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

Area of Science:

• Energy Science
• Materials Science
• Computational Intelligence

Background:

• Proton Exchange Membrane (PEM) fuel cells (FCs) are crucial for sustainable energy due to their high efficiency and environmental benefits.
• Enhancing PEM-FC output power is vital for their widespread industrial adoption and transition to clean energy.
• Accurate modeling and optimization are key to maximizing PEM-FC performance.

Purpose of the Study:

• To improve the output power density of PEM-FCs.
• To develop and validate an Adaptive Neuro-Fuzzy Inference System (ANFIS) model for PEM-FC simulation.
• To optimize PEM-FC input parameters using modern metaheuristic algorithms, particularly the Salp Swarm Algorithm (SSA).

Main Methods:

• An ANFIS model was developed using empirical data to simulate PEM-FC output power density.
• Key factors like pressure, relative humidity, and membrane compression were considered in the ANFIS model.
• The Salp Swarm Algorithm (SSA) was employed to optimize three input control parameters for maximum power density, with comparisons to Particle Swarm Optimization (PSO), Evolutionary Optimization (EO), and Grey Wolf Optimizer (GWO).

Main Results:

• The ANFIS model demonstrated high accuracy, with training and testing Root Mean Square Error (RMSE) values of 0.0003 and 24.5, respectively.
• Coefficient of determination values for training and testing were 1.0 and 0.9598, confirming the model's success.
• SSA yielded the highest average power density (716.63 mW/cm²), outperforming GWO (709.95 mW/cm²) and PSO (695.27 mW/cm²).

Conclusions:

• The study successfully enhanced PEM-FC output power density through ANFIS modeling and SSA optimization.
• SSA proved to be a highly effective algorithm for optimizing PEM-FC operational parameters.
• The findings support the advancement of PEM-FC technology for efficient and clean energy generation.