Improved butterfly algorithm optimizing ELM network parameters and its application in tennessee-eastman process

accurate diagnosis and identification of chemical system faults is of great significance to ensure the long-term safe operation and high-quality production of the system. Aiming at the problems of low fitting degree and parameters easily falling into local optimum in ELM model prediction, an improved butterfly algorithm (FBOA) is proposed to optimize ELM network parameters (FBOA-ELM) to improve its performance. Then the fault diagnosis of standard Tennessee Eastman (TE) process is carried out. The results show that the FBOA-ELM method can effectively classify the 16 kinds of faults selected in TE process. Compared with the other four algorithms, the accuracy is significantly higher than the four algorithms, which indicates that the FBOA-ELM model has better classification effect and higher accuracy.