Application Of Predictive Control In Coordinated Control System Of Ultra Supercritical Units

Due to China’s increasing attention to environmental protection,ultra supercritical power generation technology has become an important direction of technological transformation and upgrading of thermal power generation industry in the future because of its outstanding advantages in saving energy and avoiding environmental pollution.According to the characteristics of nonlinearity,large delay and multi constraints of the coordinated control system of ultra supercritical units,a dynamic matrix control turbine boiler coordinated control strategy based on model switching is proposed in this paper.Firstly,based on the nonlinear model of the coordination system of 1000 MW ultra supercritical unit,six typical operating points are determined based on the clearance calculation.Then,the nonlinear model is expanded by Taylor series at the steady-state operating point by using the partial derivative method,so as to obtain the piecewise linearized model of the system.Finally,the linearized simulation model of the controlled object is built in the visual simulation software,the step response curve of the object is obtained,and the dynamic characteristics of the controlled object are analyzed.Then,the dynamic matrix control strategy of turbine boiler coordination model switching is proposed to improve the load response performance of the unit.Based on the unit load,the prediction model is updated online.The initial condition model and the target condition model are used as the prediction model of the controlled object.The step disturbance experiment of controlled quantity,the disturbance experiment of control quantity and the simulation experiment of continuous variable conditions from condition #1 to condition #6 are carried out,and a new model switching strategy is summarized.The control system has better load tracking ability when the model switching strategy is applied.Finally,aiming at the practical problem that the weight of controlled quantity and control quantity in multivariable DMC control system is difficult to determine and has a significant impact on the control effect,two optimization algorithms,genetic algorithm and sparrow search algorithm,are used to optimize the parameters respectively,and obvious results are achieved.The relative error integral of controlled quantity in the variable working condition process of the system from working condition #1 to working condition #2 is reduced.