Extremum-seeking control of a microbial fuel cell power using adaptive excitation

Microbial fuel cell (MFC) is a novel bio-renewable energy source, whose maximum produced power is, like most other renewable energy sources (photovoltaic panels, wind turbines, etc.) highly dependent on external disturbances. Hence, an appropriate real-time optimization method should be used so that the MFC always operates at its optimum. Extremum-seeking control (ESC) can be applied to optimize this type of system. However, when the optimal operating point of the MFC varies very fast due to external disturbances, the slow convergence of the ESC induces a lack of precision in the tracking of the optimal power point. In this paper, it is proposed to use adaptive excitation signal amplitude in the ESC scheme to improve the precision of the tracking. The amplitude adaptation is performed using a neural-network (NN) model which estimates in real-time the error between the optimal and the actual point of operation of the MFC based on the external disturbance measurements. The improved ESC performance in terms of convergence speed and precision will be demonstrated at the level of simulation in the case of a comparative study between classic ESC and ESC with adaptive excitation applied to an MFC model subject to variations of a measurable external disturbance, namely, the inlet substrate concentration.