Prediction of the air temperature and humidity at the outlet of a cooling coil using neural networks

The main objective of this study is to predict air temperature and humidity at the outlet of a wire-on-tube type heat exchanger using neural networks. For this purpose, initially the heat exchanger was coupled to a refrigeration unit and placed in a wind tunnel. Afterwards, its performance was tested under various experimental conditions. We measured nine input parameters, namely, temperature and humidity of the air entering the coil, air velocity, frost weight, the temperature at the coil surface, mass flow rate of the heat transfer fluid and its temperatures at the inlet and outlet of the coil along with ambient temperature. Additionally, we measured temperature and humidity of the air leaving the coil as the output parameters. Then, a feed-forward neural network based on backpropagation algorithm was developed to model the thermal performance of the coil. The artificial neural network (ANN) was trained using the experimental data to predict the air conditions at the outlet of the coil. The predicted values are found to be in good agreement with the actual values from the experiments with mean relative errors less than 1% for outlet air temperature and 2% for outlet humidity. This demonstrates that the neural network presented can help the manufacturer predict the performance of cooling coils in air-conditioning systems under various operating conditions.