method

From the beginning of human use tools,it is now possible to manufacture largescale equipment to help people to finish the work which the human beings can not accomplish by themselves.With the advancement of research in various basic disciplines,the cross-integration of various disciplines,people's manufacturing capabilities continue to improve.At the same time,with the development of social economy,people have a higher pursuit,and have higher requirements for product quality.In order to be able to accomplish the work with higher precision and to ensure the quality of products,the equipment or system structure in industrial production is more and more complicated,and in order to improve production efficiency,the scale of industrial systems is also growing.Once the system fails,it will cause serious accidents,resulting in economic losses and even casualties.Therefore,it is necessary to take certain fault diagnosis and fault-tolerant control measures for industrial systems to ensure the reliability and safety of industrial systems.When the fault diagnosis is first proposed,it is necessary to have a very good understanding of the process mechanism of the system,and then establish a model of the system based on the working mechanism.In the case where a precise mechanism model can be established,the fault diagnosis has a very good effect.However,as mentioned above,industrial systems are becoming larger and larger,and structures are becoming more and more complex.Therefore,it is more and more difficult to establish accurate mechanism models,and process data in industrial systems becomes very abundant due to the large number of applications of sensors.The process data of these industrial systems already contain the operating characteristics and laws of industrial systems.The appropriate rules can be used to analyze the system rules.This topic firstly studies the coprime factorization technology and Youla parameterization theory.Based on this theory,it builds the integrated method of fault diagnosis and fault-tolerant control of industrial systems without changing the existing controller of the system.Next,the data-driven fault diagnosis module and fault-tolerant control module in the integrated method are implemented.The core of fault diagnosis is the identification of stable kernel representation of the system.Firstly,the study and analysis of the stable kernel representation of the system are carried out.The identification method of the system stable kernel representation of the closed-loop system is used to calculate the residual signal online,and then the statistical method is used to make the decision.In the fault-tolerant control part,faulttolerant control is mainly implemented by adding the Youla parameter matrix.At the end of the paper,this paper introduces the three-tank system,and uses its simulation model and offline historical data to verify the effectiveness of the designed data-driven integrated method for fault diagnosis and fault-tolerant control.