一种高压开关柜温度预测的改进方法

对一种Elman神经网络进行高压开关柜温度预测的方法进行改进，融合小波包分析的方法对温度样本进行硬阈值平滑优化，与Elman神经网络松耦合组成松散型小波神经网络进行超短期温度预测。仿真结果表明，改进方法相比单一的Elman神经网络有效预测比率提高，均方根误差下降，其结果与实测温度值更加吻合，提高了预测的准确率和精确程度。

An improved method of temperature prediction of high voltage switchgear

Based on the improvement of current Elman Network in temperature prediction of High Voltage Switchgear, a model called relaxing wavelet network, which loosely combines Elman neural network with wavelet package analysis and sample hard threshold denoising, is proposed to make an ultra-short term prediction. Samples can be simplified and smoothed by this method. Simulation results show that the optimized method can improve the valid rate of High Voltage Switchgear temperature prediction and reduce the root-mean-square error. Prediction values show better coincidence to the measured ones. Compared with the single Elman neural network predic- tion, this method can improve accuracy of prediction and its preciseness.