Different results with same configuration in fluent analysis

  • 159 Views
  • Last Post 16 January 2020
msalvo4 posted this 14 January 2020

Hello,

I'm working with ANSYS 2019 R3 student license to calculate a force generated on a surface in a fluid flow (fluent) analysis. My system consists of 2 bodies assembled in space claim geometry. The first body is fixed to the referent system, the second body moves in one direction (this direction is my input parameter). My aim is calculate the force generated on the second body in relation of my input parameter. The problem is that if I change my imput parameter and update design point in parameter set window I calculate a force value about 10000 - 12000 N (that is impossible) but then if I open setup cell and run calculator manually with the same option and setup configuration without change my imput parameter I obtain a force value about 1,7 - 2 N (that is a truthful result).

For example the difference between 2 results is shown in velocity contour in results cell.

The first image is the correct solution (1,7 - 2 N)

In both cases when i open results cell I have the error message "No locators provided in Location parameter"

Order By: Standard | Newest | Votes
rwoolhou posted this 14 January 2020

The second set of results are nonsense: what's the convergence like on that model? I'd also suggest checking you've got enough mesh to resolve the flow, what's the speed of sound in your system?

msalvo4 posted this 14 January 2020

I don't know where verify the convergence of my model. Considering the trend of the residuals graph during the calculation the convergence is not so far, however i don't modify my setup from atomatic calculation to manual calculation, so i shuld have the same results. I did not modify manually my speed of sound, i only set my air density like "real gas" and my air viscosity like "kinetic theory" in materials properties. I chose the mesh sizing with maximum number of nodes allowed by student license.

rwoolhou posted this 14 January 2020

Why are you using real gas? Please review the tutorials/YouTube to see how convergence is checked. 

msalvo4 posted this 14 January 2020

I'm using real gas because i must consider a real system with atmospheric air, including losses pressure values in internal pipelines

rwoolhou posted this 15 January 2020

Read the definition & use of real gas carefully: are you trying to capture the JT effect?

msalvo4 posted this 15 January 2020

I read the ansys online guide and chose the Aungier-Redlich Kwong real gas model to define my air properties. I'm not considering the JT effect

rwoolhou posted this 15 January 2020

Try again with ideal gas: unless you're at a very high pressure and/or expecting to need JT or phase change I'd not use real gas. 

msalvo4 posted this 15 January 2020

I have high pressure value in my system, so I can't consider air like ideal gas. however I tried to set air like ideal gas and run calculation, but i have always the same problem (different results if I run design points automatically or manually)

rwoolhou posted this 16 January 2020

Do you have pressure boundaries on both ends of the domain?  If so, the mass flow becomes part of the solution which makes the solution much more dependent on the initialisation values.  Looking at your "good" results, how much mesh have you got to resolve the 350m/s velocity region?  What do you count as "high pressure"? 

msalvo4 posted this 16 January 2020

I have a pressure input boundary and a atmospheric output boundary (I set outlet-vent). I have 3 atm pressure value in my system.

rwoolhou posted this 16 January 2020

Outlet vent has some additional parameters, if the outlet is to atmosphere pressure outlet may be more suitable. 

3atm isn't an overly high pressure: flow will probably choke but that's about it. If you'd been talking about 60-500 bar then I'd agree it was a high pressure application. 

Close