Intelligent Optimization For The Coordination System Of Supercritical Unit Based On Condensate Throttling

Under the power grid centred real-time unit load scheduling mode, it is normal for a large-capacity supercritical(SC) unit to participate in deep peak-load regulation and gird frequency modulation with automatic generation control(AGC) and primary frequency modulation functions being put into operation. It raises higher requirements for the load control quality of a large SC power unit under dynamic variable-load conditions. Because the coordinated control system(CCS) needs to balance both the control deviations of load and main steam pressure, and the boiler itself is with large inertia, large time delay characteristics, a large SC power unit is apt to fail in meeting the rapid load-changing requirement from the power grid load dispach center, and easy to cause larger fluctuations for main steam pressure and temperature, and so on.Condensate throttling technique, as a developing new technolongy used for rapid unit load adjustment,has attracted much attention of power plant engineers and researchers in recent years. Therefore, condensate throttling characterisitics are deeply investigated by using a full-scope power plant simulator of a 600 MW supercritical unit. On this basis, the load optimization scheme for load-changing process based on dynamic adjustment of deaerator water level set point is designed and demonstrated by simulation tests. In order to further improve the condensate throttling control effect, a neural network turbine unit load predictive model considering condensate throttling is established. Based on this model, an intelligent condensate throttling load optimization control scheme is designed and demonstrated. The study show that it is completely feasible and effective to apply condensate throttling technique to supercritical unit load control optimization to improve the unit load adaptability of large SC power unit under AGC control.