基于GRU-NN模型的电力用户能耗预测研究

Energy Consumption Prediction of Power Users Based on GRU-NN Model

针对现有预测方法在对电力用户进行能耗预测时,存在预测精度低、预测时效性差的问题,引入GRU-NN(Gate Recurrent Unit-Neural Networks,门控循环单元-神经网络)模型,开展对电力用户能耗预测方法的设计研究。采集电力用户用电信息,并从归一化处理后的信息中提取用电特征。利用GRU-NN,构建能耗预测模型。结合均方误差和拟合优度的概念,对该模型进行训练。利用训练后的模型预测电力用户能耗,模型的输出即为预测结果。通过对比实验,证明新的预测方法预测结果更加接近实际,且预测耗时短,具备较高的时效性,值得广泛应用。

In view of the problems of low prediction accuracy and poor prediction timeliness of existing prediction methods when predicting power users'energy consumption,GRU-NN(Gate Recurrent Unit-Neural Networks)model was introduced to design and research the energy consumption prediction methods for power users.The power consumption information of power users is collected,and the power consumption characteristics are extracted from the normalized information.The energy consumption prediction model is constructed by using GRU-NN.The concept of mean square error and goodness of fit is combined to train the model.The trained model is used to predict the energy consumption of power users,and the output of the model is the prediction result.The comparison experiment proves that the prediction result of the new prediction method is closer to the reality,the prediction time is short,and the timeliness is high,so the is new prediction method is worthy of wide application.