Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Assessment of URANS and LES Methods in Predicting Wake Shed Behind a Vertical Axis Wind Turbine

Armin Sheidani, Sajad Salavatidezfouli, Giovanni Stabile, Gianluigi Rozza

TL;DR

This work assesses the capability of URANS and LES to predict the wake shed behind a lift-driven VAWT by analyzing the wake with Proper Orthogonal Decomposition. It uses an H-shaped Darrieus turbine as the test case, conducts grid-converged simulations, and validates against experimental data, revealing that RANS captures large-scale wake features while LES resolves finer, more energetic structures. POD is applied to compare the two turbulence approaches in terms of spatial modes, energy content, and temporal evolution of the wake, showing that LES provides a more faithful depiction of unsteady wake dynamics though RANS can still inform global wake metrics. The results offer guidance on when to employ LES versus RANS for VAWT wake studies and highlight the practical impact for design and array configuration where wake interactions matter.

Abstract

In order to shed light on the Vertical-Axis Wind Turbines (VAWT) wake characteristics, in this paper we present high-fidelity CFD simulations of the flow around an exemplary H-shaped VAWT turbine, and we propose to apply Proper Orthogonal Decomposition (POD) to the computed flow field in the near wake of the rotor. The turbine under consideration was widely studied in previous experimental and computational investigations. In the first part of the study, multiple Reynolds-Averaged Navier-Stokes (RANS) simulations were performed at the Tip Speed Ratio (TSR) of peak power coefficient, to select the most accurate turbulence model with respect to available data. In the following step, further RANS numerical simulations were performed at different TSRs to compare the power coefficient against experimental data. Then, Large Eddy Simulation (LES) was applied for multiple TSR conditions. The spatial and temporal POD modes along with modal energy for the RANS and LES results were extracted, and the performance of the turbulence models was assessed. Also, an interpretation of the POD modes with respect to the flow structures was given to highlight the most significant time and length scales of the predictions considering the different dynamical levels of approximations of the computational models.

Assessment of URANS and LES Methods in Predicting Wake Shed Behind a Vertical Axis Wind Turbine

TL;DR

This work assesses the capability of URANS and LES to predict the wake shed behind a lift-driven VAWT by analyzing the wake with Proper Orthogonal Decomposition. It uses an H-shaped Darrieus turbine as the test case, conducts grid-converged simulations, and validates against experimental data, revealing that RANS captures large-scale wake features while LES resolves finer, more energetic structures. POD is applied to compare the two turbulence approaches in terms of spatial modes, energy content, and temporal evolution of the wake, showing that LES provides a more faithful depiction of unsteady wake dynamics though RANS can still inform global wake metrics. The results offer guidance on when to employ LES versus RANS for VAWT wake studies and highlight the practical impact for design and array configuration where wake interactions matter.

Abstract

In order to shed light on the Vertical-Axis Wind Turbines (VAWT) wake characteristics, in this paper we present high-fidelity CFD simulations of the flow around an exemplary H-shaped VAWT turbine, and we propose to apply Proper Orthogonal Decomposition (POD) to the computed flow field in the near wake of the rotor. The turbine under consideration was widely studied in previous experimental and computational investigations. In the first part of the study, multiple Reynolds-Averaged Navier-Stokes (RANS) simulations were performed at the Tip Speed Ratio (TSR) of peak power coefficient, to select the most accurate turbulence model with respect to available data. In the following step, further RANS numerical simulations were performed at different TSRs to compare the power coefficient against experimental data. Then, Large Eddy Simulation (LES) was applied for multiple TSR conditions. The spatial and temporal POD modes along with modal energy for the RANS and LES results were extracted, and the performance of the turbulence models was assessed. Also, an interpretation of the POD modes with respect to the flow structures was given to highlight the most significant time and length scales of the predictions considering the different dynamical levels of approximations of the computational models.
Paper Structure (10 sections, 7 equations, 13 figures)

This paper contains 10 sections, 7 equations, 13 figures.

Table of Contents

  1. Introduction
  2. NUMERICAL METHODS
  3. Problem Description
  4. Grid Generation
  5. Turbulence Modelling Techniques
  6. Proper Orthogonal Decomposition Method
  7. Validation Study
  8. Results and Discussion
  9. Conclusion
  10. Bibliography

Figures (13)

  • Figure 1: Dimensions of the cell zones of the computational domain
  • Figure 2: Overview of the LES mesh (a) around the turbine and consequent cell zones b) around the blade c) on the surface of the extruded blade.
  • Figure 3: Contour of index quality for the LES mesh
  • Figure 4: Power coefficient validation with respect to the experimental work of Battisti_2018 and numerical results of Franchina_2019
  • Figure 5: Validation in the wake region located at 3.5D behind the rotor with respect to the experimental results of Battisti_2018 a) TSR1.5, b) TSR2.4 and c) TSR3.3
  • ...and 8 more figures