How to measure water level in a vacuum...

[drbitboy]

Again, nice idea and I appreciate it, but drilling through the vessel is a no-no.

I thought that might be the case.


So the WIKA uses a wireless signal?

 

[Mas01]

I thought that might be the case.


So the WIKA uses a wireless signal?

No, it's not wireless. The cable feeds through a gland of a specially designed "porthole" and is routed to the PLC. The breather tube stays inside the vessel though.

 

[danw]

A differential pressure (DP) transmitter with dual remote seals and capillaries (one remote seal at the bottom of the tank, one remote seal near the top in the vapor area) was used to measure level in a vessel where vacuum was used to boil off the water in oil/water mixture. The accuracy of the level measurement was always in question because a hydrostatic pressure measurement for level ASSUMES a known density/Specific Gravity of the liquid which needs to be used as a compensating factor in calculating the level.



As the water boiled off in this application, the density of the remaining solution decreased continuously, violating the requirement for a known density for correcting the pressure-to-level calculation.


If your solution is water only, a DP with remote seal capillaries would work because the water's density doesn't vary, but would probably be 2x as expensive as radar.

 

[drbitboy]

No, it's not wireless. The cable feeds through a gland of a specially designed "porthole" and is routed to the PLC. The breather tube stays inside the vessel though.

is there any room in the gland for a small tube with water?

If the breather tube went way up then back down, and finished open downward, perhaps with an inverted funnel, that might take care of all problems except condensation.

The other option would be to have the breather tube face up, somehow ensure it was always filled with water, extend it some distance above any expected level, and reverse the connections. The measured ΔP would be proportional to the distance between the level and the top of the tube.

 

[drbitboy]

How does the drain to the pump exit the vessel? Would it be possible to run a small line into that? Or add a pressure instrument to the drain just outside the vessel?

 

[Bit_Bucket_07]

Use a capacitive level probe.

 

[Mas01]

Use a capacitive level probe.

My understanding of these devices is very limited, but I think this would entail drilling into the vessel roof, and the rod/probe itself would need to be about 4 metres long to dip into the water.

 

[Mas01]

How does the drain to the pump exit the vessel? Would it be possible to run a small line into that? Or add a pressure instrument to the drain just outside the vessel?

Not sure I understand the question.
I've sketched a poor picture of the pump relative to the vessel. Pump sits on the floor, not partially buried like in the sketch. There's a drain cover plate, above which the level sensor sits.

 

[parky]

Probably the best solution is to get Erwin SICK representative in to discuss the requirements, I would have thought an analogue rod probe type if it could be mounted would be the best, not used them on a vacuum but do not see any reason why it could not be used.

 

[Mas01]

Probably the best solution is to get Erwin SICK representative in to discuss the requirements, I would have thought an analogue rod probe type if it could be mounted would be the best, not used them on a vacuum but do not see any reason why it could not be used.

Thanks, We discussed it with the WIKA representative who visited our facility. The rep assured me it would work in partial vacuum conditions, and until the other day, it worked fine. I'll be calling him to ask him about the breather tube problem (or at least, we suspect it's the breather tube at fault).

 

[parky]

The probes I was thinking about is a rod type do not need breathers or alike it is just a rod like a level probe but instead of just detecting level it will give out an analogue level based on how much of the probe is covered, vacuum should not affect it but would need to be confirmed.

 

[Mas01]

The probes I was thinking about is a rod type do not need breathers or alike it is just a rod like a level probe but instead of just detecting level it will give out an analogue level based on how much of the probe is covered, vacuum should not affect it but would need to be confirmed.

Someone mentioned this earlier in the thread. My understanding is that this would mean drilling into the vessel wall which is 6 inch steel, we can't make any holes.

 

[drbitboy]

Not sure I understand the question.
I've sketched a poor picture of the pump relative to the vessel. Pump sits on the floor, not partially buried like in the sketch. There's a drain cover plate, above which the level sensor sits.

The idea was to run a pressure tap through the side of the drain and up through the drain into the vessel.

But we may be looking at this wrong. If there is a vacuum of 20Torr above the liquid level in the sump, then that adds 37.2cm to the target minimum level used when there is no vacuum. So if the drain pump comes on at 50cm, and stops when the level drops to 10cm whenever the pressure above the liquid is atmospheric (0Torr; no vacuum) to avoid cavitation, then when there is 20Torr of vacuum, the pump should come on at 50cm* but go off at 47.2cm (10cm + 37.2cm).

What we really want to do is stop the pump when its suction (intake) gauge** pressure goes below some point, and we don't need the level to measure that pressure, and presence or abscence of a vacuum does not change what that pressure is. Then, when the pump is off, the gauge pressure at the suction (with no flow) at which to turn the pump back on, which pressure corresponds to a level of 50cm, can be calculated as long as the vacuum above the level is known.

So we don't need a level measurement at all; all that is needed is the gauge (atmosphere-relative) pressure measured at the pump suction, and the vacuum (again atmosphere-relative) measured above the liquid inside the vessel.

* I assume the 50cm level specification is geometric i.e. the level gets close to some piece of hardware in the vessel that should not be submerged in the water.

** "gauge pressure" means "relative to atmospheric."

 

[Mas01]

The idea was to run a pressure tap through the side of the drain and up through the drain into the vessel.

But we may be looking at this wrong. If there is a vacuum of 20Torr above the liquid level in the sump, then that adds 37.2cm to the target minimum level used when there is no vacuum. So if the drain pump comes on at 50cm, and stops when the level drops to 10cm whenever the pressure above the liquid is atmospheric (0Torr; no vacuum) to avoid cavitation, then when there is 20Torr of vacuum, the pump should come on at 50cm* but go off at 47.2cm (10cm + 37.2cm).

What we really want to do is stop the pump when its suction (intake) gauge** pressure goes below some point, and we don't need the level to measure that pressure, and presence or abscence of a vacuum does not change what that pressure is. Then, when the pump is off, the gauge pressure at the suction (with no flow) at which to turn the pump back on, which pressure corresponds to a level of 50cm, can be calculated as long as the vacuum above the level is known.

So we don't need a level measurement at all; all that is needed is the gauge (atmosphere-relative) pressure measured at the pump suction, and the vacuum (again atmosphere-relative) measured above the liquid inside the vessel.

* I assume the 50cm level specification is geometric i.e. the level gets close to some piece of hardware in the vessel that should not be submerged in the water.

** "gauge pressure" means "relative to atmospheric."

You've lost me in the science there, I am a mere mortal, but maybe we can start the pump using water pressure alone.
Btw 760 Torr is atmospheric pressure. Zero Torr is a perfect vacuum.
To add more detail, there's a rotating fan above the water. Spray bars are used to blast the blades with water, but the blades must not pass through the accumulating water inside the tank. See picture.

 

[drbitboy]

You've lost me in the science there...

Look up NPSH (Net Positive Suction Head). One statement of NPSH is "The margin of pressure over vapor pressure, at the pump suction nozzle."

• Vapor pressure is a function of temperature only, and it is an absolute pressure.
• The absolute pressure of the liquid at the pump suction nozzle in our case is the sum of four quantities, all converted to pressure, at the pump suction:
  • the absolute pressure at the surface of the liquid, i.e. 760Torr absolute when the vessel is at atmospheric pressure; 740Torr (760 atm - 20 vacuum) when a 20torr vacuum is pulled
  • Static head(positive, typically): the vertical difference in head (distance) from the height of the pump suction to the height of the surface level of the liquid in the vessel, converted to 1Torr by a factor of 1.36cm per Torr
    • 1Torr is equivalent to 1mmHg i.e. 1mm of head (vertical column) of mercury
    • The specific gravity of mercury is ~13.6, i.e. is is 14.6 times more dense than water
    • 13.6 Hg/Water * 1mmHg/Torr * 1cmWater/10mmWater = 1.36 cmWater/Torr
  • Dynamic head(negative): The difference in dynamic pressure due to difference in flowrate between the surface and the pump suction,
    • i.e. Bernoulli's equation: kV²/2; V is velocity
      • At the air/water interface in the vessel, the liquid is not moving, so V is 0,
      • At the pump suction, V is some positive value, Vs
      • So the difference is [k0²/2 - kVs²/2] = -kVs²/2
    • Since water is incompressible, to first order, this difference will be more or less constant i.e.
  • Friction loss(negative): the pressure loss due to friction of the water flowing through the pipe
    • This is also related to the velocity, but to the length and diameter of the pipe as well.
    • Again, to first order, this will be constant when the pump is running

So

• NPSH = (P_vacuum + ΔP_Static + ΔP_Dynamic + ΔP_Friction) - P_partial

The only non-constant quantities in there are the absolute pressure at the surface of the liquid i.e. the vacuum, and the height difference between the surface and the pump suction, or mor specifically, the level of the surface, because the height of the pump suction is constant.

If there is no vacuum and the tank vapor is at atmospheric pressure i.e. constant, that fixes the absolute pressure at the surface, and the only non-constant quantity is the level of the surface. So in normal non-vacuum operation, the level of the surface, the value we have been trying to measure, is a proxy for the NPSH. If there is a vacuum, that is no longer true, and it may be simpler to measure the NPSH directly.