In order to ensure the safe operation and high-quality production of chemical systems,it is particularly important to accurately identify the faults of chemical processes.In order to solve the prob-lems of Tennessee Eastman(TE)process fault indistinguishability,neural network and other methods in fault diagnosis,such as low classification accuracy and unstable classification,a TE process fault diagno-sis model with optimized and improved extreme learning machine(ELM)was proposed.Firstly,the ker-nel principal components analysis(KPCA)method was used to reduce the dimensionality and extract fea-tures of the TE process data,then the improved butterfly optimization algorithm(IBOA)was used to opti-mize the weights and thresholds of the ELM,and finally the adaptive boosting algorithm integrates the classifier to complete the fault classification.The simulation results show that IBOA has better optimiza-tion ability than other optimization algorithms,and the improvement effect is significant,and the Ada-Boost-IBOA-ELM model can accurately classify different faults in the test set,and the final classifica-tion accuracy is as high as 98.5％,which further verifies the rationality and superiority of the model by comparing with other networks.
维普
万方数据