Method for online identification of low-frequency oscillation mode of electric power system based on random response signals

Abstract

The invention discloses a method for online identification of a low-frequency oscillation mode of an electric power system based on random response signals. The method comprises the following steps of: firstly, reading a section of angular speed signals of different power generators when the electric power system runs normally; then performing mean removal processing on the read signals to obtain zero-mean random response signals; processing the zero-mean random response signals by using a natural excitation technology to obtain free oscillation signals; and finally, performing mode identification on the obtained free oscillation signals by adopting a PRCE method to obtain a frequency, a damping ratio and an oscillation mode state of the low-frequency oscillation mode. The method disclosed by the invention can identify the low-frequency oscillation mode of the electric power system in a wider time range, is better in identification precision, speed and noise immunity, and is more capable of identifying an oscillation mode state, thereby providing a brand-new way and method for low-frequency oscillation analysis of the electric power system.

Description

technical field [0001] The invention relates to the field of power system stability analysis, in particular to an online identification method for a low-frequency oscillation mode of a power system based on a random response signal. Background technique [0002] Interregional low-frequency oscillation is an inherent phenomenon in interconnected power systems, and weak or negative damped oscillation modes will have adverse effects on the safe and stable operation of the system. In view of the seriousness of this problem, both IEEE and CRGRE have set up special working groups to study it. Accurate and rapid identification of low-frequency oscillation modes is of great significance to the analysis and control of low-frequency oscillations in power systems. [0003] For large-scale interconnected power systems, the model-based low-frequency oscillation analysis method has many limitations, such as the large number of state variables leading to the problem of "curse of dimensiona...

AI Technical Summary

Problems solved by technology

[0005] At present, the methods for low-frequency oscillation identification based on random response signals under environmental excitation mainly include the stochastic subspace method (SSI), which can use both free oscillation signals and random response signals as its input, but the SSI method requires a lot of Singular value decomposition leads to slow calculation speed and easy to generate false modes; stochastic decrement technology combined with TLS-ESPRIT method can realize low-frequency oscillation mode identification between power system regions under environmental excitation conditions, but this method cannot identify oscillation modes, and Oscillatory modes are also important parameters in low frequency oscillation analysis

In addition, due to the rough estimation of the free oscillation signal by the random decrement technique, the identification accuracy of this method is not high, and the noise immunity is not strong

[0006] It can be seen that the existing technology method cannot calculate the oscillation mode, the noise immunity is poor, and the identification accuracy is not high, and the calculation speed is slow

Images