Author |
Message |
Eli Mamula (spitfire)
New member Username: spitfire
Post Number: 1 Registered: 01-2008
| Posted on Friday, January 04, 2008 - 09:20 am: | |
Hello, I'm looking to use PDE Flex to generate solutions to the Reynolds Lubrication Equation (RLE) to determine stiffness and dampening coefficients for a bearing we wish to use. I'm attaching a general writeup of the approach we're taking, which also includes info on the RLE.Can you give me any sage advice, words of wisdom, etc on how to approach this problem? Thank you. |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 1044 Registered: 06-2003
| Posted on Friday, January 04, 2008 - 01:07 pm: | |
The paper concerns itself mostly with techniques for reducing the complexity of the model, resulting in a system of some 24 cells and 104 degrees of freedom. This is a ridiculously small target for modern computers, and hardly worth all the effort. What is not clear to me from the paper is what the physical equations are that you want to solve, and what equation reductions are advisable from an interpretation standpoint, as opposed to simply trying to reduce the computational load. With modern computers, it is not sensible to spend human time reworking equations to reduce the load on the computer. Junk all that and concentrate on what the equations really are.
|
Eli Mamula (spitfire)
New member Username: spitfire
Post Number: 2 Registered: 01-2008
| Posted on Monday, January 07, 2008 - 03:04 pm: | |
Here is some specifics on how the RLE is used to calculate bearing stiffness and dampening coefficients. The process is summarized like this: 1) Use the RLE to sole for the pressure distribution p=p(tau,z) 2) Integrate the distribution p=p(tau,z) to get forces in x and y directions. 3)Calculate radial load and angle. 4)Repeat solving RLE for ranges of values bounded by bearing's physical design. This should generate a nonlinear curve (or set of curves). 5)To get stiffness (k's) and dampening (c's) coefficients based on orientation with the x & y axis, five slightly different solutions of the RLE are needed (see attachment). Typically, mechanical engineers use the attached algorithm to determine the k's & c's using numerical finite difference method to get the equilibrium RLE solutions & perturbed RLE solutions. They then use these solutions to calculate the partial derivatives resulting in the desired quantities. One software package that is used to perform these calculations is called COJOUR. I think that PDE Flex can be used to get to the same end result with less chance of error and in a more efficient manner. What I am looking for is an opinion/pointers/etc on the best way to start modeling a simple journal bearing and then progress to a more complex multipad bearing.
|
|