Abstract:This paper introduces and compares the various techniques for identification and analysis of low frequency oscillations in a power system. Inter-area electromechanical oscillations are the focus of this paper. After multiresolution decomposition of characteristic signals, physical characteristics of system oscillations in signal components are identified and presented using the Fourier transform, Prony method, Matrix Pencil Analysis Method, S-transform, Global Wavelet Spectrum and Hilbert Huang transform (Hilbert Marginal Spectrum) in time-frequency domain representation. The analyses were performed on real frequency signals obtained from FNET GridEye system during the earthquake that triggered the shutdown of the North Anna Nuclear Generating Station in the east coast of the United States. In addition, according to the obtained results the proposed methods have proven to be reliable for identification of the model parameters of low-frequency oscillation in power systems. The relevant analyses are carried out in MATLAB coding environment.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to identify and analyze low - frequency oscillations in power systems, especially inter - area electromechanical oscillations. Specifically, the paper aims to:
1. **Compare different techniques**: Various techniques for identifying and analyzing low - frequency oscillations in power systems, including Fourier Transform, Prony method, Matrix Pencil Analysis Method, S - transform, Global Wavelet Spectrum, and Hilbert - Huang Transform (HHT).
2. **Multi - resolution decomposition**: Identify and present the physical characteristics of system oscillations by performing multi - resolution decomposition on characteristic signals.
3. **Verify practical applications**: Verify the effectiveness of these methods using the real - frequency signals recorded by the FNET/GridEye system during the August 23, 2011, earthquake on the East Coast of the United States. This earthquake led to the shutdown of the North Anna nuclear power plant and caused a 1,600 - MW power generation loss.
4. **Evaluate method reliability**: Evaluate their reliability in identifying the parameters of low - frequency oscillation models in power systems by comparing the results of different methods.
### Main problem summary:
- **Background and motivation**: The power system is a complex dynamic system that is vulnerable to various perturbations. These perturbations may lead to system collapse, so effective monitoring and control measures are required to ensure system stability.
- **Research object**: The paper focuses on low - frequency oscillations caused by major perturbations (such as earthquakes), especially inter - area electromechanical oscillations.
- **Method selection**: Multiple time - frequency analysis methods are selected to improve the ability to identify non - stationary signals.
- **Application scenario**: Verify the effectiveness and reliability of the selected methods through a practical case (the 2011 East Coast earthquake in the United States).
### Formula example (taking the Prony method as an example):
The Prony method is used to approximate equally - spaced sampled data as a linear combination of exponential functions. For a real signal, it can be expressed as:
\[
x(nT)=\sum_{k = 1}^{p}a_k e^{\alpha_k nT}\cos(2\pi f_k nT+\theta_k)
\]
where:
- \(n = 1,2,\ldots,N\),
- \(N\) is the signal length (number of samples),
- \(p\) is the number of exponential components (model order),
- \(T\) is the sampling period (seconds),
- \(a_k\) is the amplitude of the \(k\) - th component,
- \(\alpha_k\) is the damping factor (1/s) of the \(k\) - th component,
- \(f_k\) is the sine frequency (Hz) of the \(k\) - th component,
- \(\theta_k\) is the initial phase (radians) of the \(k\) - th component.
Through the application of these methods, the paper shows how to effectively identify and analyze low - frequency oscillations in power systems, thereby providing strong support for the stable operation of power systems.