基于AGPF的滚动轴承性能衰退趋势预测

Prediction of Declining Performance of Rolling Bearing Based on AGPF

摘要:

针对粒子滤波算法中粒子退化和计算复杂度问题,提出了一种自适应遗传粒子滤波(AGPF)算法;该算法采用遗传算法代替传统粒子滤波中的重采样方法,并根据粒子数与滤波误差方差之间的关系,自适应调节滤波过程中的粒子数;通过预测滚动轴承的性能衰退趋势,对该方法进行验证,结果表明,AGPF算法能够在保证预测精度的条件下,减少滤波粒子数,更加适用于滚动轴承的性能衰退趋势预测。

Abstract:

Aiming at the problem of particle degradation and computational complexity in particle filter algorithm, an adaptive genetic particle filter (AGPF) algorithm is proposed. The algorithm uses the genetic algorithm instead of the resampling method in the traditional particle filter, and adaptively adjusts the number of particles in the filtering process according to the relationship between the number of particles and the variance of the filter error. The results show that the AGPF algorithm can reduce the number of filtered particles under the condition of ensuring the prediction accuracy, and is more suitable for the prediction of the performance of the rolling bearings.