基于改进的MGM(1,n)模型的旋转机械故障预测方法研究

Trend Prediction Methods Study of Rotary Machinery Based on Revised MGM(1,n) Models

在旋转机械灰色多变量MGM(1,n)预测模型基础上,引入马尔可夫链预测理论,建立灰色多变量马尔可夫MGM(1,n)预测模型,该模型既从系统角度对各特征参数进行统一描述,又通过状态转移概率矩阵的变换提取数据随机波动响应,能够描述不同类型的故障及其不同的发展阶段。通过实例验证表明该模型与灰色多变量MGM(1,n)模型相比具有更高的预测精度。

Based on the multi-variable grey model （ MGM （ 1, n） ） of trend prediction for the Rotary Machinery, the Markov chains prediction method was presented,then the grey markov MGM （ 1, n ） model for predicting the trend of the Rotary Machinery was built, which either takes the changes in each index to be described from the views of systems, or extracts data of random waver esponse through statet ransfer probability matrix, this model can describe different types of vibration failures and its different development stages . The example shows that the grey markov MGM （ 1, n） model often have a higher prediction accuracy than the MGM （1 ,n） model.