Design of an aging prognostic model to predict failure of a PEM

Abstract

Proton Exchange Membrane Fuel Cells (PEMFCs) are promising clean energy technologies for transportation and stationary power applications; however, their large-scale commercialization is limited by durability and lifetime uncertainties arising from complex degradation mechanisms. This study presents the development of a data-driven aging prognostic framework for PEMFC systems using MATLAB, integrating voltage degradation, impedance evolution, neural network prediction, and Remaining Useful Life (RUL) estimation. A feedforward neural network trained with the Levenberg–Marquardt algorithm was employed to model the nonlinear degradation behavior of the fuel cell membrane using historical voltage and impedance data. Simulation results demonstrate strong predictive performance, with the trained model achieving close agreement between actual and predicted voltage profiles, as indicated by a regression coefficient approaching unity and prediction errors concentrated near zero. Over a 10-hour operating period, the PEMFC voltage exhibited an initial rise from approximately 1.20 V to 1.25 V due to membrane