一种基于LSTM神经网络的电力负荷预测方法

A Power Load Data Prediction Method Based on LSTM Neural Network Model

设计了一种基于长短期记忆(LSTM)神经网络的电力负荷预测模型,在TensorFlow框架下使用Python语言编程实现;使用西班牙2018年一整年的电力负荷数据对模型进行训练,得到的模型可准确预测电力负荷数据的日变化、周变化规律,模型损失值可达0.2,验证了模型的有效性;与RNN模型对比证明了LSTM模型的长期依赖学习能力更为优越。提出的模型是一种有效的电力负荷数据预测方法,可为电力系统的负荷预测提供依据。

The analysis and predication of power load is an important part of power system operation and planning,and has great significance.This paper proposes a prediction model based on LSTM neural network and the model is implemented by Python language with the TensorFlow framework.The model is trained by using Spanish power load data in the year 2018.In contrast to prospective measured data,the trained model can accurately predict the daily and weekly patterns on power load data with the maximum error only 0.2.Moreover,The comparison with RNN model proves that LSTM model has superior long-term dependent learning ability.As a result,this model shows that it is an effective method to predict power load data and can contribute much to power system.