Will a Venturi flow tube in a filter deliver accurate flow under negative head conditions?

[operatorjoe]

I have noticed that the gravity flow filters in our water treatment plant have been allowed to run off the scale on loss-of-head gauges and I am wondering if a venturi is accurate if the flow is actually being pulled by vacuum (or very little head)rather than a pressure flow. If it does cause flow to be wrong what would you predict the outcome, an error high or low.

 

[bimr]

I believe that the pressure taps are on either end of the venturi tube. If you have a full pipe, it should read correctly.

 

[BigInch]

Venturi meters generally indicate velocity based on the differential pressure in the pipe, therefore that pressure is not referenced to atmospheric, or gauge pressure, at all, so if you have a negative reading on a differential pressure gauge, flow is in the reverse direction from that intended. If you are measuring the difference in flow via two atmospherically referenced pressure gauges one upstream, one downstream, you could see "vacuum" readings that correspond with flow in the correct direction.

Depending on the precise shape of the venturi, it may have a more rapid reduction in size in one direction when compared to the other; most I have seen have a quicker reduction from pipe diameter upstream with a more gradual transition back to the original pipe diameter in the downstream direction, so it may or may not give equivalent readings for equivalent velocities in opposite directions, but I would not expect them to be equivalent without a calibration to specifically verify that supposition.

Independent events are seldomly independent.

 

[operatorjoe]

I appreciate both answers (BIMR and BIGINCH) and I believe both are valid as stated. Please consider that as I understand a venturi it is assumed that there is force/pressure before the restriction and the second port level is the remaining level effected by the velocity. (If there is no flow the two ports will be equal.) After reviewing some basic tutorials on the subject I suggest that if there is a siphoning flow through the venturi the second port will have an error that will reduce its level. The difference in water columns will indicate a higher flow than is actual because the differential pressure transmitter is a closed/sealed loop between the two ports. Are my points without merit? To get our operations crew to change their methods I will have to present a convincing argument. Thank you very much for your time and help.

 

[BigInch]

No. The taps on a venturi meter are reading only static pressure on both sides of the meter. VELOCITY head is directly measured with a pitot tube. A venturi meter therefore relates the drop in static pressure to the change in velocity, [Δ]P = ([Δ]V^2)/2/g. It does not tell you the actual velocity on either side. A pitot tube on the other hand, measures velocity head directly and can be related proportionally to the actual velocity of fluid at that point.

Your basic understanding of fluid flow is somewhat off as well. No fluid is " pulled by vacuum (or very little head)rather than a pressure flow." as you contend. All fluid motion is by higher energy (static head plus velocity head) on one side, lower on the other side. If you convert all pressures and velocities to absolute pressures and heads, obtain the EGL (energy grade line), you will ALWAYS see higher energy levels on one side of a fluid in motion, lower pressure to where it is going. The venturi doesn't care how flow is generated, "pressure or vacuum" is the same in absolute terms. Vacuum pressures do not convert to negative absolute pressures. Within the variation caused by velocity head, flow is always from higher pressure to lower in absolute pressure terms.

You may need to pay more attention to what your operation crew is telling you, rather than what you want to tell them. Listen a bit closer. Tell us what "loss of head" means in real terms to the venturi gauge location and come back with any more questions.

Independent events are seldomly independent.