基于差分自回归移动平均模型的电气设备温度预测

The substation equipment temperature prediction based on time sequence ARIMA model

针对电气设备温度监控及预测这一关键性问题,以某电气设备实际温度数据为样本,探讨基于非平稳时间序列的差分自回归移动平均模型ARIMA(p,d,q)来描述设备温度变化的可行性。使用Eviews6.0初步构建模型,进而采用枚举法得到最优预测模型。通过Matlab仿真,表明模型ARIMA(5,1,2),在预测误差可接受的范围内,能很好地拟合设备温度变化趋势,较准确地预测温度。并对ARIMA模型和BP神经网络模型进行了简单的对比,得到了ARIMA模型更加适合电气设备温度预测的结论。

The aim of this paper is to study the temperature monitoring and prediction for electrical equipments.The feasibility of Autoregressive Integrated Moving Average Model（ p,d,q） based on non-stationary time sequence is analyzed through taking the electrical device temperature data as the sample.Eviews6. 0 is used to build a preliminary model,and then by applying the enumeration method,the optimal prediction model can be established.Through matlab simulation,the result shows that the model of ARIMA（ 5,1,2） can well fit the temperature trends of electrical devices and accurately predict the temperature with acceptable error.On the ARIMA model and BP neural network model for a simple comparison,ARIMA model is more suitable for electrical equipment temperature prediction.