旨在构建基于遗传BP神经网络的严寒地区办公建筑采暖能耗预测模型;通过分析建筑热工计算模型,确定影响寒地办公建筑能耗的建筑参量,应用Grasshopper平台建构能耗模拟参数模型,获得2000组能耗模拟数据,应用遗传BP神经网络建模技术,以建筑形态参量为输入参量,以耗热量指标为输出参量,建构严寒地区办公建筑采暖能耗预测模型,应用模拟数据进行网络训练与测试。 The purpose of this paper is to build a forecasting model of heating energy consumption for office buildings in severe cold area based on genetic BP neural network. By analyzing the building thermal calculation model and determining the building parameters that affect the energy consumption of office buildings in the cold area, the Grasshopper platform is used to construct the energy consumption simulation parameter model 2000 sets of energy consumption simulation data, the application of genetic BP neural network modeling technology to building form parameters for the input parameters to heat consumption indicators as output parameters to build office buildings in harsh winter heating energy consumption prediction model, the application of simulation data for network training and test.