摘要
日负荷曲线预测是电力市场运营的基本内容。而短期负荷预测应用中较为成功的人工神经网络方法ANN(artific ial neural network),在很大程度上取决于训练样本以及输入变量的合理选取,它关系到算法的收敛性、计算速度以及预测的精度。通过对长春地区日负荷数据与日气象数据的基础分析,提出了选用多时段气象数据以及日类型作为相似日判别要素,并运用灰色关联理论,计算出预测日和诸多历史日的关联度,来确定ANN的训练样本,从而建立起适应性较强的日电量的预测模型。然后由日电量预测的结果,采用96点的波形系数,求出日各点的负荷预报值,经滚动预测检验证明,该方法能较好地满足实际电力系统的负荷预测要求。
Next-day load curve prediction is the important items of electricity market operation system. The algorithm of artificial neural network, which is applied in short-term load forecast successfully, depends on how to select the trained samples and input variables to a great extend, as it has great relation with the convergence, calculation speed and calculation precision. Based on overall analysis of daily load data and daily meteorological data in Changchun, this paper proposes the set of multi-intervals meteorologicals data and date type involved is considered as the criterion of determination of similar days. With Grey incidence theory, the incidence degree of temperature variation curve in historical days and that of forecast days is obtained. The days which has similar meteorological condition to future day is selected as the sample of ANN ,and daily forecast model of electricity amount which has good adaptive characteristics is constructed. By wave coefficient of daily load curve, the estimated load magnitude of 96 points per day is gotten. Verified by continuous prediction in practice system, the result is satisfactory.
出处
《继电器》
CSCD
北大核心
2005年第23期41-45,共5页
Relay
关键词
日负荷曲线预测
相似日
多时段气象数据
灰色关联理论
波形系数法
next-day load curve prediction
similar days
daily multi-intervals meteorological data
Grey incidence theory
wave coefficient