基于CFD仿真和SVM算法的建筑能耗预测系统  

Building energy consumption prediction system based on CFD simulation and SVM algorithm

作　　者：付川琪 刘清惓[1,2,3] 杨杰 丁枫 陈高颖 袁宇 FU Chuanqi;LIU Qingquan;YANG Jie;DING Feng;CHEN Gaoying;YUAN Yu(Jiangsu Key Laboratory of Meteorological Observation and Information Processing,Nanjing University of Information Science and Technology,Nanjing 210044,China;Collaborative Innovation Center of Atmospheric Environment and Equipment Technology,Nanjing University of Information Science and Technology,Nanjing 210044,China;Jiangsu Technology and Engineering Center of Meteorological Sensor Network,Nanjing 210044,China)

机构地区：[1]南京信息工程大学江苏省气象探测与信息处理重点实验室,江苏南京210044 [2]南京信息工程大学大气环境与装备技术协同创新中心,江苏南京210044 [3]江苏省气象传感网技术工程中心,江苏南京210044

出　　处：《现代电子技术》2022年第8期75-79,共5页Modern Electronics Technique

基　　金：国家公益性行业(气象)科研专项项目(GYHY200906037);国家公益性行业(气象)科研专项项目(GYHY201306079);国家自然科学基金项目(41275042);江苏高校优势学科Ⅱ期建设工程项目(PAPD-Ⅱ)。

摘　　要：为了在不同环境条件下对建筑能耗进行较为准确的预测,文中提出一种基于计算流体动力学(CFD)仿真和支持向量机(SVM)算法的建筑能耗预测系统。首先利用CFD方法建立三维建筑模型并进行仿真,获得若干输入输出样本;然后将得到的样本按3∶1的比例分为训练集和测试集,利用SVM算法对训练集样本进行学习训练,获得一个能耗预测模型;最后将测试集样本放入该模型中,对模型的准确性进行验证。结果表明:SVM能耗预测模型的结果与仿真结果相比,误差百分比在[-1.133%,1.132%];经过实际建筑模型测试,实物测试能耗值与预测能耗值误差百分比在[-6.211%,8.118%]。当环境条件改变时实物测试能耗值和预测能耗值变化趋势一致。与现有一些预测模型相比,文中预测模型使用的SVM算法不需要太多的训练样本,且结合CFD仿真方法,能够使建筑能耗预测结果具有较高的准确性。In order to accurately predict the building energy consumption under the different environmental conditions,a building energy consumption prediction system based on computational fluid dynamics(CFD) simulation and support vector machines(SVM)algorithm is proposed. The three-dimensional building model is established and simulated by means of CFD method,and several input and output samples are obtained. Then the samples are divided into training set and testing set according to the proportion of 3∶1,and SVM algorithm is used to learn and train the training set samples to obtain an energy consumption prediction model. The testing set samples were put into the model to verify the accuracy of the model. The results show that in comparison with the simulation results,the error percentage of SVM energy consumption prediction model is [-1.133%,1.132%],after the actual building model test,the error percentage between the physical test energy consumption value and the predicted energy consumption value is [-6.211%, 8.118%], and the change trend of the physical test energy consumption value and the predicted energy consumption value is consistent with changing environmental conditions. In comparison with some existing prediction models,the SVM algorithm used in the prediction model does not need too many training samples,and can increase the prediction results of building energy consumption by combining with CFD simulation method.

关 键 词：建筑能耗预测 CFD方法 SVM算法 输入输出样本 训练集 测试集 能耗预测模型 实物测试 

分 类 号：TN919-34[电子电信—通信与信息系统] TP311[电子电信—信息与通信工程]