Interharmonics Parameter Estimation Based on Improved SVD Algorithm

Ruijiao Shen,Honggeng Yang
DOI: https://doi.org/10.1049/cp:20061680
2006-01-01
Abstract:Singular value decomposition (SVD) harmonic detection algorithm operates the matrix composed by the sampling data directly, extrapolates the data, it greatly improves the frequency resolution, and is propitious to detect frequency of interharmonics. However, there will be bigger warp at the of frequency estimation when noise exists. Through analyzing the noise error, this paper presents an improved algorithm, this algorithm resolves the minimum norm problem considering the noise interference at both sides of the matrix equation. This interharmonics frequency estimation method can obtain the accurate frequency response coefficient by SVD of a new augmented matrix and calculation of the right singular matrix. The paper also proposes a simple and useful method to determine the amount of main singular value related to SNR. After gaining the signal frequency and denoising to a certain extent, the amplitude and phase of interharmonics are estimated by the method of triangle regression. The simulated result shows that the new algorithm is possessed of good anti-noise performance and number stability, can extract the frequency of the signals precisely under the condition of lower SNR ratio. (7 pages)
What problem does this paper attempt to address?