[berkmm]

Hello everyone, I was meshing a long geothermal well and soil(earth ground) around this well. Due to the long 100m soid diameter and 109.5mm well diameter feature, meshing can be a complex. Also I dont know well about the meshing techniques and tricks so I couldnt mesh well the fluid domain in my geometry. Therefore, I need some help about the meshing this geometry.

My geometry consist of 3 domain. I've attached real scale meshed and magnified scale of my domain. You can see in Figure 1 and Figure 2.

Even though I opened "Capture Proximity" function the well side has 1 tetra element. Because of that reason it lead to a divergence in fluent. Meshing of well side can be seen in the Figure 3.

The one I want to do is meshing like in Figure 4. It has dense mesh in well of geothermal and capture the physics better due to small feature meshing. I tried defetaure closed but it didnt even worked. How can I make a structured mesh like this in ANSYS Meshing?

Also some people say that they used ICEM CFD to mesh it in 0.1mm to 1m tetra elements. Does ICEM CFD better mesh tool than ansys meshing? Is it easy to mesh in this tool? Hope to see you again soon.

[mCiFlDk]

Originally Posted by berkmm

Even though I opened "Capture Proximity" function the well side has 1 tetra element. Because of that reason it lead to a divergence in fluent. Meshing of well side can be seen in the Figure 3.

The one I want to do is meshing like in Figure 4. It has dense mesh in well of geothermal and capture the physics better due to small feature meshing. I tried defetaure closed but it didnt even worked. How can I make a structured mesh like this in ANSYS Meshing?

Also some people say that they used ICEM CFD to mesh it in 0.1mm to 1m tetra elements. Does ICEM CFD better mesh tool than ansys meshing? Is it easy to mesh in this tool? Hope to see you again soon.

Hi Berkmm,

I can see several problems that I suffered from when I started meshing in ANSYS Meshing.



Capture Proximity allows to control the growth and distribution of mesh in important regions of high curvature or close proximity of surfaces. The problem I see here is that is a very useful tool when analysing regions plenty of edges and surfaces. ANSYS is detecting that your domain is basically an empty parallelepiped, so it assumes that any region needs to be refined apart from the edges as we can see in Fig. 2. To achieve a smaller mesh you'd need to:
a) Reduce globally the element size, and apply "Inflation" for a much finer result near certain surfaces
b) Create a structured mesh (to reduce drastically the amount of useless elements that the option a) would cause)



You also say that a structured mesh could be a good option for you, so I'll try to explain it a bit. A structured mesh needs first of all a bit of literature to understand it, but if I should have to explain it would be: dividing the domain semi-manually to allow the domain to be divided into hexaedron, with almost total control over it.

After having understood it (reading some literature), the "how" is not that complex, but it's tricky. The mesher basically needs that you parameterize each and everyone of the edges of the domain to tell ANSYS how the structured blocks will be placed. Every block must have 4 vertices and 4 edges (for 2D) or 8 vertices and 8 edges (for 3D). It can be done without ICEM, using the same mesher you're using (ANSYS Meshing), but it requires that you understand the tool you're using since several parameters are dependent on others, and some features deactivate others (I haven't used it for several years but I remember it was tricky). I'll recommend you to divide the blocks in a simple paper, and read a bit about ANSYS Meshing's possibilities.

I attach a PDF, I hope you find it useful to understand a bit in which tool you're working.

PDF: http://drahmednagib.com/CAD_2018/Lec..._Meshing_2.pdf



Regards