Does gas flow from low to high pressure?

In summary, the video demonstrates the Bernoulli's Principle through the use of a Venturi Tube. The pressure in the third manometer appears to be higher than in the middle one, but this is due to the air having higher velocity laterally in the constricted section. The manometers are measuring the static pressure and not the dynamic pressure of the moving air. The pressure in the middle chamber is lower than the other two, and the fluid velocity is higher in the restriction. This results in a pressure drop across the constriction and the gas flowing from low to high pressure. However, there is some ambiguity in the numbering of the chambers and the direction of flow, leading to confusion about the pressure readings.
  • #106
Sailor Al said:
I have specified the temperature in chamber A is 0°C 273 °K.
The pressure can be determined using hydrostatics from the heights of the water columns provided i the problem statement in post #65
Whatever you say…It’s apparent I’ve already wasted my time. Good day sir.
 
Physics news on Phys.org
  • #107
erobz said:
Whatever you say…It’s apparent I’ve already wasted my time. Good day sir.
You asked for the pressure and temperature in chamber A.
The pressure, from hydrostatics is 1017.8 x 103 dynes/cm2
The temp is 0°C or 273°K
 
  • #108
Sailor Al said:
I see a huge amount of working, but I still don't see an answer to the stated problem: to determine the mass flow rate through the system.
This is a really disheartening thing to say, especially because in his very long post #100 @erobz gives in his equation 10 literally his answer to your question. It literally gives you his equation of the mass flow rate, he's only asking you to do a bit of effort and plug in the numbers yourself... Don't you see that?
 
  • Like
Likes russ_watters
  • #109
Arjan82 said:
This is a really disheartening thing to say, especially because in his very long post #100 @erobz gives in his equation 10 literally his answer to your question. It literally gives you his equation of the mass flow rate, he's only asking you to do a bit of effort and plug in the numbers yourself... Don't you see that?
They are actually saying I didn't go far enough with that "huge working". I think it's their presentation of "you're wrong and I'm right, but I can't tell you why" is what strikes a nerve with everyone they seem to interact with...

That being said, In this case (after getting some sleep) It dawned on me that we have another equation that starts the whole process over again. This time it's the cliff notes version.

Assumption:
The gas is undergoing an adiabatic expansion from A to C ( ? I have less confidence that there is no heat transfer with the surroundings between A and C as I do in gong from A to B )

Hydrostatics:
$$ P_C = P_A + \rho_w g ( l_A - l_C ) $$

First Law:
$$ 0 = h_A - h_C + \frac{V^2_A- V^2_C}{2}$$

$$ \implies V_C = \sqrt{2 \left( h_A(T_A) - h_C(T_C) \right) + V^2_A} $$

Adiabatic Expansion:

$$ T_C = T_A \left( \frac{P_C}{P_A} \right)^{ \frac{1-\gamma}{\gamma}} = T_A \left( \frac{ P_A + \rho_w g ( l_A - l_C ) }{P_A} \right)^{ \frac{1-\gamma}{\gamma}} $$

Conservation of Mass (Steady State):

$$\dot m = \rho_C A V_C = \rho_A A V_A $$

$$ \implies \rho_C V_C = \rho_A V_A $$

Sub all that in along with the Ideal Gas Law:

$$ \frac{P_A + \rho_w g ( l_A - l_C )}{R T_A \left( \frac{ P_A + \rho_w g ( l_A - l_C ) }{P_A} \right)^{ \frac{1-\gamma}{\gamma}} } \sqrt{2 \left( h_A(T_A) - h_C(T_C) \right) + V^2_A} = \frac{P_A}{RT_A} V_A $$

Where:

##T_C = T_A \left( \frac{P_C}{P_A} \right)^{ \frac{1-\gamma}{\gamma}} = T_A \left( \frac{ P_A + \rho_w g ( l_A - l_C ) }{P_A} \right)^{ \frac{1-\gamma}{\gamma}}##

##T_B = T_A \left( \frac{P_B}{P_A} \right)^{ \frac{1-\gamma}{\gamma}} = T_A \left( \frac{ P_A + \rho_w g ( l_A - l_B ) }{P_A} \right)^{ \frac{1-\gamma}{\gamma}}##

##V_A = \sqrt{ \frac{ 2 \left( h_B(T_B) - h_A(T_A) \right) }{ 1 - \left( \frac{P_A}{P_B} \right)^{ \frac{2}{\gamma}} \left( \frac{A}{a} \right)^2 } }##

## P_B = P_A + \rho_w g ( l_A - l_B) ##

## h_B(T_B) = \int_{0}^{T_B} c_p(T)dT ##

## h_C(T_C) = \int_{0}^{T_C} c_p(T)dT ##

"In theory" "you might be able to determine ##P_A##, and then ##\dot m##, but ##P_A## is not only buried in a sea of symbols, but also some of those symbols ##h_B(T_B), h_C(T_C)## it finds itself buried in a non-linear function, which is then buried in a limit of an integral...It's a real pickle of an equation.

At this point I fully expect @Sailor Al to say , "but I still don't see an answer to my question", To which I say :partytime:
 
Last edited:
  • #110
Sailor Al said:
You asked for the pressure and temperature in chamber A.
The pressure, from hydrostatics is 1017.8 x 103 dynes/cm2
The temp is 0°C or 273°K
Just out of curiosity...Are you saying you have solved for the pressure independently and this is what you find from your workings? Or are you saying, "just say the pressure is X"?

After further consideration I believe you are correct that for a particular state of the three manometers taken as a system of measurements (under adiabatic assumption) the initial state in chamber A will be fixed given the temperature of the gas in chamber A. I "imagined" it's the case that if we change the initial state, the state of all three manometers would change accordingly. In other words, I think there is a one-to-one correspondence between the state of the manometers and the initial state of the flow.

So...Short of solving the equation developed in #109 above, we should be able to specify a pressure ##P_A## and use it to determine the manometer C reading iteratively adjusting ##P_A## making sure ##P_C + \rho_w g l_C =P_A + \rho_w g l_A## is satisfied.

Taking the kids fishing! Maybe tomorrow.
 
  • Like
Likes russ_watters
  • #111
erobz said:
Just out of curiosity...Are you saying you have solved for the pressure independently and this is what you find from your workings? Or are you saying, "just say the pressure is X"?
I believe the pressure solution is straightforward hydrostatics.
Here's my working:
With the air tap open, the water levels are 0, 9.5 and 5.3 cm above datum. [1]
When the air tap is closed and the system in equilibrium, the chambers are filled with air at, shall we say NTP (293°K, 1033 cm H2O) and so , since there's a fixed amount of water in the tubes, the tubes will have the same water level of 4.933 cm (average of [1]) . Let's round that it to 4.9
With the tap closed, the pressure at the datum level in the water will be the sum of the Atmospheric pressure and the hydrostatic pressure : P = 1033 + 4.9 = 1037.9 cm of H2O
With the tap open then, in cm of H2O, the pressures at the datum level will be:
PA = 1033 + 4.9 = 1037.9 (no water, so pressure due entirely to air pressure)
PB = 1033 - 4.6 = 1028.4
PC = 1033 - 0.4 = 1033.0
And 1037.9 cm of H2O converts to 1017.8 dynes/cm2
I was deeply worried by the apparent simplicity of the solution, but it seems ok to me.
A couple of thoughts.
1.) If, instead of having two wide chambers separated by a narrowed one, we had 10, 100 or ##\infty## chambers and the appropriate quantity of water to get the level at 4.9 cm in all of them, when we turned on the air tap, would we see the pattern of 0, 9.5, 0, 9.5, 0, 9.5 ....repeated indefinitely? I think so.
2) If, in this thought experiment, we then turned on the air tap, how long would it take for the the system to settle down? We'd probably have to decide the length of each section and the length of each nozzle/diffuser. I think the answer would be very similar to the answer to the piston in the infinite tube.
[ADD] 3) And what difference would it make if the system terminated with a wide chamber or a narrow chamber? [/ADD]
But let's press on with this one.
I'm deeply mired in attempting to balance mass flow rate - which I now identify as flux ##\phi##, flux density B, and linear density ##\mu##. I'm working strictly in CGS units as the pressure and Ideal Gas constants get too confusing. I am re-phrasing PM = dRT, to P = cT, where c = R/M or 2.88 106 cm2/sec2 for Air with a Molar mass of 28.95

Since this thread has passed the 100 post mark, should I request it be closed and start afresh?
 
Last edited:
  • #112
Sailor Al said:
Since this thread has passed the 100 post mark, should I request it be closed and start afresh?
No. A new thread would imply a new question/topic, but that is not the case here. We have no hard limit on how many posts a thread can have. The question is (a) whether you will ever be satisfied with any answer anyone else gives you, (b) whether you will find an answer that satisfies you on your own, and (c) whether, if you do find an answer on your own, it will actually be correct.

Based on the thread so far, I'm very pessimistic about (a) since you have already been given valid answers and have not accepted them. I am not even going to try and speculate about (b) and (c).
 
  • #113
PeterDonis said:
No. A new thread would imply a new question/topic, but that is not the case here. We have no hard limit on how many posts a thread can have. The question is (a) whether you will ever be satisfied with any answer anyone else gives you, (b) whether you will find an answer that satisfies you on your own, and (c) whether, if you do find an answer on your own, it will actually be correct.

Based on the thread so far, I'm very pessimistic about (a) since you have already been given valid answers and have not accepted them. I am not even going to try and speculate about (b) and (c).
OK, I just suspect the length of the thread might discourage the involvement of new members.
And, if my may, with respect point out that the answers so far have shown workings, but no solutions. The problem posed was to derive the mass flow rate in Kg/sec from the presented experiment.
Please be assured that I will be satisfied with (a). I have not yet found (b) and (c) if I do, I will certainly subject it to the forum for scrutiny to determine if it is correct.
 
  • #114
Sailor Al said:
I just suspect the length of the thread might discourage the involvement of new members.
That might well be the case. You should perhaps consider that as a sign that you have gone astray, and that rather than continuing to post lengthy calculations that no one else will read, you should look at the valid answers you have already been given and see whether they might not be sufficient, so that the thread would not need to be prolonged further.

Sailor Al said:
with respect point out that the answers so far have shown workings, but no solutions
You're quibbling. You have already been pointed at explicit formulas for the quantities of interest. If you are unable or unwilling to do the work yourself of plugging numbers (all of which you already have) into those formulas to get numerical answers, then this thread will indeed be closed forthwith. It is not any other poster's job to hold your hand and do all your work for you. We expect you to be able to take valid equations and work with them yourself to get a final numerical answer.

Sailor Al said:
The problem posed was to derive the mass flow rate in Kg/sec from the presented experiment.
And you have been given an explicit equation for the mass flow rate. Use it.
 
  • Like
Likes russ_watters and boneh3ad
  • #115
Sailor Al said:
Please be assured that I will be satisfied with (a).
You can say that all you want, but your actual behavior in this thread has been to ignore valid answers because the people who gave them too you assumed you had enough intelligence and common sense to know how to plug numbers into formulas for yourself. Is that not the case?
 
  • Like
Likes russ_watters
  • #116
PeterDonis said:
No. A new thread would imply a new question/topic, but that is not the case here.
I realise that at post #65, I did indeed pose a new question. The original question was how the gas appeared to be flowing from low to high pressure despite the (now clearly incorrect) assumption that, like water flowing uphill, gas would only flow from high to low. That was comprehensively and satisfactorily resolved by pointing out that my assumption was wrong. Water does flow uphill in a syphon or a U-bend. Gas does flow from low to high pressure in a venturi.
The question I then posed in #65 was to work out the actual mass flow rate in the venturi experiment presented in the video. That was a different question and maybe it is appropriate to present it in a new thread.
 
  • #117
Sailor Al said:
I believe the pressure solution is straightforward hydrostatics.
Here's my working:
With the air tap open, the water levels are 0, 9.5 and 5.3 cm above datum. [1]
When the air tap is closed and the system in equilibrium, the chambers are filled with air at, shall we say NTP (293°K, 1033 cm H2O) and so , since there's a fixed amount of water in the tubes, the tubes will have the same water level of 4.933 cm (average of [1]) . Let's round that it to 4.9
With the tap closed, the pressure at the datum level in the water will be the sum of the Atmospheric pressure and the hydrostatic pressure. In cm of H2O
P = 1033 4.9 cm
With the tap open then, in cm of H2O, the pressures at the datum level will be:
PA = 1033 + 4.9 = 1037.9
PB = 1033 - 4.6 = 1028.4
PC = 1033 - 0.4 = 1033.0
And 1037.9 cm of H2O converts to 1017.8 dynes/cm2
I was deeply worried by the apparent simplicity of the solution, but it seems ok to me.
A couple of thoughts.
No,No, and No.

You are trying to tell me that lab pressure is 0.06 psi gage... and there is a flow in this apparatus that causes 100% thermal loss of pressure in that short distance! There is about 4 cm H2O difference in pressure between chambers A and C ~0.06 psi. I'm sorry, this is just absurd. The beginning of the outlet would be vacuum...

What do you think happens if lab pressure is 30 psi gauge(which it is far more likely to be), what would the manometers read? Imagine this setup, but with a valve at the outlet initially closed. The initial equilibrium pressure is 30 psig. All three initial columns of water are (for all intents and purposes) unchanged from the last scenario, but the initial state is 30 psig, not 0.05 psig. So how do you reconcile that? It's like you are looking at that system and saying, the columns are equal, must be atmospheric pressure inside those chambers. Once the system is open it doesn't matter where the valve is placed.

Once the system is open, you cannot say anything about the pressure inside the manometers until you've done the analysis I proposed.
 
Last edited:
  • #118
erobz said:
you are trying to tell me that lab pressure is 0.06 psi gage
0.06 psi is 4136.854 dynes/cm 2
My calculations give the over-pressure in chamber of 4802 dynes/cm which is 0.069 psi, so we're in the right ballpark.
[EDIT] there will be a pressure drop as the air enters chamber A from the supply.[/EDIT]
What's absurd about 4cm pressure difference between A and C?
I think, for the experiment, the gas pressure was seriously throttled to get the water level exactly to the base of tube A.
P.S. did you catch any fish?
 
Last edited:
  • #119
Sailor Al said:
The question I then posed in #65 was to work out the actual mass flow rate in the venturi experiment
Yes, and you have been given a formula for that. That question is not identical to the question in the OP, but it's closely enough related that there doesn't need to be a new thread on it (particularly since it's already been discussed in this thread).
 
  • #120
Sailor Al said:
0.06 psi is 4136.854 dynes/cm 2
My calculations give the over-pressure in chamber of 4802 dynes/cm which is 0.069 psi, so we're in the right ballpark.
What's absurd about 4cm pressure difference between A and C?
I think, for the experiment, the gas pressure was seriously throttled to get the water level exactly to the base of tube A.
P.S. did you catch any fish?
1685402512509.png


We have the same system with a valve on the outlet, what is the pressure ##P_{ref}## at the base of the manometers?
 
  • #121
I don't know, because I don't know the pressure in the system now. There's no gas flow, there's no pressure gradient.
 
  • #122
Sailor Al said:
With the air supply closed. The pressure in the system is Atm - say 1033 cm water. The pressure at ##P_{ref}## is 1033 +##l_{C}## = 1033 + ##l_{C}##
No, the valve is at the end of the glass tube. The entire system is pressurized to whatever lab pressure is. What is ##P_{ref}##?
 
  • #123
erobz said:
No, the valve is at the end of the glass tube. The entire system is pressurized to whatever lab pressure is. What is ##P_{ref}##?
Yes, sorry, I corrected my answer after you replied.
I should have said:
I don't know. There is no gas flow. There is no pressure gradient.
 
  • #124
1685403519525.png

There is a throttle on the air supply.
Pure speculation, but I suspect, for the video, he adjusted that to get the water level at tube A to zero. (I bet that took a bit of fiddling, and water squirting through the tubes which would explain the bead of water at the top of tube C!)
 
  • #125
Sailor Al said:
I don't know. There is no gas flow. There is no pressure gradient.
Exactly. It could be 100 psi, could be 5 psi, could be 0.01 psi. Whenever the valve is opened, would the pressure in the system drop to near atmospheric in chamber A regardless of the system pressure?
 
  • #126
Sailor Al said:
View attachment 327210
There is a throttle on the air supply.
Pure speculation, but I suspect, for the video, he adjusted that to get the water level at tube A to zero. (I bet that took a bit of fiddling, and water squirting through the tubes which would explain the bead of water at the top of tube C!)
Multiple Choice: From the picture, what's the effective area of the "pinched tube"? If the tube has area ##A##, is it
a) 80% ##A##,
b)50% ##A##,
c) 0%##A##
 
  • #127
@erobz I have responded to your thought experiment in #120.
Would you like to have a look at mine: 1), 2) and 3) in #111?
I think it may throw some light onto the problem.
 
  • #128
Sailor Al said:
@erobz I have responded to your thought experiment in #120.
Would you like to have a look at mine: 1), 2) and 3) in #111?
I think it may throw some light onto the problem.
No, I think we are about to make some progress here if you answer #126 and #125. Which is exactly why you are desperately attempting to deflect the topic.
 
  • #129
erobz said:
From the picture, what's the effective area of the "pinched tube"? If the tube has area ##A##, is it 80% ##A##, 50% ##A##, 0%##A##?
I think I specified that in the original question. I suggested the diameter of A was 1.6 cm and B, 0.5 cm, so the cross-section area of A is 2.01 cm2 and of B is 0.196 cm2. So B is ~10% of A.
 
  • #130
Sailor Al said:
I think I specified that in the original question. I suggested the diameter of A was 1.6 cm and B, 0.5 cm, so the cross-section area of A is 2.01 cm2 and of B is 0.196 cm2. So B is ~10% of A.
I'm talking about the throttle you are pointing out. If the feed tube has area ##A_{tube}##, what is the area of the "pinched" section.

Also, you have not answered #125.
 
Last edited:
  • #131
erobz said:
I'm talking about the throttle you are point out.
I'm sorry, there's no way of determining that. And even if we did know that, we still don't know the pressure behind that nozzle/diffuser. It's just adding another set of nozzles and diffusers to the problem.
It's a variation on my question 1) in #111.
I only pointed out the throttle to allay your concerns about the small size of the apparent pressures in the tube that prompted my question in #118: "What's absurd about 4cm pressure difference between A and C?"
 
  • #132
Sailor Al said:
I'm sorry, there's no way of determining that. And even if we did know that, we still don't know the pressure behind that nozzle/diffuser. It's just adding another set of nozzles and diffusers to the problem.
It's a variation on my question 1) in #111.
I only pointed out the throttle to allay your concerns about the small size of the apparent pressures in the tube that prompted my question in #118: "What's absurd about 4cm pressure difference between A and C?"
The absurdity is a tiny guage pressures we expect almost no flow, and consequently almost no losses ( as a percentage) between A and C. You are trying to tell me that the entirety 100% of the internal energy stored as pressure in A is converted to heat by the time it reaches chamber C under the action of the tiny guage pressure you propose.
 
  • #133
erobz said:
Exactly. It could be 100 psi, could be 5 psi, could be 0.01 psi. Whenever the valve is opened, would the pressure in the system drop to near atmospheric in chamber A regardless of the system pressure?
This is another thought experiment, and reaches to the heart of the problem.
Here's my thought experiment 4):
a) If the valve was opened abruptly, then, whatever the supply pressure, the build-up of pressurised air would blast the water out of the tubes!
b) If it was opened slowly, to keep the water being pushed out of the top of tube B, then eventually, when fully opened, the water levels would return to the 0, 9.5, 5.3 of the video.
So yes and no. The pressures would return to near atmospheric pressure like I indicated in #111:
PA = 1033 + 4.9 = 1037.9
PB = 1033 - 4.6 = 1028.4
PC = 1033 - 0.4 = 1033.0
There is a pressure gradient across the system: All around atmospheric pressure but all slightly different.
 
  • #134
erobz said:
The absurdity is a tiny guage pressures we expect almost no flow, and consequently almost no losses ( as a percentage) between A and C. You are trying to tell me that the entirety 100% of the internal energy stored as pressure in A is converted to heat by the time it reaches chamber C under the action of the tiny guage pressure you propose.
Can I postpone responding to this till you have had a chance to review my #133 and perhaps #97 where I talk about trading the components of the gas energy - pressure, density and temperature:
Sailor Al said:
I think the following statements are a bit wooly and can be tidied up, but I think they are important.
In a venturi, a low pressure occurs in the narrow section when the fluid is either liquid or gas.
With liquid, the pressure change is due to the "trading" of potential energy and kinetic energy by a change in momentum of the fluid. Since liquid is incompressible, i.e its density is constant, and so the mass per unit volume is constant, the momentum change arises from a change in the speed of the fluid.
With gas, the low pressure is due to the "trading" of the components of the gas energy, i.e. PV and nRT, and occurs through a change in density, pressure and temperature . This "trading" can be handled mathematically using the density version of the Ideal Gas Law : PM = dRT .
That's the theory I'm using to work up the answer.
 
  • #135
Here's what's being traded.
  • Compared to A, In B the pressure falls and the density and temperature rise.
  • Compared to B, In C the pressure rises and the density and temperature fall.
In each of the three chambers the energy in the gas: PV and/or nRT is constant.
The challenge is to use that in the density version of the Ideal Gas equation P = cdT where c: for air is 2.88 106cm2sec-2
The equation has two variables: density and temperature. We know the pressures.
The three cross-section areas are defined (2.01, 0.196, 2.01 cm2).
We can make reasonable assumptions about the ambient temperature and pressure (NTP?) and the temp of the air in A: 0°C
It should be soluble for the mass flow rate ##\phi## gm/sec, I haven't got here yet.
 
  • #136
Sailor Al said:
Here's what's being traded.
  • Compared to A, In B the pressure falls and the density and temperature rise.
  • Compared to B, In C the pressure rises and the density and temperature fall.
In each of the three chambers the energy in the gas: PV and/or nRT is constant.
The challenge is to use that in the density version of the Ideal Gas equation P = cdT where c: for air is 2.88 106cm2sec-2
The equation has two variables: density and temperature. We know the pressures.
The three cross-section areas are defined (2.01, 0.196, 2.01 cm2.
We can make reasonable assumptions about the ambient temperature and pressure (NTP?) and the temp of the air in A: 0°C
It should be soluble for the mass flow rate ##\phi##, I haven't got here yet.
For near ambient gauge pressures in chamber A I expect this:

1685408948106.png


Not This:

1685408448595.png
 
Last edited:
  • #137
erobz said:
For near ambient gauge pressures in chamber A we expect this:

View attachment 327211

Not This:

View attachment 327212
But that IS what we're getting!
lA < lB
lB >lC
lC >lA
It's just you are surprised just how large are the height differences, caused by such small pressure differences.
Remember Air pressure is 1033 cm
We're looking at 5 cm differences in the water level in the tubes . 5 cm is only 0.5% of atmospheric pressure.
It only takes a tiny amount of pressure to lift quite a lot of water.
 
  • #138
Sailor Al said:
But that IS what we're getting!
##l\{A}## < ##l\{B}##
##l\{B}## > ##l\{C}##
##l\{C}## > ##l\{A}##

It's just you are surprised just how large are the height differences, caused by such small pressure differences.
Remember Air pressure is 1033 cm
We're looking at 5 cm differences in the water level in the tubes . 5 cm is only 0.5% of atmospheric pressure.
No, my concern is not for absolute differences. Its for the relative loss between chamber A and C. Its an indication of high thermal generation. Thats not going to be the case in a near ambient chamber A.

Anyhow, tomorrow I'll solve my equation...somehow.
 
  • #139
erobz said:
No, my concern is not for absolute differences. Its for the relative loss between chamber A and C. Its an indication of high thermal generation. Thats not going to be the case in a near ambient chamber A.

Anyhow, tomorrow I'll solve my equation...somehow.
Sorry we're 17-odd hours apart (I'm UTC+10), it makes the conversation a bit fragmented.
I edited my last post which makes it more readable.
I please, can I get you to consider my thought experiments 1), 2) and 3) before returning to the aerodynamic solution. It's not the relative differences in play here, it's the absolute differences.
Very small pressure changes - very small temperature and density changes.
 
  • #140
Yeah, I have considered it.

Let me get things scaled write in the context of this experiment:

Small pressure chamber A implies:

1685410193694.png


However, In the experiment we have this:

1685409788193.png


This is indicative of significant relative thermal losses. You aren't going to see this without significant pressure gradient.
 

Similar threads

Back
Top