基于混沌理论的设备状态监测应用研究

以汽车变速箱中齿轮系统疲劳实验为例，充分利用混沌方法与小波方法各自在状态监测与故障诊断领域中的优点，进行机械系统状态监测与故障诊断。通过对实验所得振动信号进行小波包分解与重构，得到反映目标部件的特征信号，借助于混沌方法，对信号进行了二次处理，进而对所得特征量的变化趋势进行比较研究，确定了各监测部位的运行状态，并实现了故障定位。计算结果和实验结果相吻合，证明了该方法的可行性。

Application of Chaos Dynamics Method to Processing Filtered Signals

This paper takes gearbox fatigue test as an example to introduce the chaos method. It can make full use of advantages of wavelet packet and chaos method in condition monitoring and fault diagnosis. Vibration signals are decomposed and reconstructed by the wavelet packet method, and the characteristic signals of target parts are extracted. Then, signals are processed again by means of the chaos method. It can not only distinguish normal conditions from faulty conditions, but also locate the faults. Experiments prove the correctness of the computation results, which means that the method introduced in this paper is effective.