2026-04-17T14:03:00Zhttps://recercat.cat/oai/request

oai:recercat.cat:2117/4444562025-10-26T12:29:16Zcom_2072_1033col_2072_452951

Verification of the dependability of an improved negative-sequence based shunt reactor protection in a power grid with inverter-based resources (IBR) Tohamy, Youssef E. Àrees temàtiques de la UPC::Enginyeria elèctrica Electric networks -- Protection Electric power systems -- Protection Smart power grids Negative-sequence current; shunt-reactor protection; Turn-to-turn fault; Inverterbased resources; Grid-following inverter; Negative-sequence injection; Short-circuit ratio; Dependability; Sensitivity; Moving-average filter Xarxes elèctriques -- Protecció Sistemes de distribució d'energia elèctrica -- Protecció Xarxes elèctriques intel·ligents This thesis validates a simplified negative-sequence–based protection scheme for turn-to-turn faults (TTFs) in shunt reactors and benchmarks its dependability in both conventional synchronousmachine grids and inverter-based resource (IBR) environments. A 500 MW transmission test bench—mark connected to a 50 MW offshore wind farm through land/sea cables, and shuntreactor compensation—was modeled. The farm Model was implemented in three different ways: (i) a conventional three-phase voltage source (Model I), (ii) a grid-following inverter (Model II), and (iii) a grid-following inverter with negative-sequence injection (Model III). The protection algorithm measures terminal currents, extracts their negative-sequence component, converts it to per-unit, and issues a trip when a fixed threshold is exceeded. Dependability was assessed for two fault severities (TTF = 1 % and 0.065 %) and two grid strengths (SCR = 10 and 1). Results show that the terminal negative-sequence current |I2|—and hence sensitivity—declines with decreasing SCR, yet remains above the detection threshold for SCR ≥ 0.1. At TTF = 1 %, all models tripped reliably; Model I generated the largest |I2| in every case, while Model III surpassed Model II in weak grids. At TTF = 0.065 %, detection margins narrowed: Model II edged Model III in strong grids, but Model III regained superiority under weak grids despite Moving-Average-Filter limitations. A practical lower-sensitivity bound was identified at TTF ≈ 0.035% for SCR = 5. Parametric sweeps confirmed near-linear |I2| trends versus both TTF %, with Model I exhibiting the steepest gradients. Overall, Model I remains the most robust, Model III is the best IBR solution—approaching Model I in weak grids or at higher fault levels—and Model II is adequate only in strong grids. These findings demonstrate that the proposed algorithm can dependably detect TTFs down to 0.065 % even in inverter-dominated, low-SCR networks and highlight the value of negativesequence injection for future IBR protection strategies. Outgoing 2025-10-26T12:29:16Z 2025-10-26T12:29:16Z 2025-10-26T12:29:16Z 2025-06-27 Master thesis http://hdl.handle.net/2117/444456 Open Access Universitat Politècnica de Catalunya