A Simplified Disturbance Detector Approach Based on Self-Adjusting Thresholds

The increasing prevalence of Inverter-Based Resources (IBRs) in modern power systems introduces new challenges to conventional fixed-threshold fault detection methods, which often lack adaptability and robustness under varying system topologies, noise levels, and signal resolutions. In this context, this paper presents a novel disturbance detection methodology that combines the simplicity of a low-cost signal processing technique with an adaptive thresholding-based detection logic. The method relies on the modulus of incremental current and employs self-adjusting thresholds derived from Mathematical Morphology. Extensive validation is conducted through simulations on a detailed real-world wind farm model comprising 120 full-converter turbines and 27,900 distinct fault scenarios. The simulation set includes variations in fault type, resistance, and inception angle, as well as the operating power factor and generation level of the wind farm. The proposed approach consistently achieves high accuracy, above 99%, across all scenarios, maintaining resilience under challenging signal-to-noise ratios as low as 40 dB and reduced sampling rates down to 16 points per cycle, without requiring system- or scenario-specific threshold tuning. Comparative analyses with conventional fixed-threshold techniques, including Sample-by-Sample, Cumulative Sum, and Current-Slope Based methods, highlight the superior robustness, accuracy, security, and deployment potential of the proposed method. Furthermore, processing benchmarks demonstrate a computational time reduction of approximately 99% compared to more complex techniques such as Wavelet Transform and entropy-based analysis, reinforcing its suitability for embedded and real-time protection systems.