Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Health state prediction and analysis of SOFC system based on the data-driven entire stage experiment
Ist Teil von
Applied energy, 2019-08, Vol.248, p.126-140
Ort / Verlag
Elsevier Ltd
Erscheinungsjahr
2019
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
•The kilowatt scale SOFC stack of natural gas systems were operated for 250 h.•Data-driven approach—Elman neural network was used for fault prediction of SOFC system.•The faults of heat exchanger breakage and the SOFC stack voltage jitter inside the system were analyzed and handled.•Application background is household generation system.
For the distributed household generation technology, the operation status identification and prediction of multiple startup and shutdown solid oxide fuel cell (SOFC) system are of great significance to the optimization of power generation efficiency and long-term operation. The existing SOFC system research focuses on traditional physical models rather than actual SOFC system, which makes the system research too idealized and ignores the actual problems, such as health state prediction, voltage fluctuation and failure analysis of key equipment such as heat exchanger. Therefore, based on the data-driven actual natural gas SOFC system, this paper builds Elman neural network state prediction model under the entire stage (including system start-stop, long-term operation, hot-standby) to predict the future voltage, and infers the potential fault information of the system from the residual voltage. On this basis, combined with the dynamic response of the system, the essential causes of voltage fluctuation and heat exchanger breakage of the actual SOFC system are found out. And the improvement scheme of the system is proposed. The results show that the typical faults of the SOFC system can be accurately identified by combining Elman neural network state prediction model with multivariable dynamic response analysis of SOFC system.