Wireless Sensor Network-Driven Fault Localization in the 11kv Trans-Amadi Radial Distribution Feeders, Port Harcourt

Uncategorized

Authors: Dubon Tombari, Hachimenum.Nyebuchi Amadi

Abstract: The provision of reliable 11 kV radial distribution feeders is important in the continuous supply of electrical power to industrial, commercial and residential loads in Port Harcourt. Radial feeders are susceptible to faults which can cause voltage fluctuations, lead to high current, long duration outages and identification challenges in locating faults in spatial availability limited protection scenarios with visual manual patrols. This study presents a method of design of a WSN-driven 11 kV radial distribution feeder fault-locator model. This methodology involved mathematical models of the feeder measured voltage and current levels, WSN distributed measurement nodes, wireless data-rate requirement analyses, normalized fault detection, impedance-based fault-distance calculation, and performance evaluation based on detection error, detection time, network communication latency and packet-loss. The model was simulated using a 12-bus, 16.5 km radial distribution network that had the steady state feeder voltage decrease from 11.0 kV at the source to about 10.3 kV at the demand while downstream branch currents decreased from about 70 A to about 20 A. Simulated measurements were used for the voltage and current disturbance scenarios while the proposed communication model was simulated for sampling rates of 10-1000 samples/s, three measurement channels, 16 bit resolution and 20% protocol-overhead. Single-line-to-ground, line-to-line, double-line-to-ground and three-phase faults were used. The normalized fault-detection index was used to identify abnormal conditions with a threshold based decision criteria while impedance-based location was used to estimate the actual fault distance in 0.04-0.07 km of the control evaluation data set of ten. The estimated fault distances was used to estimate the location error range while the detection times of the model ranged from 16 to 24 ms. The simulated communication latency ranged from 46 to 55 ms. Over 200 simulation runs, the average detection, communication and processing times of the implemented model were 12.498, 11.255 and 12.055 ms respectively. The average packet-deliver-error of the communication model was 97.778%. This study presents a mathematical and computational model of distributed radial network monitoring and the quick location estimation of faults. Its application can enhance network outage-minimization, operational visibility and fault resolution capacity objectives in distribution network management. The simulation-based analysis provides a basis for future field deployment studies.

× How can I help you?