Lightning strikes are known to cause a significant number of supply interruptions to our customers. In our Scottish Network, lighting strikes are the second highest cause of customer interruptions and minutes lost and in our Southern Network it is the fifth highest cause. Therefore, there is a need to reduce the impact that lighting related faults have on our customers.
Objectives
1) Develop a ‘point in time’ lightning analysis tool that can be used to locate lightning protection equipment in the most optimal way i.e. integrate various data sets and update visual display as described in phase 1a and 1b above.
2) Install lightning protection equipment in ‘optimal’ locations provided by the lightning analysis tool.
3) Monitor and analyse fault data to confirm effectiveness of lightning protection.
4) Update internal policies and procedures if the project is successful.
5) Share learnings with wider audience.
Learnings
Outcomes
The table below shows the total IIS improvement from 2019 – 2023, on the circuits (NRNs) where surge arresters were installed vs 10 years and 3 years prior to lightning protection being installed. It is more accurate to compare against 3 years worth of data, as network investment of some description over 10 years could likely cause improvements in IIS costs that were not related to the installation of surge arresters e.g. undergrounding, fusing of spurs, etc. could result in a reduction in lightning related faults and these investments are more likely to have taken place over a 10 year period vs a 3 year period.
Although some circuits did experience poorer performance there was an average 50% improvement in IIS costs of those circuits where lightning protection was installed.
The control group experienced a 3% improvement in IIS costs over the same time period, so you could reduce the 50% to 47% to account for this improvement, which has little impact on the overall outcome.
The key observation here is on payback period. 16 years is a long time to justify continued installation. However, we did take a blanket approach to installing surge arresters i.e. they were installed on every 3rd pole. The payback could potentially be improved by installing on every 6th pole and in locations where lightning strike density is highest. However, this may also impact IIS savings, which may in turn increase the payback period, so the right balance needs to be found.
See Table 1
Lessons Learnt
An interesting early finding was that the circuits deemed at highest risk of experiencing lightning faults by the data analytics phase were significantly different from those circuits that historically have been recorded as experiencing lightning faults in our fault database. This could possibly mean:
1) Labelling of faults as being caused by lightning may be incorrect. It may be that extreme weather and water ingress are the problem rather than lightning, therefore it may be more prudent to alter maintenance/replacement cycles associated with overhead circuit assets rather than protect circuits against lightning.
2) Lightning may cause weakness in assets, which do not immediately fault, but fault at a later date.
3) The analytics algorithm could be incorrect and therefore not be correctly identifying the highest risk circuits.
It is possible that both points one and two are correct. With regards to point 3, the analysis was done on actual lightning strike location data, whereas fault reporting for lightning does not use this data, it is done by field engineers surveying the site. There will likely be a number of cases where we do misreport lightning as a problem. We double checked the trial circuits with a control group of circuits to reduce this. A key lesson is that control groups should always be used when analysing fault performance, especially with lighting analysis where the number of strikes can alter dramatically on an annual basis.
In this case we picked a control group of five circuits (the same number of circuits we installed surge arresters on) that experienced high number of lighting faults and compared the change in IIS costs over the same years that surge arresters were installed vs the three years prior to the surge arresters being installed.
