摘要
现今电力系统短期负荷预测的点预测方法多种多样。为弥补传统点预测方法结果过于单一的问题,结合温度及日历序列因素对负荷的影响,提出基于交替条件期望(ACE)的短期电力负荷概率密度预测方法。以温度及日期序列为负荷影响因子,建立基于交替条件期望理论的非参数回归模型,计算历史负荷与影响因子的非线性回归方程;考虑日类型及星期类型等多重因素,利用模糊聚类方法选取相似日;以回归方程为基础,根据所得相似日及预测日影响因子,进行负荷回归值计算,并利用核密度估计(KDE),得到负荷概率密度曲线。利用某市的实测数据,进行负荷概率密度曲线预测,并选取概率密度众数作为负荷点预测值,与其他负荷预测方法结果相比较,仿真结果表明该方法的精度高、可靠性好。
At present,there are a variety of point prediction methods to forecast the short-term load of power system.To compensate for the simple results of traditional point prediction methods,the paper,considering the influence of temperature and day type on the load,put forward a short-term power load probability density forecasting method based on alternating conditional expectation（ACE）.With temperature and date sequence as factors influencing load,a nonparametric regression model is established based on the alternating conditional expectation theory to produce nonlinear regression equation of historical load and influence factor.Considering the factors such as the type of day and the type of week,the paper selects the similar days by using the fuzzy clustering method.The regression equation was used to calculate load regression value based on the influence factors of the similar days and the forecast day and obtain the load probability density curve by kernel density estimation（KDE）.Based on the measured data of some city,the paper predicts the probability density curve of load and selects the mode of probability density as the predicted value of load point.Compared with the results of other methods of load forecasting,the simulation results show that this method can produce results of higher precision and reliability.
出处
《华北电力大学学报（自然科学版）》
CAS
北大核心
2018年第1期58-65,共8页
Journal of North China Electric Power University：Natural Science Edition
基金
中央高校基本科研业务费专项基金资助项目(2016XS10)
电网技术国际标准研制项目(201510207-3)
关键词
概率密度预测
交替条件期望
模糊聚类
核密度估
probability density prediction
alternating conditional expectation
fuzzy clustering
kernel density esti-mation
