A Novel Fault Diagnosis Method for TE Process Based on Optimal Extreme Learning Machine

Chemical processes usually exhibit complex, high-dimensional and non-Gaussian characteristics, the diagnosis of faults in chemical is particularly important. To address this problem, paper proposes a novel fault method based on Bernoulli shift coyote optimization algorithm (BCOA) to optimize kernel extreme learning machine classifier (KELM). Firstly, random forest treebagger (RFtb) used select features, data set optimized. Secondly, new BCOA proposed automatically adjust network hyperparameters KELM improve performance. Finally, optimized feature sequence input into obtain final results. The Tennessee Eastman (TE) process have been collected verify effectiveness method. A comprehensive comparison analysis with widely algorithms also performed. results demonstrate that outperforms other methods terms classification accuracy. average rate 21 found be 89.32%.