Dynamic Identification Technology Of Reactive Power-voltage Control Characteristics And Its Application

At present,multi-stage voltage control is widely adopted in reactive power-voltage(Volt/VAr)control of power system,which uses model-based centralized optimization algorithm to set up the control target of each substation in order to adapt to the change of operation mode.With the expansion of the grid scale,the complexity of the system caused by the fluctuating power supply and the distributed power supply is increasing,and the centralized optimization control based on the traditional modeling of the power system becomes increasingly difficult.With the continuous improvement of on-line identification technology,data-driven and model-free Volt/VAr control technology has become a new concern,the core of which is the on-line identification of Volt/VAr control sensitivity and other control characteristics using WAMS and SCADA data.This paper studies the effective on-line identification method of Volt/VAr control characteristics with power system state data,and explores its application in Volt/VAr control of power system.In this paper,the on-line identification of Volt/VAr control characteristics is studied from the aspects of Volt/VAr linear sensitivity and non-linear correction of linear sensitivity.Firstly,an on-line sensitivity identification method is proposed,which is independent of the off-line system model and anti-collinearity,and realizes the on-line linear approximation of the Volt/VAr control characteristics.Then,the non-linear characterization and identification of the Volt/VAr characteristics are studied with ambient signal,and the modified sensitivity index of the nonlinear term is constructed to realize the on-line identification of the Volt/VAr control characteristics including its linear approximation and nonlinear correction.On this basis,the paper designs a gradual optimal strategy for voltage control based on dynamic identification of Volt/VAr control characteristics,which can gradually approach the control target without centralized optimization.In this paper,the dynamic identification technology for Volt/VAr control characteristics is validated in an 8-generator 36-note bus system.The effectiveness of the proposed gradual optimal strategy for voltage control is proved by comparing the simulation results of gradual optimal control and that of centralized optimal control in the situation of single running point and daily continuous operation.