引用本文: | 李扬,陈玉辰,王子健,蒋浩然.市场竞争机制下用户用电行为特性辨识模型[J].电力自动化设备,2019,39(11): |
| LI Yang,CHEN Yuchen,WANG Zijian,JIANG Haoran.Identification model of users’ electricity consumption behavior characteristics under market competition mechanism[J].Electric Power Automation Equipment,2019,39(11): |
|
摘要: |
研究市场环境下的用户用电行为,有利于优化电力市场机制,为市场的数据化运营奠定基础,对电网的安全稳定运行有着积极的作用。首先,针对不同用户在电力市场环境下对电能价格和需求响应政策等的反应机制,从购电潜力、电价敏感度、需求响应潜力3个方面构建市场行为评价指标体系,在对初始用户样本进行指标量化的基础上,采用二次聚类法确定用户初始类别;然后,基于学习向量量化(LVQ)神经网络和自组织映射(SOM)神经网络,提出具有类别增量学习功能的自适应辨识模型;最后,基于某地市的实际用户数据进行模型验证。算例结果表明,所提自适应辨识模型的辨识结果准确,能有效地辨识得到新的用户类型,同时在更新辨识模型的速度上也有较大的优越性。 |
关键词: 电力市场 用电行为 竞争机制 自适应辨识 LVQ神经网络 SOM神经网络 类别增量学习 模型 |
DOI:10.16081/j.epae.201910015 |
分类号:TM714;F426.61 |
基金项目:国家自然科学基金资助项目(动态可控负荷参与电力系统调频辅助服务理论与方法研究)(51777030) |
|
Identification model of users’ electricity consumption behavior characteristics under market competition mechanism |
LI Yang, CHEN Yuchen, WANG Zijian, JIANG Haoran
|
School of Electrical Engineering, Southeast University, Nanjing 210018, China
|
Abstract: |
The study of users’ electricity consumption behavior in the market environment is conducive to optimizing the electricity market mechanism, laying a foundation for data operation of the market, and playing a positive role in the safe and stable operation of the power grid. Firstly, according to the response mechanism of different users to electricity price and demand response policy in the electricity market environment, the market behavior evaluation index system is constructed from three aspects of electricity purchase potential, electricity price sensitivity and demand response potential. Based on the quantification of the initial user samples, the two-step clustering method is applied to determine the initial user categories. Then, an adaptive identification model with the function of category incremental learning is proposed based on LVQ(Learning Vector Quantization) neural network and SOM(Self-Organizing Maps) neural network. Finally, the proposed model is verified based on the actual users’ data of a certain city. The results show that the proposed adaptive identification model can give accurate identification results, and can effectively identify new user categories. At the same time, the speed of updating the identification model has great advantages. |
Key words: electricity market electricity consumption behavior competition mechanism adaptive identification LVQ neural network SOM neural network category incremental learning models |