特征提取是旋转机械故障分类中一个很重要的问题。为解决该问题,提出一种利用时域中已有的无量纲指标和遗传程序设计对旋转机械多类故障进行新无量纲指标提取的方法。实验结果表明,新无量纲指标比传统的无量纲指标具有更好的故障分类能力,能够对旋转机械故障进行准确分类。
Feature extraction is an important issue in rotating machinery fault~ classification. To solve it, we propose an approach to extract the new dimensionless parameter of multiple faults of rotating machinery using the existing dimensionless parameter in time domain and the genetic programming design. Experimental result shows that the faults classification ability of new dimensionless parameter is better than that of the existing ones, and the accurate classification of rotating machinery faults can be achieved.