Suynchrophasor Data-Driven Investigation of Real-World Unstable Controller Dynamics

(14.7 Hz). Following the decommissioning of a hydro power plant, a region experienced sustained oscillations. Long-term data analysis revealed a latent 10–11 Hz mode with reduced damping margin, which could result in a 14.7 Hz saturated plant response. The root cause was traced to the input stage of UPS units and verified by the data centre operator. Case 2: PV

Controller Delay (0.05 Hz). A PV plant exhibited slow, square-wave-like reactive power response following a line outage. System identification ruled out conventional nonlinearities, instead pinpointing significant feedback delays within the plant controller. Case 3: Abnormal

PV plant dynamic behaviour during sunrise and sunset (0.07 Hz). A 105 MW PV facility showed recurring abnormal dynamic response during sunrise and sunset (low active power operation). Analysis identified the cause as nonlinear inverter dynamics attempting to maintain power factor during ultra-low output periods (<1 MW), combined with disabled POI voltage control and low measurement accuracy. Overall, this paper illustrates synchrophasor measurement-based spectral analysis and system identification serve as a vital diagnosis tool to understand emerging dynamics from inverter-based assets, especially where models are absent.

By isolating specific root causes - hardware design, control delays, and nonlinear logic, the methodology enables targeted mitigation strategies for grid stability.