Bearing Fault Diagnosis Algorithm Based on Improved KELM
In this paper,an algorithm combining kernel principal component analysis(KPCA),osprey cauchy sparrow search algorithm(OCSSA)and kernel extreme learning machine(KELM)is proposed to solve the problem of non-Gaussian data and strong data coupling in the process of chemical industry.Firstly,the KPCA algorithm is employed for dimensionality reduc-tion,followed by the utilization of the OCSSA algorithm to determine the optimal values for kernel parameter γ and regular-ization coefficient C in KELM.Enhancements made to the sparrow algorithm include:augmenting population diversity through incorporating chaotic mapping technology,substituting the original sparrow algorithm's finder position update formula with a global exploration strategy derived from Osprey optimization,and replacing the follower position update formula of the origi-nal sparrow algorithm with Cauchy variation strategy.Ultimately,a bearing fault diagnosis and classification algorithm that in-tegrates KPCA,OCSSA,and KELM algorithms is established.Experimental results demonstrate that after undergoing OCSSA optimization,this approach exhibits remarkable accuracy and effectiveness in addressing bearing faults.
fault diagnosiskernel principal component analysissparrow algorithmkernel extreme learning machine
万方数据
维普
