Hi,

We are trying to simulate oil flow through tilting pad thrust bearing using CFX. We are not getting the correct result as compared to theortical results.

Problem statement;

1.One wall surface is rotating (Collar of shaft) 2.Bearing surface (wall) stationary but at an angle to rotating wall.

3.Gap (70 micron ) between above to wall is having boundray with pressure equal to 1 Atm abs .(we have taken opening boundry condition with 0 relative pressure keeping ref pressure as 1 atm)

4.Material property values for oil viscosity,density,specific heat are correctly put in material property. Oil inlet temp & viscosity variation with temp expression is also fed properly.

Meshed fluid model satisfy all the mesh quality checks given by CFX.

Model selected is : Laminar as Re < 800 Heat transfer model : total energy

Convergence value: 1e05 max residual

We are getting load carrying capacity by CFX less (About 40%) than theortical & Experimental results.

Pls letus know and guide us where we are going wrong

Regards, Julien

Hi Julien,

I am not sure what you are using for comparison for your exprimental and theoretical result but in most lubrication problems such extreme pressures are developed in the gap...often called "Druckberg" or pressure-hill that you must make the viscocity a funtion of pressure. Under such high pressures viscosity increases significantly...increasing the load capacity for a given condition.

Some of the standard refs on Hydrodynamic lubrication should have relations for viscosity as a function of pressure..... try the book by Cameron for a start.

Regards,

Bak_Flow

thanks Bak_Flow,

We are comparing CFX results with therotical(Based on Reynolds Equation).In both the results we have not taken viscosity variation with pressure, then why is the diffence in both the results. Also experimental results are matching with therotical (Based on Reynolds Equation).

if i am correct, CFX uses navier stroke equation,and reynolds equation is derived from it.

So pls let me know where are we goning wrong in CFX formulation.

Regards, julien

[elcino]

Hi, I'm working as a PhD student to simulation of EHL and general lubrication problems with CFX. So, I think Reynolds equation of lubrication is generated from N-S equations. Film Thickness and load equilibrium equation is just a physical consideration so if you can implement viscosity and density models (Ree-Eyring or Roelands model for viscosity and Dowson model for density) and formulation for load equilibrium, cavitation model and FSI (fluid structure iteration to avoid the elastically deformation of the ball) I'm pretty sure that yours and mine simulations will work. I'm trying to simulate a ball sliding on a plane but at now, after some months I could not have results.

For me the principal problem is that I'm not strong in simulations so write CEL expression for viscosity is not easy. I wrote the expression of the density and seams working. For the density let we make an example: let we consider Barus viscosity model: u=uo*exp(a*p), where uo [Pa*s] is the ambient pressure viscosity, a [Pa^-1] is the pressure viscosity index and p [Pa] is the pressure. After defining uo is easy to write the expression but my CFX seams not use it. Instead of p I use Pressure, but if I check at any iteration the value of dynamic viscosity I have Min=0.97*10^-3 and Max=1*10^-2 which for a gap of 10^6 Pa is not possible. May be I use the wrong variable "Pressure" or CFX don't do the iteration like I want.

If you use the Herschel Buckley model already implemented in Ansys the simulation works, but I want to use my model not a prepared model! How can I do? What's wrong in my formula?

For the guy who is making experiments and simulation we can talk about it by email if he want, I'm quiet strong in tribology. luca.gorasso@gmail.com

[ashish1544]

hey hi julien we have aproject on designin of thrust pad can u plz help me on this topic

[brunoc]

Julien's post is from 2006, when this forum worked differently. Don't expect him to answer you.