Methanol Reformer System Modeling and Control using an Adaptive Neuro-Fuzzy Inference System approach

Kristian Kjær Justesen, Mikkel Præstholm Ehmsen, John Andersen, Søren Juhl Andreasen, Hamid Reza Shaker, Simon Lennart Sahlin

Publikation: Konferencebidrag uden forlag/tidsskriftPosterForskning

685 Downloads (Pure)


This work presents the experimental study and modelling of a methanol reformer system for a high temperature polymer electrolyte membrane (HTPEM) fuel cell stack. The analyzed system is a fully integrated HTPEM fuel cell system with a DC/DC control output able to be used as e.g. a mobile battery charger. The advantages of using a HTPEM methanol reformer is that the high quality waste heat can be used as a system heat input to heat and evaporate the input methanol/water mixture which afterwards is catalytically converted into a hydrogen rich gas usable in the high CO tolerant HTPEM fuel cells. Creating a fuel cell system able to use a well known and easily distributable liquid fuel such as methanol is a good choice in some applications such as range extenders for electric vehicles as an alternative to compressed hydrogen.

This work presents a control strategy called Current Correction Temperature Control (CCTC), which changes the fuel cell current in order to control the flow of hydrogen to the burner that adds heat to the reforming process. The method ensures a control strategy that avoids some of the critical events for such a system, which includes too high burner fuel flows with following critical burner temperatures, and fuel cell stack anode starvation which significantly can increase the degradation of the fuel cell stack.

Modeling of the reformer dynamics is conducted using an adaptive neuro-fuzzy interference system approach (ANFIS) based on measurement results from an experimental setup. The modeling enables prediction and characterization of the important time constants of such a system.
Publikationsdato11 apr. 2012
Antal sider1
StatusUdgivet - 11 apr. 2012

Bibliografisk note

Denne poster blev tildelt "Best Poster Award".


Dyk ned i forskningsemnerne om 'Methanol Reformer System Modeling and Control using an Adaptive Neuro-Fuzzy Inference System approach'. Sammen danner de et unikt fingeraftryk.