考虑数据不确定性的中长期电力负荷预测  

Medium and long term load forecasting considering data uncertainty

作　　者：郑志杰[1] 李磊[1] 赵兰明[1] 

机构地区：[1]山东电力工程咨询院有限公司,山东济南250013

出　　处：《电力系统保护与控制》2011年第7期123-126,132,共5页Power System Protection and Control

摘　　要：中长期负荷预测时间跨度较长,其基础数据受诸多因素影响,具有不确定性和不可控性。引入蒙特卡罗算法和区间算法处理中长期负荷预测中的数据不确定性问题。根据历史年度实际情况,假定基础数据在某一范围内存在不确定性,采用蒙特卡罗算法构建了计算流程,可以得到界于某一区间的负荷预测值;采用区间算法描述基础数据的不确定性,针对区间算法固有的过估计问题,通过推导适合的公式,可以避免产生过度保守的结果,只需一次计算,就可以严格分析数据不确定性对预测结果的影响,具有节省计算时间的优点。在考虑基础数据存在不确定性情况下,通过某省电网负荷预测实例计算并与传统预测方法相比较,验证了两种负荷预测方法可以评估数据不确定性对负荷结果的影响,避免得到过度保守的负荷预测值。Mid-long term load forecasting goes through a long time and the base data are influenced by a number of factors, most of which are beyond the control and with uncertainty. In order to deal with the data uncertainty problem in mid-long term load forecasting, Monte Carlo method and interval arithmetic are introduced. According to actual situation in history, it is assumed that the base data have uncertainty in a certain range. Using Monte Carlo method to construct the calculation process can obtain the load prediction result in a certain interval. Using interval algorithm to describe the uncertainty of base data, and for reducing overestimation of interval arithmetic, suitable formula is deduced to avoid the conservative results, with the advantages of saving computing time because only once computation can seriously analyze the effect of data uncertainty on prediction results. The proposed two methods have been tested on the load forecasting of one province. In consideration of the base data with uncertainty, by comparing with conventional load forecasting method, the results demonstrate the two approaches can evaluate the effect of data uncertainty on load forecasting, and avoid getting conservative results.

关 键 词：中长期负荷预测 不确定性 蒙特卡罗 区间算法 指数平滑 

分 类 号：TM715[电气工程—电力系统及自动化]