Reliability modeling and assessment of grid-connected PEM fuel cell power plants


JOURNAL OF POWER SOURCES, vol.142, pp.264-278, 2005 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 142
  • Publication Date: 2005
  • Doi Number: 10.1016/j.jpowsour.2004.10.030
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.264-278
  • Yıldız Technical University Affiliated: Yes


Fuel cell (FC) power plants are subject to a number of possible outage and derated states due to partial or full failure of auxiliaries. Furthermore, most of the power system reliability studies reported in the literature assume mean values of the particular measure of reliability. However, the reliability indices for grids such as failure rate, outage duration, etc. vary Lot different time period due to weather conditions, variety of power demands and random faults. It is essential to obtain the estimation of reliability under all environmental, operational and loading conditions. This paper considers above mentioned seasonal variation of grid reliability indices as well as partial or full failure of fuel cell auxiliaries for grid-connected PEM fuel cell power plants (FCPPs). In the paper, a detailed state-space model of the grid-connected FCPP is presented which is a combination of proton exchange membrane fuel cell (PEMFC) power plant generation model and grid outage model. The state-space generation model of a PEMFC power plant is formed based on the failure modes of system auxiliary components. As for the grid outage state-space modeling, the effects of weather conditions such as normal and adverse weather are taken into consideration in modeling the failure and repair rates. The functional relationship between weather conditions and transition rates, namely failure and repair rates are developed based on the fuzzy set theory and embedded into Markov model (MM). Simulation results are obtained for a 5 kW grid-connected PEMFC that supplies a typical residential house using the MATLAB software package. (c) 2004 Elsevier B.V. All rights reserved.