基于混沌吸引子特征量的滚动轴承故障诊断

利用混沌吸引子特征量可以刻画滚动轴承在不同故障状态下振动特性的特点，提出一种基于关联维数、最大李雅普诺夫指数和信息熵的故障诊断方法。结合试验数据，应用支持向量机技术分析了3类特征量对滚动轴承的故障识别能力，并对比了特征量两两组合的分类效果。研究表明：3类特征量都包含着不同的故障信息，将其结合可以明显提高故障识别率。通过对实测轴承数据的故障分类研究发现，与单一特征量方法相比，该方法可以有效区分不同故障类型和故障严重程度，为滚动轴承故障的超精密诊断提供了可能性。

Fault Diagnosis of Rolling Bearing Based on Characteristic Auantities of Chaotic Attractor

The characteristic quantities have good performance in reflecting the nonlinear dynamics of rolling bearing in different fault conditions. A method of fault diagnosis of rolling bearing based on correlation dimension, largest Lyapunov exponent and information entropy is proposed. The classification abilities of each feature are evaluated by using support vector machines, as well as the combination of two quantities. The study shows that each type of quantity contains different fault information and the combination of these can significantly improve the recognition rate. The experimental results also show that these three characteristic quantities can effectively identify the different types of fault and also the same fault with different levels.