Study On Nonlinear Predictive Control Of Hydro-turbine Governing System

The large-scale grid connection of renewable energy leads to reduce the load supply of some hydro-turbine generating sets and deviate from the optimal operation area.Moreover,the wear and tear of the mechanical components of the hydro-turbine generating set during the operation process will cause the frequent mechanical vibration during the operation process.At the same time,because of the intermittent and random characteristics of renewable energy sources such as wind energy and solar energy,it will also cause the change of network end load and increase the uncertainty in the operation of the unit.Both mechanical vibration and network end load disturbance will have a negative impact on the precise control of the unit.For this problem,the prediction control algorithm is adopted to improve the control performance of hydro-turbine governing system and ensure the stable operation of hydro-turbine generating set.The research contents of this paper mainly include the following three aspects:(1)Predictive control of hydro-turbine governing system.For the accurate control of hydro-turbine governing system,the generalized predictive control based on T-S fuzzy model is proposed in the second chapter.Firstly,the discretization model is established based on the T-S fuzzy theory and the fourth-order Runge-Kutta method.Secondly,quadratic objective function,optimal theory and predictive control theory are introduced to design predictive controller.After that,the effectiveness and anti-interference ability of the designed controller are verified by considering the system without disturbance,random mechanical disturbance,system parameter changing,and the influence of controller parameters on the control performance is explored.The proposed predictive controller and PID control theory are brought into the system model for comparison.The simulation results show that the predictive controller can ensure the stability of the hydro-turbine governing system within 5 s.Compared with the PID controller,the time of the system reaching the stable state is shorter and the overshoot is smaller.Therefore,the proposed predictive control theory can provide better control performance for hydro-turbine governing system.(2)Fractional hydro-turbine governing system predictive control based on state estimator.In order to improve the anti-interference ability and control performance of the system,the Chapter 3 optimizes the predictive control algorithm from the structure of the controller.Firstly,the six-dimensional model of fractional hydro-turbine governing system is selected as the research object,and the discrete model is obtained based on G-L differential definition,T-S fuzzy and first order difference method.Secondly,a state estimator is proposed to estimate the system variables in the prediction time domain.The stability of the state estimator is analyzed by Lyapunov theory,and the stability conditions are given.Then the state estimator is solved by linear matrix inequality.After that,combined with the predictive control theory,the estimated value is used as the predicted value to calculate the control increment.Simulation results show that the system with estimator can reach a stable state within 0.5 s.(3)State feedback robust predictive control for hydro-turbine governing system.In Chapter 4,the state feedback robust predictive control of hydro-turbine governing system is proposed by combining robust constraints with predictive control theory.Firstly,the discrete four-dimensional state space model is obtained by linearizing the system with convex polyhedron optimization theory.After that,the optimization problem of predictive control algorithm can be transformed into min-max optimization problem of sampling time,and the optimal boundary of objective function can be obtained.Combined with Lyapunov theory,it is proved that the system is gradually stable.The feedback gain is solved by linear matrix inequality.Finally,the anti-jamming ability of the system is explored considering the random disturbance and sudden load change of the unit.The simulation results show that the state feedback robust prediction controller can be faster and more stable than the prediction controller proposed in Chapter 2,and the prediction controller based on the state estimator in Chapter 3 can reduce the multi-period vibration in the controlled process.