Boosting Power Density of Proton Exchange Membrane Fuel Cell Using Artificial Intelligence and Optimization Algorithms.

Autor: Ghoniem RM; Department of Information Technology, College of Computer and Information Sciences, Princess Nourah bint Abdulrahman University, Riyadh 11671, Saudi Arabia., Wilberforce T; Department of Engineering, Faculty of Natural, Mathematical & Engineering Sciences, King's College London, London WC2R 2LS, UK., Rezk H; Department of Electrical Engineering, College of Engineering in Wadi Alddawasir, Prince Sattam bin Abdulaziz University, Riyadh 11942, Saudi Arabia.; Department of Electrical Engineering, Faculty of Engineering, Minia University, Elminia 61519, Egypt., As'ad S; Renewable Energy Engineering Department, Faculty of Engineering, Middle East University, Amman 11831, Jordan., Alahmer A; Department of Mechanical Engineering, Tuskegee University, Tuskegee, AL 36088, USA.; Department of Mechanical Engineering, Faculty of Engineering, Tafila Technical University, Tafila 66110, Jordan.
Jazyk: angličtina
Zdroj: Membranes [Membranes (Basel)] 2023 Sep 28; Vol. 13 (10). Date of Electronic Publication: 2023 Sep 28.
DOI: 10.3390/membranes13100817
Abstrakt: The adoption of Proton Exchange Membrane (PEM) fuel cells (FCs) is of great significance in diverse industries, as they provide high efficiency and environmental advantages, enabling the transition to sustainable and clean energy solutions. This study aims to enhance the output power of PEM-FCs by employing the Adaptive Neuro-Fuzzy Inference System (ANFIS) and modern optimization algorithms. Initially, an ANFIS model is developed based on empirical data to simulate the output power density of the PEM-FC, considering factors such as pressure, relative humidity, and membrane compression. The Salp swarm algorithm (SSA) is subsequently utilized to determine the optimal values of the input control parameters. The three input control parameters of the PEM-FC are treated as decision variables during the optimization process, with the objective to maximize the output power density. During the modeling phase, the training and testing data exhibit root mean square error (RMSE) values of 0.0003 and 24.5, respectively. The coefficient of determination values for training and testing are 1.0 and 0.9598, respectively, indicating the successfulness of the modeling process. The reliability of SSA is further validated by comparing its outcomes with those obtained from particle swarm optimization (PSO), evolutionary optimization (EO), and grey wolf optimizer (GWO). Among these methods, SSA achieves the highest average power density of 716.63 mW/cm 2 , followed by GWO at 709.95 mW/cm 2 . The lowest average power density of 695.27 mW/cm 2 is obtained using PSO.
Databáze: MEDLINE
Nepřihlášeným uživatelům se plný text nezobrazuje