Data Quality Management of Synchrophasor Data in Power Systems by Exploiting Low-dimensional Models

This extended summary introduces our recent work on phasor measurement unit (PMU) data analysis by exploiting the low-dimensional structures in spatial-temporal blocks of PMU data. An efficient online missing data recovery algorithm is proposed to leverage the low-rank property of the Hankel PMU data matrix to recover data losses even when all the measurements are lost simultaneously. A unified approach to data reduction, privatization, and recovery is proposed. Adding noise and applying quantization enhance the data privacy and reduce the communication burden. A data recovery method from quantized measurements is proposed to approximate the actual data asymptotically. The methods are accompanied with theoretical analysis and numerical evaluations on actual PMU datasets.