Heat exchanger: tube/shell

[Anonymous]

Is it possible to simulate a tube/shell heat exchanger, either parallel or counter flow, in multiphysics by assigning mass flow/volumetric flow rates and temperatures to the fluids in the tube and shell? Video tutorial somewhere?

[Joey.X]

Autodesk multiphysics simulation's 'conjugate heat transfer" can take care the heat exchanger problem. It supports the combined fluid parts and solid parts, and the mass flow/volumetric flow rate and temperature boundary conditions are included. 

For forced convection dominated problem, use multiphysics one way fluid to thermal coupling; for mixed or free convection problems, use fully coupled fluid-thermal coupling approach. 

Refer online tutorials below,

(a)  2D, Heat Exchanger with Fluid Flow 

(b)  3D, How to Couple Fluid Flow and Steady-State Heat Transfer for Forced Convection

[Anonymous]

No Multi-physics option in the heat transfer Analysis Parameters. Using Algor Sim Pro 2011. See atttached.

[Joey.X]

 "heat transfer" analysis is basically for thermal only, but provides interface to import fluid (and other) analysis result. 

For forced convection, perform fluid analysis to fluid velocity distribution result, and then perform heat transfer analysis by "Fluid convection"(right click FEAEditor window in right hand side) to load velocity result from fluid analysis.

 For mixed or free convection problems, select Multiphysics steady coupled fluid/thermal or transient coupled fluid/thermal analysis, both fluid and thermal loadings can be applied in these analysis types. 

[Anonymous]

Joey,

Please let me to elaborate. I am interested in a 3D tube/shell heat exchanger analysis. Fluid in the tube, fluid in the shell. I know the temperatures of both fluids and want to observe the heat transfer behavior. Being able to incorporate the fluid flow of both fluids is important in the analysis. If I follow the 3D tutorial that was posted in the initial response, it demonstrates clicking on the "Multi-physics" tab. I do not have that option. Please advise what happened to it in Algor Sim Pro 2011.

Also

If I understand correctly, I can get the same result by importing the fluid flow analysis data as shown in the attachment here. It works for 3D too?

Best

[Joey.X]

hi, Jrm,

In heat transfer analysis, "Multi-physics" tab is not in the UI in V2012. There was some overlaps in Multiphysics loading setup UI. We have consolidated overlapped UIs in recent two releases, fluid convection was appearing in either "loads from file" or global parameter window(your post in last post),  the later has been cleaned in V2012. 

For your second question,  "yes" to load the result in fluid analysis.

[Anonymous]

Joey,

I get where Algor can be manipulated by running fluid analysis first, then loading that load step into a heat transfer analysis.

I don't get where either velocity or flow rates can be assigned to fluid in the multi-physics mode, should one want to go that route in lieu of the loading fluid analysis into heat transfer analysis method. Perhaps some screen shots can be attached here. Using Algor Sim Pro 2011.

[Joey.X]

In fluid flow or Multiphysics analysis, they can add surface based boundary conditions such as volume flow rate or velocity. 

As an example, attached is the flow rate boundary condition UI, you can specify from selecting surface/fan surface/external exhaust or external intake.

[Anonymous]

Joey,

How is this assigned? "multiphysics one way fluid to thermal coupling"

[Joey.X]

Perform fluid flow first to get velocity result (including velocity distribution in fluid domain), then perform thermal analysis by applying fluid result as convection loading, do this by right click in RHS FEAeditor graphics window, select "fluid convection" and pick the fluid simulation result. 

[Anonymous]

I've done that and got results. Can't say I have faith in the results, but got results. However I am still puzzled how I would perform "multiphysics one way fluid to thermal coupling" in the attached analysis type. In multiphysics, I have set up a fluid in a conduit, applied tet mesh, applied surface BC's to the fluid (inlet/outlet, flow rate) for the fluid. Applied surface temperature BC to the entrance (fluid) and heat sink to the conduit, etc. It throws back "Error: Velocity update failed. Quit."

[Joey.X]

What you attached is fluid/thermal two-way coupling which covers the   free convection (buoyancy force). The error is possible came from the questionable model setup including boundary conditions and meshing. 

Generally speaking in numeric analysis, one way coupling is much more stable, depending on your application, if it is convection dominated, one way coupling is an acceptable assumption. But it have no-ignorable buoyancy force effect, you may need to use two way coupling.

[Anonymous]

Joey,

Just to get this straight...

Heat transfer with forced fluid flow is only accomplished by 1. Fluid analysis load step then 2. Loading that load step into the heat transfer analysis (i.e. like a heat exchanger with fluid forced by a pump)

Multiphysics is used ONLY for static fluid (buoyancy convection) where the fluid flows ONLY by providing a heat differential, no pump.

Maybe it's too intuitive to think that Multiphysiscs should be able do both heat transfer scenarios with dynamic fluid AND buoyancy convection.

[Anonymous]

Please see attached from message 2. I think this is where the confusion is.

[Joey.X]

Multiphysics simulation generally implies solving coupling phenomena in multiple physical disciples, which typically includes thermal, fluid, structure, mass and electric in engineering applications. One way, two-way and multiple-way coupling are considered in this category depending on the practical problem and reasonable model simplifications. One can refer any other engineering simulation software or Wikipedia for the terminology and coverage.

Anyway, back to your very first question, it’s for the approaches and examples to simulate tube/shell heat exchanger. Thanks for your questions and discussion. 

[Anonymous]

I for one can say I do not care for patronizing. I have advanced FEA training and used ANSYS for several years. I am very aware of coupling, but your answer would suggest Algor's multiphysics performs miracles. It only performs coupled fluid thermal analysis (well sort of if one can get it to do anything at all). While ANSYS has a lot to be desired on the GUI, Algor seems weak at best if one can even get reasonable or any results at all out of the post processor. I am trying to see if Algor is less user hostile than ANSYS. What I have seen, it is just as bad or worse, not to mention not able to receive straight answers on the discussion group.

Directing a 25-year Autodesk product user to Wikipedia is just unacceptable. For one, Wikipedia is not a reliable resource as its own creator has warned against its credibility.

Maybe John Holtz would like to elaborate because at least his answers are direct and detailed.

[John_Holtz]

Hi everyone,

I hope I don't make things more confusing, but it certainly can become confusing between different degrees of terminology and different versions of the software ! I will try to avoid the terminology, and make comments starting with the newest posts and working backwords in time.

Sorry if the times are given in the users local time zone. Maybe I should convert all of these dates and times to Coordinated Universal Time (UTC)!

related to post of 08-26-2011 09:20 PM (related to post of 08-19-2011 08:36 AM) regarding the terminology "multiphysics one way".

The "forced convection dominated problem, use multiphysics one way fluid to thermal coupling" is not true multiphysics. A better term might be "uncoupled". Please see my P.S. below.

related to post of 08-26-2011 08:46 PM regarding using coupled analysis for pumped fluids.

When bouyancy effects are important in the fluid analysis, there is only one suitable workflow: "Analysis > Analysis Type > Multiphysics > X Coupled Fluid Flow and Thermal", where "X" is either "Steady" or "Transient". With this analysis type, one model includes the solid and fluid parts, the thermal and fluid flow loads. Essentially, two analyses are performed on one model, so the results show thermal results (temperture, heat flux, etc) and fluid flow results (pressure, velocity, etc).

The reason the Multiphysics analysis type should be used is because the temperatures have an effect on the flow pattern, and vice versa.

When bouyancy effects are not important, there are two suitable workflows:

1. the one mentioned above ("Analysis Type > Multiphysics")
2. separating the two phenomena into two different design scenarios.

• Design scenario 1 is "Analysis > Analysis Type > Fluid Flow> X Fluid Flow", where "X" is either Steady or Unsteady. The model consists of the fluid parts and fluid flow loads only.
• Design scenario 2 is "Analysis > Analysis Type > Thermal> Y Heat Transfer", where "Y" is either "Steady-State" or "Transient". The model consists of the solid and fluid parts but only the thermal loads. One of the thermal loads is reading the velocity profile from the results in design scenario 1. See the page "Help > Autodesk Simulation > Setting Up and Performing the Analysis > Setting Up Part 2 > Thermal Analyses > Loads and Constraints > Fluid Convection".

The reason the second workflow is valid is because the flow pattern is (assumed to be) independent of the temperature distribution. The first workflow can be used, but it is less efficient (requires a longer runtime). Theoretically, the first workflow (multiphysics) should give the same answers as the second workflow if the buoyancy effects are turned off (uncheck the checkbox on the Analysis Parameters) in the Multiphysics analysis. (It may not give "exactly" the same answer depending on how the equations are written in the solver.)

related to post of 08-20-2011 05:49 PM regarding how to apply a flow rate.

Just to clarify how to apply a volumetric flow rate, select the surface where the flow rate is know, right-click, and choose "Add > Fan Surface". Set the "Operation" to fixed flow rate and enter the volumetric flow rate, etc.

This type of load can be added in "Analysis > Analysis Type > Fluid Flow" and "Analysis Type > Multiphysics".

related to post of 08-19-2011 02:43 PM regarding the "Multiphysics" tab under the Analysis Parameters.

The document on the algor.com website is outdated, so no wonder you do not have a "multiphysics" tab. The fluid convection portion of that image is now the fluid convection dialog that I mentioned above.

Also, I think there is some confusion in the post. (oh really? ) Instead of the "Multiphysics" tab, I believe the reference was to the "Thermal" tab being removed from the 2012 version of the software. However, both Simulation 2011 and Simulation 2012 show the "Thermal" tab under the "Analysis Parameters" when the analysis type is Steady Fluid Flow. This is a software bug: the Steady Fluid Flow processor does not support buoyancy effects. Maybe we will get it fixed in 2013!

related to post of 08-17-2011 01:47 PM regarding the feasibility (the first post!)

The answer is yes. This documentation page goes through an example step-by-step: Help > Autodesk Simulation > Examples > Fluid Flow > Heat Exchanger with Fluid Flow. The equivalent page on the help wiki is here, but since the wiki is for the 2012 software, the instructions will not help you directly.

Confused? Don't hesitate to ask for clarification.

P.S. For better or worse, we use "multiphysics" to mean that multiple "analysis types" are somehow linked together. And rather than going any deeper into the discussion about one-way, two-way, uncoupled, or coupled analysis, those who are interested can review this page from the documentation: "Help > Autodesk Simulation > Analysis Types > Multiphysics".

[Joey.X]

Using “uncoupled” an aka of one way coupling sounds confusing me.

The purpose of Multiphysics is used to differentiate “single physics” especially multiple physical fields involved; personally I prefer to take this category method which is from named simulation software (not mention name here).

• Multiphysics
• • Uncoupled
  • One way coupled, aka weakly coupled
  • Strongly coupled (two way or multiple ways)
• Single physics

For further reference to clarify, I would strongly refer a key note slide from NAFEMSs (2007, not too old dated) MULTI-PHYSICS MODELLING AND SIMULATION: PROGRESS AND CHALLENGES

Wait a moment, who really cares the terminology?  In reality of engineering simulation software world, there are many commercial packages (typical CFD codes having fluid and thermal I am more familiar) doing some Multiphysics work but not explicitly declaim the name. A typical example is “conjugate heat transfer”, which has coupled fluid in fluid domains and heat transfer in fluid and solid domains.

As in forum, I just think people make their own justification.

[Anonymous]

John,

Uncle!

I have ran Algor with multiphysics transient (coupled without buoyancy). Please see the attached settings. I have a conduit with multiple turns (think copper Archimedes coil). And a fluid in the conduit (tube). No shell at this time as I am trying the baby steps to get heat transfer from the fluid to the tube. There is an inlet assigned, an outlet assigned, a velocity assigned on the inlet surface of the fluid and applied temperature assigned to the inlet face of the fluid. Node default temperature is set to 70 F. The inlet is oriented in the X-direction, so the velocity is set x=value, y=0, z=0. I am using the default global coord. sys.

I get a result with temperature and heat flux distribution but common sense and tribal knowledge tells me the results are wrong. It's like the distribution is "stuck" and does not fill out the rest of the tube. It fills out about 1/4 of the conduit.

I have tried buoyancy, adjusting tolerances, applying convection and temperatures to the outside wall of the tube, adjusting the nodal temperature, increasing/decreasing the applied temperature, increasing/decreasing the velocity, all kinds of combinations.

BTW, I have ran steady state multiphysics and both transient and steady state uncoupled (fluid analysis loaded into a heat transfer analysis). Still not satisfied the results are real world. I do think the transient multiphysics is the best choice, in theory, for a heat exchanger.

Any ideas?

[hupn]

Hi,

I have not go through the long thread, but based on your attached file, I have several suggestions. For transient fluid/thermal coupling analysis, to consider the "buoyancy" effect, you need to check the option from the analysis parameter screen of the coupling analysis. Otherwise, the heat transfer has no effect to fluid flow and the run is similar to fluid-> thermal one directional coupling. Also for coupling analysis, the k-epsilon turbulence model is not supported and only LES turbulence model is available currently. Make sure you set the turbulence model on in the analysis parameter screen of the coupling analysis too if needed.

I attached a figure for your reference.

hup