Research On Fault Diagnosis Methods Of Coal Mine Ventilator Bearings

At present,it is in the key stage of intelligent construction of coal mine.As the "lung of coal mine",the ventilator is responsible for exhausting the harmful gas and bringing fresh air.If the ventilator breaks down,it will cause huge economic loss to the enterprise,and seriously threaten the life safety of the underground personnel.So the research on the fault diagnosis of coal mine ventilator is of great significance.This article takes the ventilator bearing as breakthrough point,and analyzes the main fault forms of the current ventilator bearing.Aiming at the two aspects of feature vector extraction and fault type identification in the fault diagnosis process,the existing fault diagnosis model is improved.(1)Aiming at the problems in empirical mode decomposition,the improved extreme point symmetric extension method is used to improve the endpoint effect,and the ensemble empirical mode decomposition(EEMD)is used to improve the modal aliasing.Aiming at the problem of the large number of false components in the ensemble empirical mode decomposition,the screening criteria of the intrinsic mode function were improved and introduced into the complete ensemble empirical mode decomposition with adaptive noise.(2)Aiming at the problem of poor convergence speed and precision of genetic algorithm(GA),the genetic operator of genetic algorithm is improved.Aiming at the problem that the support vector machine(SVM)parameters are not easy to determine,improved genetic algorithm(IGA)is used to improve the SVM parameter optimization process,and an IGA-SVM model is constructed.The results show that the model has better convergence speed and convergence accuracy,it improves the accuracy of fault identification.(3)Aiming at the problem that samples training time of SVM model is long,an extreme learning machine(ELM)is used instead of SVM.Aiming at the problem of premature convergence and local optimization of particle swarm optimization(PSO),the inertia factor of the particle swarm optimization algorithm is improved,and a similar S-shaped decreasing inertial factor curve is constructed.The improved particle swarm algorithm(IPSO)and IGA algorithm are used to construct the IGA-IPSO-ELM model.The results show that the model has the characteristics of short training time,high accuracy of fault recognition,and strong robustness.There are 59 figures,16 tables and 77 references in this paper.