Visualization of Micro Damage in SOFC Using Terahertz Waves

SOFC effective area for power generation is reduced when delamination occurs. Hence, non-destructive evaluation methods for detecting delamination of SOFC are required. We focused on Terahertz-Time Domain Spectroscopy (THz-TDS) as a SOFCs’ non-destructive evaluation method. THz-TDS has been used as an effective and safe method for non-destructive evaluation of non-metal materials. However, THz-TDS generates a huge amount of measurement data for one batch test. In this study, we propose a THz-TDS based method to visualize the internal damage of SOFC by using machine learning. Each waveform was discretely vectorized, and a Self -Organizing Map (SOM) was utilized to examine for its similarity. Therefore, we found that the waveforms have been changed because of the wave interference induced by the generation of air layer associated with the occurrence of delamination.