Measuring Instantaneous and Total Fuel Flow with a Pelton Sensor and Digital Readout

Yeah PUT, your math looks good! lmao... (stupid spell check)

 

Here is a little followup on this flow meter setup.

It has worked flawlessly until my last melt when my fuel flow just suddenly and completely stopped mid-melt. I determined that the fuel pump I had upstream of the meter encoder (a pelton wheel type) had given up and so I quickly bypassed the pump and meter and raised the fuel can six feet or so to establish some pressure in the fuel supply to my siphon burner. By guess and by golly I was able to tune the furnace to high heat and finished the melt losing only about 20 to 30 minutes time---consumed by making the bypass (and feeling very happy to just happen to have enough hardware laying around to do it and not have the crucible either have to be dumped of its 60 pounds of liquid iron) and retune by eye and ear the burn rate.

Today I installed a new higher quality pump, a Holley "Black", that should be more reliable even though the one I had purchased previously was made in USA and rated for diesel. When I fired the Holley up, there was poor fuel flow and the meter was reading NO flow. Oh-oh.

Well, I found out that when it gave out, the pump diaphragm must have disintegrated and a bit of rubber jammed the encoder wheel. The encoder is made to be serviceable and so I was able to pull it apart and found that bit of rubber in the encoder wheel chamber nicely lodged between the wheel and the exit hole. Pulled it out and I am back in business! Most of the time I am not so lucky in a situation like this to actually find the offending bit, but there it was.

I am so relieved that the meter and encoder had not themselves crapped out as I have really become much more adept at tuning the furnace since becoming able to actually measure flow rates. And I am posting this info just in case someone else is actually using a similar device. And, yes, I think I am going to install a second filter in the fuel line just upstream of the encoder. It is a delicate assembly and I'd rather not have any more drama related to it.

Denis

 

Some thoughts for you guys. I am a controls guy, I do automation for a living and spent almost the last 16 years doing pumps, pressures and flows in waste treatment plants that I designed.

That pump shown on this system appears to be a solenoid pump, which makes measuring flow with a flowmeter difficult because of the pulsations where the flow goes very high then zero. Actually if you have this kind of pump you don't even need a flowmeter, just count pulses of the pump per unit of time as that is how we do it with diaphram solenoid pumps and it is very accurate. For really low flow rates that may also pulse you want a positive displacement flow meter, typically using a pair of oval shaped gears. Not cheap!

The roller pumps mentioned are called peristaltic pumps and they are extremely tolerant of dirty liquids. In fact my first casting was the housing for that kind of pump because we needed a lot of them and they were too expensive for the very intermittent use we had pumping cleaning chemicals.

Efector IFM makes a pretty bulletproof airflow meter that uses thermal dispersion to measure air velocity which converts to volume using the ID of the pipe. About $150 bucks, which if you know anything about air flow measuring is really cheap.

I'd do this kind of DIY automation using an Arduino microprocessor, even though if this was for real money not hobby I would suggest a PLC. Absolutely hundreds of cheap sensors available that talk to an Arduino. The boards are stackable and you can get a cheap touch screen display that plugs right on. Even IR sensors that possibly could measure the furnace temperature and adjust the fuel rate to control the temperature.

 

Gary, could he just setup a restricted by pass and take his readings from that? Sort of like watching the amount of water being pumped out of a leaky boat to measure how much it's taking on?

 

Gary,

I very much appreciate your comments on the monitoring fuel flow. You are right that the pump I originally used and seen in the photos is a solenoid pump but it also has a bypass so that when flow is restricted it does not go to excessive pressure. So counting the number of positions will indeed indicate the amount of fuel that flow through the pumping chamber of the pump but it would not indicate the amount of net flow because of the bypass within the pump itself.

One thing you might not have noticed in my set up Is that vertical iron pipe approximate 2 inches in diameter naproxen 6 inches in length. That is a flow pulsation reduction (anti hammering) device. In other words there is air in that pipe and fuels from the pump enters from below and through a second fitting that fuel exits from below. Compression of the air acts like a spring to smooth out pulsations. I did that for two reasons. One was so that the fuel flow into my furnace would be more even and therefore quieter and in fact that turned out to be the case and secondly it does smooth fuel flow through the Pelton wheel. I have measured actual flow into a graduated cylinder and compared it to the flow measured by the wheel and counting device. The correlation is extremely close as you may see in a couple of earlier posts in this thread.

I have also just replaced the original pump shown in the picture with a Holley rotary fuel pump which does not cause pulsation in the fuel flow.

There may be many sensors available for measurement of airflow and I will investigate that. In fact if you had a link of to one that you thought was particularly good I’d appreciate it. However I have found that low flow rate sensors for fuel are few and far between. In fact this is the only sensor that I have been able to find it on my search of the web. High flow sensors are everywhere. Perhaps since this is your business you have better sources of information that than I do and once again if you have other sensors that might be very useful I think it be a service to post those here for my benefit and more importantly for others who might be following this thread thinking of doing something similar.

I am certainly open to any suggestions as to how this might be done more simply and reliably or with less expense.

With respect to Jason‘s question of just checking the bypass, that would require a sensor and we’re right back to needing a sensitive and reliable sensor. I’m not sure exactly what methodology you have in mind to measure that bypass. Perhaps I’m missing your point. In essence I am in fact measuring the bypass of the pump since that is the net fuel flow to my burner. The new pump in question, for instance, is rated at pumping more than 100 gallons of fuel per hour at 9PSI and in practice I am using only 2 1/2 gallons or 3 gallons per hour.

Denis

 

BYW, I did find a few cheapie sensors on eBay and grudgingly tried one prior to plunking for the German-made sensor I am using. The cheapie was worthless. It would not even turn at the low flow rates shown in its “specs.” This one is very responsive at even lower flows than I use. It came with useful documentation and it’s build quality is excellent.

 

If you know what the pump puts out in terms of volume... And you know what the bypass amount is, the difference is the consumption. You can't measure fuel flow if it's not flowing, hence a constant bypass during operation. Take your car for example. Fuel pump feeds the injectors right? Well not all of that fuel is used. Probably half of it is actually returned to the fuel tank. It gets used mostly for cooling components. If I took a measurement off that return and knowing what the pump puts out on it's own, again, the difference is the consumption. No hightech stuff needed, just a flowmeter... I have a FF transmitter off the airplane and I can fart through it and get the vanes to move. From there it's just a variable resistor. Look for an old airplane part.

I'd try one of these. https://www.ebay.com/itm/1-EA-USED-...469191?hash=item488fd0c547:g:grIAAOSwwo1XfAVW

 

I see. I could abandon the currently in-service device that is very compact (probably 1/10 the volume of the unknown detector), well-calibrated, reliable, well-documented and buy an used, maybe functional, undocumented flow meter designed for an unknown aircraft of some sort, try to figure out its signal and power requirements (airplanes are usually not 12v) and pin out and then couple that to a counter which then would be converted to a flow reading. How does that put me ahead? What I am using works very well. If it should fail, I can replace it with an identical unit in a matter of a few days.

“Aircraft electrical components operate on many different voltages both AC and DC. However, most of the aircraft systems use 115 volts (V) AC at 400 hertz (Hz) or 28 volts DC. 26 volts AC is also used in some aircraft for lighting purposes.”

Added a few mins later:


"ACTUATION METHOD ANY ACCEPTABLE

FSC APPLICATION DATA ENGINE

MEASUREMENT RANGE +200.0/+5000.0 POUNDS PER HOUR"

You see, that the sensor you linked will read in a 200 pounds to 5000 pounds per hour range. Like I said, sensors for high flow rates (new, documented, and inexpensive) are commonly available on eBay. But sensitive low-flow sensors like I am using that really work at low flow, 10-20 pounds per hour, are rare.



OVERALL HEIGHT 3.500 INCHES MAXIMUM

OVERALL LENGTH 8.290 INCHES NOMINAL

OVERALL WIDTH 2.000 INCHES MAXIMUM

TERMINAL QUANTITY 1

TERMINAL TYPE CONNECTOR, RECEPTACLE



Denis

 

No, I think I would just need one---on the line going to the burner. I really don't care about bypass and adding this and subtracting that. I just need to know what is entering the burner. So, I'll take one if your lunch box has an empty space.

Denis

 

Once again, accurate low low flow is hard to come by even at a price or a five-finger discount.

Denis

 

Dennis, what you have put together is perfect, if you're always using the same fuel.

I agree with you... why count was returned to the fuel tank or what's gone through the bypass valve. Modern vehicles don't do this. They only count what is being used.

I personally don't think there's anyway to do this cheaper than the way you've done it.

I have a specifically designed diesel fuel pump that is 24 V DC. It is diaphragm actuated and you can vary the voltage a small amount 15-28vdc. to get some small regulation, but it's not very reliable.

I have come up with a solution but it's not cheaper. It's cheaper for me because I can take things off the shelf. But... I would need charts that I would have to work from. Since I am using dual fuel, I would need two different charts with fuel temperatures as a moving target. Using a gear pump and a three phase motor with a Vfd I could tune it in perfectly using the hertz or RPM read out on the VFD. The static coefficient should always be the same unless the length of the fuel delivery lines changed. If the fuel was maintained always at a constant temperature, that moving target of viscosity could be eliminated.

 

The definitive method would be to have a fuel tank sitting on an electronic load cell hooked up to an arduino / micro of your choice and weigh the tank every so often and calculate the fuel flow from the weight loss. You would need to have a known volume and full/empty weights to begin with but it should be accurate and handle any flow rate possible with the particular tank.

 

Here's one for 20Kg that you could suspend the tank from....that may be less sensitive to ground vibrations but may swing in the wind.

https://www.ebay.com.au/itm/Load-Ce...321266?hash=item42144b4ef2:g:-m8AAOSw3gJZKPeJ

 

Mark, that would work perfectly with a hit reset button. Total weight change of fuel, divided by time, averaged out after each reset hit.

 

I have a 50 pound electronic scale that a 5 gal can will sit on. I have checked the fuel flow using the scale, the stopwatch and calculator in my phone. Does that count? Well not counting pulses from a sensor I guess. But you know what I mean.

The load cell and Arduino combo is conceptually easy. Noise issues can be mitigated by not doing simple point to point measurements, but by averaging multiple measurements of change over a convenient time interval. And some care and imagination could result in fabrication of a setup ready to take the abuse inherent in a foundry setting. I have spent a few hundred hours programming Arduinos and such a project would be pretty easy. I do like the pre-packaged plug-and-play aspect of the Pelton sensor and digital counter/display though. To each his own. Arduino projects can feel very satisfying when the code runs as planned and motors start turning and displays start spitting out data. Usually takes a fair bit of tuning and debugging to get there in my experience.

Denis

 

It's certainly not an off the shelf solution although one of those hardware counting scales would go close: you could "tare" the scales with an empty tank like Matt suggests and then fill a known amount of fuel and "teach" the scales that it represents so much volume of fuel as a quantity. So long as the counting scales don't auto power off it should count off the fuel volume down to zero fuel....would be tedious to program every single use though. It would only be a fuel gauge however without consumption rate.

 

I have two tanks (each weighing) about 30 pounds full. If a weigh scale could have a reset button that zeroed out a timer and weight and displayed an average would be really neat, like grams lost per minute averaged.

This could be all encompassed in to a really small fish hang scale.