基于等维新息指数平滑法模型的中长期负荷预测  

Medium/long term load forecast of exponential smoothing method based on information and equal dimensional operators

作　　者：叶宗斌[1] 周步祥[1] 林楠[2] 黎祚[1] 程寅[1] 

机构地区：[1]四川大学电气信息学院,四川成都610065 [2]四川电力职业技术学院,四川成都610071

出　　处：《电力系统保护与控制》2012年第18期47-51,共5页Power System Protection and Control

摘　　要：指数平滑法是电力系统中长期负荷预测的重要方法之一,该预测方法的精确性主要取决于平滑系数,因此提出了等维新息数据处理的指数平滑法,实现了变平滑系数。首先,给出指数平滑法的预测模型,根据重近亲远的原则,给各期绝对平均误差值赋权。接着,采用0.618优选法优选出平滑系数。在此基础之上,引入了等维新息数据处理的思想,实现了变平滑系数,使预测结果能够更加合理地反映负荷发展规律,提高了预测精度,优化了指数平滑法的预测模型。Exponential smoothing method is one of the main medium/long term load forecasting methods for power system. The accuracy of the method depends on smoothing coefficient. So this paper puts forward exponential smoothing method based on information and equal dimensional operators, achieving the variable smooth coefficient. Above all, the forecasting model of exponential smoothing method is presented and according to the principle of close relatives with heavy far, MAE of each period is weighted. Then, optimum smooth coefficients are chosen by 0.618 optimum seeking method. On this basis, the idea of information and equal dimensional operators are introduced, the variable smooth coefficient is realized, which enables prediction results to reflect load development law more reasonably and the accuracy of the predictions can be improved. Finally, the forecasting model of exponential smoothing model is optimized.

关 键 词：指数平滑法 中长期负荷预测 等维新息 变平滑系数 0.618优选法 

分 类 号：TM715[电气工程—电力系统及自动化]