Model-based observer for direct methanol fuel cell concentration estimation by using second-order sliding-mode algorithm

Accurate estimation of the real-time methanol concentration of direct methanol fuel cell (DMFC) stack is a key technique for its feedback control and optimization. However, existing data-based methods as well as voltage fluctuation methods require a large amount of data to estimate methanol concentration, which increases the burden of embedded systems. What needs to be concerned is the methanol concentration inside the stack, because it directly affects its output power, while the existing research pay more attention to the methanol supplied concentration. To this end, a model-based observer based on second-order sliding-mode (SOSM) algorithm, is proposed to estimate the real-time methanol concentration inside DMFC stack utilizing the easily measured signals, such as DMFC stack voltage, current, and temperature. To validate the proposed method, the simulation comparisons between proposed SOSM observer and existing first-order sliding-mode (FOSM), extend Kalman filter (EKF) observer were carried out under certain operation conditions. Further, experimental verifications were implemented by using a real commercial DMFC system data to verify the performance of proposed approach. The comprehensive results demonstrates that proposed SOSM observer could estimate DMFC methanal concentration with robustness and accuracy.

Automated summary

A model-based observer using second-order sliding-mode algorithm is proposed to estimate the real-time methanol concentration inside DMFC stack. The observer utilizes easily measured signals such as stack voltage, current, and temperature. Simulation and experimental comparisons show that the proposed observer can estimate DMFC methanol concentration with robustness and accuracy.