峰谷电价下基于长短时记忆网络的配电网短期负荷预测研究

Short-Term Load Forecasting for Distribution Network in The Presence of TOU Price Based on Long-Short-Term Memory Network

随着智能电网的不断发展,电力市场的利益主体逐渐呈现出多元化的趋势,电价机制也得到调整和更新,峰谷电价政策应运而生.为提高峰谷电价下的配电网短期负荷预测精度,文中通过用户心理学原理构建电力用户的实际需求响应模型,拟合动态峰谷电价下的负荷曲线,量化多维预测输入特征,构建基于长短时记忆网络的短期负荷预测模型.通过实际算例分析,验证提出预测模型的可靠性.

With the development of smart grid,the,the interests of the main body in the power market gradually diversified,electricity pricing mechanism has been adjusted and updated,and TOU price is coming into being.In order to improve the accuracy of short-term load forecasting considering TOU price,the paper described the customers’fuzzy demand response mechanism based on the theory of customer psychology,fitted the load curve under TOU price,quantified multidimensional prediction input features,and proposed a short-term load forecasting model based on long-short-term memory networks.Through practical examples,the proposed method has higher prediction accuracy showed in experimental results.