基于并行预测策略的公共楼宇日前空调负荷预测

发布时间：2018-05-27 13:34

本文选题：公共楼宇空调系统 + 日前负荷预测　；参考：《电网与清洁能源》2016年11期

【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。
[Abstract]:With the development of urban construction in China, the power consumption of public buildings is increasing rapidly. In order to strengthen the energy consumption management and reduce the energy consumption level, it is the basis of the work to forecast the daily load of air conditioning system in public buildings. In view of the large accumulated error of daily load forecasting in the air conditioning system of public buildings, a parallel forecasting strategy is proposed to establish the load forecasting model alone for 24 hours before the day. Then the data are preprocessed by combining principal component analysis and fuzzy C-means clustering to form training data of appropriate scale and variable dimension, which are used as input of support vector machine (SVM) prediction model. Particle swarm optimization (PSO) is used to optimize the model parameters of SVM. Based on the historical data of air conditioning load in real public buildings, the proposed algorithm is compared with the serial prediction strategy and the SVM prediction algorithm based on the traditional cross-validation test parameters. The results show that the proposed method makes full use of the characteristics of air conditioning load in public buildings and has high prediction accuracy and high speed.
【作者单位】：国网江苏省电力公司;河海大学能源与电气学院;江苏省配用电与能效工程技术研究中心;南京河海科技有限公司;
【基金】：国家自然科学基金项目(51577051) 国家电网公司科技项目(SGJS0000YXJS1501044)~~
【分类号】：TU831.2

【相似文献】