基于遗传算法改进的建筑能耗预测模型研究  

Study on Improved Building Energy Consumption Prediction Models Based on Genetic Algorithm

作　　者：张朝 谭勇 ZHANG Zhao;TAN Yong(School of Instrument Science and Electrical Engineering,Jilin University,Changchun Jilin 130061,China;State Grid Jilin Electric Power Co.,Ltd.Dehui Power Supply Company,Dehui Jilin 130300,China)

机构地区：[1]吉林大学仪器科学与电气工程学院,吉林长春130061 [2]国网吉林省电力有限公司德惠市供电公司,吉林德惠130300

出　　处：《北方建筑》2024年第2期12-16,共5页Northern Architecture

基　　金：应急管理部安全生产重特大事故防治关键技术科技项目(jilin-0032-2018AQ)。

摘　　要：建筑行业作为国民经济支柱性产业,在推动社会发展的同时也消耗了大量的能源,加速建筑行业的节能减排,对实现我国2030年CO_(2)排放达到峰值的节能减排目标具有重要意义。因此,本文针对人员相对集中的高校建筑能耗进行了分析,通过收集校园内大量建筑能耗数据,充分挖掘建筑能耗特点,采用遗传算法优化BP和LSTM建筑能耗预测模型,通过实验对比选取最符合校园内的预测模型,以此预测未来校园内建筑能耗使用情况,为校园内后续碳排放预测和节能措施的制定提供基础数据。通过实验对比可以发现,GA-BP模型相较于GA-LSTM模型的平均绝对误差、均方误差、均方根误差、平均绝对百分比误差分别降低8.83%,4.80%,2.43%,10.2%,GA-BP模型误差更小,拟合程度更高,对高校建筑能耗的预测更符合实际,可用于高校能耗预测。As a pillar industry of the national economy,the construction industry consumes a large amount of energy while promoting social development,accelerating the energy conservation and emission reduction in the construction industry is of great significance to realize China’s energy conservation and emission reduction target of achieving peak CO_(2) emissions in 2030.Therefore,this paper analyzes the building energy consumption of universities with relatively concentrated personnel,and through collecting a large number of building energy consumption data on campus,fully excavates the characteristics of building energy consumption,and optimizes the BP and LSTM building energy consumption prediction models with genetic algorithm,through experimental comparison,the most suitable prediction model on campus is selected to predict the future use of building energy consumption on campus,and provide basic data for the subsequent carbon emission prediction and energy saving measures on campus.Through experimental comparison,it can be found that compared with the GA-LSTM model,the GA-BP model has a mean absolute error,mean square error,root mean square error,and mean absolute percentage error reduction of 8.83%,4.80%,2.43%,and 10.2%,respectively,the GA-BP model has smaller errors and higher degree of fitting,which is more realistic for the prediction of building energy consumption in universities,and can be used for the prediction of energy consumption in universities.

关 键 词：遗传算法 GA-BP模型 建筑能耗 建筑节能 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]