Active Disturbance Rejection Control And Its Application In A Kind Of Thermal Control System

Thermal power generating units have the characteristics of large time delay, big inertia, nonlinear,strong disturbance, multivariable coupling and so on, making control system design a great challenge. With the rapid development of Chinese power industry, thermal power generating units move in the direction of high-gain, large capacity and highly automated, where control quality becomes more and more important. Now Active Disturbance Rejection Control(ADRC), an alternative control technology studied in this thesis, is increasingly applied to the thermal power generating units.In this thesis, a novel tracking differentiator(TD) is first presented based on a hyperbolic tangent-based acceleration function. Compared with several commonly used tracking differentiators, the one here doesn't apply piecewise function and signum function, making theoretical analysis easy. After stability analysis, the thesis gives the reasons why choose hyperbolic tangent function as the accleration function of TD and conclude several parameter tuning rules. At last, the method is applied to some typical input signal, and simulation results show that it can provide good track and differentiation.The parameter tuning of ADRC is difficult, especially the plants with large time delay, nonlinear and multi-control. To solve the problem, the thesis studies an adaptive genetic algorithm(AGA).Upon this, an improved AGA based ADRC parameters optimizing method is presented. The method firstly ascertain the parameters of tracking diffenentiator,extended state observer and nonlinear state error feedback according to several tuning experience, making the scope of parameters smaller; the choice of the target function consider the criterion of ITAE, the constraint of control signals and the limitation of control energy and so on.Finally, aiming at a class of thermal power plants, an ADRC solution based on the methods discussed above is proposed. For three transfer function models of different loads in the circulating fluidized bed temperature control system, and for the three different load transfer function models in a 600MW supercritical boiler superheater control system, the ADRC solution is tested in simulation. usingthe novel tracking differentiator and the optimization methods in parameters tuning of the extended state observer and nonlinear error feedback control law.Compared with PID controllers, simulation results show that active disturbance rejection controller provides better robustness and disturbance rejection ability, achieving full control of working conditions under the step response, robustness experiments and perturbation experiments.