Sudy On Intelligent Control Of Boiler-turbine Power System

Coal-fired power as the main force of power industry, its power generation efficiency is increasingly becoming the concerned focus by people. Power generation unit coordinated control directly affects the normal operation of the power plant. The practical and feasible control system is a powerful guarantee to improve the efficiency of power generation.Boiler-turbine power system is an important component of the thermal power units, the system has the characteristics of multiple input and output, nonlinear and strong coupling, etc. For boiler-turbine system, respectively, the terminal sliding mode controller and robust adaptive controller are designed. On the basis of this, research the higher order nonlinear complete mathematical model (including boiler, turbine and generator model) of power units, which has dynamic uncertainty then design the controller.First of all, the classic nonlinear inverse system method is proposed to realize the feedback linearization and decoupling of boiler-turbine power system. Considering the influence of modeling error and the parameters of the system can change in a wide range of operating condition, combined with the radial basis function (RBF) neural network method to identify the inverse system and reduce the modeling error through online learning. Then design the terminal sliding mode controller for the decoupling system in order to achieve the limited time convergence. Using lyapunov method to analytical the stability of the system, the simulation results show that the control system can work well in the wide range operating condition, which better than the classical inverse system control method.Secondly, put forward a kind of state feedback H-infinity robust control method according to the boiler-turbine power system which based on particle swarm algorithm. This paper come the H-infinity robust control design question down to solve the Riccati inequality algebra solution. Using the improved particle swarm algorithm combined with the linear matrix inequality (LMI) to solve the Riccati matrix algebra equation, get the state feedback controller which can guarantee the system is asymptotically stable. The simulation results show that the system has strong robustness for external disturbance, has good decoupling effect and tracking ability.Finally, the backstepping dynamic surface control method is used to design a robust adaptive controller for the boiler-turbine power generation unit. The system is decomposed into two smaller subsystems then the controller design was carried out on the two subsystems. In inverse design process to join the low-pass filter, which makes the design method simplified. Theoretical analysis and simulation results show that the controller can guarantee the closed-loop system global asymptotic stability, and has strong robustness for system parameters uncertainty, which can make the system stable with occurs breakdown.