面向用电侧电能质量监测的时空压缩感知方法

A Power Quality Monitoring Oriented Time-Space Compressed Sensing Method

针对当前电能质量监测采样数据量庞大、前端数据计算处理负荷激增的问题,提出了一种自检测电能质量数据时空压缩感知方法。该方法采用压缩感知理论突破了基于奈奎斯特采样定律的传统采样方式需要大量传输数据的瓶颈;并充分考虑多个同类型测量对象的数据相关性,将不同小区的用电数据投射为二维矩阵,通过二维离散余弦变换稀疏基变形推导实现时空压缩感知;且针对冲击脉冲信号重构识别不稳定提出脉冲自检测机制与压缩感知模型结合;最后经实验证明该方法能实现电能质量数据的准确重构、提高传输数据压缩率、缩短采样重构周期从而增大控制中心可处理数据容量,提高工作效率。

Investigated here is a time-space compressed sensing method to reduce the data acquisition amount and front-end traffic load in power quality monitoring system. Sparse sampling is adopted based on compressed sensing as an alternative option to the traditional Nyquist sampling. Considering the data correlation among user areas of a same type, the electricity data are projected into a two-dimensional matrix and then the time-space compressed sensing is achieved through two-dimensional discrete cosine transform. Furthermore, a self-detection mechanism is proposed to recognize the shock pulse signal. Experiment results verify that the method could reconstruct the data accurately with high compression ratio and short reconstruction period. The data processing capacity is increased, and the operating efficiency is improved.