Artificial neural network analysis of liquid desiccant regenerator performance in a solar hybrid air-conditioning system

In this paper, experimental tests are carried out to investigate the performance of a counter flow regenerator using lithium chloride (LiCl) solution as the desiccant. A single and multilayer artificial neural network (ANN) is used to predict the performance of the regenerator. Five parameters are used as inputs to the ANN, namely: air and desiccant flow rates, air inlet humidity ratio, and air and desiccant inlet temperatures. The outputs of the ANN are the temperature, humidity ratio, moisture removal rate (MRR), and the effectiveness. ANN predictions for these parameters are compared with the experimental values. The results show that the optimum testing model for MRR in the regenerator was the 5-5-5-1 structure with R2 = 0.93, whereas the optimum testing model for effectiveness was the 5-11-1 structure with R2 = 0.95. The maximum temperature and humidity ratio difference between the ANN model and experimental are 1.4 °C and 2.1 g/kg, respectively. The MRR and effectiveness of regenerator increase slowly as function of air inlet temperature. It was found that the MRR and effectiveness increased about 0.79% and 1.1%, respectively. The moisture removal rate decreased with increasing air inlet humidity ratio and increased with desiccant inlet temperature.