Machine learning for FPSO turret maintenance
Floating Production Storage and Offtake Vessels (FPSOs) have a large turret within the vessel where the subsea risers connect, and the vessel rotates around a bearing on this turret.
Maintenance on the turret is potentially a multimillion dollar headache for FPSO operators as it involves taking the vessel off station to dry dock and an associated period of no production.
Our customer who operates an FPSO off the coast of Africa wanted an accurate projection of the usable life of the turret and the ability to predict the required maintenance in advance.
Working with a limited dataset, harnessing machine learning and custom built algorithms, our team developed a tool to predict how the condition of the bearing will change over time, taking into account radial clearance, bearing height and vessel heading.
The outputs from the tool enabled the justification to defer the planned time based maintenance, resulting in large cost savings for the customer.