基于压缩感知的电力系统故障选线研究

Research on Power System Faulty Line Selection Based on Compressed Sensing

中文关键词:  故障选线  压缩感知  高斯随机矩阵  分割增广拉格朗日收缩算法(SALSA)  小波分解

英文关键词:faulty line selection  compressed sensing  Gaussian stochastic matrix  split augmented Lagrangian shrinkage algorithm (SALSA)  wavelet decomposition

中文摘要:

      为了解决电力系统故障选线中信号的采样、传输和存储问题，提出了一种全新的基于压缩感知理论的信号压缩的方法。该方法的采样频率不用考虑奈奎斯特采样频率。采样的信号是有选择性的部分信号。并通过设计重构算法来准确恢复该全部信号。考虑到一般条件下信号稀疏度不确定性，采用一种分割增广拉格朗日收缩算法（SALSA）来重构这些稀疏度不确定的信号。通过采用快速傅里叶变换基与高斯随机矩阵并且和SALSA相结合能够很好地实现信号压缩重构。对重构信号采用小波分解，获取重构信号的主要特征,分析零序电流模极大值的极性，找出其中一条与另外两条零序电流模极大值极性不同的线路，从而确定此线路为故障线路。

英文摘要:

      A new method of signal compression based on compressed sensing theory is presented to solve the problems of signal collection,transmission and storage in power system faulty line selection.The sampling frequency of this method did not need to consider the Nyquist sample frequency.The sampling signal was a selective partial signal.The reconstruction algorithm was designed to recover the full signal accurately.Considering the uncertainty of signal sparsity under the general conditions,the split augmented Lagrangian shrinkage algorithm (SALSA) was used to reconstruct these signals.By using the fast Fourier transform basis,the Gaussian stochastic matrix and the SALSA,the signal can be compressed and reconstructed.The wavelet transform was used to obtain the main characteristics of the reconstructed signal,and the polarity of the zero sequence current modulus maxima was analyzed to find one of the lines that was different from the polarity of the other two of the zero sequence current modulus maxima.A different one is determined as the faulty line.