Modeling and control system design of an MCFC system

A molten carbonate fuel cell (MCFC) system is modeled and a control strategy is proposed. A spatially distributed two-dimensional dynamic model of single cell direct internal reforming MCFC was developed as a numerical process by reducing the PDE model to a set of ODE´s using the cubic spline collocation method and the finite difference method. Mean cell temperature, T_s;m, maximum temperature difference over a cell, ΔT_s;max, oxygen conversion, X_O2, and hydrogen mole fraction at the anode outlet, χ_H2;o, were considered as controlled variables (CV´s) under the assumption that their target values are determined by an optimizer in the upper level. Among the CV´s, T_s;m and ΔT_s;max have much slower dynamics than the other two and keeping ΔT_s;max below a certain safety limit is critical. Hence T_s;m and ΔT_s;max were designed to be separately controlled under model predictive controller (MPC). On the other hand, X_O2 and χH_2;o have fast dynamics and were designed to be regulated by single loop PID controllers with feedforward compensation. The manipulated variables (MV´s) for each CV group were determined through a system analysis using the singular value decomposition and relative gain array. The performance of the control scheme was evaluated against load (power demand) changes.