Hi All

Is it possible to use an accelerometer to detect water flow in a pipe?

I do not need to have accurate flow rates(LPM) unless is easily achieved, but just need to be able to detect that something is flowing.

If I am correct this is possible from the vibrations?

 

I have read about several ways to measure a flow rate using accelerometers and vibration but most are complex using multiple accelerometers and lacking in accuracy. If you simply want to detect i9f there is or is not flow in a pipe, depending on medium, a simple flow switch is likely the easiest and most economic solution. So is it possible? Yes, it is possible.

Ron

 

I would like to avoid cutting pipes and sealing them up. Some pipes i would need to detect are between 4-6 inches, therefore the flow sensors would be costly.

 

In theory you could inject ultrasound through the pipe wall into the fluid at one point and detect the Doppler frequency shift with a sensor at a downstream point. Acoustic coupling and signal processing might be challenging in practice.

 

Or you could bond a resistor to the pipe using high temperature epoxy, run some current through the resistor and monitor the temperature of a thermistor or other temperature sensor. When water flows, the temperature would drop. Not as certain as a flow switch but hey...if you have sufficient flow and the pipe's wall is not too think it should work.

 

Hi All

Is it possible to use an accelerometer to detect water flow in a pipe?

I do not need to have accurate flow rates(LPM) unless is easily achieved, but just need to be able to detect that something is flowing.

If I am correct this is possible from the vibrations?

Some of the most accurate flow sensors use accelerometers and the Coriolis effect of water making a turn in a curved, vibrating pipe (not the earth's rotation). But that is a research project in itself and design/calibration of curved castings of pipe. Google: Coriolis meter or MicroMotion flow meter
Describe your specific piping, tubing pump, flow rate, back pressure situation.

 

Need more specific information about the medium - there are many types of "water", clean, dirty, fresh, saline - all important.
Also the same for pipes - sizes, material - PVC, steel, etc.,
You say you want to detect flow. Be more specific - do you want to measure flow or detect the presence of water? Or is the water moving?
There shouldn't be any problem in cutting /sealing pipes, it's done all the time.
Depends again on the pipe. Slip joints are used to gain access and and reinstate pipes without cutting chunks out of them.
Access points and saddles allow inspection or even the introduction of instruments.

 

I would like to avoid cutting pipes and sealing them up. Some pipes i would need to detect are between 4-6 inches, therefore the flow sensors would be costly.

It is also worth noting that a simple flow switch is comprised of a magnetic reed switch which is notoriously unreliable. I had never thought of using an accelerometer to sense flow but I am now. I have had success using 2 pressure transducers, each at opposite ends of a run and calibrating static pressure. Then I can easily sense flow on the amount of pressure differential through the liquid circuit. If the pipe run is short it may be necessary to install a restrictor, or have a very high resolution PLC to detect such slight differentials. Even so, without a restrictor on a short run between transducers turbulent flow will cause chatter and increase the instances of false negatives.

I'm thinking that an accelerometer would indicate a false negative flow condition at low flow rates unless perhaps the pipe section the accelerometer was attached to had an internal reed attached to the accelerometer mounting point. In such a case the downstream eddys should stimulate it enough to be sensed even in lower flow conditions but there will always be a point where flow is positive but not sensed.
I suppose whether the application so feasible depends on how much flow constitutes no flow and how much flow constitutes flow in your application. A flow meter is a much more useful tool than a switch and if you are industrious enough to make a switch out of an accelerometer you should be able to make a DIY flow meter.

 

Pipes also make noise, you could try a simple microphone and some filtering of non-important frequencies. The 5 to 50Hz are likely most interesting range if your microphone can sense that sufficiently.

 

Need more specific information about the medium - there are many types of "water", clean, dirty, fresh, saline - all important.
Also the same for pipes - sizes, material - PVC, steel, etc.,
You say you want to detect flow. Be more specific - do you want to measure flow or detect the presence of water? Or is the water moving?
There shouldn't be any problem in cutting /sealing pipes, it's done all the time.
Depends again on the pipe. Slip joints are used to gain access and and reinstate pipes without cutting chunks out of them.
Access points and saddles allow inspection or even the introduction of instruments.

The water will be ideally clean, may have a little limescale, but generally, it will always be clean.

Pipes will be plastic pipes.

I would like to initially measure moving water and when water stops flowing. I know that I can simply place various types of sensors in a pipe, but I would like to avoid the plumbing aspects of handling a pipe.

Initially its measuring water, but in the future, if I was to consider various parameters, it would be good that it could produce some results of water flow rate, but this is not important at this stage but something in the distant future.

If the accelerometer was to produce some reading, maybe by considering a pumps parameter, pipes size etc, you could estimate the flow rate.

 

It is also worth noting that a simple flow switch is comprised of a magnetic reed switch which is notoriously unreliable. I had never thought of using an accelerometer to sense flow but I am now. I have had success using 2 pressure transducers, each at opposite ends of a run and calibrating static pressure. Then I can easily sense flow on the amount of pressure differential through the liquid circuit. If the pipe run is short it may be necessary to install a restrictor, or have a very high resolution PLC to detect such slight differentials. Even so, without a restrictor on a short run between transducers turbulent flow will cause chatter and increase the instances of false negatives.

I'm thinking that an accelerometer would indicate a false negative flow condition at low flow rates unless perhaps the pipe section the accelerometer was attached to had an internal reed attached to the accelerometer mounting point. In such a case the downstream eddys should stimulate it enough to be sensed even in lower flow conditions but there will always be a point where flow is positive but not sensed.
I suppose whether the application so feasible depends on how much flow constitutes no flow and how much flow constitutes flow in your application. A flow meter is a much more useful tool than a switch and if you are industrious enough to make a switch out of an accelerometer you should be able to make a DIY flow meter.

It would be good if you would like to do a joint project. You can PM if you like.

In regards to false-negative results, I am presuming we would need to perform some type of calibration on each installation to indicate when water flow is off and when it is flowing the max, and then anything between would simply be something in the middle.