Control of jump linear systems: application to the steam generator water level

This paper addresses the control of jump linear systems applied to a generic problem: the water level regulation of a steam generator. This problem captures many realistic control features. The main ones are sensor failures and time-varying behaviors. In this context, the output feedback mode-dependent controller synthesis for linear systems subject to Markovian jumps appears well-suited. This synthesis problem is cast as a nonconvex optimization problem under linear matrix inequalities (LMI) constraints. This problem is numerically solved using an efficient cone complementarity linearization algorithm. Finally, the controller is validated on a parameter-dependent simulator of the plant.