Event Identification Based on Synchronized Waveform Using Composited Quantum Kernel Support Vector Machine

Inverter-Based Resources (IBRs), such as Photovoltaic (PV), wind power generation, and Battery Energy Storage Systems (BESSs), are widely integrated into low-carbon power systems. However, due to their limited inertia support, they are prone to inducing broadband oscillations and hyper-harmonics, which increase the risks of generation tripping and load shedding. To effectively identify these anomalies, this paper proposes a novel method for detecting events in low-inertia systems. First, a distributed simulation model with 14 PV nodes is developed based on the IEEE 123-bus test feeder to simulate representative abnormal operating conditions. Subsequently, voltage and current measurements from these nodes are collected to construct a synchronized waveforms-based events database. Thereafter, a Composited Quantum kernel Support Vector Machine (CQSVM) method is proposed. By introducing a weighted kernel function fusion strategy, the expression ability of nonlinear features and anomaly recognition performance are improved. Finally, experimental results on six types of event datasets demonstrate that the proposed method exhibits better recognition accuracy and robustness compared with state-of-the-art recognition algorithms.