A NOVEL HIERARCHICAL CONTROL STRATEGY FOR FASTER LOAD-CYCLING OF A THERMAL POWER UNIT

The paper presents a novel hierarchical approach to plant-wide control by making effective use of a physical plant model. It incorporates a two-level control structure consisting of lower-level conventional PI regulators and a higher-level non-linear model predictive controller (NMPC). The lower-level PI loops help stabilise the unstable drum-boiler dynamics and allow faster governor valve action for power and grid-frequency regulation. The higher-level NMPC is based on an EKF estimator and a linear model obtained by continuous linearisation and discretisation of the non-linear physical plant model. The NMPC provides an optimal load demand (or set-point) transition by effective handling of plant-wide interactions and system disturbances. Use of a physical state-space plant model facilitates application of constraints on state variables in addition to input and output variables. Metal temperatures of heat-exchanger tubes are modelled as state variables in the non-linear plant model. For the first time, it is clearly brought out through simulation results that constraints need to be applied on the metal temperatures rather than steam temperature for faster load-cycling. Effectiveness of the constrained control scheme in ensuring larger rate of load changes without thermal constraint violation, is shown during the variable pressure mode of plant operation. Simulation results are obtained by testing the control strategy on a simulation of a 200 MW oil-fired power plant at Ballylumford in N. Ireland.