Assessment of smart grid fault isolation under high renewables penetration: field measurements

Thesis Abstract

The rapid expansion of renewable energy sources within power systems has intensified the challenge of reliable fault isolation due to reduced inertia, bidirectional power flows, and accelerated dynamics in network protection settings. This study addresses the problem of maintaining robust fault isolation performance in smart grids operating with high penetration of renewables, where conventional protection schemes experience degraded selectivity, nuisance tripping, and delayed isolation. The aim is to evaluate, validate, and enhance field-based fault isolation mechanisms under realistic high-renewables conditions, with objectives to (i) quantify the impact of renewable-induced variability on fault detection and isolation times, (ii) assess the performance of adaptive protection schemes and distributed energy resources (DER) communications in isolating faults, (iii) identify key indicators from field data that reliably predict isolation errors, and (iv) propose a data-driven framework for real-time fault isolation decision support. The study adopts a mixed-methods approach grounded in systems theory and control engineering, integrating empirical field measurements with simulation-based validation. The population comprises medium- and distribution-level networks in a regional grid with renewable penetration exceeding 60% of hourly generation during multiple seasons. A stratified sample of 12 substations, 36 feeders, and 8 DER clusters is selected based on variability in topology, protection configurations, and communication latency. Data collection combines (1) high-fidelity phasor measurement unit (PMU) datasets capturing voltage, current, frequency, and protection trips at 1 ms resolution; (2) protective relay event logs and protection coordination settings; (3) SCADA-derived operational states and DER output data; and (4) field surveys documenting protection device ages, upgrade histories, and communication network reliability. Instrument validity is ensured through calibration against known fault scenarios and cross-validation with recorded disturbance databases. Data analysis employs a multi-tiered methodology descriptive statistics to characterize baseline protection performance; time-series analysis and event-based correlation to link isolation delays with renewables variability; regression analysis to quantify the influence of wind/solar fluctuations and DER contributions on false trips and missed faults; survival analysis to model time-to-isolation under different grid states; and machine learning classifiers (random forest, gradient boosting) to identify features most predictive of isolation anomalies. A modular analytical framework aligns with a consolidated model that integrates a theoretical basis in networked control systems and two named theories complex adaptive systems theory to explain emergent protection behavior under high DER penetration, and the unified power system protection theory to interpret coordination failures. The study further develops a simulation-augmented validation using EMT-type time-domain models and digital twins to reproduce field-operating conditions and test adaptive mitigation strategies. Expected findings include (i) quantification of a threshold renewables penetration level beyond which conventional protection loses reliability, (ii) evidence that adaptive protection schemes with fast communications and DER-aware settings reduce mean isolation time by 25–40% and decrease nuisance trips by 30–50%, (iii) identification of critical indicators such as rate-of-change of regional frequency, DER curtailment events, and latencies in protective relays as predictors of isolation errors, and (iv) a data-driven decision-support framework for real-time isolation reconfiguration. The study contributes to knowledge by providing empirical field-based benchmarks for fault isolation performance in renewables-rich grids, integrating theoretical constructs with practical insights to guide protection modernization and DER coordination. The main conclusion posits that, with targeted adaptive strategies and enhanced situational awareness, fault isolation reliability can be substantially maintained in high-renewables environments. Recommendations include adopting DER-aware protection setting guidelines, implementing low-latency communication protocols for critical protection signals, deploying digital twin-enabled testing of protection schemes, and pursuing policy-and-infrastructure support for field data sharing to foster continuous improvement of smart grid resilience.