Cost-Effective Bad Synchrophasor Data Detection Based on Unsupervised Time-Series Data Analytic

In modern smart grids deployed with various advanced sensors, e.g., phasor measurement units (PMUs), bad (anomalous) measurements are always inevitable in practice. Considering the imperative need for filtering out potential data, this paper develops a novel online PMU data detection (BPDD) approach regional concentrators (PDCs) by sufficiently exploring spatial-temporal correlations. With no costly labeling or iterative learning, it performs model-free, label-free, and non-iterative BPDD power from new data-driven perspective of nearest neighbor (STNN) discovery. Specifically, spatial-temporally correlated acquired PMUs first gathered as time series (TS) profile. Afterwards, TS subsequences contaminated identified characterizing anomalous STNNs. To make whole competent processing streaming an efficient strategy accelerating STNN discovery is carefully designed. Different existing solutions requiring either offline dataset preparation/training computationally intensive optimization, can be implemented highly cost-effective way, thereby being more applicable scalable practical contexts. Numerical test results on Nordic system realistic China Southern Power Grid demonstrate reliability, efficiency scalability proposed monitoring.