System intelligence for identifying environmental threats to distribution networks

Lead Research Organisation: University of Strathclyde
Department Name: Electronic and Electrical Engineering

Abstract

Transmission and distribution network asset health can be threatened by surroundings that can impose highly localised weather effects, missed through regional meteorological observations. Diversity of plant utilisation can add further complexity to the relation between true asset health and the environment it is installed in. The challenge posed in monitoring these assets in a more representative manner is in drawing together diverse data streams to model the relation between established weather stations, specific measurement locations (such as on-tower weather measurements or LV substation outbuilding environmental monitoring) on the network and using this as a means of generating prioritisations of maintenance or inspection. Improved indication of network pinch points and their likely location can direct limited field resource to potential causes of network disruption in shorter timeframes.

To capture the potentially complex relation between the influencing factors involved, a machine learning approach will be developed to automatically identify where and how network assets may be challenged and how this can be reported in an actionable manner. These analytics will need to be designed to unlock additional value from existing data streams that network operators already hold with no additional monitoring equipment required owing to the potential difficulties associated with deploying remote or invasive sensing on certain power system assets. Analysis and re-purposing of these data streams will facilitate development of:
- Understanding of actual asset utilisation and potential operational headroom for planning purposes
- Metrics from utilisation characteristics to infer asset health index and highlight network 'hotspots'
The immediate operational value of these analysis tools is likely to lie in the ability to prioritise maintenance according to condition rather than time which could make more effective use of limited field staff resources from a network reliability perspective. The immediate operational value of these analysis tools is likely to lie in the ability to prioritise maintenance according to condition rather than time which could make more effective use of limited field staff resources from a network reliability perspective.

People

ORCID iD

Ekua Osei (Student)

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/R512205/1 01/10/2017 30/09/2021
2092551 Studentship EP/R512205/1 01/10/2017 30/09/2021 Ekua Osei