Adaptive hybrid ANFIS-PSO approach for Indirect Evaporative Cooling Performance Prediction

Nowadays, Sustainable strategies for reducing energy consumption are vital for green development, with indirect evaporative cooling (IEC) systems offering an efficient alternative for maintaining indoor thermal comfort. This study focuses on modeling the performance of an IEC energy recovery counter flow system using intelligent neuro-fuzzy techniques. Experimental data, collected under typical operating conditions in an air conditioning laboratory, were enhanced with feature engineering based on inlet air temperature, relative humidity, and air velocity. Two advanced models, standard Adaptive Neuro-Fuzzy Inference System (ANFIS) and ANFIS optimized by Particle Swarm Optimization (ANFIS-PSO) were developed and compared. Results demonstrate that ANFIS-PSO significantly outperforms the classic ANFIS, achieving a test R2 of 0.9916, MAE of 0.0151, and RMSE of 0.0212. In contrast, the standard ANFIS achieved a test R2 of 0.8136, MAE of 0.0666, and RMSE of 0.0999. These findings highlight the effectiveness of PSO-based optimization in improving model accuracy, emphasizing the practical value of the ANFIS-PSO model for efficient IEC system design and operation in demanding climates.