Study On Condensate Throttling Based Intelligent Coordinated System Optimization For An Ultra-supercritical Power Unit

With current rapid development of new energy power generation technologies,including windpower and solarpower,more and more new energy power sources have been put into grid-connected operation,which relieves the dependence of power industry upon traditional coal resources.But the random characteristics of wind and solar energy also bring great impact on the stability of power grid.In order to adopt more new energy power and to ensure power grid safety,regional power system proposes very high requirements for the operation flexibility of the large-scale coal-fired power units with the wide implementation of "two rules".It requires a large ultra-supercritical power unit to enhance its performance for deep peak load regulation with fast load-tracking capability.Coordinated control system(CCS)is an essential module of a coal-fired unit,which works to meet the external load demand and to maintain energy balance between the boiler side and turbine side of the unit.Therefore,it is of great practical significance to adopt advanced control strategies to optimize the CCS and improve its control quality.In this paper,the operation characteristics and dynamic characteristics of a 1000 MW ultra-supercritical unit are deeply analyzed,and the basic structure and control strategy of the CCS are described in detail.On comprehensive understanding of neural network theory and nonlinear system modeling methods,BP neural network models for load prediction with "condensate throttling" technology and main steam pressure prediction are proposed,the fitting accuracy and good dynamic performance of the models are verified online.Based on the established prediction models,a classic optimization theory--simplex method and a modern intelligent optimization theory--particle swarm optimization(PSO)are selected as algorithms for CCS optimization,and the coordination optimization strategies based on them are studied.In this paper,the dual optimization control scheme based on load prediction model and the dual-model combined compensation optimization strategy are studied,and an online optimization program is developed in MATLAB platform.Detailed coordinated optimization control simulation experiments are carried out through two-way real-time data communication in the full-scope simulator of a 1000 MW ultra-supercritical power unit.The simulation results indicate that the optimization results with two different coordination optimization schemes both are effective in meet the engineering requirements.The control strategy and optimization scheme proposed in this paper can effectively improve the rapidity and stability of unit load response,greatly reduce the fluctuations of main steam pressure during variable load,and improve the quality of coordinated control.