提出一种新的预测方法-基于EMD分解的时间序列模型,利用EMD分解将采集来的矿井风机振动烈度值分解成若干个固有模态函数(IMF)分量和一个残余项分量,运用恰当的时间序列模型(AR、ARMA)分别对各阶IMF进行预测,将各阶预测值重构,得到振动烈度预测值,并与单独运用时间序列模型的预测结果进行比较。结果证明:运用基于EMD分解的时间序列模型对矿井风机振动烈度进行预测比单独运用时间序列模型的预测精度有明显提高,表明提出方法的可行性、有效性。 A new prediction method is proposed based on the EMD decomposition time series model. EMD decomposition is used to decompose mine vibration severity into several IMFs and a residual term component. By using the appropriate time Sequence models (AR, ARMA) were used to predict the IMFs of each order separately, and the prediction values ​​of each order were reconstructed to obtain the forecast values ​​of vibration intensities, which were compared with the prediction results of the time series models alone. The results show that using the time series model based on EMD decomposition to forecast the mine fan vibration intensity obviously improves the prediction precision of the time series model alone, which shows that the method is feasible and effective.