Smart Grid Advanced Control Modules for Hybrid Fuel Cell/Gas Turbine Power Plants

摘要

FuelCell Energy (FCE) has developed a novel paradigm for advanced control of power plant processes with specific adaptation to hybrid fuel cell/gas turbine systems and smart grid dispatch. The paradigm involves development of intelligent control algorithms based on feedforward and feedback dynamic structures. It comprises a combined model-based robust controller and neural network supervisor (NNS) that together manipulate the plant actuators to suboptimally control the hybrid system during load ramping. The NNS performs optimal dynamic scheduling by providing unique setpoint and feedforward signal profiles to the plant controller for prescribed dispatch load profiles. Optimization routines are performed off-line. Data from off-line dynamic optimization, which utilizes a dynamic model of the hybrid fuel cell/gas turbine system, is then used to train the neural network supervisor. A robust feedback control strategy is used to compensate for setpoint deviations caused by suboptimal feedforward control moves and to counteract modeling errors and process disturbances.