Line and Generator Outage Identification using Synchrophasor Measurements

Identification of outages in power systems is crucial to raise situational awareness and prevent cascading failures. As a result of wide adoption of phasor measurement units (PMUs), synchronized measurements with high temporal resolution provide new opportunities for event detection and identification in electric power systems. This thesis focuses on two types of outages: line and generator outages. For line outages, an identification method based on the ac power flow model and voltage phasor measurements is proposed. The results demonstrate improved accuracy compared to the commonly used dc power flow approach and the significant benefits of having a higher PMU coverage. By using the proposed rejection filtering techniques, the misidentified rate can be further reduced, which is crucial to the operators. The results show relatively high identification accuracy with a small amount of PMUs. For generator outages, three algorithms for outage localization are proposed utilizing (1) frequency and ROCOF measurements, pre-outage generations and inertia constants; (2) line flow measurements and linear sensitivity factors; and (3) voltage measurements, system admittance matrix, generator internal reactances and generator injections prior to the outage events. Unlike in previous works, the frequency-based algorithm has been tested using frequency measurements generated by a realistic PMU model (proposed and tested against an SEL 451 relay). It is shown to be limited in accuracy. In contrast, the LSF-based method is able to correctly localize the origin of outage even when there are cascading generator outages in the external system. However, the existence of (nearly) indistinguishable generators may create numerical problems. A QR decomposition based clustering technique is proposed to group generators based on their impacts on available synchrophasor measurements so that the outage can be localized to the originating cluster with zero misidentification. The measurement selection problem (or PMU placement problem) can also be solved based on the same. Experimental results also demonstrate that the third proposed method, based on post-outage voltage estimation, can successfully exploit additional information (if available) to achieve even better localization. All of the proposed identification algorithms can be implemented for online application in utility control centers.

Department dc:contributor.department

Electrical and Computer Engineering

Year dc:date.issued

2019

Author dc:creator

• Dai, Zhen

Advisor dc:contributor.advisor

• Tate, Joseph E

• Outage detection
• Power systems
• Synchrophasor

Handle dc:identifier.uri

http://hdl.handle.net/1807/97391

OAI identifier oai:identifier

oai:utoronto.scholaris.ca:1807/97391