Mechanical Fault Diagnosis Based On Particle Swarm Opitimization And Wavelet Theory

Particle swarm optimization(PSO) is an optimization technique based on swarmintelligence. The algorithm provides efficient solutions for optimization problemsthrough intelligence generated from complex activities such as cooperation andcompetition among individuals in the biologic colony． In the optimization,themathematical model is structured according to the engineering problems and thespecific meaning of the mathematical model and the scope of the parameters in themodel should be fully understood.After that,the parameters involved in the particleswarm opitimization are initiated,and then,the fitness function would be proposedbased on the engineering practice.Then the fitness function is used to evaluate theparameters whether the parameters are suitable or not.Compared with the traditionaloptimized algorithm,the particle swarm algorithm in the multi-dimensional functionoptimization,the dynamic goal seeks the excellent aspect to have the convergence rateto be quick,the solution quality is high,and so on merits,so it has been widely used inengineering practice.Based on the research of particle swarm optimization and thewavelet theory,this method has been applied to diagnose the gearbox fault conditions.Based on the nonlinear and non-stationary of gearbox fault vibration signal, themathematical model of the vibration signal based on the adaptive Morlet wavelet isproposed to extract the gearbox fault characteristics information.The adaptive wavelettransform(AWT) ensure that the model parameters optimized by the particle swarmoptimization correspond to the adaptive wavelet parameters and has goodperformance in both time and frequency domain.For the gearbox fault diagnosis based on the particle swarm optimization and thewavelet theory,firstly,the original signal is analyzed by the Hilbert transform to get thesignal envelope.Then the parameters involved in the adaptive wavelet areproposed.And then the method based on the particle swarm optimization and the leastmean square error(LMSE) are used to optimize the parameters in the modelling whichthen would be plugged into the model formula.At last,the continuous wavelet transform based on the adaptive wavelet are evaluated the fault conditions of thegearbox.The results of the experimence demonstrates that the proposed method isfeasible and effective.Integrated the optimal ability of the PSO and the fine resolution in time andfrquency domain of the wavelet theory,the method of fault diagnosis based on theparticle swarm optimization and wavelet theory is proposed and used to detect thefive gear fault conditions.The diagnostic results show the feasibility and reliability ofthe proposed fault diagnosis methods by the identification for different gear damagelevels under identical crack fault category.The results of research provide a novel wayto promote the improvement of the PSO algorithm along with the wavelet theory andthe application to other fields.