[iorozco86]

Hi every body. Is it possible in cfx increase the size of the domain each time step. For example if i have 1 x 1 domain at t=0 then have 1 x 2 domain at t=1 and so on. Its not like deforming it but like adding grids. Any ideas guys?

[ghorrocks]

You can use dynamic remeshing to add more mesh.

Alternately you can do an initial simulation, then stop. The start a new simulation with the 2x1 mesh and interpolate the first run results onto the second as an initial condition. This much simpler, and can be scripted to run automatically if you want. It is of course a bit limiting in what it can do.

Another alternative as you say is to use moving mesh.

Which option is best will depend on exactly what you are trying to do.

[singer1812]

Or if it is really just discreet addition, you can put in all of the additional mesh in, connected by GGI, and have conditional GGI interfaces close versus time.

No remeshing, or anything needed.

[iorozco86]

thanks guys for your replies:

I am trying to simulate the oil well drilling process. So in few words its an annular space with inlet and outlet (multi phase flow; drilling mud fluid and solid cuttings from the drillbit). In Unsteady state the drill bit (inlet) advance into the formation so the pipe (annular space) increase its length with time. So i tried mesh deformation but when the inlet advance it deforms, and i don't want that, i just want to change its position adding meshes :S

[ghorrocks]

Don't forget you can model this in the absolute frame of reference (ie a stationary earth and the drill bit advances) or in the drill bit frame of reference (ie the earth moves past a stationary drill bit). If there are no accelerations then these two frames of reference are equivalent.

I would have thought modelling the drill bit as stationary would mean you do not need to add mesh.

[iorozco86]

in fact, the acceleration is neglected because the rate of penetration of the drill bit is so small like 50 - 100 ft/hr, but since I m modeling the cuttings bed (sedimentation of the solid cuttings at bottom of the well) and the propagation of it in the length of the well as a function of time I want to consider the length that adds to the initial length as the drill process exist in the time.

So, if I want to move the inlet (drill bit) each time step how can I do it guys? im really lost in this case, cuz I got move the inlet as I want (at a rate of penetration on the formation) but it move part of the inlet, and part of the wall ( of the annular space that shares with the inlet) doesn't move so the inlet deforms and the vectors of the flow at inlet moves.

[ghorrocks]

in fact, the acceleration is neglected because the rate of penetration of the drill bit is so small like 50 - 100 ft/hr,

That has nothing to do with it. Accelerations are rates of change of velocity - so if the velocity does not change then there is no acceleration. It does not matter if the velocity is big or small.

Advancing the drill is a simple application of moving mesh. After a while you will need either a dynamic remesh or a stop/restart to a new mesh when the old mesh is too stretched.

But why model this transient? If the rate of advancement is small then why not model this as a series of steady state simulations? It sounds like you can decouple the advance velocity to the fluid flow and that greatly simplifies the simulation.