Study On Power System Oscillation Of Hydropower Units With High Penetration

With the implementation of the overall optimization of resources allocation and construction of the southern power grid,the permeability of hydropower energy in the system continues to improve,and there are both low-frequency oscillation and ultra-low frequency oscillation.The low frequency oscillation signal of power system is a typical multi-component mixed noise signal,the ultra-low frequency oscillation signal often occurs in the power grid with high permeability of hydropower,and the frequency of oscillation is low.The generation of multiple types of oscillation signals poses a new challenge to the identification of characteristic parameters.It is very important for the safe and stable operation of power system that how to accurately extract and effectively suppress multiple types of complex oscillation signals.This paper describes the research background and significance of power system low-frequency oscillation and ultra-low-frequency oscillation,introduces several typical methods commonly used in frequency oscillation signal identification,the traditional vibration signal preprocessing algorithm is easy to cause modal aliasing,loss,distortion and other phenomena when decomposing the vibration signal,while the anti-noise performance is poor.In this paper,the variational mode decomposition method is introduced,which has a strong theoretical basis and the processing ability of non-stationary signals,and effectively avoids the phenomenon of mode aliasing,loss and distortion when decomposing signals.However,the value of mode number and penalty factor need to be set artificially,which will make the decomposition result deviate to a certain extent,and for the oscillation frequency of multiple types of oscillation signals Because of the large rate span,it is difficult to consider the multi-modal characteristics at the same time with a single penalty factor,which can not ensure the fidelity of the signal.In order to extract the dominant modes of the characteristic parameters accurately,the adaptive variable mode decomposition(AVMD)algorithm and Hilbert transform are used to identify the characteristic parameters of the oscillating signals.Based on the energy loss coefficient(ELC)and the multiple matching penalty factor(MMPF),the number of modes and the penalty factor value are determined.The dominant mode of the vibration signal is extracted adaptively and effectively,and the characteristic parameters of the vibration signal are identified by the Hilbert transform.For the ultra-low frequency oscillation in the system,two strategies are adopted to suppress it.One is to adjust the PID governor parameters to reduce the negative damping in the system,the other is to use the additional damping controller to increase the positive damping provided by the load,so that the total damping torque of the system is positive.Through the simulation analysis,the inhibition effect of the two strategies is compared.