基于EEMD和ARIMA模型的汽轮机故障趋势预测

由于汽轮发电机转子振动状态监测数据具有非线性和非平稳性,采用普通时间序列预测模型时预测的精度较低。研究通过分析振动信号的频率成分,融合EEMD分解平稳化处理和ARIMA预测模型的思想,建立一种混合预测模型。结果表明:该方法能够适应振动状态监测数据特征,反映了振动状态的主要变化趋势,具有较高的预测精度以及更大的应用范围,其预测趋势对进一步进行振动状态分析具有一定的参考价值。

Turbine Fault Trend Prediction that Based on EEMD and ARIMA Models

Because of the nonlinearity and non-stationarity of the vibration condition monitoring data of the steam-turbine generator rotor,it has the low accuracy if it takes ordinary time series prediction model to make prediction.This papers analyzes the frequency components of the vibration signal,and integrate the thought of EEMD decomposition stationary processing and ARIMA prediction model to establish a mixed prediction model.Experimental results show that this method can adapt to the data characteristics of vibra-tion condition monitoring,which has reflected the main trends of vibration state.It has higher accuracy and greater range of applications,and its forecast trends has a certain reference value for further analysis of the vibration state.