Effects of turbulence modelling in AD/RANS simulations of single wind & tidal turbine wakes and double wake interactions

  • Linlin Tian
  • , Yilei Song
  • , Ning Zhao*
  • , Wenzhong Shen
  • , Chunling Zhu
  • , Tongguang Wang
  • *Corresponding author for this work

    Research output: Contribution to journalJournal articleResearchpeer-review

    Abstract

    A comparative analysis of the RANS-based generalized actuator disc method (AD/RANS) with different turbulence models for simulating single wind & tidal turbine wakes and double wake interactions is performed. The tested turbulence models include the Realizable k-ε, SST k-ω, Linear Reynolds-stress (RSM) and two proposed RSM models with modified closure constants. The evaluation is based on comparing their results against five sets of published experimental data for various types of turbines operating under a wide range of inflow conditions. Special emphasis is put on both horizontal and vertical distributions of velocity deficit and turbulence intensity throughout the whole wake region. The results identify that the wake flow prediction is particularly sensitive to the turbulence model implemented, with a maximum difference reaching up to 40%. Overall, the proposed Mod2 RSM model (along with a new derivation of turbulence intensity) is recommended as a robust choice for capturing wake behaviors, in both near and far field from the turbine. Additionally, the weaknesses and strengths of other models are discussed, aiming to give a guideline for choosing the right model for the right job as well as to provide a reference study about the accuracy improvement of wake predictions.
    Original languageEnglish
    Article number118440
    JournalEnergy
    Volume208
    Number of pages14
    ISSN0360-5442
    DOIs
    Publication statusPublished - 2020

    Keywords

    • Wind turbine wakes
    • Turbulence models
    • AD/RANS method
    • Wake development
    • RSM model

    Fingerprint

    Dive into the research topics of 'Effects of turbulence modelling in AD/RANS simulations of single wind & tidal turbine wakes and double wake interactions'. Together they form a unique fingerprint.

    Cite this