Model Based Start-up Optimization of a Combined Cycle Power Plant

This paper deals with the modeling and the optimization of the start-up of a combined cycle. It decomposes the transient in several phases that are optimized separately. For each phase a simple model is used to minimize its duration under constraints. Several state and control input constraints are given in the paper to impose a trajectory that respects the stress limitation and ensures the damping of the responses. Each constrained minimum time optimization problem is solved by a SQP algorithm available in Matlab™. The optimal trajectories are tested on a complete dynamic model which consists of a process modeled by first principle equations and a sequence described by a finite state machine. This validation model is developed in Simulink™ and Stateflow™.