Learning Forum

[Jian Liu]

I'm doing things about laser melting materials, and the models include the gas phase as well as the metal phase. The VOF, melting and solidification models were opened in the FLUENT software, and a surface heat source was used between the two phases, taking into account the shallow depth of the laser action.The model is shown in the figure.

1.According to the information in the forum, I tried writing UDFs, adding a surface heat source at the two-phase interface (z=0) and including a gradient that causes my heat source not to act on the gas phase , can I set the heat source to act only on the metallic phase, even though I have added the statement (z<=0&&C_VOF(c, sec_th)>0.05 && C_VOF(c, sec_th)<1), but it doesn't seem to work?

2.Whenever I calculate a certain time step, it always pops up: Divergence detected in AMG solver: temperature Divergence detected in AMG solver: temperature Error at host: floating point exception

Any tips or ideas on how to use this in ANSYS Fluent are greatly appreciated, thank you very much!

 

 

[Rob]

You can set the heat source to be in cells that are wet, or have some value of the liquid(solid) material. However, you can't then stop heat being convected/conducted into adjacent cells. Yes, the free surface may slow down heat transfer, but if you're adding it into a cell with some gas present you've already got hot gas. 

Divergence can be caused all sorts of things. Check the mesh, source terms etc and review the flow & energy fields as the divergence occurs. That'll hopefully highlight what's going on. 

[Jian Liu]

Okay, thank you very much.

I'll keep trying.

[Dubey92]

You have to use gradient of volume fraction to convert the surface heat sources into volumetric heat sources. For applying the source in the cells at the interface of gas and metal, you have to write a separate UDF to track the cells at the interface because when the melting starts, the interface continuously changes.

[Jian Liu]

Thank you very much for your visit!
I also want to communicate with you:
1. At present, it is true that the surface heat source is converted into a volume heat source through the "magnitude of gradient of volume fraction", and of course there is normal heat transfer in the air domain.
2. According to what you said "separate UDF to track the cells at the interface", I can track the two-phase interface and apply a heat source through the "magnitude of gradient of volume fraction" described in 1 combined with the if statement "(C_VOF(c, sec_th)>0.05 && C_VOF(c, sec_th)<1)", and does it work? (In fact, I read your post carefully before, not sure whether your subsequent work has improved in udf)

really thankful!

[Avinash Kumar]

Sir,

My problem is simillar to yours Can you suggest the solution for given problem below

I am trying to simulate the problem of laser surface melting in fluent . I have created a domain of 11*5*2 mm cuboid shape with 4,44,000 elements and mesh is fined along on of the faces along the length direction through the centroid of the cuboid along the length and laser source is applied as heat flux with gaussian distribution and it is applied as heat flux on the face and as user defined function and in model and boundary conditons are top face and all side face is convected and bottom face is adiabatic .i am sharing the details of the problem in for of photos and i am trying to run calcultions its not converging #questions #disscussion
I would greatly appreciate it if you could kindly assist me in resolving this issue as soon as possible.