基于平衡策略的支持向量机在电力系统中期负荷预测的应用  

Support vector machine based on balance strategy applying in forecast mid-long term load of electrical power system

作　　者：肖志刚[1] 徐宏[1] 杨素林[2] 王岩[1] 孙玉梅[1] 

机构地区：[1]河北农业大学机电工程学院,河北保定071001 [2]河北农业大学信息科学与技术学院,河北保定071001

出　　处：《河北农业大学学报》2006年第5期123-126,共4页Journal of Hebei Agricultural University

摘　　要：针对电力负荷中期预测比较困难并且存在较大误差的问题,提出了基于径向基核函数与基于平衡策略的Sequential minimal optimization(SMO)改进算法相结合的负荷中期预测方法,结合EUNITE网络提供的实际数据,研究了日最大负荷的前后期关系、日最大负荷与节假日的关系和当日与对应星期数的相关性,建立了相应的电力负荷中期预测模型。并对预测结果进行了分析。算例表明,该算法具有运算速度快、精度较高的优点。To the power system, mid - long term load forecast was difficulty and big error exists. Mid - long term load forecast method based on radio basis kernel function and improved algorithm based on the balance strategy SMO was put forward. Using the actual data provided by EUNITE network, the relationship between prophase and anaphase the daily maximum- load, the relationship between daily maximum-load and holiday, the relativity between that day and related weekday were studied. The Mid - long term load forecast model was presented. The forecast results show that the method has some advantages such as high speed and efficiency etc.

关 键 词：负荷预测 SVM SMO算法 

分 类 号：U223.54[交通运输工程—道路与铁道工程]