摘要
基于正交设计和灰色系统理论,提出一种预测年电力负荷的新方法。采用新陈代谢技术和加权最小二乘参数辨识法对标准GM(1,1)模型进行改进。以背景值系数α、建模所需数据个数m和加权参数q作为可控因素,根据专家经验设计了三因素三水平正交表。以平均绝对百分比误差为输出目标,通过信噪比分析,得出最优参数水平组合,并通过方差分析,进一步得出各可控因素对预测效果的影响程度。对2个电网的负荷进行预测,结果验证了所提方法的可行性和有效性。
A new annual power load forecasting method is proposed based on orthogonal design and grey system theory in this paper. The traditional GM (1,1) is improved by information renewal and weighted least-square estimation. The background value coefficients, thenumberofmodelingdatapointsm, and weighted parameter q are chosen as controllable factors. Based on experts' experiences, an orthogonal array with three factors and three levels is designed. By taking the mean absolute percentage error as output object, optimal combination of parameters is acquired through signal to noise analysis, and the impact degree of each factor is obtained through an analysis of variances respectively. The feasibility and effectiveness of the method proposed is verified by load forecasting in two power grid systems.
出处
《电网与清洁能源》
2010年第2期28-32,共5页
Power System and Clean Energy
基金
北京市教委科技成果转化与产业化项目
关键词
正交设计
灰色模型
负荷预测
最优参数组合
orthogonal design
grey model
load forecasting
optimal parameters combination