Know where to look.
And when to act.
An explainable machine learning dashboard turns fibre-optic measurements into maintenance priorities, helping network teams assess which connections need attention and how soon.
Fibre-optic networks supporting high-voltage infrastructure
The challenge
Fibre-optic connections support communication and monitoring across high-voltage networks. When a connection starts to deteriorate, maintenance teams need to know where it is happening and whether action can wait.
The measurements are available. Turning them into a clear maintenance decision is harder.
Teams must interpret changes over time, investigate potential problems and decide which connections deserve attention first.
Which connection needs maintenance next, and what evidence supports that decision?
The solution
Common Sense AI is developing a dashboard that combines historical and current measurements to identify signs of degradation and estimate when maintenance may be needed.
Explainable machine learning models help teams understand three things:
- Where to look: which routes, cables or splices show signs of deterioration.
- When to act: the estimated number of days before intervention may be needed.
- What to check: the measurements behind the prediction and the uncertainty around it.
Teams can move from a network-wide overview to an individual connection. The underlying data stays available, so technical experts can check the prediction before deciding what to do.
What it enables
The dashboard is designed to give maintenance teams a clearer basis for planning their work.
It brings potential problems, maintenance priorities and supporting evidence into one view. Teams can focus their investigations and use their technical judgement to decide when an intervention is warranted.