UNRESOLVED: Inlet/outlet geometry not allowing transfer of flow URGENT HELP

UNRESOLVED: Inlet/outlet geometry not allowing transfer of flow URGENT HELP

Anonymous
Not applicable
9,866 Views
52 Replies
Message 1 of 53

UNRESOLVED: Inlet/outlet geometry not allowing transfer of flow URGENT HELP

Anonymous
Not applicable

Dear Autodesk Help.

 

In regards to my last post (this is becoming very urgent as I have tried multiple different ways) I rebuilt my Inventor model to first extrude holes for the inlets, and then extrude solids to mark the inlets. In CFD, when highlighting using volumes, these inlets are highlighted as one volume. The flow is still contained in this volume when the model is run. I have attached my latest inventor and CFX files, as well as another print screen of what i am experiencing in this model.

 

Assistance would be much appreciated, as I am going the same way as described and found previously and have no idea why this fluid isn't crossing over.

 

Kind Regards,

 

JAck

0 Likes
9,867 Views
52 Replies
Replies (52)
Message 21 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hi Jack, 

Sorry to bring you so much of confusion. Below is my reply:

 

In this model you haven't stated any volume flow rates at the outlets?

Yes. This is the first model I ran for testing the boundaries.

 

So How do I know for sure if any Carbon monoxide is leaving at the P = 0 outlet? I feel like I have so much conflicting information about what to do now, and did you do a conversion to increase the flow rates of the scalar and hence reduce its value? ( you have the units as m^3/h). 

If the scalar is bigger than 0 on the P=0 boundary, there is CO leaving. Does this make sense to you?

This model is just for my testing. The value of volume flowrate or diffusion coefficient is not the exact number.

 

I have also been able to get vectors at the P = 0 boundaries, but I dont know how much this means. I need to be able to make sure the amount of CO in the car park is less than 0.00005 scalar, or doesn't exceed this amount in any area. 

In regards to how many inlets/outlets, i thought this may be applicable as there is a massive amount of CO in my model regardless of how i set up the initials.

I have graphed your model after the first few iterations, as you can see the majority of the area is already 5e-05, in other words everyone in that carpark is suffering from carbon monoxide poisoning - attached is the screenshot.

After calculation gets converged, check the scalar result if is bigger than 0.00005. Scalar result during the run has no point.

 

The way these systems work in real life is that a fan pulls (for this case 100 000 L/s) out of the car park, this is 100 000 litres every second. All this model is doing is throwing in carbon monoxide into a box, allowing it to slowly leak out an opening. Can autocad CFD do exhaust modelling?

Yes. By using volume flowrate=100000l/s at the outlet will let this amount of mixture being pulled out.

 

I have checked the latest model you uploaded, below is my suggestion:

You don’t need to extend the volume flow rate boundary. The extension at P=0 may not need to be that long. 1/3 of the length should be enough.

Besides, it is an incompressible analysis so please don’t use compressible analysis. As there is no temperature change in the model, I don’t think you need to enable heat transfer.

I suggest you run a steady state analysis first and check the result after converges.



Xiaojin Zhang
0 Likes
Message 22 of 53

Anonymous
Not applicable

Hi XiaoJin,

 

This is the problem I am having - the model converges, or the lines basically go straight - and the whole model is RED, even to the point where the red starts infiltrating the volume flow rate inlets.

 

Are you sure that the flow rate at the outlet works? you have only modeled a P = 0 here? If i input a flow rate leaving the model - do i state a scalar or anything else or just a flow rate? I have done this before, but the result still looks exactly the same, a massive box full of red, that just slowly keeps filling and filling.

 

when it comes to entering boundary conditions, say I have 8 outlets, they all have 1000 applied, is that 1000 across all 8, or 1000 each? 

 

I have run models where I put ridiculous amounts of supply air and exhaust air, all to have the model just fill entirely with CO. This is very frustrating.

 

I have extended the extrusions because the vector magnitude at the top of these wasn't' showing arrows directly upwards, it still isn't happening so I'm not sure what that's about, I'm more worried about actually being able to model something of worth using this program...

 

I will run this model with exaggerated exhaust flows to see if anything happens. Will upload CFZ file when finished.

 

Thank you

 

Jack

 

 

 

 

 

 

0 Likes
Message 23 of 53

Anonymous
Not applicable

Attached is my latest file, after 100 or so iterations it still seems to be exploded with carbon monoxide.

 

the WORST CASE scenario ppm for the carbon monoxide after 1 second is as follows:

 

generation rate: 2.42 g/s

 

in one second (given volume of the space is 54 400m^2)

 

concentration of CO is (2.425/54 400) = 4.46e-5 which is actually 45 ppm.

 

Now this is easily the worst case scenario, because there is going to be tens of thousands of liters of air being blow into the space, and even more exhausted, which in theory brings this value down even further. in my model (I have inflated the volumetric inputs and outputs greatly to see if they yield a more worthwhile result) I am still generating large amounts of unexpected CO. This is starting to look like a dead end with this program. Please advise.

 

PS I am 100% sure changing mesh and extrusion length isn't going to magically make less CO collect in this volume, I've tried every possible way of doing this, its starting to really get me worried

0 Likes
Message 24 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hello Jack,

It is a cfdst file. I could not open it.

I will try to run your model on my side. I will get back to you after I receive the result.

Please kindly wait for a moment.

Thank you!



Xiaojin Zhang
0 Likes
Message 25 of 53

Anonymous
Not applicable

Hi Xiao,

 

I have already left work, I was meant to save that as a CFZ file, that's quite upsetting as I won't be in till tomorrow :(. no worries thank you very much for your help, I am absolutely lost now

0 Likes
Message 26 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hi Jack,

May I confirm with you about the number of the volume flow rate? If it is true, the P=0 boundaries are acting as inlets and also pulling air into the model.

I am wondering if the total volume flowrate of the air is 73238l/sX2?

 

Capture.JPG

 



Xiaojin Zhang
0 Likes
Message 27 of 53

Anonymous
Not applicable
Hi Xiao,

p= 0 as an inlet?! No as i have previously stated. there is a total of 73 000 entering air
there is 2.12 l.s carbon monoxide entering also
exiting should be 93 000 l.s but we cant add this in. we can only add p= 0 as an outlet. i thought we already concluded this? i also drew a diagram! please let me know if this cant be done.
0 Likes
Message 28 of 53

Jon.Wilde
Alumni
Alumni

Hi Jack,

 

To try to be clearer here, you have 93,000 l/s being pulled out.

You have assigned 72,900 l/s in, split over 5 inlets at 8100 l/s each.

(Ignoring the minimal scalar 1 flows as they are near zero).

 

This means that you still need to add an additional 20,100 l/s to the model for there to be a mass balance. This is why CFD is treating those P=0's at inlets.

 

We need to have a mass balance. Could you help us to better define the model to guarantee this?

Something sounds wrong, where does the other 20,100 come from?

 

Thanks,

Jon

0 Likes
Message 29 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hi Jack,

Just in case that you didn't notice my private message. Jon and I tried to call you just now but no one answered.

Could you please confirm if the number you input is correct? Please send me your number by private message if it is not correct.

We are glad to call you tomorrow. Jon and I will be available to from 8:00AM German time. Could you please let me know which time works best for you?

Thank you!

 



Xiaojin Zhang
0 Likes
Message 30 of 53

Anonymous
Not applicable
Hi John,
In that case I tried to model a negative like I said before. I can't upload my newest one till tomorrow :(.

In the models that I did today, I tried modelling just the inflow ( 93 000 l/s), with P=0 inlets, but the scalar still seemed to pour into the model too rapidly, with high outputs. is there a way to drive the flow and measure the work output of the prescribed device if so?

I also tried a negatively pressure at the outlets to try this -500pa each (assuming all selected will apply that to the outlets) which seemed to have no effect.)

You mentioned resistances to help drive the flow out? how would I go about that? it would tell me a volume flowrate equal to the inputs in this case?

Cheers for the help, Shall keep trying

Regardis,

Jack
0 Likes
Message 31 of 53

Jon.Wilde
Alumni
Alumni

No worries, I am running with what we have so far, which should be OK in principal.

 

I suggest we skip resistances for now, get what we have working and then add complexity if you deem it necessary.

These are more to model leakage paths - not necessary yet.

 

What is actually happening in reality where you have the CO being emitted?

What is the flow rate in reality? 

Is it what you have assigned or did you drop it down to just introduce 100% CO from this surface at a super low rate?

Hopefully Xiaojin can reach you tomorrow to get some more details 🙂

 

Thanks,

Jon

 

PS:

 

Pulling air out seems to at least give a stable solution with no outlet recirculation (of course, as it is now a flow rate), although I do have questions about the scalar values as above. I guessed the values for now as there is still an open question over what flow is going where to get a mass balance, which we must have.

 

Pulling flow out.png

0 Likes
Message 32 of 53

Anonymous
Not applicable

Hi Jon,

 

Sorry about my last few messages, they were sent from my phone, and I was at training when you called, I could hear you but you couldn't hear me! 

 

I shall upload another latest one soon, I think the longer outlets were helping with said re circulation, although sometimes the corner ones' velocity vector seems to point downward.

 

I wanted to know about the resistances to try and force more extraction so the overall CO value doesn't get unreasonably high.

 

In reality? There is a mixture of Air/CO that is leaving the exhaust ports at an increased rate to what air is being pumped into the space.

 

Usually when we design systems for these sort of problems, we have deemed-to-satisfy conditions the car park must adhere to, usually this is by doing a calculation which yields the Carbon monoxide generation rate, and in-turn a Flow rate needed as exhaust from the space - This is the 93 000 (note that number is actually for a car park double the space). The flow rate in is a percentage of the airflow in, understanding that air isn't entering/leaving at the same rate or at least this is how the fans are sized up.

 

In this case, we have:

 

8730 g/h, or 2.12 L/s Carbon monoxide generation.

 

The exhaust that was calculated using the deemed to satisfy clauses is 93 000, this number is rather arbitrary, It should be half of this to exhaust the amount of CO desired. I'm working with the larger number to try and get this model to work.

 

In reality I guess it is hard to tell what each flow rate will ACTUALLY be, but the fans would be sized for 93 000 exhaust and 73 000 Supply, I have never been involved with commissioning these systems or anything so I'm not sure how close the installed flow rates are compared to the theoretical.

 

If what you mean by realistic flowrate is my CO generation, then 2.12 L/s is this value (that would be scalar of 1 I assume), some of the other other ones I have split into multiple inlets and just divided the scalar by 500, hence the total CO generation would be 1060.

 

In regards to the mass flow, how are we to assign a flow rate out of the system that ensures balance? we have too many unknowns, the equation being as follows:

 

x(73 000) + y(2.12) = FLOWRATE OUT(x + y)

 

note that the volume of the entire space I have calculated as 54 400, so just roughly, the max concentration would be (2.425g/544000) or 44ppm.

 

I have noticed the first iteration of the scalar reads a 80 or so ppm reading. Could this be right as an initial value? 

 

In terms of mass flow, I thought we were going to assign the large flow inwards, and have pressure = 0 outlets only to ensure flow leaves the volume through these outlets.

 

When you say pulling air out, do you mean you assigned a flow rate leaving the system through the exhaust extrusions? Doesn't this mean that the mass balance isn't going to be balanced? Referring back to my equation with the unknowns.  If you did have a flow rate out, how did you make sure it balanced? Would it be better to assign outlets as having a leaving flow (say, the exhaust flow rate of 93 000), and have the inlets as the pressure value instead? This way it will balance the flow at the inlets to whatever it needs to be (less than exhaust), and more importantly, would allow a fixed amount of exhaust to leave the system.

 

Hope to hear from you guys soon, thanks.

 

Jack.

 

 

0 Likes
Message 33 of 53

Anonymous
Not applicable

Dear All,

 

I'd like to upload the following screenshot, 

 

Upon modelling a flow rate out of the model, and leaving inlets as P=0 so the model can balance itself, in the convergence plot window, it can be seen from the graph that there is a minimum value of scalar of 1.16e-2, this equates to 11 600 ppm, is this value showing me that the results have already exploded and something isn't right? as you can see using my scale the 50 ppm has already been exceeded.

 

This is why i went back to check a simple mass/volume for CO (equating roughly to the 44ppm as described above). getting this large value seems to be signifying something is wrong with the model, how is so much CO getting displaced when it's flow is so low? This happens regardless of whatever I try for my boundary conditions, very frustrating!

 

Having said this, assigning the flows out the top of the system and leaving the inlets as P = 0 is most likely my current best solution, despite it still inflating. Can I also verify - If i highlight 6 outlets and apply a 15 500 L/s, this is 15 500 L/s out of EACH inlet? not all the inlets combined, correct?

 

Regards,

 

Jack

0 Likes
Message 34 of 53

Anonymous
Not applicable

Atttached is my result for the P = 0 Inlets and a prescribed volumetric flow rate out of the model adding up to 93 000 L/s.

 

This model converged after said 446 iterations, as you can see the scalar variable for 50ppm is everywhere, increasing this to 200ppm shows red everywhere also. 

 

This is where I'm stuck: This 93 000 is for a car park double the size, it appears like none of the scalar is exhausted in the initial stages of the run, or just not enough is being exhausted fast enough. the value of 93 000 should be more than enough to dilute the mixture down to reasonable levels.

 

As I said earlier, the scalar variable shot to 400ppm instantly as a minimum value in the convergence window (I'm not sure how relevant this is, but it seems to be as the model continuously explodes).

 

I have tried every combination of boundary conditions I can think of. This is becoming very critical now as i have to model 6 total models for this study (2 halves of this current car park, times by 3 levels). Becoming very worried, need to know if this program is capable of what I am modelling. Every which way I have tried has failed considerably.  Attached is a CFZ of this current model.

 

Regards,

 

Jack.

0 Likes
Message 35 of 53

Anonymous
Not applicable

Dear All.

 

Attached is another half a days worth of modelling:

 

I have modeled in this one a negative pressure at the outlets, and P = positive pressure at the inlets, to try and drive the flow.

 

As can be seen, the software has exploded with values of CO again.

 

Please advise on what I should be trying next to try and model an exhaust system

 

Regards,

 

Jack

0 Likes
Message 36 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hi Jack,

Thank you very much for your reply. I understand than you have much concern about the scalar result. However, I would like to make sure all settings are correct before we dig in to it.

I want to make sure the followings first:

  1. Mass flowrate balance

The mass balance equation should be:

Supply air (73000) + CO (2.12) =Exhaust (93000) + Flow from P=0

Now we have more exhaust getting out from the model than the supply air, so there need air coming from the P=0 boundary.

I could not understand your equation. May I ask what do x and y stand for?

               x(73 000) + y(2.12) = FLOWRATE OUT(x + y)

  1. CO volume flowrate

If I understand correctly, the volume flowrate of CO is 2.12l/s. May I confirm how is CO discharged to the model? Is it pure CO or mixed with air?

  1. PPM calculation

You are calculating the max concentration by 2.425/544000. I would like to bring your attention that It is not the Max concentration but the average value.

As the scalar of CO is set to 1, the max value of the model should be scalar=1 at or near the inlet.

 

I look forward to your reply.

Thank you!

 



Xiaojin Zhang
0 Likes
Message 37 of 53

Anonymous
Not applicable

Hi Xiao,

 

I am not sure why neither of you guys can hear me on the phone, is there a contact number I can contact you guys on? 

 

"the mass balance equation should be:

Supply air (73000) + CO (2.12) =Exhaust (93000) + Flow from P=0"

 

I agree with this mass balance, the equation I wrote is just to show basically what you have shown, with x and y being the concentrations (please ignore this equation, It was merely as a description)

 

Have we concluded that having a flow in/and a flow out of the model does not work as an exhaust system for this program? It did not work for me.

 

You are correct with the 2.12 L/s of carbon monoxide, This is pure carbon monoxide. If you have an email I can send you a report that explains a lot of this and where I get my logic from. This is pure CO entering the model at the bottom 4 inlets that I have assigned it to for every model so far.

 

And yes I understand that it will be scalar 1 at the outlet, I only calculated the concentration in the volume after one second as a ball-park figure to see if what CFD is producing is plausible. 

 

Do you have an email I can send a report to? It won't let me attach it over the standard message function.

 

Regards,

 

Jack.

 

P.S I could also set my skype up if that would help you guys hear me, not sure at all why this isn't working. 

 

 

0 Likes
Message 38 of 53

Xiaojin.Zhang
Autodesk Support
Autodesk Support

Hi Jack,

Thank you for the clarification. I sent an E-mail to you just now. Did you receive it?

Do you mind using teamviwer for the conversation?

Below is the download link:

http://upandready.typepad.com/up_and_ready/2012/08/teamviewer-quick-support-to-the-rescue.html

 

As regards to the scalar result, as the volume flow rate of the CO is very small, we need very fine mesh and very tight convergence criteria to get an accurate result. I am now trying about this. I will let you know how my result looks liker later.

 

 

Thank you!

Xiaojin



Xiaojin Zhang
0 Likes
Message 39 of 53

Anonymous
Not applicable

Hi Xiao,

 

I recieved your email in regards to your missed call, I can use the team viewer but I will be leaving the office shortly. shall leave the office machine on with team viewer and also get the program for my home computer.

 

Okay i understand a fine mesh is needed, but I don't think that will make a difference to the value exploding? I have tried with a very fine mesh, the EXACT same thing happens, hence why i do not believe it has to do with the mesh or anything in the geometry. I could be wrong.

 

Shall hope to hear from you soon.

 

Regards,

 

Jack 

 

 

 

0 Likes
Message 40 of 53

Anonymous
Not applicable

Screenshot and latest CFZ file of runs with differing pressure at inlets/outlets. This was the way that this was modeled (in a different program) in an earlier stage of the project I'm doing

 

Results have exploded all over the model. The pressure differential was the way that this other model had been done.

 

This newest model I have taken one of the outlets away and made it a P = 0 for balancing, please note before this I just had all the outlets as the 66.7Pa.

 

 

Also attached is a screenshot of what has been deemed to be a solved system, the red spots are 100 000 ppm (people being poisoned) - this is one of the reasons why I think the models just explode the amount of CO, that guy got a solution similar to me, He's just played with the legend. I am not happy to fudge the results like this man has done because this system is meant to keep people safe and there is a liability involved. 

 

I am trying everything, please advise in writing if this program cannot model this exhaust system, I need to know this now as it is getting very late. Seeing from the other models on this forum I don't think this program can do what we want. This isn't a university project. It needs a result that can scale into real life and be relied upon. 

 

Regards,

 

Jack

 

0 Likes