[Jaguar]

Hello Forum,

I am new, this is my first post, and I am happy to join this great forum.

I have had experience with the FEM methodology, but mainly structural. I am new to the fluid world.

I have a problem.
I want to simulate a Box (with its vents), which inside has several Fr4 boards, with some heat generators.

Main heat generator has thermal-pad on it and aluminum heat sink. Intention is to evaluate the assembly condition, to find out the optimum ventilation and heatsink size.

I have been dealing with unconvergences, and most of the time my model crash. I only obtain the | ERROR #004100018 has occurred in subroutine FINMES. |
| Message: | Fatal overflow in linear solver.



My model is basically a Box, with adiabatic floor, 4 walls with constant temperature, and the top has an opening BC, with constant temperature, zero gradient turbulence, and static entrain pressure of 1 atm.

I am using a modified Air material properties, for wich I used the variable density. (it varies with temperature).

This is the variable I am using:
Dens: (360.77819 [kg m^-3 K^1.00336] )*(T)^-1.00336

These are air properties I am using:

-----------------
MATERIAL: Air Multi Temperature
Material Description = Air from 100K to 1600K at 1 atm
Material Group = Air Data, Constant Property Gases
Object Origin = Default
Option = Pure Substance
Thermodynamic State = Gas
PROPERTIES:
Option = General Material
Thermal Expansivity = 0.003200 [K^-1]
ABSORPTION COEFFICIENT:
Absorption Coefficient = 0.01 [m^-1]
Option = Value
END
DYNAMIC VISCOSITY:
Dynamic Viscosity = 1.913E-05 [kg m^-1 s^-1]
Option = Value
END
EQUATION OF STATE:
Density = Dens
Molar Mass = 28.96 [kg kmol^-1]
Option = Value
END
REFERENCE STATE:
Option = Specified Point
Reference Pressure = 1 [atm]
Reference Specific Enthalpy = 0. [J/kg]
Reference Specific Entropy = 0. [J/kg/K]
Reference Temperature = 40 [C]
END
REFRACTIVE INDEX:
Option = Value
Refractive Index = 1.0 [m m^-1]
END
SCATTERING COEFFICIENT:
Option = Value
Scattering Coefficient = 0.0 [m^-1]
END
SPECIFIC HEAT CAPACITY:
Option = Value
Specific Heat Capacity = 1.005E+03 [J kg^-1 K^-1]
Specific Heat Type = Constant Pressure
END
THERMAL CONDUCTIVITY:
Option = Value
Thermal Conductivity = 2.71E-02 [W m^-1 K^-1]
END
END
END
-----------------

I am using double precision, 1 atm reference pressure. I don't have inlets or outlets.

I have tried several meshes, from coarse to fine (From 4M elements to 17M) and sometimes converges sometimes it doesn't ....

I can't find what is wrong...

I am using turbulence K-e .... buoyancy activated

Any help or advise would be appreciated.....


Thank you very much!

[ghorrocks]

Why are you using that weird density function? Why not just use incompressible bousinesq buoyancy? Or ideal gas for buoyancy?

The linear solver overflow means your model has diverged big-time. You need smaller time steps, better mesh quality, correct your physics or use a better initial condition. I should write an FAQ about this some day.

[Jaguar]

Thank you very much for your response, it's appreciated.


My intention with the function was to capture the density change as the air is being heated up... but it seems it's giving me problems to converge.

For the mesh, I will improve. I am preparing a new mesh, that I want to try.

Other thing I will try is to reduce the time step.

So, my next model will run with improved mesh, smaller time step and with bousinesq buoyancy....

I will let you know about the results...

Thank you again!

[Jaguar]

Hello Glenn,

Already followed your advise. Thanks.

I used ideal gas for buoyancy, and adjusted the mesh. Everything seems ok now, (It converges normally) but the problem I have is that temperature values of the ICs are way much hotter than expected. (like 2.3 times higher)

I had a validation run, which is only one IC on open ambient, and this validation matches the experimental result. When I run with all the ICs, they tend to get very hot. It seems that they are overheating.

Interface is conservative interface flux, please check info of IC and GAS interface:

FLOW:
DOMAIN: IC_Domain
BOUNDARY: IC_GAS Side 1
Boundary Type = INTERFACE
Interface Boundary = On
Location = F58.56,F61.56,F60.56,F57.56,F62.56
BOUNDARY CONDITIONS:
HEAT TRANSFER:
Option = Conservative Interface Flux
END
END
END
END
DOMAIN: GAS_Domain
BOUNDARY: IC_GAS Side 2
Boundary Type = INTERFACE
Interface Boundary = On
Location = F123.124,F120.124,F121.124,F122.124,F119.124
BOUNDARY CONDITIONS:
HEAT TRANSFER:
Option = Conservative Interface Flux
END
WALL INFLUENCE ON FLOW:
Option = No Slip
END
WALL ROUGHNESS:
Option = Smooth Wall
END
END
END
END
END

Also, Heat generation is a subdomain, with volumetric generator:

FLOW:
DOMAIN: IC_Domain
SUBDOMAIN: HEAT
Coord Frame = Coord 0
Location = IC_Domain
SOURCES:
EQUATION SOURCE: energy
Option = Source
Source = 2304000.0 [W m^-3]
END
END
END
END
END


And material properties:

LIBRARY:
MATERIAL: Air Ideal Gas
Material Description = Air Ideal Gas (constant Cp)
Material Group = Air Data, Calorically Perfect Ideal Gases
Object Origin = Default
Option = Pure Substance
Thermodynamic State = Gas
PROPERTIES:
Option = General Material
ABSORPTION COEFFICIENT:
Absorption Coefficient = 0.01 [m^-1]
Option = Value
END
DYNAMIC VISCOSITY:
Dynamic Viscosity = 1.831E-05 [kg m^-1 s^-1]
Option = Value
END
EQUATION OF STATE:
Molar Mass = 28.96 [kg kmol^-1]
Option = Ideal Gas
END
REFERENCE STATE:
Option = Specified Point
Reference Pressure = 1 [atm]
Reference Specific Enthalpy = 0. [J/kg]
Reference Specific Entropy = 0. [J/kg/K]
Reference Temperature = 25 [C]
END
REFRACTIVE INDEX:
Option = Value
Refractive Index = 1.0 [m m^-1]
END
SCATTERING COEFFICIENT:
Option = Value
Scattering Coefficient = 0.0 [m^-1]
END
SPECIFIC HEAT CAPACITY:
Option = Value
Specific Heat Capacity = 1.0044E+03 [J kg^-1 K^-1]
Specific Heat Type = Constant Pressure
END
THERMAL CONDUCTIVITY:
Option = Value
Thermal Conductivity = 2.61E-2 [W m^-1 K^-1]
END
END
END
MATERIAL: Si
Material Group = CHT Solids, Particle Solids
Option = Pure Substance
Thermodynamic State = Solid
PROPERTIES:
Option = General Material
EQUATION OF STATE:
Density = 2200 [kg m^-3]
Molar Mass = 1 [kg kmol^-1]
Option = Value
END
REFERENCE STATE:
Option = Specified Point
Reference Specific Enthalpy = 0 [J/kg]
Reference Specific Entropy = 0 [J/kg/K]
Reference Temperature = 25 [C]
END
SPECIFIC HEAT CAPACITY:
Option = Value
Specific Heat Capacity = 7.40E+02 [J kg^-1 K^-1]
END
THERMAL CONDUCTIVITY:
Option = Value
Thermal Conductivity = 1.38 [W m^-1 K^-1]
END
END
END
END


I don't see any problem on the data.. just that temperatures are way much higher than expected... do you have any idea? Could you give me some advise? or any checkpoint I may be missing?

Thank you very much for your help. This problem is driving me nuts...

[ghorrocks]

How hot is it getting? How hot should it get?

[Jaguar]

Component should be on the 100°C order....

From the simulation, I am obtaining (Plotting conservative value):

When plotting Air domain temp on the IC component surface: ~ 240°C
When plotting IC domain temp on the IC component surface: ~ 350°C

Do you have any Idea if what could I have been missing?

My guess is that I am wrong on my input data for Heat Generation.

I receive the W spec (For example, 1 Watt power consumption) and I take the volume of the IC, and I use volumetric heat generation. [ W / m3 °K] ~ I divide power over volume.

My understanding is that not all the Wattage power is turned into heat, some of it is transformed into work and some other is loss as heat (Here is where efficiency of the IC comes into play), but I don't have that Spec, so I got the request to use 100% of the Watt to have the worst condition. How valid would this be?

Do you have any previous experience in aproximate efficiencies?

I could check the expected value, the temp I am having... make the division, and obtain an efficiency value. Modify my input, and make it run again. But I feel that I would be uncorrectly handling the efficiency value, because it could be covering up any other possible error.

My problem is closed enclosure, with forced high temperature (Heat chamber) and the experimental value, was done on regular environment (21°C), no enclosure. Validation experiment, fits ok using 100% of the power. Intended simulation, it seems that doesn't fit equal....

I am confused .

[Jaguar]

Glenn,

Thank you for helping me.

[ghorrocks]

First of all work through this FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

And looks like you have some homework to do to confirm what the heat inputs really are. If you don't know the heat input you are just guessing.

Also don't forget that with temperatures like this radiation may well start becoming important. It may be necessary to add a radiation model.

[Jaguar]

Hello Glenn,

I can make the model run stable, and I am monitoring the temperature at 0.1 mm from the volume of each source of heat. (I have several sources).

What I see, is a constant up. My understanding is that I should have asintotical curves..(At some point)

Steady state simulation should allow me to see this asintotical behaviour on the air temperature that is close to the heat generation correct?


My model basically is a heat generation volume, attached to a PCB. There is air surrounding. then I have the enclosure, and air outside. Air in and Air out are the same fluid domain, connected by the enclosure vents. Enclosure is made with a plastic solid domain (not just inner walls) because one point of interest is the temperature outside the enclosure.

Fluid domain is a big square with the enclosure in the middel. My boundary conditions: Adiabatic wall on the bottom of the air domain (Enclosure is floating on it) Walls with 40°C constant temperature on the sides, and opening condition on top, with static Entrain Pressure of 0, constant temperature of 40°C.

Buoyancy is applied, using Ideal Gas. I am doing it steady state... but I am afraid that is not stabilizing... is adding heat forever.... Should I change the boundary conditions?

On the other hand, about radiation. I don't have any information on it... so I wouldn't know how to properly set it up...