transient heat transfer

[Anonymous]

Hi all,

I am working on a model, where I aim to simulate heat transfer through the thickness of a masonry wall. This type of an analysis is completely new to me and I'm progressing bit blindly.

I constructed the model, selected the isotropic thermal brick element options, meshed it. I went through the analysis parameters based on the information provided at http://download.autodesk.com/us/algor/userguides/mergedProjects/setting_up_the_analysis/Thermal/****...

At this stage of my work, I selected two surfaces and assigned controlled temperatures (but normally what i need to do is to define the ambient temperature on both sides of the wall rather than assigning temperature values to the surfaces of the wall. This is going to be a new message I guess) and tried to run the model. 

The resulting temperature distribution, which does not make any sense, is attached.This is a three layer masonry wall with different material properties for stone, mortar and rubble. Apart from the strange temperature values, don't you think the results seem a bit too homogeneous? 

Please help!!

yasmin

[Anonymous]

Hi,

Just realized that in the analysis that I talked about in my previous message, the controlled temperature load was applied with the "varies with time" option activated with a very simple loading curve (in terms of multiplier,time: 0,0 - 1,0.5 - 1,1), and at the end of the analysis I receive an error message saying "Error occurred with module: C:\PROGRA~1\Autodesk\SIMULA~1\srun11.exe".

When I carry the same analysis deactivating the "varies with time" option, on the other hand, I obtain the results as indicated in the attached figure, which do not seem meaningful, as once again the temperature distribution is too smooth. Please note that he modelled wall is a highly heterogeneous structure, composed of three different materials, each of which with different material properties.

please help..

y

[AstroJohnPE]

Hi Y,

If I understand your loading (10 C forced on one end; 2 C forced on the opposite end), then the results look correct to me. There is a very strict limit to what the temperature gradient can be through the different material; it is forced to be 10 C to 2C regardless of the material! You might be able to see a slight difference if you use a 10 color scale in the legend instead of the 128 smooth colors.

Also, what is the variation in the material properties? It may not be large enough to cause an obvious effect.

Another way to think of your problem is a bunch of springs connected between two rigid plates. Most of them are in parallel (equivalent to different materials side-by-side in your thermal model) and some are in series (equivalent to different layers through the thickness). Some springs are strong (equivalent to low thermal conductivity) and some are weak (high thermal conductivity). One plate is stretched 10 units, and the other plate is stretched 2 units. When looking at the displacements, they will look relatively smooth because all of the springs connected between the plates stretch between 10 and -2 units. The variation in displacement (temperature) depends on the stiffness (thermal conductivity). BUT, the variation in force will be more noticable because it is related to the displacement (about the same in all springs) and the stiffness (different in each spring). I expect the same to be true in your thermal model. The temperature change is fairly consistent through different materials, but the heat flow or heat flux through each section will be different.

Your case is more complicated than my spring analogy because you have heat flow in 3 directions from one face to the opposite. My spring analogy only has results "flowing" in one direction.

[Anonymous]

Hi John,

Thanks a lot for your reply,

I actually did change the material properties to accentuate the difference, and was able to see the not-so-smooth temperature transitions so I am convinced that my model is working.

However,

1. I could not find anything related to the error message that I mentioned in my previous message. Ultimately, what I am trying to simulate in this model is a time-dependent temperature variation both indoor and outdoor (consider that my wall belongs to the outer facade of a building) to be able to see the temperature profiles through the thickness of the wall. So, I need to use the time-variation option.

2. I need to define indoors and outdoors temperatures, rather than the surface temperatures. Am I going to need to model two additional volumes on both sides to represent the air?

3. Instinctually, I guess the model size is important for this type of analyses (as the wall is normally continuous on four sides). How can I be sure that my model represents a sufficiently big part of the actual wall to simulate the temperature profile?

Sorry for a lot of questions.

yasemin