Deepsteel™
Corrosion Detection
90%
Quicker than trained human assessors
Deepsteel Keen AI’s Corrosion Detection and Forecasting Service
Corrosion Detection on transmission fittings
Steel electricity pylons are the literal backbone of the world’s electricity grid powering the lives of billions of people across the globe. They are all however corroding at different rates depending upon their immediate atmospheric conditions and if or how well they have been painted. In many parts of the world, including North America, Europe and the Pacific region, corrosion can reduce the expected useful life of a pylon from 100 years to 50 years.
Network Operators routinely survey pylons for signs of corrosion or damage to paintwork as by spotting corrosion early or the need to paint they can dramatically reduce maintenance costs and extend the life of a pylon. Existing processes are problematic as they rely on human visual inspection which are time consuming, subjective and introduce bottlenecks. The subjectivity of human assessments makes it difficult to spot trends and forecast rates of corrosion across a network.
The Industry’s Most Accurate and Reliable Corrosion Detection and Forecasting Solution
50%
Reduction in tower lifespan due to corrosion
Cost Reduction
Months of manual processing is now just days. Quicker decisions to protect the asset, instead of replacing it, can lead to an overall saving
Increased Lifespan
Identifying the location of corrosion early and regular painting can significantly increase asset's useful life to over 100 years.
Objective Assessments
Different analysts produce different assessments for the same assets; our AI is consistent throughout your entire inventory
Deepsteel identifies, grades, localises current levels of corrosion on a pylon and predicts its likely future. This is done through a four stage AI pipeline: Identifying steel lattices from a scene, detecting corrosion on steel, mapping corrosion and a forecasting model which combines historic trends and atmospheric. Read our National Grid case study here.
The results are as accurate as a trained human assessor, 90% quicker and crucially consistent across structures and time. This spatial and temporal consistency underpins the training data for the forecasting model which generates reliable forecasts of future state. By targeting pylons with higher rates of corrosion and reducing the survey frequency for low corrosion towers, in the UK it reduces the number of surveys undertaken by 8.3% per year. This saves up a typical operator hundreds of thousands of dollars per year in data collection costs.
Talk to us about Deepsteel™
