Abstract

Energy consumption in buildings is either directly or indirectly related to HVAC systems. As buildings increase in size their corresponding cooling loads also increase, which leads to the necessity of having evermore efficient cooling systems. Control systems and energy management strategies play an important role in improving the overall performance of the system, also allowing the reduction of energy consumption with existing equipment. However the design and implementation of these strategies is not trivial. It is the purpose of this study to develop and implement a series of neural networks (NNs) for energy management control strategies (EMCS) in a model. These Neuro EMCS are: start and stop lead times, temperature base economy cycle, supply air and water temperature reset, and gain selector for the PI controller. To determine the potential benefits and possible energy savings a comparison is made between the developed NNs and a previously developed and tested EMCS algorithms (Base Case). The results show that the adaptive NNs perform very well and as such they are considered as good candidates for implementation in real building systems.