Dynamic Estimation and Fusion of Frequency and ROCOF Data via Local Gossiping

Phasor Measurement Units (PMUs) constitute the backbone of many monitoring and control applications in power systems. Frequency and Rate of Change of Frequency (ROCOF) of ac voltage waveforms are closely related to system stability. Therefore, these parameters are commonly monitored by PMUs to trigger appropriate protection or control actions when either the frequency or the ROCOF values exceed predefined critical thresholds. However, previous studies revealed that PMU measurement uncertainty, particularly for ROCOF, drastically grows during fast transient events, making the reported values less reliable. In this paper, this problem is mitigated by an Adaptive Kalman Filter (AKF), potentially running on each node. Under stationary conditions the AKF assumes the ROCOF to be zero and the frequency to be constant and it relies solely on local PMU frequency and ROCOF data in the update step. In nonstationary conditions instead (i.e., when some anomalous event is detected) the local PMU values are merged with those “gossiped” by the PMUs installed at adjacent nodes used as an additional input to the model. A basic consistency check is performed to filter possible and bad data. The algorithm performance is analyzed through simulations using the IEEE 5-bus test system as a case study in an Under-Frequency Load Shedding (UFLS) recovery scenario.