[StudentinANSYS]

Hello!

I just registered on this forum as I am going to do lots of work now and in the future in my studies.

I am new to ANSYS CFX and this may be a small question to ask in an entire new thread.

The problem:

I want the mean temperature in a solidified body of air and make the assumption that the air in another body is 10 °C lower somewhere else.

This because I have to assume a convective heat loss from the model which is dependent on a temperature somewhat lower than within my geometry which is a small area of the big surveillance pod we're simulating.

How do I make an expression of the temperature in a certain point or mean body temperature which I can use in the boundry condition on the outer surface? I need an outside temperature along with my heat transfer coefficient.

Furthermore, where can I find syntax for ANSYS to get these, probably already implemented funcitons in ANSYS?

I'm quite new to this and the question might be stupid but I'm glad for all the help I can get.

Regards
Jakob

[ghorrocks]

I want the mean temperature

volumeAve(T)@volume

make the assumption that the air in another body is 10 °C lower somewhere else

I do not understand this question. Please explain more fully what you trying to do.

You seem to be asking how do you set a temperature to be a function of a temperature somewhere else - in concept this is easy, just set a CEL expression to abc = voluemAve(t)@Region, and then set the wall temperature of the other surface to variable abc. But I think you will find this is numerically unstable and will require additional treatment to converge. But more importantly, this does not sound physically realistic, and physically unrealistic things are always very difficult to converge (because they are impossible ).

So I recommend you think about what is really happening here. Or if you want us to help you work out a better way of modelling it please explain what you are doing more fully. Include images and a description of what it does - and why you think the temperature of one region is linked to the temperature somewhere else.

[StudentinANSYS]

Originally Posted by ghorrocks

volumeAve(T)@volume



I do not understand this question. Please explain more fully what you trying to do.

You seem to be asking how do you set a temperature to be a function of a temperature somewhere else - in concept this is easy, just set a CEL expression to abc = voluemAve(t)@Region, and then set the wall temperature of the other surface to variable abc. But I think you will find this is numerically unstable and will require additional treatment to converge. But more importantly, this does not sound physically realistic, and physically unrealistic things are always very difficult to converge (because they are impossible ).

So I recommend you think about what is really happening here. Or if you want us to help you work out a better way of modelling it please explain what you are doing more fully. Include images and a description of what it does - and why you think the temperature of one region is linked to the temperature somewhere else.

I have a problem with a power switch where I am qualificating it to withstand mach 2 and have calculated the heat on the planes outer surface. The power switch is connected to a heat sink which transports most heat to the outer surface where the convection is high.

The trapped air acts like a solid, only conduction is modelled. But the air which is not trapped is to assume a temperature. Because the volume of this air is bigger and because the convection maybe could be neglected on these surfaces we just want to approximate the temperature in this air to make an expression for the convection. And the only thing I can tell is that it should be lower than the trapped air. To make convection I need a temperature difference.

Thanks for volumeAve! It solved the problem. Also found more functions in the meny in ansys really convenient

[flotus1]

A hint for further questions: a sketch is worth a thousand words. Literally

[StudentinANSYS]

Originally Posted by flotus1

A hint for further questions: a sketch is worth a thousand words. Literally

Thanks I will remember this

Anyways thanks for all response, highly appreciated.

[ghorrocks]

To back up Alex's comment - after reading your description in post #3 I am even more confused.

Why is the switch going at Mach 2? That does not seem normal switch behaviour And what do you mean by trapped air acts like a solid?

Please attach an image of what you are trying to do. Here is the way to attach an image: https://www.cfd-online.com/Wiki/Ansy...n_the_forum.3F

[StudentinANSYS]

Originally Posted by ghorrocks

To back up Alex's comment - after reading your description in post #3 I am even more confused.

Why is the switch going at Mach 2? That does not seem normal switch behaviour And what do you mean by trapped air acts like a solid?

Please attach an image of what you are trying to do. Here is the way to attach an image: https://www.cfd-online.com/Wiki/Ansy...n_the_forum.3F

This is a picture of the pod and its power-distribution system: http://imgur.com/a/RaQjx

Acknowledge it is really simple and out of proportions.

According to Holman J.P, Heat transfer 10th edition, 2010 trapped air can be modeled as a solid if a new effective k-value is calculated according to the circumstances. This is also confirmed by our professor.

Mach 2 is the speed the pod is required to function during flight mounted on a SAAB GRIPEN.

As we're not modelling the entire pod I had to estimate an air temperature for the air underneath the topmost box mounted on the inside of the pod's frame. As it doesn't heat flux much from the inner frame of this box to the inside air in the pod (the temperature is about the same here as inside the box of the power converters) I just wanted to estimate it somewhat and therefore used volumeave(T)@inside air in box and subtracted 10 instead of neglecting it.

The mechanical attachment to the frame of the pod is acting as a heat sink and is transporting most of the heat out of the power switch. The convection on this surface is really high and is cooling it off efficiently.

[ghorrocks]

I still have little idea of what you are modelling. The image is very simple and is not labelled.

From what I can guess you are doing:
Your assumption that the trapped air can be modelled as a constant temperature rise compared to somewhere else sounds very dodgy. Why not just model it as adiabatic? Or why not model it as normal air anyway?

[StudentinANSYS]

Originally Posted by ghorrocks

I still have little idea of what you are modelling. The image is very simple and is not labelled.

From what I can guess you are doing:
Your assumption that the trapped air can be modelled as a constant temperature rise compared to somewhere else sounds very dodgy. Why not just model it as adiabatic? Or why not model it as normal air anyway?

No, I'm not modelling the trapped air to have an temperature relating to something else. I'm modelling the air underneath the box to have the air temperature somewhat the same as the air close to the temperature of the outer frame of the pod, just to get some convection there and not neglecting the heat flux to the other side of the power switch.

http://imgur.com/a/sb5zU

these are some result. Temperature will always be the same as I am stopping sims at a given qualification temperature, just interested of the time the pod can run. Estimating it around 30 minutes with different elevated temperatures

/Jakob

[ghorrocks]

Can you please imagine you are explaining this to your grandmother. But only that your grandmother happens to have a reasonable amount of general experience in CFD . I have no idea what the pod is, where the outer frame is, where the power switch is, or why convection is relevant. Also I have no idea what relevance the qualification temperature is or how this relates to how long the pod can run.

I suspect at this stage is that the originally posted question is an XY question (http://xyproblem.info/) and that the correct approach for this is totally different. But as I don't have much of an idea of what is being proposed in this thread I can't suggest anything useful yet.

[StudentinANSYS]

Originally Posted by ghorrocks

Can you please imagine you are explaining this to your grandmother. But only that your grandmother happens to have a reasonable amount of general experience in CFD . I have no idea what the pod is, where the outer frame is, where the power switch is, or why convection is relevant. Also I have no idea what relevance the qualification temperature is or how this relates to how long the pod can run.

I suspect at this stage is that the originally posted question is an XY question (http://xyproblem.info/) and that the correct approach for this is totally different. But as I don't have much of an idea of what is being proposed in this thread I can't suggest anything useful yet.

This became a little bit unclear I suppose. But the original question was only that I needed some syntax in ansys and that solved my question. I'm not allowed to share the entire assignment. It's big and I understand now that it is hard to explain but because you've helped me so much I feel obligated to explain it.

I'm qualificating an electrical system on what temperatures it should withstand. The warmest part is the power switch. It has to be designed to withstand the temperature I figure out according to an initial stabilized outside temperature. After recieving the critical temperature in the power converters in the switch, I run it at elevated temperatures as an sensitivity analysis to check how long it can withstand running if it flips from idle to load.

[ghorrocks]

Thanks. So why not just use a normal cht simulation with an air domain with no special treatment, just flowing past at mach 2 on the outside and mixing and connecting inside the device?