Performance analysis of a liquid desiccant air conditioning system based on a data-driven model

Liquid desiccant dehumidification system (LDDS) has emerged as an energy-efficient approach for air dehumidification. In this study, an adaptive RBF neural network model, which describe the relationship between the system inlet and outlet variables, for the desiccant regeneration subsystem in LDDS is proposed. The air, desiccant solution, hot water inlet temperature and flow rate, air inlet humidity and desiccant solution inlet concentration are chosen as the model inputs. The model outputs are the regenerator outlet air temperature and humidity. The developed models are validated, and the model predicting errors are within ±8%. Moreover, the moisture removal rate and its effectiveness are adopted as system performance indices, the effects of system inlet parameters on regeneration performance are also investigated. The change trend of system performance, which affected by the system inlet parameters can guide the system operation and control.