Seminar by Adrien Marsick

Rank-based statistics for condition monitoring

Modern condition monitoring of wind turbine shaft line relies on a wealth of fault-sensitive indicators. On a fleet-wide scale, faults are automatically detected whenever an indicator exceeds a preset threshold. Choosing an appropriate threshold is complex with compromises made between false and missed alarms. On the other hand, analyzing the trend of series indicators seems more robust and generalisable. Rank-based statistics offers a promising framework for this issue.
After a quick introduction on the subject concerning series of conditions indicators, new signal processing tools based on rank statistics will be introduced.