Free surface is below than expected..results are not good

mechovator · November 3, 2009, 01:13

Hello dears
Hope you will be having a good day. I am trying to simulate flow through reservoir of a dam. Boundary conditions were inlet, free surface (as given in bump tutorial) and outlet is spillways (which is at the top most of the dam). I did a steady state analysis but the level of the water (free surface)remained well below the spillways in the results; but physically some water should be going out of the spillways to conserve mass coming in from inlet. So physically the results are wrong. I think that problem is with in the expressions defining such as UpPres etc. I can not exactly figure out what DownH, UpH, DownPres and UpPres exactly mean in the tutorial. Height of free surface should be around 135m. I gave DownH=91.44m and UpH=51.86m. and free surface was somewhere around 55 m. Please suggest what I should do to bring it to the required level.
Regards.

Rui · November 3, 2009, 05:26

Dear,

What you should do?
You should try to be clear in the description of your problem.
DownH=91m, UpH=52m ????!!!! What's this???? And you expect a free surface at 135m????!!!!???

In the tutorial, DownH is the downstream free surface height (the free surface height at the outlet boundary); UpH is the upstream free surface height (the free surface height at the inlet boundary); DownPress and UpPress are the pressures on these boundaries (they are calculated using the hydrostatic law : rho*g*h).

mechovator · November 3, 2009, 05:31

oohhhh...okkkk... Thank you very much. You cleared my mind. In tutorial it was not written that what they meant. I will run a simulation by taking these values and am hoping for the best. Thanks again.

mechovator · November 7, 2009, 05:48

Hello thanks RUI. I ran various solutions and got good results. But my question to the forum is that if we are locating the free surface ourselves then what is the role of CFX into it?? Should it not calculate the height of the free surface itself? Our height is an initial guess but still it effects the solution so much. Whatever value should we give it should converge to the exact free surface height? Another question to clear my self is . In the bump free surface tutorial if water is falling out of a weir (in free atmosphere) then we should give 0pa of relative pressure instead of DownPrees. Is it correct?
Regards

ghorrocks · November 8, 2009, 04:22

I suspect you will need to move the downstream boundary further downstream.

AliTr · November 9, 2009, 02:48

Quote:

Originally Posted by mechovator

But my question to the forum is that if we are locating the free surface ourselves then what is the role of CFX into it?? Should it not calculate the height of the free surface itself?

a dam and a weir problem, can be solved in two ways, either you know the flow rate and looking for the water level U/S (upstream) of the weir or you know the water level and you are after the flow rate.

for the first case, split the U/S plane horizontally and create an Inlet boundary opening at your U/S plane, lower than weir crest. use a mass flow boundary (equal to flow rate times density) and solve the model, this should give you the correct water level for corresponding Flow rate.

for the second case, again split the U/S plane horizontally and create a surface as high as your water level and make it an Opening boundary with Mass and Momentum option set to "Opening Pres. and Dirn".
for relative pressure, use a CEL function defining the hydrostatic pressure at the boundary.

fab · November 11, 2009, 08:18

hi guys,i have the seem problem of the surface tension!
i have to localise the surface between water and air during the flow,my question is:how can cfx do this?i have the boundarys like described in the tutorial but i still have unaccurate results!!
what i have to do ?
thanks a lot

mechovator · November 11, 2009, 11:51

Quote:

Originally Posted by ghorrocks

I suspect you will need to move the downstream boundary further downstream.

Sorry . Did not understand what you mean by this? If I have a pressure bc at the weir how can I further move it? It is a spillway there. Plz guide me

mechovator · November 11, 2009, 11:58

Quote:

Originally Posted by AliTr

a dam and a weir problem, can be solved in two ways, either you know the flow rate and looking for the water level U/S (upstream) of the weir or you know the water level and you are after the flow rate.

for the first case, split the U/S plane horizontally and create an Inlet boundary opening at your U/S plane, lower than weir crest. use a mass flow boundary (equal to flow rate times density) and solve the model, this should give you the correct water level for corresponding Flow rate.

for the second case, again split the U/S plane horizontally and create a surface as high as your water level and make it an Opening boundary with Mass and Momentum option set to "Opening Pres. and Dirn".
for relative pressure, use a CEL function defining the hydrostatic pressure at the boundary.

Thanks for replying but; I have two questions then 1. why define as opening as it is an outlet and water can not come back from the weir after falling down? 2. why hydrostatic bc when it is going in the free atmosphere?

mechovator · November 11, 2009, 12:00

Quote:

Originally Posted by fab

hi guys,i have the seem problem of the surface tension!
i have to localise the surface between water and air during the flow,my question is:how can cfx do this?i have the boundarys like described in the tutorial but i still have unaccurate results!!
what i have to do ?
thanks a lot

Plz tell us what results are you taking about in detail? I may be able to help you.

fab · November 11, 2009, 16:17

hi,i have declared the surface tension as an equation in dependance of the temperatur at cfx-pre.all others boundary conditions are like the tutorial but with my values! and the solver is starting,it means cfx-pre is correct,right??!
the problem is: i have to get the surface tension-curve after running to know how my surface tension is changing with different temperaturs because i have 2 temperature on the bottom.how did cfx know how my cells are fulling(water /air)and did cfx using my equation of the surface tension?
thank u very much

AliTr · November 11, 2009, 19:25

Quote:

Originally Posted by mechovator

Thanks for replying but; I have two questions then 1. why define as opening as it is an outlet and water can not come back from the weir after falling down? 2. why hydrostatic bc when it is going in the free atmosphere?

defining an openning for Inlet, will keep the water level the same while you are seeking the weir capacity (flow rate) for the same water level.

hydrostatic BC is for inlet, at the Outlet use a 0 Pa pressure BC (just Air) if water falls out of the model.

AliTr · November 11, 2009, 19:33

Quote:

Originally Posted by fab

hi,i have declared the surface tension as an equation in dependance of the temperatur at cfx-pre.

how did cfx know how my cells are fulling(water /air)and did cfx using my equation of the surface tension?

Would please share the equation you used for surface tension and also describe the place you applied it.

if you want to see the surface tension curve in CFX-post, you should create a point on water surface (IsoSurface made out of water Volume Fraction of 0.5) and extract the related variavle on that point along the time.

ghorrocks · November 11, 2009, 20:04

Are you sure surface tension is significant? Modelling surface tension is difficult and reduces the stability of the simulation so only model it if it is required.

fab · November 12, 2009, 04:29

my tension equation is:sigma(t)=-0,949*t+15,858
i have declared it as an expression,is it wrong?!
how shoul i declaring this point on the surface to get tension curve at cfx-post?
thank you

fab · November 12, 2009, 04:31

and i have an constant surface tension coeffizient!! i don´t know if it true ! anb do i have heat transfer to?i think so wright?!
it´s my master thesis,yes i have to modelling the surface tension

ghorrocks · November 12, 2009, 06:24

You have put your posting on a thread discussing dam break flows and surface tension is definitely not important there. Maybe in future you can start a new thread when you start a new conversation.

From what I see in your expression you have made surface tension a function of time, and a strange one at that. Two things look strange - Your expression will end up with negative surface tension eventually and - at t=0 sigma=15.858. I don't know what unit system you are working in but this looks strange to me.

Alternately if t is temperature then your expression will always return a negative surface tension as t is in Kelvin by default.

Are you sure your surface tension equation is correct? What fluid are you modelling anyway?

fab · November 13, 2009, 02:58

hi,my expression is depending on temperatur in degre celsius not time!
i´m modelling first of all water but later i have to modelling ch3oh,and the endexpression should be :
sigma(temperatur,concentration)=a(T).exp(0,72*(x(m eoh)+b(T))-c(T)*x(meoh)).
a(T)=linear
b(T)=linear
c(T)=linear
the the values of the expression are experimentally prouved.
do u understand my problem now?
thank u lot
and sorry for my broken english:-)

ghorrocks · November 13, 2009, 04:33

What is the length scale of the simulation? Either a Re number, Weber number or some other appropriate scaling number.

Surface tension is not fully linearised in the solver so is in effect solved explicitly (to an extent anyway). CFX is an implicit solver so that means as soon as you turn on surface tension the time step required for stability will reduce enormously. Are you sure your time step is small enough.

For the ink jet printer simulations I am currently doing I start the simulation with timesteps of 1e-11s, and allow it to grow to 1e-9s as the simulation progresses. Goes to show I mean REALLY small timesteps.

And these instabilities are magnified because your surface tension is not constant (and not even linear).

fab · November 15, 2009, 08:30

Hallo,my timestep is automaticly setting,my length scale is 50 mm!
do i have a heat transfer to?well the problem is : i have to get the surfacetension-curve as funktion of the temperatur first and than of the concentration.!should i set the surfacetension coefficient as constant?and how did the solver know where my surface exactly is?

November 3, 2009, 01:13	Free surface is below than expected..results are not good	#1
mechovator Member mechovator Join Date: May 2009 Posts: 32 Rep Power: 16	Hello dears Hope you will be having a good day. I am trying to simulate flow through reservoir of a dam. Boundary conditions were inlet, free surface (as given in bump tutorial) and outlet is spillways (which is at the top most of the dam). I did a steady state analysis but the level of the water (free surface)remained well below the spillways in the results; but physically some water should be going out of the spillways to conserve mass coming in from inlet. So physically the results are wrong. I think that problem is with in the expressions defining such as UpPres etc. I can not exactly figure out what DownH, UpH, DownPres and UpPres exactly mean in the tutorial. Height of free surface should be around 135m. I gave DownH=91.44m and UpH=51.86m. and free surface was somewhere around 55 m. Please suggest what I should do to bring it to the required level. Regards.

November 3, 2009, 05:26		#2
Rui Member Rui Igreja Join Date: Mar 2009 Location: Aveiro Posts: 68 Blog Entries: 1 Rep Power: 17	Dear, What you should do? You should try to be clear in the description of your problem. DownH=91m, UpH=52m ????!!!! What's this???? And you expect a free surface at 135m????!!!!??? In the tutorial, DownH is the downstream free surface height (the free surface height at the outlet boundary); UpH is the upstream free surface height (the free surface height at the inlet boundary); DownPress and UpPress are the pressures on these boundaries (they are calculated using the hydrostatic law : rhogh). __________________ Rui

November 7, 2009, 05:48	Thanks and further confusion	#4
mechovator Member mechovator Join Date: May 2009 Posts: 32 Rep Power: 16	Hello thanks RUI. I ran various solutions and got good results. But my question to the forum is that if we are locating the free surface ourselves then what is the role of CFX into it?? Should it not calculate the height of the free surface itself? Our height is an initial guess but still it effects the solution so much. Whatever value should we give it should converge to the exact free surface height? Another question to clear my self is . In the bump free surface tutorial if water is falling out of a weir (in free atmosphere) then we should give 0pa of relative pressure instead of DownPrees. Is it correct? Regards

November 11, 2009, 16:17	free surface problem	#11
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	hi,i have declared the surface tension as an equation in dependance of the temperatur at cfx-pre.all others boundary conditions are like the tutorial but with my values! and the solver is starting,it means cfx-pre is correct,right??! the problem is: i have to get the surface tension-curve after running to know how my surface tension is changing with different temperaturs because i have 2 temperature on the bottom.how did cfx know how my cells are fulling(water /air)and did cfx using my equation of the surface tension? thank u very much

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bubblze induced turbulence at a free surface ????	thomas	Main CFD Forum	6	April 21, 2012 09:48
Free surface height v/s time	ashish	CFX	2	October 4, 2005 14:13
free surface display	novice	Siemens	5	August 4, 2004 01:07
Free Surface Flows / Ocean Waves / Far-field boundary conditions	Sundar Prasad	Main CFD Forum	5	May 4, 1999 10:25
Code with reliable free surface capability in a rotating system	Subhasish Roy Choudhury	Main CFD Forum	1	September 4, 1998 18:47

November 3, 2009, 05:31		#3
mechovator Member mechovator Join Date: May 2009 Posts: 32 Rep Power: 16	oohhhh...okkkk... Thank you very much. You cleared my mind. In tutorial it was not written that what they meant. I will run a simulation by taking these values and am hoping for the best. Thanks again.

November 8, 2009, 04:22		#5
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	I suspect you will need to move the downstream boundary further downstream.

November 11, 2009, 08:18		#7
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	hi guys,i have the seem problem of the surface tension! i have to localise the surface between water and air during the flow,my question is:how can cfx do this?i have the boundarys like described in the tutorial but i still have unaccurate results!! what i have to do ? thanks a lot

November 11, 2009, 20:04		#14
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	Are you sure surface tension is significant? Modelling surface tension is difficult and reduces the stability of the simulation so only model it if it is required.

November 12, 2009, 04:29		#15
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	my tension equation is:sigma(t)=-0,949*t+15,858 i have declared it as an expression,is it wrong?! how shoul i declaring this point on the surface to get tension curve at cfx-post? thank you

November 12, 2009, 04:31		#16
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	and i have an constant surface tension coeffizient!! i don´t know if it true ! anb do i have heat transfer to?i think so wright?! it´s my master thesis,yes i have to modelling the surface tension

November 12, 2009, 06:24		#17
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	You have put your posting on a thread discussing dam break flows and surface tension is definitely not important there. Maybe in future you can start a new thread when you start a new conversation. From what I see in your expression you have made surface tension a function of time, and a strange one at that. Two things look strange - Your expression will end up with negative surface tension eventually and - at t=0 sigma=15.858. I don't know what unit system you are working in but this looks strange to me. Alternately if t is temperature then your expression will always return a negative surface tension as t is in Kelvin by default. Are you sure your surface tension equation is correct? What fluid are you modelling anyway?

November 13, 2009, 02:58		#18
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	hi,my expression is depending on temperatur in degre celsius not time! i´m modelling first of all water but later i have to modelling ch3oh,and the endexpression should be : sigma(temperatur,concentration)=a(T).exp(0,72(x(m eoh)+b(T))-c(T)x(meoh)). a(T)=linear b(T)=linear c(T)=linear the the values of the expression are experimentally prouved. do u understand my problem now? thank u lot and sorry for my broken english:-)

November 13, 2009, 04:33		#19
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,703 Rep Power: 143	What is the length scale of the simulation? Either a Re number, Weber number or some other appropriate scaling number. Surface tension is not fully linearised in the solver so is in effect solved explicitly (to an extent anyway). CFX is an implicit solver so that means as soon as you turn on surface tension the time step required for stability will reduce enormously. Are you sure your time step is small enough. For the ink jet printer simulations I am currently doing I start the simulation with timesteps of 1e-11s, and allow it to grow to 1e-9s as the simulation progresses. Goes to show I mean REALLY small timesteps. And these instabilities are magnified because your surface tension is not constant (and not even linear).

November 15, 2009, 08:30		#20
fab New Member Join Date: Jun 2009 Posts: 13 Rep Power: 16	Hallo,my timestep is automaticly setting,my length scale is 50 mm! do i have a heat transfer to?well the problem is : i have to get the surfacetension-curve as funktion of the temperatur first and than of the concentration.!should i set the surfacetension coefficient as constant?and how did the solver know where my surface exactly is?