This thesis details the development and validation of a fully coupled aero-hydro-servo-elastic-mechanical multi-body simulation (MBS) model with gearbox dynamics for a bottom fixed offshore wind turbine based on a reference 10 MW turbine.
A state of the art review of wind turbine modelling theory, publications and modelling software is presented. From this review, a gap in our current literature was identified in the coupling between environmental loads (wind and waves) and internal mechanical drivetrain excitations (such as gear meshing) within the same simulation model. The most common approach is based on running the external and internal simulations separately using the output of one as the input to the other. This is referred to as decoupled modelling. In a decoupled model, the drivetrain is reduced to an equivalent two mass model.
A coupled model is a useful resource to investigate the impacts of various environmental loads and operating conditions directly on the drivetrain. Failures on wind turbine drivetrains cause the longest down-times and by increasing modelling fidelity, preventive maintenance can be planned based on simulation results.
An onshore decoupled wind turbine model was first implemented in the MBS software SIMPACK and validated against data from the proposed 10MW reference wind turbine from the Technical University of Denmark (DTU). The aerodynamic (AeroDyn) and turbulent wind (TurbSim) modules used are developed by the American National Renewable Energy Laboratory (NREL) as part of the open source package Open-Fast. An in-depth guide to the model development is presented along with components and full system validation based on available data and code-to-code comparison with Horizontal Axis Wind turbine simulation Code (HAWC2) simulation results. A dedicated pitch and torque control system was developed using co-simulation with SIMULINK. The simulations conducted are based on wind turbine industry standards developed by the International Electrotechnical Commission (IEC) and the results shows good agreement with the reference data for steady and turbulent wind cases.
The validated onshore wind turbine model was subsequently placed on top of a bottom fixed monopile foundation and stochastic hydrodynamic wave loads were implemented, using the HydroDyn module from NREL. The design of the monopile is based on a study specifically conducted for the DTU 10MW turbine. The effects of adding the waves on the dynamics at the top of the tower were found to be negligible in this configuration but still included in the final simulations.
Lastly, the fully coupled model was generated by creating a higher fidelity drivetrain model. This model is derived from a proposed medium speed gearbox model for the DTU 10MW turbine. The equivalent two mass drivetrain was replaced by three massless gearbox stages connected with rotational spring dampers to maintain the same characteristics as the de-coupled model. A selection of load cases corresponding to various operating sequences were chosen to investigate the output of the coupled model. These include: startup, emergency shutdown, pitch fault, lack of sun gear lubrication and planet gear misalignment. The results presented are a combination of internal drivetrain dynamics and wind turbine structural response. An analysis of gearbox gear interactions is conducted to illustrate how the data obtained could be used for gear teeth bending and contact stress calculations.
Using a coupled approach is computationally demanding and increased the simulation times up to 100 times compared to the decoupled environmental model. This is because of the high frequency associated with gear meshing compared to environmental loads. It is concluded that if drivetrain dynamics are of interest, this high frequency is required regardless. Environmental forces can therefore be included in the same model, thus saving overall simulation and processing time.
A coupled model will also provide drivetrain/gearbox engineers and researchers a centralized model to study the impact of wind turbine operating conditions on the drivetrain dynamics. A list of modelling recommendations are given in order to increase coupled model fidelity for this purpose.