Calling all geeks

[unixronin]

We have a 500-gallon outside propane tank. It's maybe 30 feet from the house.  It has a gauge on top, under an armored cover, but the tank is frequently on the far side of (and frequently covered in) significant amounts of snow.

Does anyone happen to know of, or have any suggestions for, a way I can remotely monitor the level of propane in the tank?  Obviously anything that would involve retrofitting a sensor of some kind inside the tank is pretty much out.

Edited to add:  I need to get data back in a form that can just be logged and charted automatically 24 hours a day, preferably without having to do image recognition on webcam images of the gauge.

Comments

[gdmusumeci.livejournal.com]

How about a flow meter?

[unixronin.livejournal.com]

Yeah, I pondered the possibility of monitoring flow at the point the line enters the house. I suspect it's a more difficult problem than just monitoring the level, though.

[n5red.livejournal.com]

A web cam with a macro lens?

[johnkzin.livejournal.com]

Or a remotely placed, weather protected (like in a clear pelican/otter box), web cam.


He said that the gauge is on the wrong side of the tank... and frequently covered in snow. What about a small roof for the tank, and well placed mirrors so that the location/orientation of the gauge isn't a problem, and it doesn't get covered in snow?

[unixronin.livejournal.com]

Well, no, the gauge is on top of the tank. But it's under a hinged steel cover about 3/16" thick designed to protect the gauge and valves. If I had a webcam out there, I'd have to have a light for it so it could see the gauge, and I'd have to have a way to supply power to both.

[glitch25.livejournal.com]

I take it you mean this is out? http://www.amazon.com/Robertshaw-Wireless-Propane-Level-RS-228/dp/B000IXILAK

[unixronin.livejournal.com]

That looks like it might have possibilities, actually. Doesn't look like there's any easy way to extract a digital reading from it, but at least it'd give us a remote monitor inside the house.

Not having much joy looking for anything on how it works.

[rbos.livejournal.com]

Throw rocks at it and learn what "empty" sounds like.

[glitch25.livejournal.com]

Or use a remote actuated mallet and a sensor array to "hear" what empty sounds like. :-)

Tank level

[ewen]

The first thing that occurred to me would be to measure the tank weight from below. Scales with an electronic output are relatively easily obtained for scientific purposes. There's a potential niggle, though, that the weight may vary depending on temperature as well as volume of contents (as the proportion of gas and liquid changes), but possibly there is an algorithm to compensate for that if you also have the temperature figure (which is even more easily obtained).

The other option is that the guage on the outside is presumably a pressure guage. I'd imagine it should be possible to purchase a pressure guage with electronic output at some (probably non-trivial) expense. It does require ensuring the tank is then pressure sealed again, which simply sitting the tank on scales doesn't.

Ewen

PS: Given more convenient weather conditions I believe one can observe the levels from the outside by the proportion of condensation on the outside of the tank. But lack of convenient weather conditions seems to be a major sticking point here...

Re: Tank level

[unixronin.livejournal.com]

Yeah, when the weather conditions are conducive to observing condensation it's probably warm enough that we're not using the heating. Replacing the gauge isn't an option, as it is very unlikely Eastern Propane would allow us to modify the tank.

Re: Tank level

[cipherpunk.livejournal.com]

The weight is fixed. Whether the temperature is 50C or -50C, if the tank is dry weight plus 10,000L of propane at STP, the tank will still be dry weight plus 10,000L of propane at STP.

The volume of gas contained within the tank will vary according to local temperature, but that's easily determined from Boyle's Law once you have the contents determined.

A piezoelectric weight gauge is, IMHO, the best option here.

Re: Tank level

[ithildae.livejournal.com]

The snow on the tank with throw the weight off. The snow will also distort the sonics of a percussive form of measurement. The pressure will be dependent on Temperature. I am stumped for a reasonable way of looking inside a metal container. Some sort of flow measure would be the best I can think of, off the top of my head.

[Thought] Is it possible to introduce a radio signal? There may be a frequency that propane liquid is opaque to. The chamber should have a resonant frequency based on the gas chamber volume.

Re: Tank level

[cipherpunk.livejournal.com]

Density of propane: 508 kg/m^3
Density of snow: 50-150kg/m^3

The relative densities are such that it will take a lot of snow to throw off the calculations by an appreciable amount.

You can also take the weight of the snow into account. Look at how much the tank weighed before the blizzard, then look at how much it weighed after the blizzard. The difference will be a fairly good approximation of the weight of the snow on the tank, which can then be factored into ongoing calculations.

Piezo is simple, effective, and accurate enough. It's still the one to beat, IMO.

Re: Tank level

[ithildae.livejournal.com]

I will take your word for piezo accuracy. I would have thought that a 10 - 30% density decrease for snow would still make a noticeable difference to the weight. (But I considered an error > 5% from theoretical to be a personal insult in my college labs.)

Would it be possible to measure under just one anchor point?

Re: Tank level

[cipherpunk.livejournal.com]

Well, like I said, you can account for most of the accumulated snow weight just by measuring tank weight before and after the blizzard. Admittedly, some propane will get used during the blizzard, but as far as first-order estimates go it's not a bad one.

I sympathize with you on your within-5% instincts. It's an excellent ideal to live up to. In this case, though, we can live with 50% experimental error. If unixronin measures the tank as being 10% full, then it's time to refill the tank regardless of whether it's 5% full or 15% full. Quick and dirty is good enough for our purposes.

I think it's possible to measure under one anchor point, yes, esp. if it's done under the geometric center of mass.

Re: Tank level

[unixronin.livejournal.com]

There's actually two primary goals here:

(1) Get an actual metric for the rate we're using propane in various combinations of use
(2) When the delivery slip says 300 gallons, be able to verify that the driver really did put 300 gallons in.

Apparently a previous delivery driver in this area was caught over-billing customers and putting the excess propane into his own tank. He got free propane, while the customers on his route paid for propane they weren't getting.

Re: Tank level

[unixronin.livejournal.com]

Snow on the tank will throw the weight off, yes, but not by much. The maximum snow load on the tank might be the equivalent of a gallon or two of propane.

[jhetley.livejournal.com]

Remote-reading gauges exist, but probably don't solve your immediate problem.

[unixronin.livejournal.com]

Yeah, I doubt very much that Eastern Propane would allow us to actually modify the tank.

[dafydd.livejournal.com]

Have you talked to Eastern Propane to see if they have any solutions?

[unixronin.livejournal.com]

Not yet, but I plan on asking them. I don't expect them to have much to offer in the way of help, but they may be willing to let me put a remote sender onto the gauge.

[cipherpunk.livejournal.com]

Piezoelectric pressure sensor. Armed with the dry weight of your tank and the current weight you can figure out the current contents. Liters of propane at STP can be converted into whatever other metric you want.

[unixronin.livejournal.com]

Yeah, a load cell would be the best solution. I don't know the dry weight of the tank, but I could derive calibration points from the readings at various known gauge levels. The problem is I'd need to put the load cells at the tank mounting points, and that's going to be a bit of a challenge on an already-mounted tank that already contains over a thousand liters of liquid propane.

Another approach I thought about is using an ultrasonic transducer to measure the depth of liquid in the tank, either from the top or from the bottom. In principle, I should be able to bounce a ping off the liquid-gas interface and, again, determine calibration points relative to the gauge.

[cipherpunk.livejournal.com]

I'm assuming it's on a four-piling mount? Put a fifth support piling in the geometric center. Mount the piezo on that. Use a car jack to lift it into position, then slide the rest of the piling underneath. Take a baseline measurement when you have 1000 liters. The liquid density of propane is mostly temperature insensitive, so it shouldn't be too hard to extrapolate from there to get the numbers you need.

[unixronin.livejournal.com]

I'm assuming it's on a four-piling mount?

Presumably, yes. I won't have another chance to check and make sure until the next thaw, unless I want to crawl around on my knees in the snow, which doesn't seem like a good idea considering how much the cold seems to be bothering them. Another likely possibility is a bracket and pad at each end.

[jhetley.livejournal.com]

Working on _anything_ outside will present problems until maybe April . . .

[jhetley.livejournal.com]

One technical consideration, also related to weather. Remember that the ground moves a lot -- freeze/thaw, frost-heave, moisture expansion of clay soil, whatever, and that your tank probably moves irregularly because of that. Support under one end of the tank may freeze or thaw before the other, etc...

Unless the whole thing sits on a slab.

[unixronin.livejournal.com]

Which I probably won't be able to check until ... oh, about April. ;)

[http://users.livejournal.com/crossfire_/]

Most propane tank level sensors I know are made out of temperature sensors. Propane + metal absorbs heat better than just metal, so the fill line is the point where the external temp drops on the propane tank.

These tend to get a little fuzzy as the tank level gets very low, but you'd have historical data to look at to determine that your tank was actually low as opposed to the fuzzy data being due to equipment malfunction.

[novel-tharkun.livejournal.com]

Any knowledge of the sensor type on there's method? Does it have a cutoff valve below it? If so, it's probably a pressure (possibly temperature-compensated) gauge, and there are several pressure gauges with digital transmission. Batteries are generally cost semi-reasonable and can last from 6 mos to a few years. Issues might be if it needs a thermostat & heater, and that your supplier might demands a explosion proof or hazard class I device, which are pricier.