PMU-Driven Fault Location in Distribution Systems with DERs: Impedance-Based and Multi-Agent System Strategies

Efforts are increasingly directed toward the adoption of low–carbon generation technologies to address environmental concerns arising from greenhouse gas emissions. Modern active distribution networks pose significant challenges due to their complex configurations, characterized by feeder heterogeneity, low reactance-to-resistance ratios, and unbalanced loads. These factors complicate system awareness and observability, making fault detection and isolation essential for rapid reconfiguration and service restoration. The growing demand for power quality, security, and reliability has driven distribution system operators to deploy advanced measurement and communication technologies. In this context, fault location, isolation, and service restoration (FLISR) systems supported by advanced distribution management systems enhance distribution network reliability by enabling optimized grid reconfiguration under non-ideal operating conditions. The integration of renewable energy resources and distributed generation further compounds these challenges, introducing protection and stability issues, unpredictable power flows, and increased difficulty in fault detection, particularly for high-impedance faults (HIFs). These complexities highlight the necessity of accurate fault location techniques to support efficient repair dispatch, minimize outage impacts, and maintain overall grid reliability. This research aims to enhance system observability and stability in distribution system (DS) by providing accurate fault location estimates to service providers. With the increasing deployment of PMUs in distribution networks, this study leverages synchrophasor measurements for fault location, addressing gaps in the existing literature. Conventional fault location methods often rely on centralized architectures, creating a single point of failure due to their dependence on central computing facilities. To mitigate this limitation, the study investigates decentralized, knowledge-based approaches specifically MAS for effectively managing complex outage and load scenarios. The work focuses on two complementary strategies. First, it develops an extended impedance-based fault location algorithm for DS with DERs, utilizing synchrophasor data from PMUs. Particular attention is given to accurately locating HIFs, which are especially critical due to their potential role in wildfire initiation, while considering the practical constraints of limited synchrophasor availability. Second, the study explores the design and implementation of a MAS-based fault location and isolation framework, where PMUs act as intelligent agents capable of local decision-making and inter-unit communication across the network. For the research, initial step was to develop a test system and towards this a DS test case with DERs in PSCAD/EMTDC platform is developed. IEEE 34 bus distribution system is considered as the test system, with loads and lines modeled as close to the original feeder as possible. A type 2 wind turbine generator model is considered and placed at two of the bus to modify the base system to a multi-source network. For obtaining synchrophasors for the algorithm, a 3 phase PMU model is developed in PSCAD. Since the aim is to use the existing PMUs in the network, for the scope of this study, optimum PMU placement for DS monitoring purpose is considered. Towards that, a study on OPP is conducted and an approach is developed with slight modifications to incorporate the DERs in the system. Taking into consideration the ease of implementation and cost effectiveness, first, a novel impedance based method is proposed to locate the fault by using the highly accurate time-synchronized voltage and current phasors obtained from distribution phasor measurement units. The synchrophasor measurements obtained from the substation and various feeder segments are used in a two-step algorithm based on the apparent impedance calculation to locate the exact source of the event. The algorithm uses phasor estimates to first identify the faulted feeder sub region and later uses measurement from a remote end device to eliminate pseudo-faulted points to obtain the actual fault location. The effectiveness of the proposed method is realized using PSCAD/EMTDC and MATLAB platform on an IEEE 34 bus system. Based on different fault types simulated at various parts of the system, the algorithm accurately estimates fault location in the range of ±1% of the line length. Sensitivity analysis is performed on the developed impedance algorithm for parameters like fault resistance, load variation, effect of DERs and distance to fault. The algorithm was more sensitive towards fault resistance and was successful in locating all faults within the range of 0.01 Ω to 10 Ω. The proposed method is effective in locating faults for any type of network and topologies, with as many or as few (minimum 2) phasor measurement units in the system. A hybrid multi-agent system architecture for fault location and isolation in both conventional and DER-based distribution networks, utilizing a minimal number of PMUs is proposed. The developed multi-agent framework uses a decentralized hierarchical agent scheme and consists of heterogeneous agent types, with the PMUs defined as the main computational agents. The agents receive their local synchrophasor measurements and perform the algorithm by communicating with their neighboring agents to locate the faulted bus with minimal communication messages. The new fault location algorithm is a two-step process: first, the agents identify the faulted section using impedance comparison and isolate the faulted zone; second, the corresponding agents within the zone identify the faulted bus using voltage comparison and send the bus information to the control center. Studies are conducted by interconnecting the IEEE 34-bus network with distributed energy resources developed in PSCAD/EMTDC and a multi-agent system implemented on the JADE platform. Simulations of test cases verify the efficacy of the proposed multi-agent scheme in accurately locating various faults with fault resistance upto 100 Ω occurring in the distribution network. The average computational time for the agent implemented system is found to be 19 ms, within the acceptable range. The algorithm also effectively locates high impedance faults for fault currents ranging 10 - 50 amps (approximately 1,500 Ω) upon receiving disturbance detection alarms. The fault location method on an active distribution systems using PMUs offers an excellent solution to electrical power utilities to improve reliability. In this context, the proposed algorithms, methods, and research findings in this thesis would help implement fault location techniques in DS to realize reduced outage length and repair costs.

Name thesis:degree_name

Doctor of Philosophy (Ph.D.)

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Electrical Engineering

Grantor

University of Saskatchewan

Year dc:date.issued

2025

Author dc:creator

• Chandran, Sandhya

Advisor dc:contributor.advisor

• Gokaraju, Ramakrishna

Committee members dc:contributor.committeemember

• Bretas, Arturo
• Teng, Daniel
• Faried, Sherif
• Jamali, Nadeem
• Mehr, Aryan Saadat

• Distribution systems, Fault location, Phasor measurement unit, Synchrophasors, FLISR, Impedance based, Multi-agent system, Voltage comparison, Optimum PMU placement

Language dc:language.iso

en

Handle dc:identifier.uri

https://hdl.handle.net/10388/17248

OAI identifier oai:identifier

oai:harvest.usask.ca:10388/17248