Real-Time Event Detection Based on STA/LTA Method Using Field Synchrophasor Measurements
Zhilin Chen, Hao Liu, Junbo Zhao, Tianshu Bi
Abstract
Real-time detection of frequency disturbances with varying time scales is crucial for maintaining situational awareness in power systems. Synchrophasor measurement units (PMU) can provide synchrophasor measurements but might be impacted by data quality issues, including data loss, anomalies, and noise problems, yielding misdetection and false detection. In this article, a real-time event detection method based on the Short Term Average to Long Term Average (STA/LTA) triggering algorithm is proposed. The method includes feature and scoring functions for STA/LTA calculation that consider the characteristics of fast (transient events) or slow dynamics (oscillatory and excursion events). This enables the detection of events at different time scales simultaneously and enhances the method's robustness to data quality issues. To distinguish real disturbances from frequency changes caused by non-disturbance factors, the events are detected collaboratively using all valid frequency data from multiple devices, which further increases accuracy. The proposed method is validated using field PMU measurements.