Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Piston Ring CFD/FEA/FSI
Posted Apr 3, 2018, 11:57 p.m. EDT Computational Fluid Dynamics (CFD), Structural Mechanics Version 5.3 1 Reply
Please login with a confirmed email address before reporting spam
Goal
My main goal is to define the pressure-flow (PQ curve) past piston rings for a known oil and temperature where the rings are not in motion. Laminar flow and small displacement should suffice. I need help with some Comsol modelling decisions.
Background
I am modelling an application very similar to pistons rings in engine although my application has some differences. My rings are not in constant motion like in an engine. My rings are stationary and move sometimes in practice; for this simulation, I am going to ignore their motion.
The fit between the bore diameter and the rings is an interference fit. In the undeformed state, the piston rings interfere with the bore. Below are some pictures to familiarize yourself with the geometry and loading.
Needed Help
I am having trouble figuring out the setup. The rings are bigger than the bore and thus I need load the rings so they are within the bore diameter before I start the Fluid Structure Interaction. This is turning out to be tricky. If I "form union" between the fluid body and the solid parts, I end up with thin slivers that won't mesh (see picture below). If use the free tetrahedral mesh, I get warning and lots and lots of elements trying to resolve this. Therefore, I suppressed the fluid body at first so that it didn't split the ring geometry. The ring geometry is a prime candidate for a swept mesh. I was able to get a swept mesh (shown below) of the rings and solve the contact problem. Now, I don't know what to do next.
- I need to build a fluid mesh that works with the solid mesh. Is this a good place to use "Identitiy boundary pair" between the fluid and the ring?
- I need to start the Fluid Structure Interaction from the deformed state of the rings when they fit in the bore. How do I do this?
- I plan to use "Automatic remeshing" of the fluid body for when the gaps between the ring and bore get to small. I am going to run into problems with this? Can I do a restart analysis if I get a failure?
Thanks in advance. Learning FSI has been fun.
-------------------Jason Nicholson
Please login with a confirmed email address before reporting spam
I came up with an idea that might work.
- In CAD, draw the bore diameter (diameter of the fluid body) bigger than the piston rings. In Comsol, use "form union" at the end of the geometry node. If the fluid body is bigger than the piston rings, Comsol won’t split the piston ring boundary surfaces (this alleviates the meshing problems of the rings). This should allow me to make a “swept mesh” on the piston rings thus reducing the degrees of freedom in the solid bodies while maintaining high quality solid mesh.
- In step 1 of my study, disable CFD. Solve for the deformed shape of the rings that fit within the bore. Use a "Moving Mesh" interface driven by the displacements of the rings so that it deforms the fluid mesh appropriately. Turn on Automatic remesh if needed.
- In step 2 of the study, the initial mesh is the solution of step 1. In step 2 of the study, use a "Deform Geometry" interface to shrink the diameter of the bore (fluid body diameter shrinks). Automatic remesh is necessary in this step because the fluid geometry/mesh will undergo large deformations; large deformations will greatly distort the mesh and automatic remesh is needed to maintain the fidelity of the elements. I need to better understand weak constraints of the boundary conditions in the “Deform Geometry” interface because I may need them to prevent transferring incorrect boundary conditions to later steps.
- In step 3 of the study, the initial mesh and geometry is the solution of step 2. Enable CFD, disable deform geometry, enable structural mechanics, enable moving mesh. Automatic remeshing will be needed as the gap between the fluid body the rings shrinks or grows. Try to maintain 4-6 elements through the thinnest part of the fluid body. Couple the structural mechanics and CFD. Ramp the inlet pressure and suppress backflow on the outlet. Plot the outlet flow vs. inlet pressure. The results of step 3 should meet my goal for this simulation.
I should make a simpler courser model to verify my steps and figure out the necessary solver settings in each step. This will allow me to iterate and understand the settings quickly. I won’t run the CFD in step 3 on the course model because convergence will probably be an issue. I will aim to keep the degrees of freedom at around 100,000-200,000 or less in the course model. This size of model solves in minutes in my experience. In the finer model, I will aim at 1,000,000-5,000,000 degrees of freedom for fidelity. This model will take hours to days to solve.
If you have any further suggestions, I welcome your comments.
-------------------Jason Nicholson