COSMOS Newbie: Thermal Stress Analysis?

[Dravvy]

I am currently modelling an engine cylinder head in SolidWorks and originally only planned at investigating mechanical stress across key components.

However I now realise that I will also need to consider thermal stress. For instance the expansion of a component which is constrained by surrounding material.

Does anyone know if this is possible within COSMOS or will I need to use a different piece of software, for instance Abaqus.

I have searched the SolidWorks web page and have found brief details but nothing concrete.

http://www.solidworks.com/sw/products/cfd-flow-analysis-software.htm

Thanks in advance.

Regards,
Dan

 

[JMirisola]

Base COSMOS, which is actually called Simulation now, won't do it. You'd need to upgrade to Flow Simulation, as you've already discovered.

Jeff Mirisola, CSWP
Design Manager/Senior Designer
M9 Defense
My Blog

 

[SteveMartin]

I haven't actually had cause to do this, but you should be able to do thermal stresses with just Simulation (not Flow), assuming you don't intend to model any of the fluids in the cylinder.

Run a thermal analysis on the solid to get the temperature distribution, then import those temperatures in to a static analysis to get the resultant displacements & stresses. The mesh for the two studies has to be identical. Easiest way to ensure that is to mesh one of the two studies, and then copy the mesh into the other one.

 

[EngAddict]

Yeah I believe it is only the Simulation Professional that you will require.

However you might find that the temperature of the piston is almost the same as its surroundings and hence will not experience thermal stress. An F1 engine has to be heated up to operate, it is essentially seized when cold due to the tolerances. When heated up the stress from its surroundings would be reduced.

I am no expert but I believe you would need a larger temperature variance to generate any significant stresses or greatly different thermal expansion coefficients.

 

[Dravvy]

Thank you for the replies, although an old thread it's still a current topic for my dissertation. I must be honest I have only been concentrating on ensuring the mechanical stresses are correct although the next step is to look into thermal stresses which occur.

I have calculated an approximate peak cylinder pressure and temperature so I now just need to apply this to my geometry. I have investigated a professional journal studying a similar topic and the heat transfer which was used was convection.

I have had a play with a thermal analysis within simulation and it seems fairly comprehensive although I'm not sure the steps in which I should follow. Following the study advisor I haven't come across running a thermal analysis to acquire a temperature distribution and I’m unaware of how to start my study.

Apologies for sounding clueless about the subject


Regards,
Dan

 

[kellnerp]

You need the flow package because you have to account for the cooling effect of the fluid flowing through the head and the hot exhaust exiting the head. You can do two fluid simulations in FloWorks as long as they are not free surface problems. Expect very long mesh and run times. Problems with heads are frequently due to dead spots in the flow of the coolant.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

http://www.linkedin.com/in/engineeringtransport

"Node news is good news."

 

[Dravvy]

To grasp the fundamentals of the procedure I plan to first ignore the cooling effect of the engine coolant and look solely at the temperature distribution across the solid body and then hopefully investigate the stresses at the valve heads and combustion chamber.

Using my simple cylinder head model (without cooling jackets) I have applied the relevant materials and added a peak temperature at the combustion chamber and valve surface. I then went on to apply a convection load and applied a heat transfer coefficient although could not pick the entire body, only the surface. Surely this coefficient needs to be across the entire body?

Within the analysis menu there is a 'connections advisor'. Currently my model has a global bonded contact and I'm unaware what this should be set to, is this a feature to define which parts are in contact with each other and if so do I need to seal the combustion chamber by modelling a simplified piston head and block assembly?

I realise there is a bundle of questions there and I may have not been specific, any help with this issue would be greatly appreciated.

 

[kellnerp]

You want to use a convection coefficient in the combustion chamber, on the head, on the exhaust ports and on the coolant passages. They will all be different and are applied to the surfaces.

You don't apply a temperature directly to the head. You may give it a reference temperature, but that is all you will need.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

http://www.linkedin.com/in/engineeringtransport

"Node news is good news."

 

[Dravvy]

Thanks for the advice, I have managed to acquire a simple temperature distribution which seems reasonable.

I guess my next question is how to import this distribution into a static study to investigate stress, or if this functionality is available?

Regards,

 

[kellnerp]

I do believe there is an option in the stress analysis to bring in temperature results. Look in Properties for the analysis. You might look in help for the exact procedure.

TOP
CSWP, BSSE

http://www.engtran.com

http://www.niswug.org

http://www.linkedin.com/in/engineeringtransport

"Node news is good news."