Fault diagnosis of chemical processes based on the SVM optimized by fuzzy rough sets and a whale optimization algorithm

Chemical processes are complex and it is not easy to establish corresponding accurate mathematical models. In order to solve the problem that it is difficult to ensure the accuracy and speed of fault diagnosis due to the large amount of fault data and lots of attributes, a fault diagnosis method for chemical processes based on the support vector machine (SVM) optimized by fuzzy rough sets (FRS) and a whale optimization algorithm (WOA) was proposed. By analyzing the historical data of the chemical process, the fault types can be identified. Firstly, the fuzzy rough sets was used to select the features of the discretized process data, and the minimum fault feature set was obtained by attribute reduction. Then, the whale optimization algorithm, a new meta heuristic algorithm, was used to optimize the parameters of the SVM, and a fault data classifier was constructed according to the global optimal fitness function. Finally, the data set after attribute reduction was input into the SVM fault classifier optimized by the WOA, which formed a FRS-WOA-SVM fault classifier, so as to realize the fault diagnosis of chemical processes, and the results were compared with those of fault classifiers optimized by the conventional genetic algorithm and optimization algorithm. The Tennessee Eastman (TE) process particle swarm optimization was used as an example. The results show that the method proposed has high accuracy and fast diagnosis speed, and can effectively diagnose the faults in chemical processes. © 2022, Editorial Office of Journal of Vibration and Shock. All right reserved.