Methanol concentration estimation in a direct methanol fuel cell system

Abstract

An inferential sensor based on state estimation techniques is proposed for measuring the methanol concentration in an open-cathode direct methanol fuel cell. The strategies considered include full-state and reduced-state Luenberger observers along with a proportional-integral observer. The estimators are designed using linear dynamics derived from a comprehensive nonlinear dynamic model of an experimental open-cathode direct methanol fuel cell. A representative simulation study is used to assess the performance of the estimation schemes deployed on both the linear and nonlinear dynamic models of the system. All three estimation methods are initially tuned to deliver adequate performance on the linear model. The simulation results show that the proportional-integral state observer outperforms the other designs when the dynamics are driven away from the linear regime. It is argued that the additional robustness of the proposed proportional-integral observer design holds significant promise for use as the basis for developing an effective inferential methanol composition sensor.