Model-based optimal and robust control of renewable hydrogen gas production in a fed-batch microbial electrolysis cell.

Microbial electrolysis cell (MEC) is a fundamental type of bio-electrochemical system. MEC is a novel and emerging renewable energy technology that is based on biomass. The behavior of the MEC system is highly nonlinear due to the complexity of its dynamics. For the desired optimal production of hydrogen, feedback control of MEC processes is necessary. Due to the novelty of MEC, limited research is available on its control. Studies on linear model-based robust control of MEC processes are missing. In this article, we develop a nonlinear dynamic model for MEC, linearize this model, and calculate a linear time-invariant transfer function. Based on this linearization, a fixed-structure, optimal and robust controller is proposed to achieve a fast-settling time exhibiting no overshoot and having zero steady-state error. The robustness of the developed controller is evaluated for parameter uncertainty, measurement noise, and disturbance rejection. Batch biomass processes are fed only at the start of each process cycle. The output does not follow the desired response when the substrate or biomass is consumed. Then, the error accumulates, and it causes the control effort to increase unboundedly. The existing literature on control of fed-batch MEC processes does not consider this integral windup phenomenon. In this article, we also develop an anti-windup control strategy to eliminate the integral windup error and to avoid any possible instability or destruction. The overall conclusion of our study is that the developed robust controller achieves a faster and more robust response than the existing controllers. We provide an anti-integral windup solution to eliminate windup errors in feedback control of fed-batch MEC processes. [Display omitted] • We design an optimal robust controller for a dynamic model of microbial electrolysis cells. • We linearize the non-linear model and calculate a time-invariant transfer function. • We evaluate uncertainty, noise, and disturbance of the controller. • We develop a strategy to eliminate windup error and avoid instability. • Our controller achieves a more fast and robust response than the existing controllers. [ABSTRACT FROM AUTHOR]