Subsynchronous oscillation tracing method based on improved STFT and convolutional neural network

CN119312071APending Publication Date: 2025-01-14NORTH CHINA ELECTRIC POWER UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310854382.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-07-12
Publication Date
2025-01-14

AI Technical Summary

Technical Problem

The existing sub-synchronous oscillation traceability method fails to effectively consider the random volatility of the wind farm, and the method based on transient energy flow is mainly aimed at steam turbine units, which fails to fully explain the wind power grid-connected oscillation mechanism, and lacks positioning accuracy.

Method used

The improved short-time Fourier transform is used to obtain high-resolution time-frequency distribution images of generator set ports, and a convolutional neural network is used to build a sub-synchronous oscillation source recognition model to mine the connection between time-frequency distribution and disturbance source position to realize online Accurate positioning.

Benefits of technology

Through accurate acquisition and synchronous extrusion transformation of instantaneous frequency, the time-frequency resolution is improved, and the online identification of sub-synchronous oscillation sources is achieved in convolutional neural networks. It has engineering application prospects and is suitable for a variety of system operation modes and unit control strategies.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119312071A_ABST
    Figure CN119312071A_ABST
Patent Text Reader

Abstract

The invention discloses a subsynchronous oscillation tracing method based on improved STFT and a convolutional neural network, and belongs to the field of power system safety and stability. The method comprises the following steps: calculating short-time Fourier transform of active power of each unit port, and carrying out instantaneous frequency ridge line monitoring; fourier series fitting is used to obtain a reconstructed instantaneous frequency component after smoothing processing; a multi-overlapping component separation problem is considered, and path recombination is carried out according to the change rate of the components in the intersection interval to obtain a precise instantaneous frequency; performing synchronous extrusion short-time Fourier transform to obtain high-resolution time-frequency distribution; and taking the time-frequency distribution image of each unit port as the input of the model, constructing a subsynchronous oscillation source identification model based on the convolutional neural network, and mining the relation between the time-frequency distribution and the disturbance source position. According to the method, unit port active power high-resolution time-frequency distribution identification and subsynchronous oscillation source accurate positioning can be realized, and the method has important engineering significance for maintaining safe and stable operation of a power system.
Need to check novelty before this filing date? Find Prior Art