面向需求响应的数据驱动零售电价定价策略

Data-driven Electricity Retail Pricing Strategy for Demand Response

需求响应旨在削峰填谷和改善负荷轮廓,以提升电力系统运行的安全性和经济性。合理的电力零售定价有助于引导用户调整用能行为,促进需求响应。在此背景下,提出一种用能行为学习模型和一种个性化零售定价方案。首先,分析用能行为特征,利用基于多类注意力机制的长短期记忆网络提取特征关联性,构建用能行为学习模型。然后,考虑价格风险,以最大程度开发用户需求响应潜力为目标,建立个性化零售定价模型。算例仿真结果验证了所采用的学习模型相对传统深度学习模型的优势,提出的个性化零售定价方案可开发时变需求响应潜力,且能有效规避零售商收益风险和用户购电成本风险。

Demand response aims at shifting peak load to valley load and improving the load profile,so as to enhance the safety and economy of power system operation.Reasonable electricity retail pricing enables to guide users to adjust their energy consumption behaviors,thus facilitating demand response.In this context,an energy consumption behavior learning model and a customized retail pricing strategy are presented.Firstly,characteristics of energy consumption behaviors are analyzed,and long short-term memory networks based on multiple attention mechanisms are exploited to extract feature correlations and construct the energy consumption behavior learning model.Secondly,considering the price risk,a customized retail pricing model is proposed with the goal of fully developing users’demand response potential.Case simulation results demonstrate the superiority of the proposed learning model compared with traditional deep learning models,and the proposed customized retail pricing strategy is capable of developing time-varying demand response potentials while it is effective to hedge the profit risk for the retailer and the cost risk of electricity purchase for the user.