High-order gas-kinetic scheme for numerical simulations of wind turbine with nacelle and tower using ALM and IBM

TL;DR

This paper introduces a novel high-order gas-kinetic scheme that integrates ALM and IBM for detailed 3D wind turbine simulations, accurately capturing wake interactions and turbine components using GPU acceleration.

Contribution

The integration of ALM and IBM into a high-order GKS for wind turbine simulation with nacelle and tower components is a new approach, enabling accurate and efficient large-scale turbulent wake modeling.

Findings

Accurately predicts turbine wake flow and vortex interactions.

Validates numerical accuracy with turbulent flow benchmarks.

Demonstrates effective GPU-based parallel computation for large-scale simulations.

Abstract

For the first time, the actuator line model (ALM) and the immersed boundary method (IBM) are integrated into the high-order gas-kinetic scheme (GKS) to simulate the wind turbine with the nacelle and tower. The high-order GKS is extended to the simulation of three-dimensional weakly compressible isothermal flows within a well-developed two-stage fourth-order framework. For the wind turbine, the rotor blades are represented by a group of actuator points in ALM, and the nacelle and tower are represented by a group of Lagrangian points in IBM. Both ALM and IBM are integrated through an external body force added in the momentum equation within the high-order GKS. The high-order GKS is implemented on Graphics Processing Units (GPU) to achieve the parallel computing capabilities for the large-scale simulation of turbulent wakes. Turbulent channel flow and turbulent circular cylinder flow are firstly simulated to validate the numerical accuracy of weakly compressible high-order GKS. The NREL 5 MW reference wind turbine is simulated using ALM without the nacelle and tower. Furthermore, the NTNU Blind Test 1 wind turbine is simulated with nacelle and tower using IBM. The current method yields the periodic power and thrust coefficients of the rotor blade due to the blade-tower interactions, while the steady coefficients are obtained when the tower is omitted. Compared to the turbine wakes without the tower, the interaction between tower vortex and tip vortex causes an earlier transition. The high-order GKS with ALM and IBM also well predicts the asymmetric mean flows of turbine wake, including the time-averaged streamwise velocity and turbulent kinetic energy, which are in good agreement with NTNU experimental data. The multiple-GPU enabled high-order GKS integrated with ALM and IBM offers an accurate and efficient approach for realistic wind turbine simulations.