Deep learning artificial intelligence framework for sustainable desiccant air conditioning system: Optimization towards reduction in water footprints

Desiccant evaporative cooling systems pave the path towards energy and environmental sustainability in buildings especially; however, direct coolers such configurations result high water consumption. The application of modern computational intelligence tools, including artificial meta-heuristic optimization algorithms, can improve operational comprehension desiccant while addressing minimization total footprints with maximization capacity. contribution/objective this research is to address gaps understanding through deep learning, genetic algorithm, multicriteria decision analysis applied a system working under real transient experimental conditions building located Austria. Within methodology, calibrated, experimental, validated data monitoring displaying desiccant-enhanced adapted generate set input-output sets. includes ambient temperature, humidity, regeneration supply airflow rate, return rate yielding capacity system. results learning algorithm using an neural network have suggested that architectures 5-[6]-[6]-1 5-[12]-[12]-1 are best accurately predict coefficient determination 0.98856 0.99246, respectively. Secondly, “white-box model” used develop digital twin model which helps replication earlier conditions. optimized 45.17 kg/h 3.32 tons refrigeration. These optimal values found combination design variables temperature 28 °C, relative humidity 52.0%, 2.13 kg/s, flow 2.35 70.0 °C. It concluded data-driven models extend interpretation participate its performance enhancement.