Thermodynamically consistent reduced dimensionality electrochemical model for proton exchange membrane fuel cell performance modelling and control

JOURNAL OF POWER SOURCES

卷:454

Article

Kravos, Andraz¹  Ritzberger, Daniel²  Tavcar, Gregor¹  Hametner, Christoph³  Jakubek, Stefan²  Katrasnik, Tomaz¹ 

[1] Univ Ljubljana, Fac Mech Engn, LICeM, Ljubljana, Slovenia

[2] TU Wien, Inst Mech & Mechatron, Vienna, Austria

[3] TU Wien, CD Lab Innovat Control & Monitoring Automot Power, Vienna, Austria

关键词: PEM fuel cells;    Electrochemical model;    Performance modelling;    Reduced dimensionality models;    Parameter sensitivity;    Observers;   

DOI  :  10.1016/j.jpowsour.2020.227930

来源: Elsevier

【 摘 要 】

The reduction and prevention of degradation effects of proton exchange membrane fuel cells calls for precise online monitoring and control tools such as coupled virtual observers. To present significant progress in the area of computationally fast electrochemical models used in observer applications, this paper provides the derivation of a zero-dimensional thermodynamically consistent electrochemical model for proton exchange membrane fuel cells performance modelling and control. The model is further extended to accommodate the transport of gaseous species along the channel and through gas diffusion layer, yielding a quasi-one-dimensional electrochemical model. In addition, the presented work features the determination of an optimal set of calibration parameters proposed and based on mathematical and physical rationale, which is further supported with parameter sensitivity analysis. Multiple validation steps against polarisation curves at different operational points confirm the capability of the newly developed model to replicate experimental data. Furthermore, investigation in models generalisation capabilities shows that the model exhibits very good extrapolation capabilities for operation points outside the calibrated variation space of parameters. Additionally, the newly developed model can be successfully parametrised with little effort on a small calibration data set. These features position the proposed modelling framework as a beyond state-of-the-art model for virtual observers.

【 授权许可】

Free