基于BIM-DB仿真和LS-SVM的建筑能耗预测

Building Energy Consumption Prediction Based on BIM-DB Simulation and LS-SVM

建筑外围护结构性能对建筑能耗有十分显著的影响,其中影响较大的参数包括:外墙综合传热系数、外墙太阳辐射吸收系数、屋面综合传热系数、屋面太阳辐射吸收系数、外窗综合传热系数和窗墙比。为了在设计阶段实现建筑能耗预测,本文将建立的BIM模型导入Designbuilder能耗分析软件,针对以上六个外围护结构设计参数进行正交试验并利用能耗仿真获取数据样本集;在此基础上利用最小二乘支持向量机(LS-SVM)对建筑能耗数据进行预测。为了验证预测的可靠性,将预测结果分别同BP人工神经网络、小波神经网络以及SVM预测结果进行对比分析,结果表明LS-SVM对建筑能耗进行预测具有明显优势。

Performance of building envelope structure has a very significant impact on building energy consumption,the influential parameters include:external wall comprehensive heat transfer coefficient,external wall solar radiation absorption coefficient,roof comprehensive heat transfer coefficient,roof solar radiation absorption coefficient,external window comprehensive heat transfer coefficient and window to wall ratio.In order to realize the prediction of building energy consumption in the design stage,the BIM model is imported into the Designbuilder energy consumption analysis software.Orthogonal tests are conducted for the above six exterior envelope design parameters,and the data sample set is obtained by energy consumption simulation.On this basis,the least squares support vector machine(LS-SVM)is used to predict the building energy consumption data.In order to verify the prediction reliability,the predicted results are compared with BP artificial neural network,wavelet neural network and SVM,the results show that LS-SVM has obvious advantages in building energy consumption prediction.