Development of a Python-Based Tool for Correcting Pressure-Induced Errors in Fiber-Optic Temperature Measurements in Fuel Cell Stacks

Abstract

In the area of sustainable mobility, proton exchange membrane (PEM) fuel cells are preferred choice because of their exceptional efficiency and minimum emissions. Nevertheless, inhomogenous thermal gradients abysmally impact their performance in the long run which leads to formation of local hotspots that potentially accelerate material degradation of the fuel cell membrane and reduced electrochemical efficiency. Despite the critical importance of temperature management, there is currently no economically viable and spatially resolved method capable of accurately monitoring surface temperature distributions across the active area of PEMFCs under realistic operating conditions. To address this issue, a temperature analysis tool is developed to characterize the temperature behavior of DLR FCCP under controlled load points using both static and time-resolved approaches. The static analysis evaluates statistical temperature metrics, including mean temperature, pulse height, accuracy error, spatial uniformity root error (RMSE), inter-pulse standard deviation, and confidence intervals. The time-resolved analysis broaden these metrics across successive time frames, enabling visualization of transient temperature behavior and identification of fiber locations exhibiting elevated heating rates within the active region. The proposed methodology provides spatially resolved heatmaps of the fuel cell active region and demonstrates the capability to detect non-uniform temperature distributions and localized hotspots with high sensitivity. These results highlight the potential of the developed tool as a cost-effective solution for in-situ temperature diagnostics, supporting improved thermal management strategies and enhanced durability of PEM fuel cells.