论文部分内容阅读
为克服无相关历史数据的困难,满足首次设计峰谷分时电价时挖掘居民用户行为规律的需要,提出一种模拟居民对分时电价需求响应规律的模型。在一定的响应行为随机分布假设前提下,首先通过问卷设计与抽样调查,统计目标地区居民在各种设定的电价情景下选择执行分时电价的频率及概率;其次结合问卷种类与统计结果设计模型的输入、输出属性并确定训练样本集合;最后在训练样本集的基础上运用LSSVM回归算法构造响应行为预测模型。该模型可在输入一定幅度内任意分时电价的情况下,输出对应的目标居民平均响应结果及标准差,从而实现了在无历史数据时,对居民在分时电价下的响应规律进行模拟,并为更多研究提供数据支持。算例仿真验证了该模型方法的合理性以及可行性。
In order to overcome the difficulty of no relevant history data and meet the requirement of mining user behavior rules when designing the TOU price for the first time, a model to simulate the resident demand response to TOU price is proposed. On the premise of a certain random distribution of response behavior, the questionnaire design and sample survey are firstly used to calculate the frequency and probability of residents in the target area to choose and implement time-sharing electricity prices under various set price scenarios. Secondly, based on the questionnaire types and statistical results design Input and output attributes of the model and determine the set of training samples; Finally, LSSVM regression algorithm is used to construct the response behavior prediction model based on the training sample set. The model can output the corresponding average response result and standard deviation of target population under the condition of inputting arbitrary time-sharing price within a certain range, so as to simulate the resident’s response under time-of-day price without historical data, And provide data support for more research. The simulation results show that the model method is reasonable and feasible.