Research Highlights
Data Denoising and Compression for Smart Grid Communication
by Dr. Jesmin Khan, Dr. Gregory Murphy, Dr. Sherif Bhuiyan and Jonathan Williams
The basic concept behind Smart Grid (SG) can be defined as the transformation of the traditional power
grid using communications, artificial intelligence, advanced automatic control, information technology
and signal processing techniques. This concept allows effective generation, distribution, communication
and consumption of energy. In the Smart Grid, the monitoring and measurement units (e.g. smart meters,
frequency disturbance recorders (FDRs), phasor measurement units (PMUs), Wide Area Monitoring Systems
(WAMSs), Supervisory Control and Data Acquisition Systems (SCADAs)) record statuses across all levels
of the grid. Accordingly, there is an overwhelming flow of data to be circulated and stored among utilities,
control centers, and customers in SG in real time.
Therefore, the communication and storage of the data is an important issue in SG that needs effective
data compression. On the other hand, denoising of the power system signal is necessary for power system
disturbance analysis, as the effectiveness of the disturbance detection techniques is greatly deteriorated by
the noises riding on the signals. We propose a complete framework based on wavelet packet decomposition
(WPD) for power system data denoising and compression in smart grid communication. We verify our
proposed method on real data from FDR, PMU and power system load; which are recorded during the
occurrence of different types of faults. We have compared our method with wavelet decomposition (WD)
and the Matlab built-in function ‘wpdencmp’. The simulation results show that the