基于自适应小波基双迭代同步挤压小波变换的电力系统间谐波检测方法研究

This project focuses on the accurate detection of time-varying inter-harmonics in dense frequency condition, and systematically analyzes the optimization and design of inter-harmonic detection based on Synchrosqueezed Wavelet Transform (SWT). The main contents include: 1) design adaptive three-parameter Morlet wavelet and take it as the primary function of SWT, to improve anti-mode mixing ability of SWT, and distinguish spectral range of inter-harmonic components more clearly; 2) introducing the idea of cross validation into inter-harmonic detection based on SWT, to improve judgment performance to true and false inter-harmonic spectral line and calculate the number of inter-harmonic components precisely; 3) constructing SWT of zoom and double iteration based on the accurate calculated inter-harmonic components, and stepwise search for precise spectral ranges, simultaneously eliminate interaction impacts among spectra of different components using band-pass filtering for more accurate detection of inter-harmonic parameters. 4) constructing Synchrosqueezed Wavelet Packet Transform (SWPT) analysis method. The SWPT and frequency doubling method are combined to make the inter-harmonic frequency deviate from the central frequency of chaotic noise, and overcome the influence of complex noise on the detection of dense inter-harmonics. An inter-harmonic detection method based on adaptive SWT of double iteration is proposed, which provides a key basis for accurate detection and control of time-varying inter-harmonics in dense frequency. Therefore, the project not only has certain scientific and theoretical significance, but also has more important application value.

本项目围绕密集频率条件下电力系统时变间谐波的精确检测展开研究，系统地分析基于同步挤压小波变换(SWT)的间谐波检测优化与设计问题。主要内容包括：1)设计自适应三参数Morlet小波作为SWT的基函数，提高态SWT的抗模混叠能力，更加清晰地区分间谐波分量的频谱范围；2)引入交叉验证思想，提高SWT对真伪间谐波谱线的判定能力，精确计算间谐波分量个数；3)构造变焦双迭代SWT，利用迭代逐步探测间谐波分量提取时的精确频谱范围，同时利用带通滤波去除分量频谱之间的交互影响，更精确的检测间谐波参数；4)构造同步挤压小波包分析法，结合倍频法使间谐波频率偏离混沌噪声中心频率，克服复杂噪声对密集间谐波的检测影响。通过研究，提出基于自适应双迭代SWT的间谐波有效检测方法，为密集频率条件下时变间谐波精确检测及治理提供关键基础。因此本项目既有一定的科学理论意义，也有较重要的应用价值。

随着风能、光伏等新能源发电以及储能设备大规模并网运行，非线性电力电子的大量使用，导致大量的冲击性负荷接入电网，使电网中产生了大量的谐波、间谐波。而间谐波不仅具有谐波的危害，还可能在传播中被放大而影响电网其他设备的正常运行，从而产生更大的安全运行隐患。由于同步挤压小波变换(SWT)具有谐波检测精度高、抗噪性能强等优点，本项目围绕密集频率条件下电力系统间谐波的精确检测展开研究。本项目的主要研究结果为：1）在原有“阈值函数的分数阶小波”的基础上对阈值函数进行改进，使其适应微电网扰动检测信号去噪要求，并利用最小熵理论确定Morlet小波参数，以提高SWT对谐波的检测精度。2）将改进的小波包分解算法引入到SWT中，结合K折交叉验证，改善多分量信号频谱混叠现象，精确确定间谐波个数。3）通过注入高频信号，结合HHT，提高对暂态扰动信号的提取精度，达到抑制模态混叠的目的，同时对高频注入HHT时频图进行同步挤压处理，利用双迭代方法优化噪声环境下挤压后时频图的聚焦能力，提高密集频率间谐波的抽取精度。4）在同步挤压小波包变换(SWPT)的基础上，引入频移理论和倍频法，使间谐波频率偏离混沌噪声中心频率，提高对高频间谐波信号的辨识精度。. 本项目联合同步挤压小波算法、小波包分解算法、频移理论构造了一种抗噪性能强、具有一定普适性的密集间谐波检测算法，通过最小熵理论和倍频法提高高频检测精度，结合K折交叉验证，改善多分量信号频谱混叠现象，精确确定谐波/间谐波个数，为分析、治理电网中的谐波/间谐波提供理论依据。