Application Of Neural Network Inverse Control To The Coordinated Control System And Superheated Steam System Of Supercritical Unit

Due to large-capacity supercritical and ultra-supercritical coal-fired units have apparent advantages in their economy, efficiency and environmental indexes, they have increasingly become the dominated units in China’s power grid. Under current load dispatch mode centred by regional load center, it is common for a supercritical or ultra-supercritical unit involved in automatic generation control(AGC).Thus the unit often need change load in large scope and work in dynamic load conditions, which means higher requirements for its overall control quality. Coordinated control and main steam temperature control are two important control systems. Thus it is necessary to optimize them jointly with advanced control strategies to ensure that the unit can meet power network requirements of fast peak load adjustment and frequency regulation, while avoiding main steam pressure, temperature and other critical parameters largely deviating from their proper values to keep the unit running safely and economically.As an advanced intelligent control strategy, neural network inverse control has been sucessesfully applied to various industrial processes in recent years. Therefore, neural network inverse control method is used for joint optimization of load, main steam pressure and main steam temperature for a 600 MW supercritical power unit. After establishing neural network inverse models for the coordinated control and superheated steam temperature objects, optimization control scheme was designed. Detailed simulation experiments were conducted with the power plant simulation system to validate the method. The results show that the coordinated control and main steam temperature control were effectivley improved when the overall optimization program was added, with faster load response, reduced fluctuations in main steam pressure and temperature compared to the original PID control.