行星齿轮系故障特征多目标进化选择方法

Multi-objective Evolutionary Selection Method for Planetary Gear Trains Fault Feature

针对行星齿轮箱复杂故障难以辨识问题,提出一种基于遗传算法的行星齿轮系多目标故障特征优选方法。从振动信号中提取多域故障特征,建立了多域故障特征优选数学模型,将特征选择问题转化为多目标优化问题;采用遗传算法求解得到最优特征子集。实验表明,该方法对高维故障特征具有较好的降维性能,能有效选择高质量的故障特征子集,使得故障诊断性能明显提高。

Because it is hard to identify complex fault in planetary gearbox,this paper proposes a new method based on genetic algorithm which is used to extract the fault signals from vibration,establishes the mathematical optimization model for multi-domain fault features,changes the feature selection problem into multi-objective optimization problem and then uses the genetic algorithm to get the optimal feature subset. The experimental results show that the proposed method has a better dimensionality reduction performance for complex fault features,and can be used to effectively select a high quality fault feature subset,thus improving the effect the of fault diagnosis obviously.