摘要
针对节能管理方面面临内部能耗状况无法准确计量、缺乏科学的能耗指标和评价标准、能耗管理考虑不够等问题,本文建立了MEA-BP神经网络能耗预测模型。基于能耗监管平台输出的历史能耗数据,对BP神经网络与MEA-BP神经网络进行训练,使用训练后的模型对建筑能耗进行预测,对比能耗预测值与能耗真实值,发现MEA-BP模型的能耗预测结果精度较高,有实际应用价值。根据能耗预测值,通过能耗监测平台对用能设施、用能方式进行相应控制与改善,提升了水、电、暖、气的节能效率,实现了最优的能耗控制和能源有效利用。
In order to solve the problems of energy saving management,such as poor measurement accuracy of internal energy consumption, inconsistency of scientific energy consumption index and evaluation standard,and insufficient level of energy consumption management,the MEA-BP neural network predicting the building energy consumption is established. Based on the historical energy consumption data output from the energy consumption supervision platform,the BP neural network and MEA-BP neural network were trained. By comparing the actual value with predicted value on energy consumption,it is found that MEA-BP model is more accurate in predicting energy consumption. The MEA-BP model combined with energy consumption supervision platform could improve the energy saving efficiency about water,electricity,heating and gas,achieving the optimal energy consumption control and efficient use of energy.
作者
滕文龙
丛炳虎
商云坤
张予宸
白天
TENG Wen-long;CONG Bing-hu;SHANG Yun-kun;ZHANG Yu-chen;BAI Tian(The First Hospital of Jilin University,Changchun 130021,China;College of Computer Science and Technology,Jilin University,Changchun 130021,China)
出处
《吉林大学学报(工学版)》
EI
CAS
CSCD
北大核心
2021年第5期1857-1865,共9页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(61702214)
吉林省科技厅自由探索重点项目(YDZJ202101ZYTS128)
吉林省科技厅国际合作项目(20200801033GH).
关键词
公共建筑
神经网络
能耗监测平台
能耗数据
能耗预测
public buildings
neural network
energy consumption supervision platform
data of energy consumption
energy consumption prediction