基于PCA-PSO-BP神经网络的住宅供热逐时负荷预测  

Hourly heating load forecasting for residences based on PCA-PSO-BP neural network

作　　者：王新雨 郭振伟[1] 于丹[2] 刘益民 崔治国 Wang Xinyu;Guo Zhenwei;Yu Dan;Liu Yimin;Cui Zhiguo(Chinese Society for Urban Studies,Beijing;Beijing University of Civil Engineering and Architecture,Beijing;China Academy of Building Research Co.,Ltd.,Beijing)

机构地区：[1]中国城市科学研究会,北京 [2]北京建筑大学,北京 [3]中国建筑科学研究院有限公司,北京

出　　处：《暖通空调》2023年第3期138-142,160,共6页Heating Ventilating & Air Conditioning

基　　金：中国建筑科学研究院青年科研基金项目“建筑供热系统仿真及控制策略评价工具的研究与开发”(编号:20200109331030019)。

摘　　要：为了准确预测供热负荷,提出了一种基于主成分分析法和粒子群优化算法改进的BP神经网络(PCA-PSO-BP)预测模型。首先利用主成分分析法融合影响热负荷的特征指标,消除指标之间的冗余性和相关性;同时采用粒子群算法优化BP神经网络的初始权值和阈值,克服了BP神经网络容易陷入局部最优的缺陷,提高了BP神经网络的预测精度。基于北京某居住建筑供热系统的实际运行数据,对模型的性能进行了验证。仿真结果表明,改进的模型预测精度提高了4.07%。In order to accurately forecast the heating load,an improved BP neural network forecasting model based on principal component analysis(PCA)and particle swarm optimization(PSO)is proposed.Firstly,the PCA is used to fuse the characteristic indexes affecting the heating load to eliminate the redundancy and correlation between the indexes.At the same time,the PSO is used to optimize the initial weights and thresholds of BP neural network,which overcomes the defect that BP neural network is easy to fall into local optimization and improves the prediction accuracy of BP neural network.Based on the actual operation data of the heating system of a residential building in Beijing,the performance of the model is verified.The simulation results show that the forecasting accuracy of the improved model is improved by 4.07%.

关 键 词：供热负荷 预测模型 BP神经网络 主成分分析(PCA) 粒子群算法(PSO) 特征变量 

分 类 号：TU995[建筑科学—供热、供燃气、通风及空调工程] TP183[建筑科学—市政工程]