Journal articles
Fault diagnosis for fuel cell systems: A data-driven approach using high-precise voltage sensors
Abstract : Reliability and durability are two key hurdles that prevent the widespread use of fuel cell technology. Fault diagnosis, especially online fault diagnosis, has been considered as one of the crucial techniques to break through these two bottlenecks. Although a large number of works dedicated fuel cell diagnosis have been published, the criteria of diagnosis, especially online diagnosis have not yet been clarified. In this study, we firstly propose the criteria used for evaluating a diagnosis strategy. Based on that, we experimentally demonstrate an online fault diagnosis strategy designed for Proton Exchange Membrane Fuel Cell (PEMFC) systems. The diagnosis approach is designed based on advanced feature extraction and pattern classification techniques, and realized by processing individual fuel cell voltage signals. We also develop a highly integrated electronic chip with multiplexing and high-speed computing capabilities to fulfill the precise measurement of multi-channel signals. Furthermore, we accomplish the diagnosis algorithm in real-time. The excellent performance in both diagnosis accuracy and speediness over multiple fuel cell systems is verified. The proposed strategy is promising to be utilized in various fuel cell systems and promote the commercialization of fuel cell technology.
Cited literature [15 references]
https://hal-amu.archives-ouvertes.fr/hal-02004101
Contributor : Zhongliang Li <>
Submitted on : Wednesday, February 12, 2020 - 11:50:03 AM
Last modification on : Friday, February 14, 2020 - 1:20:13 AM
Document(s) archivé(s) le : Wednesday, May 13, 2020 - 12:49:04 PM
Contributor : Zhongliang Li <>
Submitted on : Wednesday, February 12, 2020 - 11:50:03 AM
Last modification on : Friday, February 14, 2020 - 1:20:13 AM
Document(s) archivé(s) le : Wednesday, May 13, 2020 - 12:49:04 PM
File
Paper_RENE_unmarked_15072018.p...
Files produced by the author(s)
Licence
Distributed under a Creative Commons Attribution - NonCommercial - NoDerivatives 4.0 International License
