Design and performance prediction of desiccant coated heat exchanger using ANFIS-AI tool and dynamic model

To improve indoor air quality, enhance the energy exchange capabilities and minimize the spread of microbial pollutants, desiccant coated heat exchanger seems a promising alternative to conventional heat exchangers such as rotary wheel and adsorber beds. Thus, precise prediction of the design and performance characteristics of desiccant coated heat exchanger is vital for improving the overall air conditioning system performance. Therefore, in the present investigation, an adaptive neuro-fuzzy inference system with artificial neural network fuzzy logic has been implemented to predict the exit parameters of desiccant coated heat exchangers. The advantage of an adaptive neuro-fuzzy inference system is its structure that combines the neural network's learning ability, excellent information representation capability of fuzzy logic, and adaptive algorithm. Silica gel is taken as the coated desiccant. The impact of inlet and design parameters on the performance characteristics of desiccant coated heat exchanger during dehumidification has been evaluated using the artificial intelligence adaptive neuro-fuzzy inference system. Also, the optimal design/inlet conditions for the given operating and design parametric range are obtained. For the obtained optimal conditions, the maximum valves of dehumidification capacity, vapor pressure difference ratio, and sensible energy load ratio are 2.01 kg/h, 0.84, and 0.135, respectively. Further, a dynamic model based on the Laplace transform is developed to assess desiccant coated heat exchangers' adsorption kinetics and thermal effects. Moreover, the interactive effect of Lewis number and Stanton number on the performance parameters has been examined judiciously. Adsorption kinetics of silica gel shows that the water adsorped per mass desiccant is about 0.1129 g/g at the end of sorption time.

Automated summary

In order to improve indoor air quality, accurately predicting and optimizing the design and performance characteristics of desiccant coated heat exchangers is crucial. This study utilizes an adaptive neuro-fuzzy inference system to investigate the exit parameters of desiccant coated heat exchangers, as well as factors such as adsorption kinetics and thermal effects.