Research On Fault Diagnosis Of Control System Based On Data Driven

With the development of science and technology,the modern control system is more and more large and complicated.In order to improve the safety and reliability of the control system,we need to monitor the status of the system,discover the fault information system in time,and take corresponding measures to prevent catastrophic accidents.The complex control system is very difficult to get the accurate mathematical model,which leads to the traditional fault diagnosis method based on model in application has great limitations.Because of the rapid development of computer technology,the control system can obtain more and more information,the fault diagnosis method based on data driving is in this situation came into being.This study takes the control system as study object,and uses the method based on data driven for the fault diagnosis.First of all,we need to detect fault of control system,to determine whether the occurrence of fault.To solve this problem,a fault detection method based on feature extraction and clustering is proposed.This method uses slow feature analysis to extract the feature from normal data and fault data of control system.The fuzzy C-means clustering algorithm is used to obtain the clustering center of normal data,and the clustering radius is taken as the threshold.The difference model of the data to be detected and the clustering center is constructed to judge whether the data of the control system deviates from the normal state.Secondly,the fault library of the control system is constructed and the fault types are classified.Because the closed-loop control law is introduced in the whole system,the fault can be propagated in the system,and the fault data can be generated by different faults.Therefore,the fault feature extracted is combined with the fault characteristic of the control system to improve the distinguishing characteristic of the fault.The adaptive clustering feature of subtractive clustering is used to judge whether the current fault is the fault type trained in the faulty library.And then the fault data to be detected for pattern matching to determine the type of failure.Finally,the above method is applied to the superheated steam temperature control system of the power plant to verify the method.