MODELLING AND ANALYSIS OF AIR-COOLED CHILLERS WITH WATER MIST SYSTEM USING NEURAL NETWORKS

Artificial neural network (ANN) is receiving ever-increasing attention in the highly complex HVAC systems. The traditional approaches are not robust enough to handle the increased complexity. This study deals with predicting the performance of air-cooled chillers with water mist system using ANN techniques. The efficiency of air-cooled chillers can be improved by applying water mist to pre-cool ambient air entering the condensers to decrease compressor power. In this research, the air-cooled chiller plant coupling water mist system of an institutional building in Hong Kong was investigated with ANN. The dry-bulb temperature and relative humidity of ambient air, the dry-bulb temperature and relative humidity of the air entering and leaving the condenser coils, the condensing and evaporative temperature, chiller cooling load, chiller power, water mist generation rate and some other related parameters were measured under a wide range of operating conditions. ANN model for the air-cooled chillers with water mist system was developed and evaluated with the experimental data. The ANN model was used for predicting the performance of such chiller systems. This study reveals that neural networks can predict the performance of the air-cooled chillers with water mist system with a high degree of accuracy and can be used as an on-line optimization process tool.