基于长短期记忆神经网络的暖通空调系统能耗预测  

Energy Consumption Prediction of HVAC System Based on Long Short-term Memory Neural Network

作　　者：廖文强 王江宇[2] 陈焕新[2] 丁新磊 尚鹏涛 魏文天 周镇新 LIAO Wenqiang;WANG Jiangyu;CHEN Huanxin;DING Xinlei;SHANG Pengtao;WEI Wentian;ZHOU Zhenxin(China-EU Institute for Clean and Renewable Energy at Huazhong University of Science and Technology,Wuhan,Hubei 430074,China;School of Energy and Power Engineering,Huazhong University of Science and Technology,Wuhan,Hubei 430074,China)

机构地区：[1]华中科技大学中欧清洁与可再生能源学院,湖北武汉430074 [2]华中科技大学能源与动力工程学院,湖北武汉430074

出　　处：《制冷技术》2019年第1期45-50,54,共7页Chinese Journal of Refrigeration Technology

基　　金：国家自然科学基金(No.51876070;No.51576074)

摘　　要：建筑系统的能源消耗中,暖通空调系统能耗占大部分。降低暖通空调系统(HVAC)的能耗量对实现建筑节能具有重大意义。通过对暖通空调未来短期能耗进行预测,调整系统运行模式,可以实现有效的能耗降低。本研究使用了一种基于长短期记忆神经网络(Long Short-term Memory,LSTM)的暖通空调系统能耗预测方法,对某地供暖系统的能耗进行预测,将预测结果与真实值进行对比。最终结果表明,LSTM预测模型相比传统的预测方法效果更好。In the energy consumption of building systems, the energy consumption of heating, ventilation and air conditioning (HVAC) systems has always occupied a considerable part. Reducing the energy consumption of HVAC systems is of great significance for achieving building energy efficiency. By predicting the future short-term energy consumption of HVAC and adjusting the system operation mode, effective energy consumption can be reduced. For this purpose, a long short-term memory neural network (LSTM) based on HVAC system energy consumption prediction method is used to predict the energy consumption of a certain heating system, and the prediction results compare with the real value. Final results show that the LSTM prediction model has better prediction effect than the traditional one.

关 键 词：长短期记忆神经网络 空调能耗预测 深度学习 

分 类 号：TU83[建筑科学—供热、供燃气、通风及空调工程] TU111.195