Improving the wide-area PMU-based fault location method using ordinary least squares estimation

• Published in: Electric power systems research Vol. 189; p. 106620
• Main Authors: Mouco, A., Abur, A.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.12.2020; Elsevier Science Ltd
• Subjects: Actuation; Algorithms; Estimating techniques; Fault diagnosis; Fault location; Fault tolerance; LARS; LASSO; Least squares; Measuring instruments; OLS; Phasors; PMU; Regression analysis; Sparse estimation; Fault location; PMU; Sparse estimation; LARS; LASSO; OLS
• ISSN: 0378-7796; 1873-2046
• DOI: 10.1016/j.epsr.2020.106620

•Faults are located on any line in the transmission system using a limited number of PMU measurements.•Fault location is independent of the fault type, unknown fault resistance.•Post-fault voltages are recovered from limited period of recorded transients after the fault occurs until the relay clears the fault. This paper presents an approach which allows system-wide fault location using only a sparse set of synchronized phasor measurements. The proposed algorithm combines post-fault voltage prediction and sparse estimation. The resulting method enables practical application of the sparse estimation formulation by using Prony analysis to predict post-fault steady-state voltages at those buses equipped with phasor measurement units. The prediction requires only a short transient recording which is typically available before the actuation of protective relays. Following the prediction step, the least angle regression based sparse estimation algorithm is employed to identify the fault location. However, for particular conditions usually associated with small fault currents, the sparse estimation algorithm may not be able to correctly identify the fault location. To mitigate this problem, ordinary least squares is used to increase the robustness of the proposed method. A 3-Phase model of the IEEE 118 bus system is developed on the alternative transients program to simulate realistic fault transients and test the performance and accuracy of the proposed fault location method.