Vývoj kontrolního systému pro rychlé nalezení tvarových odchylek na povrchu bipolárních desek pro palivové články pomocí laserem promítané mřížky

Abstract

Fuel cells are a promising new alternative to internal combustion engines in the automotive industry. However, the production and examination costs are still too high, which makes this technology less attractive to manufacturers. Particularly with the bipolar plates, care must be taken to ensure that the quality is right. The goal of this work is to develop a new inexpensive and fast inspection system to detect topographic defects on bipolar plates surfaces by means of laser grid projection. This work analyzes the role of a bipolar plate within a fuel cell. It identifies the different production processes used in the industry today to point out the associated defects, reviews the published work of current surface inspection methods and compares their image acquisition and processing units. It derives a list of requirements for the inspection system and gives a proposition for a first prototype set-up for the system. Python was used for the control of the image acquisition and for the image processing. Several algorithms based on statistical properties, the Gray-Level Co-occurence Matrix and the Fast Fourier transform of the image have been tested with metallic discs and metallic bipolr plates. Finally, the system has been validated for the inspection of flat metallic components and needs to be improved for the inspection of metallic bipolar plates. Ideas on how to modify the set-up and the program are given in order to ameliorate the system.

Fuel cells are a promising new alternative to internal combustion engines in the automotive industry. However, the production and examination costs are still too high, which makes this technology less attractive to manufacturers. Particularly with the bipolar plates, care must be taken to ensure that the quality is right. The goal of this work is to develop a new inexpensive and fast inspection system to detect topographic defects on bipolar plates surfaces by means of laser grid projection. This work analyzes the role of a bipolar plate within a fuel cell. It identifies the different production processes used in the industry today to point out the associated defects, reviews the published work of current surface inspection methods and compares their image acquisition and processing units. It derives a list of requirements for the inspection system and gives a proposition for a first prototype set-up for the system. Python was used for the control of the image acquisition and for the image processing. Several algorithms based on statistical properties, the Gray-Level Co-occurence Matrix and the Fast Fourier transform of the image have been tested with metallic discs and metallic bipolr plates. Finally, the system has been validated for the inspection of flat metallic components and needs to be improved for the inspection of metallic bipolar plates. Ideas on how to modify the set-up and the program are given in order to ameliorate the system.