基于SVM的大型公共建筑能耗预测模型与异常诊断

发布时间：2018-03-05 16:31

本文选题：大型公共建筑　切入点：建筑节能　出处：《土木工程与管理学报》2017年06期 　论文类型：期刊论文

【摘要】：针对建筑运行中管理粗放、使用中能源浪费的问题,提出了基于支持向量机的能耗预测与异常诊断方法,为大型公共建筑节能提供理论支撑与实现路径。由于大型公共建筑具有极大的单位面积耗能量而得到广泛的关注。本文针对大型公共建筑能耗特点,从历史能耗数据、气候因素、时间周期因素三个方面选取11个输入参数作为样本特征,构建基于支持向量机的大型公共建筑能耗预测模型,对建筑逐日能耗展开预测。在能耗预测基础上,以测试集平均相对误差与最大误差作为判定标准进行能耗异常诊断,将该方法应用于夏季空调系统能耗异常诊断中,通过能耗预测值与实际值的对比,发现了空调系统运行中存在的不合理使用现象,为建筑节能管理运行提供参考。
[Abstract]:Aiming at the problems of extensive management and energy waste in building operation, a method of energy consumption prediction and anomaly diagnosis based on support vector machine (SVM) is proposed. This paper provides theoretical support and realization path for energy saving of large public buildings. Because large public buildings have a great amount of energy consumption per unit area, it is widely concerned. According to the characteristics of energy consumption of large public buildings, this paper analyzes the historical energy consumption data. In this paper, 11 input parameters are selected as sample features from three aspects of climate factor and time cycle factor, and then a large-scale public building energy consumption prediction model based on support vector machine is constructed, and the daily energy consumption is forecasted on the basis of energy consumption prediction. The average relative error and maximum error of the test set are taken as the criterion to diagnose the abnormal energy consumption. The method is applied to the energy consumption anomaly diagnosis of the air conditioning system in summer. The comparison between the energy consumption prediction value and the actual value is carried out. The unreasonable use phenomenon in the operation of air conditioning system is found, which provides reference for building energy saving management operation.
【作者单位】：信阳师范学院建筑与土木工程学院;华中科技大学土木工程与力学学院;
【基金】：国家自然科学基金(71571078;51778262;51708241) 国家重点研发项目(2016YFC0800208)
【分类号】：TU111.195

【相似文献】