Sulphur Hexafluoride (SF6) escape from high voltage switch gear creates operational and environmental challenges. Asset owners are under increasing pressure to reduce emissions or face penalties. Monitoring data alone is not enough to correctly and promptly characterise faults, as measurable changes in SF6 density are not always indictive of faults.
A hybrid modelling method has been developed which combines a physics-based prediction with a regression-based residual model. The developed model can accurately predict monitoring output by combining sensor behaviour with a physical understanding of the system.
The model can act as a digital twin of the associated asset allowing changes in density to be more meaningfully interpreted, resulting in more true anomalies being distinguished from normal operational variation. Additionally, consequences from changes in operational regime and environmental conditions can be anticipated using the model allowing risks to asset health to be forestalled.
The software framework embodying the developed models has now been transferred to the partners IT personnel to assess operational deployment potential.
A proactive approach to management of assets promotes higher levels of availability. Active interventions in SF6 management can mitigate regulatory intervention.
The digital twin approach taken in Networks-16 allows us to both apply understanding of asset behaviours as well as build knowledge of these in the context of new data streams unlocking their value
Fiona Irwin​, SSEN Transmission
Networks-16 outputs have been delivered in such a way that these can be added to our information system workflows with minimal modification, shortening the time to realise benefits from the models and the underlying sensor data streams.
Nicol Lumsden, SSEN Transmission