i can't see how FEA would help ... CFD might ... if there is a physical gap CFD might show that (apply a high pressure on one side of the assembly, see where the flow goes).
as i write ... if there is a physical gap in the assembly, you should be able to "find free edges" with your front-end of choice (PATRAN, FeMap, ...).
but of course it's a different question if you're wondering if structural deformation will open up a leak-path.
Hi,
yes, I can probably see what M.Gortner meant... In fact, if in the assembly there is some seal between two components which deform under load, this seal could possibly leak out given the pressure and the deformed working conditions.
In this case, FEM would give the deformed status of the assembly, but then the tightness of the seal would depend on some criteria generally given by the seal's manufacturer and mainly based on "decompression".
I wasn't sure if FEA would allow water pressure or similar boundary conditions.
CFD....would that allow this scenario:
You assemble your parts (with compressed orings). Then you apply a pressure. You could ramp the pressure and calculate the pressure where the "seal" is broken or when the fluid penetrates your assembly.
yes can can easily apply pressure to a bunch of surfaces.
if your O-ring is a reasonably linear spring, you could run with a unit pressure, determine by how much the compression in the seal has been reduced, and extrapolate to a pressure that would cause the seal to leak.
The Oring is presumably made of rubber, then its behaivour is inherently non-linear and as such does not lend itself to a linear extrapolation. On top of this you have the added complexity of the Oring being compressed when assembled before any pressure is applied. With a solver like Abaqus you would apply a first step in the analysis just to achieve the assembled condition and then apply the pressure as a second step in multiple increments. By inspection of the results after every increment in the second step, you could hopefully attain a pressure at which the seal is deemed to be broken.
has an interesting and unique element formulation that I think may be able to do what you want. I haven't used it for anything quite like this, but its element formulation lends itself to Fluid-structure interaction that may work well for this. It handles the non-linear behavior, and it's also much more reasonably priced than Abaqus.
I believe Abaqus has the capability to model pressure penetration, usually used for gaskets.
I reckon a model as suggested by johnhors would be an ok first pass, but for some situations as the pressure penetrates further (say, half way through a gasket surface) the area exposed to pressure increases enough to begin having an effect on the overall separation load.
I think this could be acomplished by structural analysis of the gaskets or interfaces in question. You can run a contact analysis, and get information from your contact elements about seperation and contact pressure. You cna then formulate an acceptance criteria (i.e. no seperation and a contact pressure of at least X). A test program would need to be set up to validate your FEA results and/or validate your accept/reject criteria.
Hi,
I agree you don't need CFD at all. First, it would be a big deal to set up a CFD analysis in which the mesh must change at each step, if you want to simulate the separation btw parts and OR. Moreover, the initial gap could be too small for the solver engine to consider it "open wall" and have the fluid leak out.
Instead, with a FEA you can do a multi-step contact analysis, as suggested by others, and change the application of boundary conditions at each step (following the deformation of the OR and the parts, the pressurized area will change). If, at a certain step, you find out that the contact pressure has dropped to zero (or near zero), then the junction is leaking under load.
Challenges: correct material definition for the OR (it's a very big problem), steps definition, BC changes = very long and tedious analysis.
If I were you, I'd run the static part of the analysis without OR and compare the relative deformation to the max allowable OR decompression given by manufacturer's tables.
