Model predictive control for gas turbine shaft speed based on model compensation using extended state observer

Abstract

Gas turbine engines have offered advanced power generation with relatively small size on marine vessels. The operational flexibility has high control requirements for shaft speed regulation. Model predictive control provides a methodology for safe and efficient operation of gas turbines. However, when working at partial load, the controller degrades due to the mismatch between predictive model in model predictive control and dynamics of gas turbine. To deal with the issue of model adaptability in model predictive control, model compensation–based model predictive control is proposed for the gas turbine operation. The dynamics of the plant is modified by the extended state observer to get close to the pure integrator. The model predictive control is devised based on the modified plant. The wide-range load regulation and disturbance rejection of gas turbine have been improved by leveraging the benefits of model compensation–based model predictive control.