Power grid harmonic detection method based on variational mode decomposition

Abstract

The invention discloses a power grid harmonic detection method based on variational mode decomposition, which is characterized in that a key parameter K of the variational mode decomposition is selected via fast Fourier transform, a frequency value at each harmonic peak distributed in the frequency spectrum is taken as a central frequency initial value for the mode decomposition, a VMD local optimization direction can be effectively limited, and the final result is enabled to be converged at the target frequency in a local optimization manner. Meanwhile, a penalty factor [alpha] is gradually increased through monitoring the difference between an average value of the mode instantaneous frequency and a target mode instantaneous frequency and the variance of the mode instantaneous frequency,thereby not only being capable of acquiring a more accurate mode component, but also being capable of effectively avoiding a phenomenon of mode aliasing, redundancy and loss.

Description

technical field [0001] The invention relates to a grid harmonic detection method, in particular to a grid harmonic detection method based on variational mode decomposition that can extract phase, amplitude and frequency information of harmonic signals. Background technique [0002] With the large-scale application of new energy power generation systems, more and more power electronic devices have been introduced into the power system, resulting in excessive grid harmonics. Power grid harmonics will cause overcurrent of power electronic equipment, and in severe cases, it will cause major accidents such as large-scale off-grid of new energy power generation systems. Theoretically speaking, power electronic equipment can adjust the control strategy to suppress overcurrent on the premise of clarifying the comprehensive information (amplitude, phase, and frequency) of the harmonic signal characteristics of various types of power grids, and avoid the occurrence of equipment off-gr...

AI Technical Summary

Problems solved by technology

If the number of modal functions K is too small, the obtained modal components will have modal aliasing phenomenon, and the correct modal components cannot be obtained; if the number of modal functions K is too large, there will be redundant The generation of modal components leads to modal redundancy, which also increases the algorithm operation time and reduces the algorithm efficiency

Similarly, when If the value is too small, the modal aliasing phenomenon will be caused, and the parameter will be increased , although it can effectively increase the separation accuracy of modes and their center frequencies, but if the value is too large, mode mixing and mode loss will occur

In addition, the initial value of the center frequency of the harmonic component in the existing variational mode decomposition is zero, and sometimes the final result will locally deviate from the target frequency

Therefore, so far, there are no relevant reports on the use of VMD for power grid harmonic signal feature detection that can increase the separation accuracy of modes and their center frequencies, and avoid mode aliasing, redundancy and loss.

Images