Topics relate to Fluent, CFX, Turbogrid and more

High Ra Natural Convection Setup

    • T-Ying Lin

      Following with a well documented paper to try and benchmark my work. Was able to get comparable results for low Ra number models but when I tried to do the same for a high Ra number case (10^9), the heat transfer coefficient produced was significantly higher than expected. I'm modeling the natural convection of a concentric horizontal cylindrical annuli with the boundary conditions seen here in this diagram. The energy balances out, I'm using a similar mesh and monitoring the temperature, velocity, and pressure to see when the model converges. I'm honestly at a loss as to why it's so much higher. Any help would be appreciated! 

    • C N
      Ansys Employee

      Hello T-Ying Lin,

      Reasons for getting higher  convective heat transfer coefficient are as follows

      1) It is known fact that rayleigh number is a product of grashoff number and prandtl number where if the value goes beyond 10^9 the flows becomes turbulent .

      2)The Heat transfer coefficient parameter is the property that combines the fluid flow and solid geometry. 

      3)The convective heat transfer coefficient for turbulent flow is higher because it has thinner stagnant fluid flim layer on the surface than the laminar flow and also increases with wall roughness as reynolds number increases and the viscous sub layer becomes thin. 

      4)The turbulent flow generate more momentum and eddies due to the higher entrophy (randomness of molecules)- this contributes to more velocity component less viscous resistance . So heat dissipation happens quickly.

      The Ansys treats this problem as turbulent flow which is different from the laminar flow mechanics.

      I am attaching the user guide link for some best practices guide lines for setting up high rayleigh number flow cases

      7.3. Operating Conditions (ansys.com) In this link refer Guidelines for Solving High-Rayleigh-Number Flows

      I hope this helps you in getting the results correctly for your high rayleigh number flow case.

      Thank You,

      Chaitanya Natraj


Viewing 1 reply thread
  • You must be logged in to reply to this topic.