KIER Develops an AI Analysis Model to Identify Fuel Cell Failure Causes

2025-01-29

A research team led by Dr. Chi-Young Jung at the Hydrogen Research & Demonstration Center of the Korea Institute of Energy Research (KIER) has developed a novel technique for rapid analysis of the microstructure of carbon fiber paper, a key material in hydrogen fuel cells, leveraging virtual space and artificial intelligence (AI) learning.

Carbon fiber paper plays a crucial role in facilitating water discharge and fuel supply within hydrogen fuel cells. Consequently, analyzing the microstructure of carbon fiber paper is essential for diagnosing the condition of fuel cells. However, conventional methods involve destructive analysis of carbon fiber paper samples followed by microscopic examination using electron microscopy. This approach is time-consuming and precludes real-time analysis.

To overcome these limitations, the research team has developed a technique that utilizes X-ray diagnostics and an AI-based image learning model for analyzing the microstructure of carbon fiber paper. For this purpose, approximately 5,000 images extracted from over 200 carbon fiber paper samples were used to train a machine learning algorithm. Furthermore, by utilizing data from the developed model, the research team elucidated the influence of design parameters such as carbon fiber paper thickness and binder content on fuel cell performance. This innovative technique enables near real-time condition diagnosis through precise analysis using only X-ray computed tomography, eliminating the need for electron microscopy.

Source: https://www.etnews.com/20241219000148