that helps tremendously. (Doesn't matter for your purpose, but I think your value of .96 for % flow was wrong but I noticed you had it at .56 on entry). Your voice is exceptionally clear and pleasant - thanks!
Hi Shivani, This is an interesting tutorial. I am just wondering if this method can also be applied for getting the pressure drop vs velocity across a thin plastic woven mesh filter? For example, a 109 x109 mesh; Opening Size=0.0059" ;Open Area=41%;Wire Diameter=0.0034".The filter will be installed on the inlet and outlet enclosure vents in a natural airflow environment(no fans).
I used this method for evaluating the pressure drop through (0 < v < 6m/s) a sample of perf metal and my A=1945 and B= -683.4. That negative B coefficient shuts down the solver.
Hi Tau Design, I tested your porous media values (A = 1945, B= -683.4) along with a dummy porosity value of 0.5 in SOLIDWORKS 2017 SP2.0 and the solver begins to solve with no warnings given. We would recommend that, if you wish to diagnose this issue further, that you contact your local VAR so they can work with you and your file set directly to come to a resolution.
GoEngineer I'm running 2016 sp4. My local reseller couldn't get it to work with those values, either and is sending it up the chain to SW. BTW, porosity is around .375 for this- unidirectional, obviously.
Thanks for this amazing tutorial! This tool is quite useful when trying to reduce the calculation time. I have some questions about the process I hope you can help me with: 1.- Considering that the characteristic length is the length of the plate to be modeled. Shouldn't the pressure drop value that is used in the calculations be measured between both faces of the plate? (in the case of v=0.1 m/s: 101336-101318=18) 2.- Does the fact that you the measurement of the average pressure is done in the solid part of the plate (and not in the whole cross section of the tunnel) affect the pressure drop accuracy? Great video!
Yes, it seems like the technique would hold. At a glance the openings are large enough to maintain the Navier Stokes’ continuum condition, and the concern is memory requirements. Memory Requirements: SOLIDWORKS Flow Developers recommend 3-5 cells across a channel for accuracy. This can make the elements quite small and hit hardware limits. Workarounds would be a shorter and narrower wind tunnel. Navier Stokes’ Continuum condition: I will copy for you what SOLIDWORKS has to say on the topic. Essentially, we are trying to avoid microfluidics. If you have other examples, you can make sure they meet these requirements. The accuracy of the Navier-Stokes equations can be defined with the Knudsen number Kn=L/D (L- free path of gas molecules, D - typical task size) less than 0.001. A free path of gas molecules can be estimated with the following relation: L~1.225*Vis*k^0.5/(Ro*a). Here Vis - dynamic viscosity; k - adiabatic power; Ro - density; a - speed of sound. Kn < 0.001 Navier-Stokes with Slip Boundary condition (zero velocity at the wall). Continuum Flow regime. We can use Flow Simulation. 0.001< Kn < 0.1 Navier-Stokes OK but with slip conditions (non zero velocity at the wall). Slip Flow regime. 0.1< Kn< 10 Navier-Stokes model not applicable. Other models are applicable (like Burnett equations, example: hypersonic flows about space vehicles in low earth orbits or flows in microchannels of MEMS devices).
When chosing pressure drop vs flowrate it ask you to fill in area in Y, X och Z direction. Is that the cross-sectional area the air passes through in each direction? Second question: I get the same results with different porosity values, between 0-1, any idea why?
When filling in pressure drop vs flowrate, there is a dropdown where we’re choosing “Unidirectional”, “Isotropic”, or “Orthotropic”. Since you’re being asked to fill in X, Y, and Z, my guess is yours is set to Orthotropic. You should be able to flip that if it’s not what you desire. Porosity (when using pressure drop vs flowrate) does not effect the overall results. The only thing it affects is the velocity within the porous media itself. Thus Velocity cut plots through the center of the media would look slightly different, or surface goals set on the media itself. If velocity is an important output, we recommend making measurement devices offset from the media - that way an incorrect porosity value doesn’t affect our velocity outputs. There is a description of porosity in the Flow Simulation help. I’ll copy that here: Please note that the porosity value is only used for calculating the flow resistance when one of the dependencies on the pore size is selected as the Permeability type. With all other permeability types the porosity is only used as a scale factor during the results processing: the displayed velocity value is divided by porosity. Porosity is also used for calculation of thermal conductivity of the porous matrix.
Hi, thanks for a great video! Two points are not clear though: 1. While finding the first point to build the graph you used 101339 Pa as pressure, but th XY plot you are using later to get the Characteristic length shows 101336. 2. It is still not clear what to use as Charachteristic length: real design length, drop distance from XY plot of the pressure, or the distance where the pressure changes is I make a Pressure crossection along the tunnel (this may be the same as previous, I am still running a batch so can't be sure). Please comment
+Serge Barshtman Hi Serge,1. I am glad you caught this--this is a typo on my part. I should not have used the average pressure (101339) in the EXCEL table. The one from the XY plot (101336) was the right one to use. 2. This is another good question. Characteristic Length is usually defined as the length of the real part that you're approximating since we wouldn't generally know the distance of the pressure drop region. (A) The benefit of using real part length as characteristic length is you can test out different geometry changes on the porous media instead. (B) Also, we have to consider that the bigger the porous media block, the more streamlines can be affected. So if there was going to be flow coming in from the sides, a bigger block would force them to act in a way that they generally wouldn't. In the video, neither (A) or (B) are factors, so I can use drop distance to make the results for perforated plate vs porous media look more similar. I hope that clarifies my thought process.Regards,Shivani Patel
+Shivani Patel Thanks for your response. After simulating and characterizing a few different Porous Media it becomes clear that small typos are not changing the global result - I think you made a similar point while choosing the value in the clip. As for your answer to my second question - if I got it right, in cases where the simulated width is larger than the real one it still should be Ok if we later define the Porous Media with unidirectional settings, which should be usually the case. One other thing caught my attention: when same material is simulated to characterize the Porous Media, but the size of the sample is changing - A and B turn to be different. Since we are defining only a linear flow velocity and measuring the pressure change - the area of the sample should not change the result at all. I was simulating different in size wall perforation (same holes array), so in this case I kept the frame/contour around the holes, and I suspect that this could be the reason why I got a little different total open/close ratio leading to different results. Would using a fairly large specimen, instead of a small pipe cross section, be a better way to find the Porous Media resistance curve avoiding "designer's set of mind" mistakes?
+Serge Barshtman Serge, When you mentioned Area affecting A&B, I am guessing the reason is because Area appears to be affecting the pressure drop. My first thought is the boundary layer from the borders of the 'tunnel' are affecting the region. Running a larger test file is a great idea to fix this issue. Check out some videos on setting up symmetry in the computational domain if you go this route--that tool can help keep things running quickly. As for your interpretation for simulated length v real length, I can get behind that. I have seen someone use porous media to simplify a complex 'mesh' of sheet metal in the middle of swirling flow. In that case, making the porous media too long would have had a detrimental affect. Regards, Shivani
Hi there, I believe you’re referring to the technical reference pdf I pull up. One method to find that is to go to SOLIDWORKS Help button in the menu, go to SOLIDWORKS Simulation, then Flow Simulation Technical Reference. You have to have Flow Simulation added in to access this. You can also manually navigate to your installation folder, and get to it that way. An example of the default file path is: C:\Program Files\SOLIDWORKS Corp 2020\SOLIDWORKS Flow Simulation\lang\english\Docs.
Nice tutorial but I'm not getting results where the created porous media matches my perf coupon in my wind tunnel model. I'm getting a much higher pressure on the face of the porous media coupon than I am on the perf. Starting to piss me off. Also, my B term has a negative value. Question: Is the porosity the volume of the coupon with holes divided by the volume of the coupon without holes or is it the volume OF the holes divided by the volume without holes?
Hi Tau Design, Porosity is defined as the fluid volume divided by the total volume of the porous media. V = Vf + Vs Porosity = Vf/V Where V is the total volume of the porous media Vf is the fluid only volume Vs is the solid only volume
Hi there, No, testing for absorption in porous media is not possible in SOLIDWORKS Flow Simulation. However, you can assume perfect absorption on any particles that collide with that porous media. When running a particle study, create a specific wall condition for the porous media faces, and apply the condition absorption.
Hi all, I have few queries related to the video. Can anyone please answer them and guide me kn right path ....?? How come the grate shown serves as porous media ....?? Is it just because of the square holes present in the plate...? Are these the same holes causing whatever the pressure drop you calculated ...?? Can we have different sizes of plates for wind tunnel testing and actual simulation..?? Sorry if the questions sounds stupid and thanks for the time. Cheers... 🙂
While calculating pressure drop, one side you have taken value at front surface of grate, but for other side u have taken atmospheric pressure 101325 Pa, instead of value on rear surface which is 101318 Pa. Can you explain about this please ...??
that helps tremendously. (Doesn't matter for your purpose, but I think your value of .96 for % flow was wrong but I noticed you had it at .56 on entry). Your voice is exceptionally clear and pleasant - thanks!
great tutorial. how do you set the length of the wind tunnel?
The Wind tunnel length is part of the model so the length is set as part of the initial model.
Hi Shivani,
This is an interesting tutorial. I am just wondering if this method can also be applied for getting the pressure drop vs velocity across a thin plastic woven mesh filter? For example, a 109 x109 mesh; Opening Size=0.0059"
;Open Area=41%;Wire Diameter=0.0034".The filter will be installed on the inlet and outlet enclosure vents in a natural airflow environment(no fans).
I used this method for evaluating the pressure drop through (0 < v < 6m/s) a sample of perf metal and my A=1945 and B= -683.4. That negative B coefficient shuts down the solver.
Hi Tau Design,
I tested your porous media values (A = 1945, B= -683.4) along with a dummy porosity value of 0.5 in SOLIDWORKS 2017 SP2.0 and the solver begins to solve with no warnings given. We would recommend that, if you wish to diagnose this issue further, that you contact your local VAR so they can work with you and your file set directly to come to a resolution.
GoEngineer I'm running 2016 sp4. My local reseller couldn't get it to work with those values, either and is sending it up the chain to SW. BTW, porosity is around .375 for this- unidirectional, obviously.
Thanks for this amazing tutorial! This tool is quite useful when trying to reduce the calculation time. I have some questions about the process I hope you can help me with:
1.- Considering that the characteristic length is the length of the plate to be modeled. Shouldn't the pressure drop value that is used in the calculations be measured between both faces of the plate? (in the case of v=0.1 m/s: 101336-101318=18)
2.- Does the fact that you the measurement of the average pressure is done in the solid part of the plate (and not in the whole cross section of the tunnel) affect the pressure drop accuracy?
Great video!
Yes, it seems like the technique would hold. At a glance the openings are large enough to maintain the Navier Stokes’ continuum condition, and the concern is memory requirements.
Memory Requirements: SOLIDWORKS Flow Developers recommend 3-5 cells across a channel for accuracy. This can make the elements quite small and hit hardware limits. Workarounds would be a shorter and narrower wind tunnel.
Navier Stokes’ Continuum condition: I will copy for you what SOLIDWORKS has to say on the topic. Essentially, we are trying to avoid microfluidics. If you have other examples, you can make sure they meet these requirements.
The accuracy of the Navier-Stokes equations can be defined with the Knudsen number Kn=L/D (L- free path of gas molecules, D - typical task size) less than 0.001. A free path of gas molecules can be estimated with the following relation: L~1.225*Vis*k^0.5/(Ro*a). Here Vis - dynamic viscosity; k - adiabatic power; Ro - density; a - speed of sound.
Kn < 0.001 Navier-Stokes with Slip Boundary condition (zero velocity at the wall). Continuum Flow regime. We can use Flow Simulation.
0.001< Kn < 0.1 Navier-Stokes OK but with slip conditions (non zero velocity at the wall). Slip Flow regime.
0.1< Kn< 10 Navier-Stokes model not applicable. Other models are applicable (like Burnett equations, example: hypersonic flows about space
vehicles in low earth orbits or flows in microchannels of MEMS devices).
When chosing pressure drop vs flowrate it ask you to fill in area in Y, X och Z direction. Is that the cross-sectional area the air passes through in each direction?
Second question:
I get the same results with different porosity values, between 0-1, any idea why?
When filling in pressure drop vs flowrate, there is a dropdown where we’re choosing “Unidirectional”, “Isotropic”, or “Orthotropic”. Since you’re being asked to fill in X, Y, and Z, my guess is yours is set to Orthotropic. You should be able to flip that if it’s not what you desire.
Porosity (when using pressure drop vs flowrate) does not effect the overall results. The only thing it affects is the velocity within the porous media itself. Thus Velocity cut plots through the center of the media would look slightly different, or surface goals set on the media itself. If velocity is an important output, we recommend making measurement devices offset from the media - that way an incorrect porosity value doesn’t affect our velocity outputs.
There is a description of porosity in the Flow Simulation help. I’ll copy that here:
Please note that the porosity value is only used for calculating the flow resistance when one of the dependencies on the pore size is selected as the Permeability type.
With all other permeability types the porosity is only used as a scale factor during the results processing: the displayed velocity value is divided by porosity.
Porosity is also used for calculation of thermal conductivity of the porous matrix.
Check your numbers for the volumes used to calculate the porosity, V_total is too big by a factor of ten.
Hi, thanks for a great video!
Two points are not clear though:
1. While finding the first point to build the graph you used 101339 Pa as pressure, but th XY plot you are using later to get the Characteristic length shows 101336.
2. It is still not clear what to use as Charachteristic length: real design length, drop distance from XY plot of the pressure, or the distance where the pressure changes is I make a Pressure crossection along the tunnel (this may be the same as previous, I am still running a batch so can't be sure).
Please comment
+Serge Barshtman Hi Serge,1. I am glad you caught this--this is a typo on my part. I should not have used the average pressure (101339) in the EXCEL table. The one from the XY plot (101336) was the right one to use. 2. This is another good question. Characteristic Length is usually defined as the length of the real part that you're approximating since we wouldn't generally know the distance of the pressure drop region. (A) The benefit of using real part length as characteristic length is you can test out different geometry changes on the porous media instead. (B) Also, we have to consider that the bigger the porous media block, the more streamlines can be affected. So if there was going to be flow coming in from the sides, a bigger block would force them to act in a way that they generally wouldn't. In the video, neither (A) or (B) are factors, so I can use drop distance to make the results for perforated plate vs porous media look more similar. I hope that clarifies my thought process.Regards,Shivani Patel
+Shivani Patel Thanks for your response. After simulating and characterizing a few different Porous Media it becomes clear that small typos are not changing the global result - I think you made a similar point while choosing the value in the clip.
As for your answer to my second question - if I got it right, in cases where the simulated width is larger than the real one it still should be Ok if we later define the Porous Media with unidirectional settings, which should be usually the case.
One other thing caught my attention: when same material is simulated to characterize the Porous Media, but the size of the sample is changing - A and B turn to be different. Since we are defining only a linear flow velocity and measuring the pressure change - the area of the sample should not change the result at all.
I was simulating different in size wall perforation (same holes array), so in this case I kept the frame/contour around the holes, and I suspect that this could be the reason why I got a little different total open/close ratio leading to different results.
Would using a fairly large specimen, instead of a small pipe cross section, be a better way to find the Porous Media resistance curve avoiding "designer's set of mind" mistakes?
+Serge Barshtman Serge,
When you mentioned Area affecting A&B, I am guessing the reason is because Area appears to be affecting the pressure drop. My first thought is the boundary layer from the borders of the 'tunnel' are affecting the region.
Running a larger test file is a great idea to fix this issue. Check out some videos on setting up symmetry in the computational domain if you go this route--that tool can help keep things running quickly.
As for your interpretation for simulated length v real length, I can get behind that. I have seen someone use porous media to simplify a complex 'mesh' of sheet metal in the middle of swirling flow. In that case, making the porous media too long would have had a detrimental affect.
Regards,
Shivani
Hi Shivani i have a cuestion, where can i find the
explanation of the first two equations F=Kpv and k=GP/L?
Hi there,
I believe you’re referring to the technical reference pdf I pull up. One method to find that is to go to SOLIDWORKS Help button in the menu, go to SOLIDWORKS Simulation, then Flow Simulation Technical Reference. You have to have Flow Simulation added in to access this. You can also manually navigate to your installation folder, and get to it that way. An example of the default file path is: C:\Program Files\SOLIDWORKS Corp 2020\SOLIDWORKS Flow Simulation\lang\english\Docs.
Nice tutorial but I'm not getting results where the created porous media matches my perf coupon in my wind tunnel model. I'm getting a much higher pressure on the face of the porous media coupon than I am on the perf. Starting to piss me off. Also, my B term has a negative value. Question: Is the porosity the volume of the coupon with holes divided by the volume of the coupon without holes or is it the volume OF the holes divided by the volume without holes?
Hi Tau Design,
Porosity is defined as the fluid volume divided by the total volume of the porous media.
V = Vf + Vs
Porosity = Vf/V
Where
V is the total volume of the porous media
Vf is the fluid only volume
Vs is the solid only volume
does flowsim is capable to solve adsorption phenomenon in porous media?
Hi there,
No, testing for absorption in porous media is not possible in SOLIDWORKS Flow Simulation.
However, you can assume perfect absorption on any particles that collide with that porous media. When running a particle study, create a specific wall condition for the porous media faces, and apply the condition absorption.
Hi all,
I have few queries related to the video. Can anyone please answer them and guide me kn right path ....??
How come the grate shown serves as porous media ....?? Is it just because of the square holes present in the plate...?
Are these the same holes causing whatever the pressure drop you calculated ...??
Can we have different sizes of plates for wind tunnel testing and actual simulation..??
Sorry if the questions sounds stupid and thanks for the time.
Cheers... 🙂
While calculating pressure drop, one side you have taken value at front surface of grate, but for other side u have taken atmospheric pressure 101325 Pa, instead of value on rear surface which is 101318 Pa.
Can you explain about this please ...??
Well that was a rather impressive tutorial. Now I just need to spend the money on the Flow simulation package , ouch.