Economic analysis of condition monitoring systems for offshore wind turbine sub‐systems. Issue 8 (1st November 2015)

• Record Type: Journal Article
• Authors: May, Allan; McMillan, David; Thöns, Sebastian
• Is Part Of: IET renewable power generation. Volume 9:Issue 8(2015)
• Page Start: 900
• Page End: 907
• Publication Date: 2015-11-01
• Subjects: wind turbines -- offshore installations -- wind power plants -- condition monitoring -- gears -- vibrations -- hidden Markov models -- power generation economics -- electric drives -- poles and towers -- blades -- foundations -- maintenance engineering -- durability -- failure analysis -- alarm systems; condition monitoring system economic analysis -- offshore wind turbine subsystem -- vibration‐based monitoring system -- gearbox -- generator -- drive train -- blades -- tower -- foundation -- maintenance plan -- wind farm -- life cycle -- hidden Markov model -- false alarm failure -- six‐month failure warning -- acoustic emission sensor -- oil sensor -- CM system failure detection rate -- operational cost; Renewable energy sources -- Periodicals; 333.79405
• DOI: 10.1049/iet-rpg.2015.0019
• Languages: English
• ISSNs: 1752-1416

Abstract

Abstract : The use of condition monitoring systems on offshore wind turbines has increased dramatically in recent times. However, their use is mostly restricted to vibration based monitoring systems for the gearbox, generator and drive train. A survey of commercially available condition monitoring systems and their associated costs has been completed for the blades, drive train, tower and foundation. This paper considers what value can be obtained from integrating these additional systems into the maintenance plan. This is achieved by running simulations on an operations and maintenance model for a wind farm over a 20 year life cycle. The model uses Hidden Markov Models to represent both the actual system state and the observed condition monitoring state. The CM systems are modelled to include reduced failure types, false alarms, detection rates and 6 month failure warnings. The costs for system failures are derived, as are possible reductions in costs due to early detection. The detection capabilities of the CM systems are investigated and the effects on operational costs are examined. Likewise, the number of failures detected 6 months in advance by the CM systems is modified and the costs reported.