Development on Diagnosing Method of Metallic Micro-particles Buried into Fuel Cell Membrane Electrode Assemblies using Electromagnetic Field Excited Oscillation

- Automatic Diagnosing System based on Machine Learning

Abstract

In the manufacturing process of Membrane Electrode Assemblies (MEA) of fuel cell, there remains a problem of metallic micro-particles buried into Gas Diffusion Layer (GDL). It leads to the performance drop-down of fuel cells. In this study, we propose a real-time diagnosing system introducing Machine Learning to EMA-LDS, which is composed of electro-magnetic impact generators and laser displacement sensors. EMA-LDS was experimentally applied to both samples of gas diffusion layer (GDL) with or without burying Fe micro-particle with a diameter of 100 μm. As an experimental result, the contaminated GDL was estimated to have a high probability of metallic micro-particles, although the normal GDL had lower probability. In conclusions, the proposed method can discriminate metallic micro-particles according to Machine learning. Therefore, EMA-LDS has an effective potential as a diagnosing system of metallic microparticle into MEA.