Optimization of an NH3-H2O Absorption Cooling System Using an Inverted Multivariate Function with Neural Networks and PSO

Absorption systems offer a practical alternative to traditional compression systems, especially when low-grade heat sources are available. Their applications range from vaccine preservation to space conditioning, making performance optimization essential. This study employed a multivariate inverse artificial neural network with multiple parameters (ANNim-mp) to simultaneously enhance the cooling load and coefficient of performance in an experimental single-effect ammonia–water absorption cooling system. Optimization was carried out using particle swarm optimization. The results showed significant performance improvements: up to 100% in cooling load and 97% in COP when optimizing two variables. With four-variable optimization, improvements reached 98.7% and 106.7%, respectively. These results demonstrate the strong potential of the ANNim-mp approach in enhancing the efficiency of absorption cooling systems.