CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Help with journal bearing simulations (

elcino May 1, 2012 05:55

Help with journal bearing simulations
Good morning everybody,

I'm working on fluid dynamic simulation of journal bearings. The domain is very simple, imagine to subtract from a cylinder another cylinder with a little eccentricity from the first and a lower diameter, and u will have the domain. The lubricant inlet are 4 radial holes and the outlet are the two side leakage. The division scheme is hexa-cooper, 15 division along the film and good aspect ratio and quality, laminar and isothermal flow. Operative pressure 101325 Pa, outlet gauge pressure 101325. Some articles state that the inlet pressure is 0.2 MPa, which, i think means that u have Gauge pressure inlet 101325 Pa and Static pressure 200000 Pa. Even if everybody will try to do this simulation, with several under relaxation factors, good or different meshes and so on: the convergence for rotational speed of the internal wall up to 1 rad/s are impossible to been reached (three months of simulations) and the residuals of the continuity equations are around 10^-2.

I've tried to use the mass flow inlet condition but i don't know the mass flow so is impossible. Some other authors state that the mass flow is the difference between the mass flow in the minor area (remember the inner cylinder have an eccentricity) and the mass flow trough the larger are which will be at the opposite side of the bearing and the journal works as a lubricant pump, which i think is kind of correct!

1) Do u think my BC are wrong?

2) Can some one explain how to impose, via UDF, the condition that the mass flow at inlet is equal at the difference between the mass flow in the minor plane and the mass flow in the larger plane (I have a brief knowledge of UDF's)?

(if someone want I've a journal for gambit to have the mesh and for fluent to run a calculation)

Thanks a lot!

LuckyTran May 1, 2012 14:33

This problem has beautiful analytical solutions to it and requires no CFD work to get velocity and thermal profile everywhere. Any reason you are using CFD? The analytical solutions are probably better than anything you could get from CFD also since fewer simplifications are made.

elcino May 1, 2012 22:41

Thanks, i'm very interested, give me some feedbacks for the reference of the analitical solution i will take a look of that but, my studies have the aim to apply CFD commercial technologies to reach god simulation of journal bearings.

LuckyTran May 1, 2012 22:58


Originally Posted by elcino (Post 358729)
Thanks, i'm very interested, give me some feedbacks for the reference of the analitical solution i will take a look of that but, my studies have the aim to apply CFD commercial technologies to reach god simulation of journal bearings.

It is a simple search. There are entire journals and divisions devoted to bearings and it shows up a vast number of texts on hydrodynamics of bearings. My personal favorite is Fundamentals of Fluid Film Lubrication by Hamrock and co.

I've also solved the finite aspect ratio journal bearing problem my self, it is not hard, just directly solve the equations of motions. For small eccentricity ratios it is short, ~1 page long. It is often assigned as a homework problem for graduate courses in fluids / convective heat transfer although.

I will not send you texts you should do your own search so do not request any. Again, tribology is an entire field of study, and journal bearings are arguably the most fundamental of bearings. The analytical solutions are immensely powerful and in most cases exceed the information that can or has been obtained with CFD. CFD offers very little advantages until you get to very complex geometries and very complicated loading scenarios (which even then CFD will do poorly compared to the modelled analytical approximations).

elcino May 2, 2012 05:40

Sure u are great in it, and sure about the great power of numerical methods.

By the way, as a complete CFD software fluent gives u the possibility to do many things, like share data with a mechanical solver, good and easy post processing of the data, create calculation loops and is available in the main companies of the world. My aim is to built a set of simulation useful for industry which automatically cover a large number of geometries, lubricant models, integrated with matlab and gambit in a single matlab program which recall fluent and gambit in batch mode. I will validate my calculations with some experiments. After that i will have a simple txt file of 100Kb that could used and modifided by anyone in the world to create is own loop with any geometry shape and lubricant properties. This is the aim.

Secondly i'm not strong in numerical methods and less in C or fortran programming like, i guess, 90% of the industrial engineers in the world! and this is the second reason why i started to use fluent as the solver for this problems.

I had many references of authors that can do this works with great results, but until now i just can't reply them so there's no sense to go with my models.

M. Deligant, P. Podevin, G. Descombes, 2011,
“CFD Model for Turbocharger Journal Bearing
Performances”, Applied thermal Engineering, 31,
pp. 811-819.
K.P. Gertzos, P.G. Nikolakopoulos, C.A.
Papadopoulos, 2008, “CFD Analysis of Journal
Bearing Hydrodynamic Lubrication by Bingham
Lubricant”, Tribology International, 41, 1190-

By the way again my question was strictly related to the fluent solver! Thanks to anyone who could help me!

stumpy May 2, 2012 17:15

Those pressure BC's don't sound correct if I've understood you correctly. If you have an operating pressure of 101325 Pa then the outlet pressure should be 0. At the inlet the Gauge Total Pressure should be 200000 Pa. The Initial Gauge Pressure shouldn't really matter, but use 200000 Pa too.

elcino May 10, 2012 23:26

Hi, so I've understand the BC input that u told me. I've tried but seams not working. Now I've create a better mesh grid and I've put inlet and outlet far from the bearing fluid domain, I hope it works but I'm not sure. I've read some articles about CFD on Journal bearing s with the use of fluent and the have amazing convergences with crapy mesh grids. If u want u can take a look at the references above, that i can give u! Not clear BC selection!!!!!

All times are GMT -4. The time now is 10:18.