Image recognition of cracks and the effect in the microporous layer of proton exchange membrane fuel cells on performance

Proton exchange membrane fuel cell (PEMFC) is considered as a potential power source for future automotive applications and microporous layer (MPL) is one of the core components. The cracks on the surface of micro-porous layer formed during the deposition process change the overall transport capacity and effect the output performance of fuel cells by regulating water management. In this work, the image recognition and feature extraction of microporous layer surface crack is proposed and elaborated in detail. The simple, fast and accurate method based on image recognition is realized by programming a regional connectivity algorithm and calcu-lating the geometric features of the circumscribed geometry of the cracks. The application of artificial intelli-gence image processing technology at the micron level provides convenience for data analysis of crack distribution and improvement of microstructure. In this paper, scanning electron microscope (SEM) images of different carbon paper samples are taken, and the samples are tested under the same conditions to verify the feasibility and accuracy of the crack identification method, and to study the effect of cracks on the performance of fuel cells. The results prove that the image recognition algorithm can quickly and accurately identify the characteristics of cracks after Gaussian filtering and grayscale thresholding. Besides, it is found that cracks have impact on fuel cells and surface slender-shaped cracks with large area can improve the performance of fuel cells at high current densities. The method proposed in this paper can be used to improve the existing PEMFC design and quality check in process production by analyzing the local crack characteristics of the same sample and statistical data. In addition, the results of this paper are instructive for further water management research on microporous layer in fuel cells and can be used as a reference for practical applications.