Mrinal Bhowmik, B. Kiran Naik, P. Muthukumar, R. Anandalakshmi,
Performance assessment and optimization of liquid desiccant dehumidifier system using intelligent
models and integration with solar dryer,
Journal of Building Engineering,
Volume 64,
2023,
105577,
ISSN 2352-7102,
https://doi.org/10.1016/j.jobe.2022.105577.
(https://www.sciencedirect.com/science/article/pii/S2352710222015832)
Abstract: Realistic performance predictions are necessary to control the liquid desiccant
dehumidification systems effectively. In the present study, the application of artificial intelligence (AI)
based artificial neural network (ANN), gene expression program (GEP) and adaptive neuro-fuzzy
inference system (ANFIS) is investigated to estimate the dehumidifier effectiveness of desiccant
dehumidifier systems while considering the effect of moisture removal rate and sensible heat factor as
performance characteristics. It is found that the AI-based GEP model has the best prediction capability
compared to the other developed AI models. In addition, the sensitivity analysis of independent
parameters on the system performance parameters is estimated using the cosine amplitude method.
The results demonstrate that the inlet desiccant temperature and specific humidity have a more
substantial influence on the dehumidifier effectiveness and sensible heat ratio. Further, an algorithm
involving the combination of multi-objective particle swarm optimization (MOPSO) with the GEP model
is developed to optimize the input process parameters and to achieve better dehumidification
performance. Lastly, based on obtained optimal operating conditions, a case study on a dehumidifierintegrated solar dryer is proposed for food/agriculture products drying applications using AI-based GEPMOPSO model. It is also observed that the inlet water temperature and solar intensity have a significant
impact on the useful heat gain of the solar collector.
Keywords: Dehumidifier; Artificial intelligent models; Moisture removal rate; Dehumidifier integrated
solar dryer; Multi-objective optimization