Abstract:Fault diagnosis has the very important role in the process of industrial production, especially for high molecular distillation, small fault will cause its purification rate, therefore a algorithm based on improved whales to optimize the fault classification method of support vector machine (IWOA - SVM)is proposed in this paper, and the reverse learning strategies and logarithmic weighting factor are added into the ordinary whale algorithm. Firstly, the reverse learning strategy (OBL) is used to replace the random initial population, and the reverse population is selected by the reverse learning strategy. On the one hand, the global search ability of the swarm intelligence algorithm can be effectively improved by OBL, and on the other hand, the diversity of the whale algorithm in repeated iteration can be refined, and make it jump out of the local optimal solution. Then, the adaptive weight factor is introduced and added to the whale optimization algorithm, and the global search ability is greatly enhanced by the dynamic change of the weight factor. Finally, the improved whale algorithm was used to optimize the parameters of SVM, and the optimized support vector machine was used to diagnosis and identify the fault data obtained in the process of scraping evaporation, and the results of IWOA-SVM were compared with WOA-SVM, SVM, PSO-SVM and GWO-SVM. The results show that the classification accuracy of IWOA-SVM proposed in this paper increases by 2%, and its accuracy remains above 98%. It is concluded that IWOA-SVM is better than other algorithms in terms of the accuracy of classification results and the robustness of the algorithm.