UK Power Networks experiences faults on Low Voltage (LV) and High Voltage (HV; Note 6.6 or 11kV) underground cables each year, with over 80% of these faults attributed to “age and wear” or “unknown” causes. A previous study on an NIA project, “Prediction of weather-related faults”, showed that there was a measurable relationship between the amount of rainfall and the number of LV and HV faults. However, the previous study also showed that further work was required to develop an operational faults forecast model for LV and HV underground cables. A review of soil characteristics in combination with rainfall is likely to provide additional detail to be able to develop an operational faults forecast model. The data on soil characteristics can also be used to develop a soil resistivity assessment tool, for secondary substation earthing studies.
Objectives
The objectives of this project include:
1. To determine the relationship between soil type, rainfall and underground cable faults
2. To build a faults forecast model for underground cables (if a meaningful relationship exists between soil type, rainfall and underground cable faults)
3. To build a soil resistivity assessment tool for desktop earthing assessments at substations
Learnings
Outcomes
A soil resistivity tool was successfully implemented in UK Power Networks’ GIS tool, GSA. This tool will enable substation designers and planners to carry out earthing assessments of distribution substations from a desktop and to enable them to determine the requirements for effective earthing of the substations. This tool is now operational in GSA.
Faults forecast models, one based on work completed by the Met Office and another based on linear regression comparing faults and weather, were implemented in GSA. Also, a model based on linear regression, comparing faults and customer calls, was implemented in GSA. However, further work will be required as part of business as usual activities to operationalise these models. UK Power Networks is carrying out ongoing work with the Met Office to review the accuracy of the more granular weather forecasts and to determine whether to base the faults forecast models on less granular weather data only. This exercise will be completed as part of business as usual activities.
Lessons Learnt
The project established that there is a weak correlation between the number of underground cable faults and geology. The stronger correlation is driven by rainfall; hence is possible to predict the number of underground cable faults (excluding faults caused by third party damage or known defects) based on rainfall only. Analyses also showed that wind speed had an impact on faults volumes, the main impact being on overhead line assets.
The conclusion was that overall fault volumes was a function of the amount of rainfall and wind speed in a given area. To operationalise a fault forecast model would however require weather forecasts (for rainfall and wind speed) for small areas and for these to be more accurate than they currently are.
Our experience on the project showed that whilst weather forecasts for larger areas e.g. London as a whole was accurate, it is a lot more difficult to provide accurate forecasts for smaller local areas. For example, on the project we utilised forecasts for 2x2km grid squares which appeared to be less accurate than the higher levels forecasts for a larger area. Further work is required to determine what minimum level of granularity e.g. 10km2, 20km2 or other, would be practical to use for an operational weather and faults forecast model.
