Measuring Height with a Pressure Sensor

SunPro 12V Fuel Gauge and Sending Unit

Calibrating a flow meter? That's a dream.

Or, do what bigg says in the next post. WOW, am I traveling back and forth in time here???

Put a transparent sight tube on the tank with a ruled gage so you can see the level. Then use a stopwatch when the level reaches a bottom and top mark of known height. If you want to put optical or conductive probes top and bottom you can, but you will have to visually confirm where they switch on rising level. Direct the inlet pipe to your 'piezometric tank' against the side of the tank to minimize turbulence, but don't submerge it as that would change the back pressure on your loop while the calibration is in progress.

Drill a hole in the tank at desired height...

"Even with your fancy set up you still need some reference to calibrate against"

That is the crux of the biscuit here.

http://cache.freescale.com/files/sensors/doc/data_sheet/MPXC2011DT1.pdf

If I understood your question correctly, the problem you are facing was solved by the industry: washing machine (water) filling sensor. It senses pressure change of the water in a turbulent environment. You couple it with a flexible hose. You set its tripping point by mounting it lover or higher. Problem solved.

If you need a (calibrated) flow sensor on the top of that for some process purpose, the always precisely filled "piezo" tank elevated above the large one, with a (set or regulated) valve takes care of the steady flow.

Sure, it is a two step process. But it is workeable, with known components.

... washing machine (water) filling sensor...
Yeah, but they do it that way because it's it's cheap, not accurate!
And it's not measuring flow, it's measuring water level.
Other than that, it's a great way to measure flow accurately.
Del
_{(I want to measure arrow speed, so I'm doing it by measuring the height of my chimney and the weight of my car... will this work?)}

Change the shape of tank to conical with a small hole (about 0.5mm) at the bottom of cone. Put a small beaker just under the hole of cone. Open gate valve for different periods and make precise measurements of water you receive in tank and beaker, set a correlation for future measurements.In future you don't need to measure water of large tank, so just a laboratory barrette or a pipette may work.

First of all, thank you very much for all the comments!

I'm amazed of how many replies! This Forum rocks!

lyn, a floating device for measuring the level sounds good but I don't think they are that accurate, thanks, though.

bigg, the piezometric tank has a sight tube and it reduces the turbulences a lot, it would be nice for an optical measurement or a capacitive measurement of level ( I guess that with 'conductive probes'). Directing the inlet pipe against the side of the tank was also recommended to us by the mechanical people (the ones asking for the project), thanks for pointing that out!

SolarEagle, there is no desired height. I need the height to get the volume and I need the volume and time (interval) to the current flow which is the output variable of the control system.

Del the cat, yeah I know it seems over complicated, but is the price for having a short budget lol. Your idea of putting a calibrated flow meter in series was our first one. The only problem is that a accurate flow meter too expensive (700USD to 9000USD for 10m^3/h). But I have to agree that it would be much simpler and accurate. Thanks!

Your idea for making the flow laminar its good, thanks. Yeah oscillations are going to be a big problem. We are going to get the transfer function of the valve and the curve of flow vs steps and with the coefficients from simulink it's possible to reduce the oscillation a little bit. We don't get to that point yet.

It's just a senior project 1, I guess we will make several mistakes, but at least not the same mistakes as the group of last semester. They used a ultrasonic sensor to get the height and their system was way too inaccurate to be useful for the mechanics. I'm not sure they passed the course.

bioramani, nice sensor. It seems very accurate. However, the datasheet says 0 to 10kpa. Thats absolute pressure? Because the atmospheric pressure is about 100kpa. Would it work? Sorry for my dumb cuestion. Thanks!

leveles, as Del the cat said, it's not accurate for measuring flow, thanks for posting!

rakesh_semwal, I didn't get you. You mean using smaller flows to calibrate the flow meter? One of the specs of the project is to deal with 10m^3/h I think I would have big overflowing problems there. Thanks btw!

bioramani, nice sensor. It seems very accurate. However, the datasheet says 0 to 10kpa. Thats absolute pressure? Because the atmospheric pressure is about 100kpa. Would it work? Thanks!

Freescale makes MEMS sensors for various ranges. Both absolute and gauge versions are available. We have been using their 3 Bar absolute presure version to monitor steam pressure in sterilizers for years now.

The 10 kPa is gauge pressure and should suit your purpose.

It looks like, it bears expanding on the subiect.

By the way, I built a university laboratory setup with such sensors, when my daughter and colleagues did some biomechanical work for NASA as undergrads.

Most people are unaware of the quality available in silicon pressure sensors. For $20 or so you get a sensor to drop in into your proiect. All the calibration, temperare compensation is already taken care at manufacturing. I would not have it any other way, as I, and presume you have no means to do a sensor calibration run.

First, the task to be accomplished. I put it down, how I would go at it. If I misunderstand something, Andre, feel free to come back.

1,. A steady level should be maintained in the smaller "header" tank to maintain exact pressure to the pipe (with resistance) leading the flow to the main tank. The level should be measured and regulated precisely, with electronic sensor.

2,. The pipe should have a settable resistance to the flow, to maintain steady flow, by volume to the big process tank. The settable resistance should be either settable by a knob, or be a replaceable module. I have a preference to the second, as it is simpler, and after calibration a well known quantity. I would use standard glass pipes of different inner diameter, bent in half for easy coupling into the piping, with short silicon hoses.

2.1,. The pressure differential accross the resistance piece can be measured / monitored via differential pressure sensor. It will provide the flow data via the calibration process.

2.2,. If somebody is handy with computer programming, all the calibration data can be applied automatically. For a lab setup, it may be simpler on paper.

3,. Calibration to the required precision and repeatability is an absolute requirement for any newly designed and built equipment. That is a fussy, precise work, but no big deal otherwise. If you do not know how, any good chemist or physical chemist can lend a hand and glassware. They need to do that routinely.

4,. If this would be an industrial process, I would recommend a set of temperature calibration charts. The measuring setup is tightly controlled. But any and all liquids show significant temperature dependence in their hydraulic parameters. In a lab, the need for it can be sidestepped by keeping all temperatures tightly controlled.

5,. Tying 1,. and 2,. you can set any flow by switching in a resistance module, and then modulating the height in the header tank, according the calibration. If the header tank is deep enough, it means a continouos coverage of flow from low to high volume.

Now, to the sensors. There are as many different qualities as price points. And I am talking silicon electronic sensors. A cheap one to set water level in a washing machine may be less than $4, a nice one for real measurement may be $20, a top of the line individually calibrated and laser trimmed may cost you $100. If you need it, that price is well worth it!

6,. I can help you to select the right one, but you have to do some heavy lifting. The first the terminology used is different: absolute, gauge, differential, analog, digital, calibrated, sefcalibrated....... Start with Wikipedia, learn all.

7,. Once you got that under the belt, Google the following:

"silicon diaphragm pressure sensor"

"mems silicon pressure sensor"

"differential silicon pressure sensor"

You will be inundated with tons of choices, as we are not selective enough. But, it will be a useful learning experience.

8,. Once you navigate these waters, and have some idea of what you might need in these terms, we can narrow the choices in the next round, if you care.

Good luck with the hunting!

Seems to me that a sight glass in the tank will eliminate the turbulence problem and permit you to use any kind of height sensor you wish: ultrasonic, optical, etc.

Your example shows a ball valve controlled by a stepper in the inlet to the flowmeter.

If you are wanting accuracy, a ball valve should not ever be used as they do not display nor provide linear flow charactristics.

Even with a precision "V-Ball" the flow loses it's linearity long before it reaches the 100% open position.

If you want accuracy, use a globe or plug valve designed for accurate flow control within the range of your application otherwise your results will be non-linear and intermittent at best.

How about a DP transmitter,

Rosemount 3051SCD2A would be suitable i guess, You may leave low tapping open and the high tapping connected to the bottom tapping of the tank.

So you should get the mm per your requirement .

I think this should work well.

Best Regards

Naveen Menon

OP, don't forget that any measurement device, whether it be steel tape or fancy, schmancy transmitter, has to itself be calibrated back to a known standard for your end product to have a known accuracy. Otherwise, GIGO. It's easy to assume a shiny $1,000 widget is perfect, but it's not.

Aw c'mon... we all know if it's got a digital display, it must be accurate.
_{(That's a joke, just in case there is any doubt. But I'm sure we've all met people who believe it!)}
Del

I agree. Rosemount make very rugged and very accurate pressure transmitters.

They can even detect the turbulence from a passing fly, if required.

I still think it's an overcomplication.
It's relatively easy to get an acurate measurement of volume and time.
Once you have set up a controlled flow (fixed head, linear flow), take your uncalibrated flow meter and spot calibrate it at various flows (achieved with mandraulic gate valve) by measuring time and volume.
You then use that flowmeter to calibrate others, the traceability is done by the volume (or weight) and time measurements.
Del
_{(I promise I'll go and hide in my secret cat nest now)}

...???...

<...measure the height of the water in the tank...without expending lots of money?...>

Sure.

Use a dipstick.

If a constant level is required, then get rid of the tank.