基于神经网络的暖通空调控制系统时间迟延的确定

Determination of Time Delay in HVAC Systems Using ANN

本文提出一种人工神经网络（ＡＮＮ）方法，确定暖通空调（ＨＶＡＣ）系统对控制信号响应的时间迟延，在研究中选用了四层网络。在网络学习过程中采用了一种加速方法，以改善梯度下降法。对制热和制冷两种工作方式的实验数据，采用ＡＮＮ和传统算法两种方法确定时间迟延，并进行比较。结果表明ＡＮＮ可以有效地用于确定ＨＶＡＣ系统的时间迟延。

This paper presents an artificial neural network(ANN) approach to determine the time delay for a heating, ventilating, and air -conditioning(HVAC)plant to respond to control actions. The ANN is a fully connected four-layer network. An acceleration technique was used to improve the General Delta Rule for the learning process. Experimental data for heatiing and cooling modes were used with both the ANN and a traditional mathematical method to determine the time delay, The results show that ANNs can be used to effectively determine the time delay for HVAC systems.