RESEARCH ON ROLLING BEARING FAULT DIAGNOSIS BASED ON IMPROVED HHO-LSTM
SHAO LiangShan
ZHU SiJia
Abstract：In order to solve the problem of rolling bearing fault diagnosis,an intelligent diagnosis model IHHO-LSTM was proposed,which combined the improved Harris hawks optimization(HHO)algorithm with long short-term memory(LSTM)network.HHO algorithm was prone to fall into local optimum and slow convergence in the solution process.Based on these problems,Cauchy distribution function and simulated annealing(SA)algorithm were introduced to expand the universality of global search and avoid falling into local optimization.The improved HHO was used to quickly determine the optimal super parameter values of LSTM model,so as to improve the accuracy of time series diagnosis.The rolling bearing experimental data of Case Western Reserve University were used for fault diagnosis experiments.The results show that IHHO-LSTM model can realize the feature extraction and fault diagnosis of rolling bearing,and the accuracy of the model is nearly 97%.
Keywords：Rolling bearingDeep learningHarris hawks optimization algorithmLong short-term memoryIndustrial big dataFault diagnosis
Publication Date：2024-02-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：7( 17-23 )