Local and global control of commercial building HVAC systems using artificial neural networks

Abstract

Discusses the results of a proof of concept experiment in which artificial neural networks (ANNs) are used for both local and global control of a commercial building HVAC system. The experiment was conducted in a laboratory in which full scale testing of multizone HVAC systems can be done repeatably. Data collected in the laboratory were used to train ANNs for the components and systems involved. For the local controller the model itself is a neural network that has been trained on actual data from a heating coil. For the global controller experiment a two zone HVAC system was used. This paper is presented in two parts. The first deals with the local control experiment whereas the second part describes the global control experiment.