Thermodynamic analysis of an open cycle solid desiccant cooling system using Artificial Neural Network

This paper examines the performance of an installed open cycle air-conditioning system with a silica gel desiccant wheel which uses a conventional heat pump and heat exchangers for the improvement of the outlet air of the system. A neural network model based on the training of a black box model with experimental data was developed as a method based on experimental results predicting the state conditions of air at the process and regeneration stream. The model development was followed by a Sensitivity Analysis performed on these predicted results. The key parameters were the thermodynamic condition of process and regeneration air streams, the sensible heat factor of the room, and the mass air flow ratio of the regeneration and process streams. The results of this analysis revealed that all investigated parameters influenced the performance of the desiccant unit. Predicted output values of the proposed Neural Network Model for Desiccant Systems are in line with results from other correlation models based on the interpolation of experimental data obtained from industrial air conditioning installations.

► A neural network model based on experimental data was developed.
► Description of the experimental setup.
► Prediction of the state conditions of air at the process and regeneration stream.
► Sensitivity Analysis performed on these predicted results.
► Predicted output values in line with correlation model based on data from industry.