Probabilistic estimation of voltage sags using erroneous measurement information

• This paper proposes a robust method to locate faults and estimate the magnitude of voltage sags using limited information.
• Statistical analysis was used to find the most likely fault location and sag magnitude at non-monitored busbars.
• Robust handling of measurement errors reduces the sensitivity of impedance based fault location algorithms.
• The method eliminates multiple fault location solutions caused by overlapping impedance paths.
• It is validated on a generic section of the UK's distribution network.

Estimating voltage sag performance is important for distribution network operators who are keen to reduce costly interruptions, plan network investment and reduce operational expenditure. This paper proposes a robust method to locate faults and estimate the magnitude of voltage sags using information from a limited set of arbitrarily accurate monitoring devices. The developed method uses statistical analysis and impedance based fault location equations to find the most likely fault location and sag magnitude at non-monitored busbars. The method robustly handles measurement errors, and helps to eliminate some of the sensitivity present in existing impedance based fault location algorithms. The method is also shown to be effective at eliminating multiple fault location solutions caused by multiple overlapping impedance paths by synthesizing information from all monitors installed in a network. The method is validated and shown to be effective on a generic section of the UK's distribution network.