A Study On Multivariable Disturbance Rejection Control Of Thermal Processes

Chinese electric power is experiencing a rapid development stage from traditional energy to new energy.As renewable energy continues to be integrated into the grid,coal-fired power plants are shouldering the burden of deep peak load regulation.Moreover,the thermal power plant is characterized by severe nonlinearity,various constraints and strong coupling of multiple variables,making it difficult for a traditional control system to achieve satisfactory results in the face of uncertainties such as fuel quality fluctuations.Therefore,it is urgent to design advanced disturbance rejection control strategy to improve the operation flexibility of thermal power units.The main contributions and advantages of this working are as follows:1.To solve the problem of strong coupling and non-minimum phase item in thermal process,a disturbance suppression strategy which can compensate decoupling is designed.Firstly,based on the full analysis of the model,a decoupling compensator which can retain the model information is designed,and the decoupler is aproximated from Taylor series expansion by ignoring the high frequency components.Then,a proportional-integral compensator which is easy to be implemented is obtained.Secondly,a model-based active disturbance rejection control is proposed for the decoupled model in each channel,where the model information is fully considered for the controller to obtain the strong disturbance rejection ability and the tracking problem is solved.Further,the anti-windup compensator is designed for the controller to deal with the input constraints.Finally,the robustness and decouling effect are analyzed during the design process.2.In view of the multivariable characteristics of the thermal process,a multivariable disturbance observer is introduced to realize the multivariable undecoupled set point tracking control.Facing the tracking problem of the system,a set point filter which can realize offset-free tracking is proposed based on the analysis of two degrees of freedom.Moreover,through the mathematical transformation of the system,the problem of non-strict proper is dealt with skillfully,and the output disturbance is equivalent to the input disturbance.Due to the actuator constraints,the integral action hidden in the disturbance observer is reavealed,which provides a theoretical basis for the design of the anti-windup compensator.Moreover,the stability of the anti-windup compesntor is guaranteed under the H? theory.3.Based on the nonlinear dynamic analysis of the boiler-turbine unit,a predictive control system with disturbance suppression ability is proposed.After analyzing the nonlinear dynamics of the Bell-Astrom model obtained from mechanism modelling,it is found that the system has bifurcation and chaos,which are related to the choice of disturbance term.Thus,the perturbation model is obtained on the basis of nonlinear dynamic analysis rather than the presupposition.Once the perturbation model is determined,a T-S fuzzy robust predictive controller is designed based on the input-to-state stability theory under the support of multi-model strategy.To further suppress the bifurcation and disturbance of the system,the robust invariant set technique is adopted to constrain the stable limit cycle of the system to the minimized disturbance invariant set.Moreover,the global control of the boiler-turbine unit is realized by using the dual-mode predictive control strategy,where the computation burden is reduced.4.In order to improve the operation flexibility of thermal power units,thermal process should be able to withstand the impact of strong unknwon disturbance.To deal with the large nonlinear and actuator constraints in a wide range of variable operating conditions,a fuzzy predictive control strategy with enhanced robustness is proposed in this paper.During the design process,multivariable fuzzy disturbance observer and auxiliary controller are jointly designed.The lumped disturbance of the system is estimated by the disturbance observer and then feedforward compensated.The observer error is compensated by the auxiliary controller,and the tracking performance of the system is realized by the predictive control.By the combination design of the observer and the auxiliary control,the constrained disturbance invariant set and the tight constraint set for predictive control are obtained.The design of predictive control is completed on the premise of ensuring the stability of the system and satisfying the recursive condition.The simulations arried on the boier-turbine unit show that the control strategy has strong disturbance suppression ability and fast tracking ability.