摘要
随着我国城市建设的推进,公共楼宇的用电能耗增长迅速。为加强能耗管理、降低能耗水平,对公共楼宇空调系统日前用电负荷进行预测是工作的基础。针对当前公共楼宇空调系统日前负荷预测累积误差大的现象,提出对日前24 h单独建立负荷预测模型的并行预测策略。然后融合主成分分析和模糊C均值聚类对数据进行预处理,形成合适规模及变量维度的训练数据,将其作为支持向量机预测模型的输入,并通过粒子群算法对SVM的模型参数进行自适应寻优。以实际公共楼宇空调负荷历史数据为基础,对比分析所提出的算法与串行预测策略及传统交叉验证试凑参数的SVM预测算法,结果表明提出的方法充分利用了公共楼宇空调负荷的特点,预测精度高、速度快。
With the development of urbanization in China,the energy consumption of public buildings is increasing at a faster pace. It is the basis of the work to forecast the electric load of the air conditioning system in public buildings for strengthening the management of energy consumption and re-duce the level of energy consumption. In view of the pheno-menon that the daily load forecasting deviation of the air conditioning system in public buildings is large, a parallel prediction strategy is put forward to set up the load forecasting model in the day ahead 24 hours. Then combine the principal component analysis and fuzzy C means clustering to pre-process the data. The algorithm forms the appropriate data dimension and data size. And the processed data is used as the input of the support vector machine model. The parameters of SVM are optimized by PSO algorithm. Based on the real historical data of air conditioning load in public buildings,comparative analysis of the proposed method and the traditional SVM algorithm. The results show that the proposed method makes full use of the characteristics of air conditioning load in public buildings,forecasting the load in high accuracy and high speed.
出处
《电网与清洁能源》
北大核心
2016年第11期80-86,90,共8页
Power System and Clean Energy
基金
国家自然科学基金项目(51577051)
国家电网公司科技项目(SGJS0000YXJS1501044)~~
关键词
公共楼宇空调系统
日前负荷预测
并行预测策略
数据预处理
支持向量机参数优化
public building air conditioning system
day ahead load forecasting
parallel prediction strategy
data pre-processing
parameter optimization of SVM
