Two outages detected, 1 excavation work damaging the cable and 1 stone penetration into the cable
Hafslund Nett owns, develops and operates the electrical distribution grid in south-eastern Norway and has roughly 720 000 customers. Norwegian society expects an uninterrupted supply of energy at all times and it is Hafslund Nett’s social responsibility to ensure safe and reliable supply of energy. In 2018, Hafslund Nett achieved security of supply equivalent to 99.98 per cent, taking into account both planned and unplanned events.
Nevertheless, they are interested in anything that could help reduce cable-related outage time. Hafslund decided to start an R&D project to learn more about online quality monitoring of power cables. Monitoring live cables on partial discharge (PD; PD- activity is a good indicator of weak spots in underground cable networks) as well as intermittent and full fault activity can help them preventing outages or shortening the outage time significantly.
Hafslund learned about DNV GL’s Smart Cable Guard (SCG) system and started their R&D pilot with SCG early 2018. The goal of the project was to test the SCG technology as a tool to monitor the ageing cable network in Oslo and ultimately reducing SAIDI even further. In this project 6 feeders are monitored with 6 Smart Cable Guard systems where each system consists of two sensor injector units (SIU) and two control units (CU), one of each placed at each cable end.
Example of a 1 system on 1 circuit visualized in the SCG web interface as a single line drawing
The 6 SCG systems monitor a total of approximately 18 km mixed cable (paper insulated cables and cross-linked polyethylene cables) and those 18 kilometers are consisting of 31 cables sections (cables sections between ring main units) with mixed cable types.
Outage 1- Stone penetrating a cable
On the 18th of September 2018, the SCG system detected a fault on one of the feeders being monitored by the SCG systems. Hafslund received an SMS and email with the information about the fault straight after the fault was detected by the SCG system.
Although the system did not pick up any PD activity before the outage, the fault feature of the SCG system did its work and indicated the fault to the spot on the cable section.
SCG web site information details on fault
Fault location indicated on single line diagram
What happened: a stone that punctuated the outer protection layer of the cable. Most probably a high-ohmic fault between phase and grounding was created. The position was 384 m from the last station. The SCG system positioned the fault in a range of 50 m from the actual fault location. Given the total length of the monitored segment of 4542 m, SCG was 1,1% accurate. “Not bad!”, was the customer reaction. There were many joints close by the spot detected by the SCG system making it more likely that a joint failed. The SCG operator who was available in this R&D project, analyzed the information and advised a joint only 2 meters away from the fault point.
Due to the R&D character, Hafslund’s normal procedures were followed to find the outage and the SCG information was compared with the actual findings only afterwards. This and the low degree of automation (e.g. remote-controlled circuit breakers) in this area of the grid, resulted in an outage time of 1,5 h before all customers had power again. Had the information from the SCG-system been taken into account, could the outage have been shortened to about 45 minutes.
The SCG systems self-test was tested after this repair as after repairs the cable was set live again and normally the SCG system should then start automatically which it did not. The SCG system has a systems self-test and in case of malfunctioning the SCG operator is warned which the contacted Hafslund. After local investigation it turned out that during the repair works the power was disconnected from the SCG sensor in that station, after connecting the sensor to power the system went live and is guarding the circuit since.
Outage 2- Excavation work
In August 2019 the same SCG system detected another fault on the cable. The following fault information was sent by email within 5 minutes after the event happened to Hafslund:
Email send to Hafslund with fault information
Screenshot of the SCG web interface showing thefault on the single line diagram
In this case the outage was created by excavation works on that location that damaged the cable. Although the system did not pick up any PD activity before the outage, the fault feature of the SCG system did its work and indicated the fault to the spot on the cable section.