摘要
短期负荷预测是针对未来一天到数天各时段的负荷预测的研究，是电力系统负荷预测工作的一项重要内容．针对传统神经网络预测模型应用于短期负荷预测的缺陷，改进了多角度数据分析和组织策略，选择不同年份相近历史日作为相似日，通过最小二乘支持向量机填补确实数据，利用聚类算法预测相似日的短期负荷；同时通过灰度关联算法，考虑气象因素作用下的短期负荷预测模型．实例证明：通过建立与负荷数据相适应的数学模型，对负荷数据进行分析与预测，通过气象因素修正预测模型，可以获得更精确的负荷数据预测．
Short-term load forecasting is a study of load forecasting for each day-to-day period, which is an important part of power system load forecasting. Aiming at the shortcomings of traditional neural network prediction model applied to short-term load forecasting, this paper improves the multi-angle data analysis and organization strategy, chooses different historical days as similar days, and fills the exact data with the least squares support vector machine and use clustering algorithm to predict short - term load of similar day. In the meantime, the gray-scale correlation algorithm is used to consider the short-term load forecasting model under meteorological factors. It is proved that the mathematical model of the load data is established, and the load data can be analyzed and predicted. The prediction model of the load data can be obtained by the meteorological factors.
出处
《数学的实践与认识》
北大核心
2018年第3期131-143,共13页
Mathematics in Practice and Theory
关键词
短期负荷预测
聚类算法
灰色关联
相关系数
short-term load forecast
clustering algorithm
gray-scale correlation
correlation coefficient
