A methodology to develop reduced-order models to support the operation and maintenance of offshore wind turbines. (1st February 2020)

• Record Type: Journal Article
• Authors: Lin, Zi; Cevasco, Debora; Collu, Maurizio
• Is Part Of: Applied energy. Volume 259(2020)
• Publication Date: 2020-02-01
• Subjects: Linearisation -- Aero-hydro-servo-elastic -- Dynamic response -- Reduced-order model -- Offshore wind turbine; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2019.114228
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: A method to develop offshore wind turbines reduced-order models is presented. The case study is a gearbox failure, monitoring the shaft torque signal. The blade modes contribute substantially to the torque dynamics above 1.5 Hz. The tower modes do not contribute noticeably to the torque dynamics below 0.5 Hz. Nonlinear reduced-order models capture the relevant dynamics at a lower cost. Abstract: From an operation & maintenance (O&M) point of view, it is necessary to model the aero-hydro-servo-elastic (AHSE) dynamics of each wind turbine but, on the other side, wind farms generally include hundreds of wind turbines. Simply using and linking several advanced, single wind turbine models of dynamics to represent a wind farm can be computationally prohibitive. To this end, this paper developed a reduced-order model (ROM), able to capture the relevant dynamics of the system for a specific failure, having a lower computational cost and therefore more easily scalable up to a wind farm level. First, a nonlinear AHSE model is used to derive the time-domain response of the wind turbine degrees of freedom (DOFs). The failure mode, its relevant DOF, and the relevant operational conditions during which the failure is likely to occur are identified. A linearisation of the nonlinear aero-hydro-servo-elastic-drivetrain (AHSE-DT) model is then carried out. Subsequently, a number of linear ROMs are developed based on the linear full-order system but excluding high-frequency states using the modal truncation (MT) method. For the targeted DOF (rotor torque signal) and the load cases simulated, the results from the linear ROMs showed that the blade modes are important to capture not only the DOF of extreme values, but also the DOF of high-frequency responses (above 1.5 Hz). The results from nonlinear ROMs showed that the ROM eliminating all the tower modes (rigid tower) is acceptable to capture the DOF of low-frequency response (below 0.5 Hz), as it has almost the same spectral responses as the full-order nonlinear model.