2

I was doing an experiment on water hammer in pipes. During the experiments, in several cases, the transducer reported negative pressure, as low as -1.9 bar gauge! But I have read that it is not possible for the pressure to practically go beyond absolute zero. I thought it might be due to some calibration issues of the transducer, and recalibrated and double checked the experimental values. I did the same experiment again. But the pressure shot down to 1.9bar or 190 Kpa below gauge, when cavitation occurred inside the pipe.

The pressure transducer i am using is a 0-30 bar transducer (forgot the company name), and i am measuring the values at 50,000 samples per second. the operating pressure inside the pipe in steady flow conditions is 1.7bar gauge and it is being shown correctly.

Can anyone give a possible explanation for the same?

enter image description here

The graph denotes the actual pressure vs time graph obtained during the analysis. the sampling frequency is 100000Hz and the Amplitude shows pressure value in bar, in gauge scale. The time scale is shown in 10microseconds value

Roopesh
  • 51
  • Please note that I am not questioning why the pressure is negative. My confusion is why it shot down below absolute pressure. I believe pressure cannot go below 0 bar absolute or -101325 Pa gauge (Unless ofcourse otherwise the dark matter is proved :D).. Here, in my case, the pressure shot down to -1.9bar below gauge – Roopesh May 08 '18 at 07:13
  • We have used a Keller PA 23-SY pressure transducer for pressure measurement. The red and white plots depict the two transducer viz, white for the transducer placed near to QAV and the red for the transducer kept at a distance of 20 m from the QAV (Quick Acting Valve). The closing time measured is around 30milliseconds. – Roopesh Jun 06 '18 at 03:53

2 Answers2

1

The sensor measures push or pull force on a plate, and, using the area of the plate, translates that into pressure. There are multiple ways to get a "pull" force on the plate without negative pressure. Inertia from moved sensor, internal pressure in the sensor,...

Sensors only perform as expected if they are used within their limitations. Imagine a diaphragm displaced by air under pressure. the amount of displacement somehow is translated into a pressure reading. Now that pressure is removed fast. The diaphragm will bounce back, and without any air to stop it, it will move further than its ususal "0 pressure"-position, for a short time, then resettle on the "0 pressure"-position. This will then be read as "negative" pressure.

The ability to answer to changing pressures, or pressures baove/below certain thresholds is one factor that drives the specialisation of sensors. (Medium, accuracy, drift, etc. are other factors.) Not every sensor will be able to measure every sort of pressure & pressure-change reliably.

bukwyrm
  • 110
  • I believe I am using a strain gauge based sensor as it was said to have a diaphragm. The transducer gives me an electrical output (in mA).

    There dosent seem to exist any adjustment screws or anything as so, where in I can adjust the inertia of the plate or internal pressure of the sensor as you mentioned. Atleast I didnt find any.

    Moreover, Please note that, in steady condition, the transducer shows the pressure which matches exactly with analytically calculated value.

     – Roopesh May 08 '18 at 05:29
  • See my edited answer. It might be that your sensor is simply not able to handle the pressure-changes correctly. think about it this way: if every pressure sensor would also be able to map pressure changes perfectly, every pressure sensor could also be used as a perfect microphone. (And even among good microphones, there are no perfect ones) – bukwyrm May 08 '18 at 09:07
  • Related? – Keith McClary May 08 '18 at 14:51
  • @Keith McClary, I did read the article you point out, but again, they are discussing about negative pressure with respect to the atmospheric datum and not to absolute datum. I am here discussing about the absolute datum point. I my experiment the pressure drops below the absolute zero reference point and not the atmospheric datum. I am not talking about the differential pressure here, but absolute pressure. Please view the Plot I have attached. The plot is referenced to gauge scale. – Roopesh Jun 06 '18 at 03:59
  • @Roopesh I found this article . – Keith McClary Jun 06 '18 at 18:02
  • @Keith, That was really helpful, thanks a lot Keith. I find that negative absolute pressure are indeed possible. I would be thrilled to study more in this. :) – Roopesh Jun 07 '18 at 03:53
  • @Roopesh please self- answer this one, there do not seem to be any correct answers now (including mine). In gases negativebpressure is impossible, but with liquids and solids it is ! TIL. Props for not being dissuaded by the naysayers. You science well! – bukwyrm Jun 07 '18 at 15:20
0

Water hammer in pipes will exhibit transient pressure spikes that go both positive and negative. When they go negative, cavitation can occur. Therefore, if you observed cavitation in the experiment, then it is evidence that the pressure in the system did indeed go below atmospheric for a brief time.

niels nielsen
  • 92,630
  • I do agree that the pressure can go negative in water hammer. I dont object negative pressure. My problem is that the pressure given by transducer is below absolute pressure.It is -1.9 bar gauge, which means around 88kPa below absolute pressure. My question is, Is it possible? If not, what might be the probable reason for the same! – Roopesh May 08 '18 at 07:10
  • @roopesh, it may be that the only significant compliance in your test system is the flexible membrane inside the pressure gauge. this means that when a shock wave develops in the system, the gauge itself becomes a victim of the water hammer effect you are studying. I recommend you remove the gauge and instead affix a strain gauge or several of them to the outside wall of your test pipe, calibrate them with a known pressure source applied to the pipe, and repeat your tests. – niels nielsen May 08 '18 at 08:00
  • we already tried removing the gauge, priming it for any gas or air entrapment, calibrating thrice to ensure the constants, and repeated the test as well. We have taken around 20 different cases, and each cases repeated 3-5 times to ensure correct values. The gauge showed correct pressure inside the pipe(same apparatus) during steady flow, verified analytically and with another gauge. Moreover, we have 3 gauges mounted at 3 locations and all shot to negative pressure beyond absolute! – Roopesh May 08 '18 at 09:07
  • can you provide a continuous gauge reading (a pressure-versus-time chart) that shows the pressure profile of your water hammer/cavitation event from start to finish? – niels nielsen May 08 '18 at 16:28
  • sorry for the delay, got some heck of a work. I have attached one of the wave form graph you have asked. The graph attached is directly obtained from the DAS, unedited and original. As you can clearly see, the graph, in its negative direction, have crossed the absolute zero condition (The plot is on gauge scale and not on absolute scale) – Roopesh Jun 06 '18 at 03:51
  • here are my thoughts. Imagine we pull negative pressure on water contained inside a vessel. When the pressure we exert on it falls below the saturation pressure of the water at room temperature, then instead of a stronger vacuum you get cavitation- and the pressure oscillations will be "clipped" off at that pressure. Your plots do not show evidence of pressure being truncated on the negative-going side, do they? if you are getting cavitation, you should have the negative peaks all clipped off at the saturation pressure and a burst of cavitation bubbles in the water at each clipped-off cycle. – niels nielsen Jun 06 '18 at 07:14
  • When we reduce the sampling frequency, we obviously do get a clipped off result of the pressure. Moreover, the sudden increase in pressure after the first peak can only be attributed to cavitation vapor collapse, as far as i know. Else why is the second and third peak higher than any other peaks? Moreover, whether or not the cavitation is occurring, my doubt is, the pressure should not go beyond zero pascal absolute! But mine is going below that, as you can see very much clearly in the graph! – Roopesh Jun 06 '18 at 07:43
  • Theoretically speaking, you are right and the pressure should not go beyond 3540 Pa at 25 centigrade. Supposedly giving a relaxation for the same, owing to probable experimental error, still, the graph should never go beyond -101325 and even if a 30% error is attributed (owing to probably limitations or uncertainties in instruments), still it should not go beyond -131722.5 bar. See the graph and you see it has gone well beyond the value! – Roopesh Jun 06 '18 at 07:46
  • this suggests a fault in the functioning of the transducer. I wish I could be there to assist in finding it- I used to be good at this sort of work when I was still in the workforce. – niels nielsen Jun 06 '18 at 07:48
  • I even considered that possibility. We re calibrated the instrument thrice to ensure the proper working. Moreover, the transducer is giving a correct pressure (average value) in a steady condition. We have given a head of 17m at inlet and the transducer shows the theoretically computed pressure value with no problem! That again confuses me! – Roopesh Jun 06 '18 at 07:53
  • And, just to note, we did the experiment in one of the Asia's leading calibration and testing laboratory for flow metering and calibration (Atleast they boast so, even though they couldn't help me find a solution for this!) I hesitate to name the same! – Roopesh Jun 06 '18 at 07:56
  • I am NOT suggesting the gauge was miscalibrated. this is NOT a calibration issue, this is most likely the gauge's spurious electrical response to a very fast, strongly negative-going transient signal spike. In other words, the gauge is being subjected to a pressure transient which causes it to produce meaningless data. time for bed now- it's 1 AM here. -Niels – niels nielsen Jun 06 '18 at 08:26
  • ROOPESH! I think I have your answer. your pressure gauge is behaving like a microphone- it is picking up the hammer shock acoustically transmitted into the pipe itself and responding to it as if it were a pressure signal produced by the water. this is easily tested! – niels nielsen Jun 06 '18 at 09:16