Operation characteristic control of direct methanol fuel cell liquid feed fuel system using adaptive Neuro-fuzzy inference system model

Currently, Direct Methanol Fuel Cell (DMFC) technology suffers from the low power density caused by slow reaction rate and undesired methanol crossover, which limits its commercialization application. At this study, Adaptive Neuro-Fuzzy Inference System (ANFIS) can predict the needed amount of methanol fuel from the relationship of current and voltage curve of DMFC under different operational conditions and keep high power density. The ANFIS is a collaborative data bank from repeated experiments results under different amount of methanol fuel liquid. The model is a control scheme for predicting of supply to a fuel cell system under dynamic loading conditions, with a high accuracy in an easy, rapid and cost effective way to regulate the concentration of a liquid feed fuel cell without any fuel concentration sensor. The control scheme uses operating characteristics of fuel cell, such as potential, current and temperature. Our previous study has presented a fuel sensorless control algorithm (IR-DTFI) to calculate the quantity of fuel liquid required at each monitoring cycle. Furthermore, we develop ANFIS to strengthen the concentration regulating process. The ANFIS had been verified by systematic experiments, and the experimental results proved that the scheme can effectively control the fuel supply of a liquid feed fuel cell with reduced response time, even while the Membrane Electrolyte Assembly (MEA) deteriorates gradually.