一种面向大规模建筑群能耗预测的BP神经网络模型配置方法

针对大规模建筑群的能耗预测问题,提出了一种基于BP神经网络的模型配置方法。该方法将复杂的建筑能耗预测相关的影响因素、参数、算法与建筑能耗实际预测功能进行解耦分离,实现任意预测模型的可配置化创建,有效地降低了建筑能耗预测系统的开发和维护成本,使得大规模建筑群的能耗预测系统易于实现。目前,可配置化的建筑能耗预测系统已被应用于多栋建筑的能耗预测,有效地降低了能源浪费。

A Model Configuration Method Based on BP Neural Network for the Prediction of Multi-Functional Buildings' Energy Consumption

In order to solve the prediction of multi-functional buildings'energy consumption,thispaper proposed a model configuration method based on BP neural network, which was to decouple the influent factors, parameters and algorithms from the actual prediction functions. The method could easily create one prediction model of building energy consumption, and greatly reduced the software system development and maintenance's costs, ensuring the prediction of multi-functional buildings' energy consumption achieved easily. Now, this method has been used to develop the prediction system of building energy consumption, which has been implemented for some buildings and help to reduce energy waste effectively.