基于OOA-BP的短期空调冷负荷预测  

Short-Term Air-Conditioning Cooling Load Prediction Based on OOA-BP

作　　者：洪方伟 戴石良 熊静 HONG Fangwei;DAI Shiliang;XIONG Jing(School of Civil Engineering, University of South China)

机构地区：[1]南华大学土木工程学院

出　　处：《上海节能》2025年第4期591-598,共8页Shanghai Energy Saving

摘　　要：建筑空调冷负荷预测对于提前调整制冷站设备参数,降低卷烟厂中央空调系统能耗具有十分重要的意义。通过DesignBuilder负荷模拟软件建立了卷烟厂联合工房的建筑模型,完成了相关参数的设置,得到了全年逐时冷负荷数据。采用灰色关联度分析法筛选出对空调冷负荷影响较大的因素作为预测模型输入,在Matlab里建立BP、OOA-BP两种空调冷负荷预测模型,采用了RMSE、MAE、MAPE、MSE四项误差评价指标。仿真结果表明,OOA-BP相较于BP负荷预测模型RMSE降低了33.1%,MAE降低了42.22%,MAPE降低了40.6%,MSE降低了55.24%。基于OOA-BP负荷预测模型精度较传统的BP神经网络模型精度有了较大的提高,具有一定的实际应用价值。Forecasting the cooling load of building air conditioning is of great significance for adjusting the equipment parameters of refrigeration stations in advance and reducing the energy consumption of central air conditioning systems in cigarette factories.The architectural model of the joint workshop of the cigarette factory was established through DesignBuilder load simulation software,relevant parameters were set,and hour-by-hour cooling load data were obtained throughout the year.The gray correlation analysis method is used to screen out factors that have a greater impact on the air conditioning cooling load as input to the prediction model.Two air conditioning cooling load prediction models,BP and OOA-BP,are established in Matlab,and four error evaluation indicators,RMSE,MAE,MAPE,and MSE,are used.Simulation results show that compared with BP load prediction model,OOA-BP reduces RMSE by 33.1%,MAE by 42.22%,MAPE by 40.6%,and MSE by 55.24%.The accuracy of the OOA-BP load forecasting model is greatly improved compared with the traditional BP neural network model,and has certain practical application value.

关 键 词：负荷预测 DesignBuilder 鱼鹰算法 人工神经网络 

分 类 号：TU831[建筑科学—供热、供燃气、通风及空调工程]