Fault location proves its value to Elvia
Two real outages during an R&D project — a stone through the cable, and an excavator. Neither was preceded by partial discharge, and both were located anyway.
A live fault placed within 50 m on a 4,542 m segment — 1.1% of circuit length
On the second fault, the location was in Elvia’s inbox within five minutes
Fault location works even when no partial-discharge activity precedes the outage
About Elvia
Elvia is Norway’s largest power operator, supplying more than 1.8 million people across Akershus, Inland, Oslo and Østfold, with roughly 900,000 customers. Norwegian society expects uninterrupted supply at all times, and it is Elvia’s social responsibility to deliver it safely and reliably. In 2018 the company achieved security of supply equivalent to 99.98%, counting both planned and unplanned events.
The challenge
At 99.98%, the remaining prize was not fewer faults but shorter ones. Elvia was interested in anything that could reduce cable-related outage time, and started an R&D project to learn more about online quality monitoring of power cables.
Monitoring live cables for partial discharge — a good indicator of weak spots in an underground network — as well as for intermittent and full fault activity, can prevent outages outright or shorten them significantly.
Our contribution
Hafslund began the pilot in early 2018. The goal was to test the technology as a tool for monitoring the ageing cable network in Oslo, and ultimately to reduce SAIDI further.
Six feeders were monitored with six systems, each consisting of two sensor injector units and two control units — one of each placed at each cable end. Together they covered approximately 18 km of mixed cable, paper-insulated and cross-linked polyethylene, across 31 sections between ring main units.
The benefits
On 18 September 2018 a fault was detected on one of the monitored feeders, and Elvia received an SMS and an email with the details immediately. A stone had punctured the cable’s outer protection layer, most probably creating a high-ohmic fault between phase and ground, 384 m from the last station. The fault was positioned within 50 m of its actual location — on a 4,542 m monitored segment, an accuracy of 1.1%. "Not bad!" was the customer’s reaction. With many joints close to the detected spot, the operator analysed the information and pointed to a joint just two metres from the fault point.
Because the project was R&D, Elvia followed its normal procedures to find the outage and only compared the monitoring data with the findings afterwards. That, and the low degree of automation in this part of the grid, meant 1.5 hours before all customers were reconnected. Had the fault-location information been acted on, the outage could have been shortened to roughly 45 minutes.
In August 2019 the same system detected a second fault, this time from excavation work that damaged the cable, and emailed the fault information within five minutes of the event. Neither outage was preceded by any partial-discharge activity — which is precisely the case fault location exists to cover.
Originally published as a DNV case study.
See your grid the way we see it.
A 30-minute walkthrough of the platform on data from a network like yours.