基于LMBP神经网络的建筑能耗预测.pdf

第25 卷 第6 期 计算机技术与发展 Vol． 25 No． 6 2015 年6 月 COMPUTEＲ TECHNOLOGY AND DEVELOPMENT June 2015 基于LMBP 神经网络的建筑能耗预测 路 阔，钟伯成 (上海工程技术大学电子电气工程学院，上海20 1620) 摘 要:建筑能耗短期预测对实时性要求较高，传统神经网络存在收敛速度慢的缺点。为此，采用LM 算法改进标准BP 神 经网络，建立了基于LM 算法的建筑能耗预测模型。首先通过理论说明该算法的先进性，然后设计一套建筑能耗数据采集 系统和建立基于LMBP 神经网络的建筑能耗预测模型，最后采集某建筑一个月的整点电量作为预测模型的实验数据。实 验结果表明，该模型明显提高了训练速度，且预测精度满足实际需求，说明了LMBP 神经网络适用于建筑能耗短期预测。 关键词:建筑能耗;数据采集;短期预测;神经网络;BP 算法;LM 算法 中图分类号:TP391． 9 文献标识码:A 文章编号:1673 － 629X (2015)06 － 0216 － 03 doi :10 ． 3969 /j ． issn． 1673 － 629X ． 2015 ． 06 ． 048 Building Energy Consumption Prediction Based on LMBP Neural Network LU Kuo ，Z HONG Bo － cheng (College of Electronic ＆ Electrical Engineering ，Shanghai University of Engineering Science ， Shanghai 201620 ，China) Abstract :The traditional neural netw ork is too slow in term of convergence speed to meet the high real － time requirements of short － term prediction of building energy consumption． Therefore ，LM algorithm is adopted instead of conventional BP algorithm to establish the building energy consumption model． Firstly through theoretical description of the advanced algorithm ，then design a set of data acquisition sy stem to monitor building energy consumption and set up the prediction model based on LM BP neural netw ork． Finally a building ’s 24 － hour pow er consumption data for one month is collected by the data acquisition sy stem as the experimental samples to verify the mod- el． Empirical results show that the LM BP neural netw ork prediction model sig