Severe pollution and harsh weather are one of the main issues for electric utilities causing flashovers and unplanned line outages. Currently, there is no pollution measurement information across the network. This project will use insulator leakage current monitoring
sensors to capture and share information remotely. This will help characterise the risk of equipment degradation due to pollution and assist with designing and maintaining Overhead Lines (OHLs) in pollution-high-risk areas of the network. With this, early design mitigation
and maintenance procedures can be carried out to prevent faults due to flashovers.
Benefits
If the system can reduce the number of unplanned outages caused by flashovers and/or the need to carry out maintenance, then this will reduce the costs for consumers. Financial penalties for faults will also be avoided. Having data regarding pollution risks could be useful information for communities to take steps to reduce pollution risks.
This project will seek to capture information that can characterise the risk of asset degradation from pollution. A short section of the 132kV OHL case site will be used as the innovation use case.
Data on the frequency of flashovers and their cause will be determined after the results of the trial, so at this stage the CBA conducted reflects the minimum investment case incorporating a risk factor using a simple probabilistic method. The risk analysis considered the below risks:
• Flashovers/faults are not caused by pollution.
• The reasons for the faults and flashovers faults were not properly diagnosed and continue to occur regardless of the implementation of the pollution monitoring projects.
• Despite information obtained on pollution and remedial design actions, flashovers due to pollution still occur.
The CBA compares the innovative monitoring system scenario with the counterfactual scenario of not installing the monitoring system at all in a short section of the extreme case site and so a high frequency of flashovers continues to occur. Benefits considered account for the cost savings from outage avoidance and ignore any other co-benefits at this moment. Therefore, the potential lifetime risk-adjusted cost benefits have been calculated as follows –
• Counterfactual (“do nothing”) base case: Outage cost (61 outages per lifetime x £12,000) = £732,000
• Innovative monitoring system case: Outage cost (8 outages per lifetime x £12,000) = £96,000
• Innovative monitoring system case: Monitoring system cost (supply £120,435 + installation £1,050) =£121,485
• Net risk-adjusted cost benefit: 66.7% x (£732,000 – £96,000) – £121,485 = £302,727
If this cost saving is capital budgeted considering our model’s discount rate and depreciation is estimated to £108,288.
Main assumptions considered:
• Asset life 45 years
• Lifetime flashover frequency scenarios for the risk analysis 10, 20, 40 faults.
• Average flashover outage cost £12k
• Average of 61 outages per lifetime at the case site
• Drop in faults using monitoring systems assumed 8 outages at case site per lifetime.
• Average outage duration 3 hours
• Probability of any kind of flashovers to occur 66.7%
The results showed that, the potential benefits considering risk are estimated to be at least £303k only from the avoidance of outages. The trial will help obtain more data on the cause of the degradation, with cost estimations being revisited on project completion. If successful, the dissemination of learnings can lead to further benefits being realised across similar schemes.
Learnings
Outcomes
Over the project period, from March 2024 to March 2026:
- One the sites has demonstrated a very low to low pollution severity through the equivalent salt deposit density (ESDD)/ non-soluble deposit density (NSDD) and the direction dust deposit gauge (DDDG) measurements. Also, for both methods, the proportion of soluble components (Type B pollution) illustrates the absence of industrial activity nearby and presence of birds’ populations; both were anticipated as pollution factors at the beginning of the project. The high frequency of rainfall throughout the project has also had a huge impact on the results: rain enhances the insulators’ self-cleaning which avoids the accumulation of pollution.
- The insulation level has been assessed, from a pollution perspective, in accordance with IEC 60815 standards, using the three insulator models. The estimated quantity of insulators closely matched the actual installed units. The only exception was one tower, where the insulation level was slightly lower than the estimated value. Nevertheless, due to the low leakage current levels observed, the insulation can still be considered adequately sized.
- Considering that two suspension towers are properly sized and show low leakage current values, it can be indicated that for suspension strings, perform slightly better the E-100P-146 insulator, which is installed on another tower.
- The monitoring part of the study has shown positive results and has corroborated the conclusions established in the overall final report conducted on the pollution severity of the site. The leakage current remained below 20 mA constantly for the three sensors. Therefore, based on the two years of monitoring, no flashover risks were detected on the specific section of OHL.
- Additional data analysis has been carried out on one tower, since it is the only tower in the project that has shown activity beyond what can be considered background noise. It has been demonstrated that the most influential parameter is abundant rainfall, which facilitates self-cleaning. Additionally, some thresholds (used to establish baselines and relevant significance in the data mapped) have been defined based on rainfall, relative humidity, dew point, and ambient temperature.
A machine learning approach has been used for the prediction of the few observed leakage current peaks (in which good results were obtained for the monitored dataset) using regression algorithms. It is worth noting that the machine learning approach used is specific to the monitored dataset, and it should be understood that no flashover risks have been observed throughout the project timeframe. Therefore, detection of high leakage current peaks cannot be guaranteed when using this approach without the collection of real data.
Lessons Learnt
Key lessons learned:
The area selected for the assessment was not optimal as the pollution levels were found to be low – locations with known higher pollution would likely yield better results. As explained in section 7, a machine learning approach was attempted to predict the small number of leakage current peaks observed. However, it was not relevant during the project on this site because pollution levels were low throughout the project. Therefore, to confirm the performance of the three insulator models, further monitoring in high pollution environments previously assessed using IEC 60815 methods, should be carried out. This would support decisions on insulation dimensioning under polluted conditions.
