Determination of optimal reformer temperature in a reformed methanol fuel cell system using ANFIS models and numerical optimization methods

Kristian Kjær Justesen, Søren Juhl Andreasen

Research output: Contribution to journalJournal articleResearchpeer-review

9 Citations (Scopus)

Abstract

In this work a method for choosing the optimal reformer temperature for a reformed methanol fuel cell system is presented based on a case study of a H3 350 module produced by Serenergy A/S. The method is based on ANFIS models of the dependence of the reformer output gas composition on the reformer temperature and fuel flow, and the dependence of the fuel cell voltage on the fuel cell temperature, current and anode supply gas CO content. These models are combined to give a matrix of system efficiencies at different fuel cell currents and reformer temperatures. This matrix is then used to find the reformer temperature which gives the highest efficiency for each fuel cell current. The average of this optimal efficiency curve is 32.11% and the average efficiency achieved using the standard constant temperature is 30.64% an increase of 1.47 percentage points. The gain in efficiency is 4 percentage points, from 23 % to 27 %, at full power where the gain is largest. The constant reformer temperature which gives the highest average efficiency is found to be 252 °C at which temperature it is 32.08%, only 0.03 percentage points lower than the maximum efficiency curve.
Original languageEnglish
JournalInternational Journal of Hydrogen Energy
Volume40
Issue number30
Pages (from-to)9505–9514
Number of pages10
ISSN0360-3199
DOIs
Publication statusPublished - Aug 2015

Keywords

  • Reformed methanol fuel cell
  • Fuel cell modeling
  • Operating point optimization
  • System efficiency optimization
  • HTPEM fuel cells
  • ANFIS modeling

Fingerprint

Dive into the research topics of 'Determination of optimal reformer temperature in a reformed methanol fuel cell system using ANFIS models and numerical optimization methods'. Together they form a unique fingerprint.

Cite this