Fault Diagnosis Of Chemical Process Based On Optimized SVM

With the development of science and technology,the complexity of the system is increasing.Fault diagnosis plays a more and more important role in the production and operation of the chemical process,which can effectively improve the reliability of the system,reduce the occurrence of accidents and the heavy losses caused by system failures.However,there are still problems in the diagnosis accuracy and speed to be improved.In this paper,a Support Vector Machine(SVM)fault diagnosis method for chemical process based on Fuzzy Rough Set(FRS),Grid Search(GS)and Whale Optimization Algorithm(WOA)is proposed.The specific work is as follows:Firstly,a fault diagnosis method based on Fuzzy Rough Set and Support Vector Machine is proposed to solve the problem that the inaccuracy of fault identification affects the speed of fault diagnosis.The Fuzzy Rough Set attribute reduction algorithm based on attribute dependence is used to select the fault features of the process data,and the minimum fault feature set is obtained,which makes the fault features more concise and accurate,and then the fault is classified by Support Vector Machine.The experiments of Tennessee-Eastman(TE)process show that the fault diagnosis method based on Fuzzy Rough Set and Support Vector Machine can reduce the diagnosis time while ensuring the accuracy of fault diagnosis.Secondly,a Support Vector Machine fault classification model based on Whale Optimization Algorithm is proposed,and the Grid Search method is added to improve the accuracy of fault diagnosis.The Grid Search method is used to roughly optimize the parameters of Support Vector Machine,and then the Whale Optimization Algorithm is used to accurately optimize the Support Vector Machine fault classification model,which can effectively reduce the possibility of the model falling into local optimization and improve the accuracy of fault classification.The comparative experiments show that the chemical process fault diagnosis model based on Grid Search and Whale Optimization Algorithm optimized SVM has higher diagnosis accuracy and can diagnose the faults in TE process quickly and effectively.Finally,a chemical process fault diagnosis platform based on improved Support Vector Machine and MATLAB App Designer is designed and developed,which solves the problem that it is not easy to operate the fault diagnosis model directly in the actual production process,facilitates the staff to quickly realize fault diagnosis,and intuitively reflects the accuracy and fault types of fault diagnosis.The function of the platform is explained and tested by TE process,and the effectiveness of the improved SVM fault diagnosis platform is verified by the fault diagnosis experiments of offshore artesian wells.