Osmose has performed many tens of thousands of direct, physical inspections on steel infrastructure over the past decade. The data and technical learnings from these inspections have been consolidated and utilized in conjunction with machine learning algorithms and existing corrosion science to build advanced predictive models of below-grade corrosion. Since environmental variables like soil pH, structure-to-soil electrochemical potential, soil resistivity, and redox potential are significant indicators of corrosion risk and activity (and are captured in Osmose's inspection process), they are among the key variables in our predictive models.
The Corrosion Onset Model predicts the number of years a specific structure will be in service before corrosion activity begins to create damage/section loss.
The Corrosion Damage Model predicts both the amount of damage that has occurred and the rate of future damage due to corrosion activity for a specific structure.
These two models are combined to create the Corrosion Risk Model.
The graphic above depicts the lifecycle of a particular galvanized steel lattice tower that is 53 years old at the time that the prediction was made. This structure began corroding at 48 years old, and the green dot shows a current section loss prediction of 0.03 inches. The future rate of corrosion is depicted by the dashed green line with the grey polygon representing the potential error.
Why Use Predictive Modeling
The Osmose Corrosion Risk Model provides structure owners with valuable information designed to aid in the decision-making process. For instance, the model can help structure owners decide when to begin inspection cycles (based on the predicted onset of corrosion), where to begin inspections based on which structures and lines are at the greatest risk for below-grade corrosion, and most importantly, which structures are likely in need of immediate attention.
For more information on predictive modeling, please contact your local Osmose professional or email [email protected].