【摘要】 目前,传统指标的用电特性研究仅仅从用户的负荷曲线本身出发,而基于需求响应的用电特性研究相对很少。从电网在紧急负荷控制、季节性调峰、经济削峰、消纳清洁能源等运行场景,以及日前、日内、实时等多时间尺度下进行负荷调节需求研究,提出了多场景下用户弹性负荷的配电网资源分类和调节特性。针对此目的,设计了基于改进支持向量机(Support Vector Machine, SVM)算法研究多类用户负荷预测技术。研究结果表明,在峰谷分时电价浮动比率变化时,该描述模型可以用来建模响应电量比率变化的一般规律,能够揭示政策向导使能的分时电价,能够体现电力用户电能变化的关系。此外,该方法有助于灵活调整分时电价政策,用时提供基础数据实施政策激励。更多还原

【Abstract】 At present, the research on the electricity characteristics of traditional indicators only starts from the user’s load curve itself, while the research on the electricity characteristics based on the demand response is relatively rare. This paper studies the demand for load regulation from the operation scenarios of power grid such as emergency load control, seasonal peak regulation, economic peak cutting and absorption of clean energy, as well as the multi-time scales such as day, day and real time, and proposes the distribution network resource classification and regulation characteristics of user elastic load under multi-scenario. For this purpose, a multi-class user load forecasting technique based on improved Support Vector Machine(SVM) algorithm was designed. The research results show that when the peak-valley time of use(TOU) power price fluctuation rate changes, the description model can be used to model the general law of response to the change of the electricity rate, to reveal the TOU power price of energy enabled by the policy guide, and to reflect the relationship between electric energy changes of power users. In addition, this method is helpful to flexibly adjust TOU power price policy and provide basic data to implement policy incentives.更多还原