40 yr old thought experiment

Well, suggest a sensor topology to use doppler effect to measure fluid velocity in a pipe.

Hmmn. Probably most of the methods that immediately come to mind (i.e. off- axis ultrasonic transducers) you'd probably protest that these were time-of- flight. I have to admit that while the _phenomenon_ of doppler shift requires no "particles" (in the ordinary sense), the _detection_ of doppler shift doesn't seem to work in this context. Lacking some kind of (fluid) discontinuity you don't get any reflections.

velocity in a pipe.

Gets interesting in the later half or so, including scope shots.

Tim

Den onsdag den 22. juli 2015 kl. 22.38.36 UTC+2 skrev Tim Williams:

I think John is right to use doppler you need to reflect off something moving, particles etc, to measure the speed

The ultrasonics are time of flight, not doppler

this is the one used in formula1:

-Lasse

Laser doppler velocimetry uses particles. A moving medium with fixed source and receiver isn't Doppler, because the frequency doesn't change, just the phase.

(The math is similar to light entering a denser medium--there's a time delay, but no frequency shift.)

Cheers

Phil Hobbs

One cool technique is CWFM, which is done by frequency sweeping/chirping a CW transmitter. The sweep can be a sawtooth or a triangle. Some WWII bomber altimeters worked that way. If you mix the transmit and receive signals, you get a tone that's proportional to round-trip time delay, namely distance or altitude. Sort of like doppler with nothing moving.

A one-way system (transmit and receive transducers spaced apart in a pipe) should be sensitive to both distance and fluid velocity. The numbers probably don't work for most flow meter apps.

I've done coherent lidars like that, using current-tuned diode lasers and noise cancellers. They work great. I have a line in the water to do something like that for vehicles, robots, and drones. We'll see.

Cheers

Phil Hobbs

Under some circumstances, the alignment of molecules would reduce the thermodynamic freedom of the liquid (i.e. reduce the heat capacity, and raise temperature). Then on leaving the capacitor gap, the heat capacity would rise (and temperature drop). The thermal signal might be easier to detect than an electrical polarization.

So would this thermal signal be sensitive to flow rate, assuming for a moment that the container was not thermally conductive & absorbing said signal? Of course any real system would be thermally lossy...

Well, it'd be zero at zero flow, by energy conservation. The trailing edge of depolarization would be a long tail, but the onset of polarization would be as abrupt as the fringe field. If the molecules have to align (liquid-crystal transition kind of alignment) it gets complicated.

'Adiabatic depolarization' is a good search string.

Feed temp changes in too fast and the display says 'fluid speed is 1 m/s or 2 or 3 or 4 or 5...' Feed them in too slow and the display says 'still waiting for data'

Feed in a recognisable pattern to avoid this.

NT

It made my head hurt.. :-) but thank you none the less.

Hmm... So, you put two halves of a capacitor, charge them, measure the energy loss (current?) required to keep the voltage stationary and compare it to known values, inferring fluid flow. Did I get this right?

This idea needs to spend a whole lot of time on the drawing board to become a tangible and correctly-working artifact.