Amirkabir University of Technology Amirkabir Journal of Mechanical Engineering 2008-6032 54 3 2022 05 22 Experimental Cathode-Anode Flooding Diagnosis of Polymer-Electrolyte Fuel Cell of Power under 300W Using Adaptive-Neuro-Fuzzy Method Experimental Cathode-Anode Flooding Diagnosis of Polymer-Electrolyte Fuel Cell of Power under 300W Using Adaptive-Neuro-Fuzzy Method 615 628 4651 10.22060/mej.2021.20293.7207 FA Ali Khanafari Department of Mechanical Eng., Sharif University of Technology, Tehran, Iran Aria Alasty Department of Mechanical Eng/ Sharif University of Technology Mohammad J. Kermani Department of Mechanical Eng., AmirKabir University of Technology, Tehran, Iran Saeed Asghari Associate Professor / Institute of Materials and Energy, Tehran, Iran Journal Article 2021 07 21 Today, due to the growing importance of polymer electrolyte membrane fuel cells in the production of clean energy, the diagnosis of this energy converter has become very important. Diagnosis can significantly increase the useful life and reliability of the fuel cell. A major part of the defects related to the polymer electrolyte membrane fuel cells is due to the disturbance of the moisture balance in them. Flooding is one of the most common defects associated with fuel cell imbalance, which is possible in both the cathode and the anode side of the cell. In previous works, the cathode has been considered as the only possible place for flooding, mainly because it is the source of water production. In this article, the anode is also considered as a possible place for this phenomenon. The method of this research is based on taking data from the stack under healthy operating conditions and trying to estimate the output parameters of stack voltage, cathode pressure drop, and anode pressure drop using related inputs using the adaptive neuro-fuzzy method. In conditions of uncertain operation in which the healthy or flooding operation of the stack is not known, comparing the deviation of the actual values of the outputs from the model with the allowable values of these deviations (0.735 [V], 0.0092 [bar] and 0.0047 [bar], respectively) can lead to determining flooding or normal conditions. Today, due to the growing importance of polymer electrolyte membrane fuel cells in the production of clean energy, the diagnosis of this energy converter has become very important. Diagnosis can significantly increase the useful life and reliability of the fuel cell. A major part of the defects related to the polymer electrolyte membrane fuel cells is due to the disturbance of the moisture balance in them. Flooding is one of the most common defects associated with fuel cell imbalance, which is possible in both the cathode and the anode side of the cell. In previous works, the cathode has been considered as the only possible place for flooding, mainly because it is the source of water production. In this article, the anode is also considered as a possible place for this phenomenon. The method of this research is based on taking data from the stack under healthy operating conditions and trying to estimate the output parameters of stack voltage, cathode pressure drop, and anode pressure drop using related inputs using the adaptive neuro-fuzzy method. In conditions of uncertain operation in which the healthy or flooding operation of the stack is not known, comparing the deviation of the actual values of the outputs from the model with the allowable values of these deviations (0.735 [V], 0.0092 [bar] and 0.0047 [bar], respectively) can lead to determining flooding or normal conditions. https://mej.aut.ac.ir/article_4651_e24edf07dae0d65ce6db5e25e13782a3.pdf