1D-3D Nesting: Embedding reduced order flexible multibody models in system-level wind turbine drivetrain models

A Wind Turbine (WT) is a complex system in which many sub-systems interact and influence one another. To properly predict its behavior through simulation, a system-level modelling approach is necessary. A system-level model can capture the complex multiphysical interactions between sub-systems. Port-based modelling tools (e.g. AMESim, Simscape) offer an appealing solution to build such system-level models, as the assembly of a model is simplified to selecting and connecting a number of blocks. The mechanical models available in these port-based tools are often limited. Rotary shafts are often represented using a lumped rotational inertia with a single Degree of Freedom (DOF). This paper presents an approach that allows embedding more detailed flexible multibody models into a port-based model. The multibody model contains geometrical details in 3D, where the port-based framework is inherently 1D; therefore this technique is called 1D-3D Nesting. This allows achieving higher modelling accuracy for critical components while retaining the advantages of the port-based approach for system-level modelling, and a reasonable computation cost for dynamic simulations. The technique is applied on a case of industrial complexity, the NREL WT drivetrain. Nested 3D models of two shafts are incorporated into a system-level model of the drivetrain which also includes the rotor aerodynamics, gearbox, generator and controller. Using this model, the response of the WT during a Low Voltage Ride Through (LVRT) is simulated. The effect of including the nested 3D models on the torsional dynamics of the turbine and the resulting bearing forces is assessed by comparing the simulation to a baseline model.

Publication year:2019

Accessibility:Open