摘要
针对内蒙古西部草原民居建筑构造简单,围护结构热工性能差,民居建筑的采暖能耗高等问题,提出基于人工蜂群(ABC)优化BP神经网络的能耗预测方法。本文根据民居建筑特点,选取11个能耗影响因素作为输入参数的样本特征,利用De ST-h模拟结果作为网络的学习和测试数据,构建基于ABC-BPNN的能耗预测模型。然后将预测结果与De ST-h模拟计算结果进行对比,得出训练样本平均误差为0. 075 W/m^2,测试样本平均误差也仅为0. 29 W/m^2。最后将基于ABC-BPNN的预测模型与基于BP神经网络的预测模型进行对比,结果表明,基于ABC-BPNN的预测模型平均误差更低。所以人工蜂群优化BP神经网络算法可以相对准确地预测内蒙古西部草原民居建筑能耗,为草原民居建筑节能提供理论依据。
Aiming at the problem of simple structure, poor thermal performance of envelope structure and high energy consumption in Inner Mongolia western grassland residential buildings, the energy consumption prediction method based on artificial bee colony algorithm(ABC) optimization BP neural network is proposed. In this paper, according to the characteristics of residential buildings, 11 energy consumption factors are selected as the sample characteristics of the input parameters, and the DeST-h simulation results are used as the learning and test data of the network to construct the energy consumption prediction model based on ABC-BPNN. Then comparing the prediction results with the DeST-h simulation results, and the average error of the training samples is 0.075 W/m 2, and the average error of the test samples is only 0.29 W/m 2. Finally, the prediction model based on ABC-BPNN is compared with the prediction model based on BP neural network. The comparison results show that the average error of prediction model based on ABC-BPNN is lower. Therefore, the artificial bee colony optimization BP neural network algorithm can relatively accurately predict the energy consumption of grassland residential buildings in western Inner Mongolia, and provide a theoretical basis for energy saving of grassland residential buildings.
作者
金国辉
魏雪
张伟健
JIN Guo-hui;WEI Xue;ZHANG Wei-jian(School of Civil Engineering, Inner Mongolia University of Science and Technology,Baotou 014010, China;China Electric Power Development Research Institute Co Ltd, Beijing 100053, China)
出处
《土木工程与管理学报》
北大核心
2019年第2期48-52,60,共6页
Journal of Civil Engineering and Management
基金
国家自然科学基金(51768053
51668051)
内蒙古自然科学基金(2016MS0516
2017MS(LH)0532)
西部绿色建筑国家重点实验室培育基地开放研究基金(LSKF201803)
关键词
内蒙古西部草原民居
建筑能耗
人工蜂群算法
BP神经网络
Inner Mongolia western grassland residential buildings
building energy consumption
artificial bee colony algorithm
BP neural network
