基于气象热舒适度的建筑能耗灰色神经网络预测

针对建筑能耗受局地气候多因素影响的特点,为了客观准确地对建筑能耗进行预测,本文引入了气象热舒适度来综合分析气候对建筑能耗的影响,并以该指数预测值、建筑能耗原始数据和日期类型作为输入层,进行复合灰色神经网络模型预测建筑能耗。该方法不仅克服了灰色模型和神经网络存在的预测缺陷,同时还考虑了气象因素对建筑能耗的影响。通过对北京某大厦的实例应用分析,取得了较高精度的预测结果,证实了该方法的合理可靠,为建筑能耗预测提供了新途径,其预测结果也将为大型建筑空调系统的再优化设计和改造提供参考。

Prediction of Building Energy Consumption with Grey Neural Network Method based on Weather Thermal Comfort

In order to predict the building energy consumption(BEC) which is influenced by the local weather conditions,the parameter of weather thermal comfort(WTC) was introduced to analyze BEC.Furthermore,the prediction value of WTC,original data of BEC,and date type were taken as input cells for grey neural network method to carry out the prediction.On one hand,this method could overcome the shortcomings of grey model and neural network in forecast process,on the other hand,the effects of weather conditions were taken into account. Based on the discussion of a real project in Beijing,the relatively accurate prediction results were obtained and the dependability of this method was proved. As a result,this method could be considered as a new strategy to predict BEC, and corresponding references for retrofitting and re-optimization design of air-conditioning systems in large-scale buildings could be provided.