基于经验模式分解的神经网络模型及其在转子系统故障诊断中的应用

为了对非平稳、非线性系统时间序列进行建模,提出一种基于经验模式分解的神经网络预测模型,研究它的有效性。通过太阳黑子数据的仿真试验,验证该神经网络结构比对应的单一神经网络结构性能优越。根据该方法组成一个多分量神经网络模型库,用于转子故障的模型诊断,这些模型可以用做一步向前预测器,对检测和诊断信号进行比较,从预测误差提取特征,能够确定机器的状态。不同故障状态的转子振动信号用来训练和检验模型。实验数据表明,这种方法用于故障诊断具有一定的工程实用性。

Neural Network Predictive Model Based on Empirical Mode Decomposition and Its Application to Fault Diagnosis of Rotor System

The effectiveness of a multi-mode neural-network predictive model based on empirical mode decomposition for the time series prediction of non-stationary,nonlinear dynamic systems has been investigated.The simulated experiment for sunspots' benchmark shows that the multi-mode architecture outperforms the corresponding single-scale architectures.Then,an observer bank of multi-mode neural-network is used for model diagnosis of rotor fault vibration signals.These models can be used as one step ahead predictors allowing comparison of signals for the purposes of fault detection and diagnosis.From the prediction error,features can be extracted and be used to determine the machine's condition.Vibration data of rotor acquired from different fault conditions are used for training and testing models.The experiment results indicate that this approach could be used to diagnose fault conditions.