A question for the asylum physicists (or engineers)

[Portofino]

Presume you suspect a leak ?
What have you seen ?

Conections and pipe s I would prove out 1st .

Sometimes if you inadvertantly over fill the plate that holds the filler neck and air vent on the top of the tank is pressured up and a seal maybe temporary leaks , it runs about and of course normally visable access is a problem to some of the surfaces /seams
You could pita top up in 100 L increments from its now assumed 10 % ( isolate it 1st ) and keep a watch tapping talk power on the seems /jionts ,visable every day ---


What's are they made of ,I reccal you said fibreglass ? In the DP .

 

[jfm]

I'm not sure about tanks of your size, but a standard test for testing tanks is the bubble test. i.e. dunk it under water and watch for bubbles not so easy when they're big. I'd have thought using air was a bit iffy in your scenario as it's more temperature volatile with moisture content etc and that an inert gas such as helium would be used.
Another trick from the DIY book is the soap bubble test .... fasten a kids balloon to the vent. When you fill with air and seal it should inflate and provide pressure, but not enough to deform the tank or put too much pressure on a pinhole (too much pressure and it wont bubble but just blow the soap water clear)

None of this definitively answers your question though

Those look like leak tests rather than pressure tests.

Typical boat fuel tank needs to be built for pressure too. You might have a 3-4m head of fuel if the vent pipe is high, then you need to allow say x3 for a 3g acceleration if boat falls off a wave, plus a safety factor.

 

[MapisM]

Ref 3rd para, a pressure test isn't a leak test (even if the procedures overlap). It is a test to ensure tank can withstand a certain pressure. The tests are normally done with liquid not air so that if the test is failed the "explosion" is very small. Pressurised liquid contains far far less potential kinetic energy than pressurised gas.

Aha, understood, thanks.
I suppose than for a bench test (as opposed to a tank already installed in a boat) it's also relatively easy to turn the tank upside down, hence testing both its top and its bottom with the same pressure of liquid.

Now, back to an ex-post test of a tank installed in a boat, am I right in assuming the following...?
1) let's start from a tank as the one in my previous pic, rated for 20 kPa (=200 mbar gauge reading, i.e. 1200 absolute).
2) the tank isn't actually an exact cubic meter, and has an irregular shape, but let's assume that it's 0.8 meter high max.
3) by topping the tank with fuel, what happens is that its bottom goes "automatically" to 1080 mbar absolute pressure. Well, a bit less actually, since diesel is lighter than water, but let's leave that aside.
4) when air is pumped in from its top, if we would pressurize air again at 1200 absolute, the tank bottom would actually reach 1280, i.e. 28 kPa, which is well above its "maximum allowed" test pressure of 20 kPa, as per its plate.
5) since obviously the idea is not to crack the tank rolleyes, what can be done is pressurize it with air at 1120 from its top, which brings the bottom pressure to 1200 total.

The drawback of this is that only the lower part of the tank gets tested at 1200, while the upper part would only reach 1120 - and turning a 57' boat upside down is not an option!

Then again, if the idea is to find out whether there is a (veeeery tiny!) leak somewhere, by leaving the tank empty and filling it only with air at 1200, potentially it could take ages before the leak becomes evident on the pressure gauge...

Any other ideas/comments on this are more than welcome.

In the meantime, thanks everybody for your contributions and sanity checks! :encouragement:

 

[Firefly625]

my brain hurts just reading this

 

[MapisM]

Typical boat fuel tank needs to be built for pressure too. You might have a 3-4m head of fuel if the vent pipe is high, then you need to allow say x3 for a 3g acceleration if boat falls off a wave, plus a safety factor.

I fully understand your point, but from this perspective 20kPa is not much, I reckon.
I mean, 20kPa (MAX, btw) after all means just a 2m head, without even starting to consider any dynamic effects.
Then again, I never heard of any F57 with tanks exploding after falling off a wave...

 

[MapisM]

my brain hurts just reading this

No worries there.
Not only mine, but also a few other folks' brains did, in the last couple of days...
...Hence this thread!

 

[BruceK]

Those look like leak tests rather than pressure tests.

Typical boat fuel tank needs to be built for pressure too. You might have a 3-4m head of fuel if the vent pipe is high, then you need to allow say x3 for a 3g acceleration if boat falls off a wave, plus a safety factor.

Aahh. Now I understand the question more. It's a slow morning here. Sorry Mapism

 

[MapisM]

What's are they made of ,I reccal you said fibreglass ? In the DP .

Yep, but that's just a coincidence.
And they are a bit different btw, structurally built and glassed inside the stringers, upon the hull bottom.

 

[MapisM]

Aahh. Now I understand the question more. It's a slow morning here. Sorry Mapism

No need to apologise - actually the crux of the matter is indeed a leak.
A very tiny one as I said, which for several reasons is extremely difficult to find.
The tank check is going to be done mostly to rule out that possibility, but we couldn't think of any other more leak-focused tests, other than through pressure...

 

[Bouba]

The problem with fiberglass is that leak on the outside may not be anywhere near the leak on the inside. Remove the fuel, put a vacuum cleaner on the tank opening. Then paint with resin every scratch, mark, scrape, pinhole that you can find on the outside. If you find the hole by chance the resin will get sucked into the leak and seal it

 

[BruceK]

The problem with fiberglass is that leak on the outside may not be anywhere near the leak on the inside. Remove the fuel, put a vacuum cleaner on the tank opening. Then paint with resin every scratch, mark, scrape, pinhole that you can find on the outside. If you find the hole by chance the resin will get sucked into the leak and seal it

That's clever. :encouragement: Used to use a similar technique when building large scale model aircraft to strengthen firewall boxes. If the resin and wicking surface is hot (~60 C) it becomes like penetrating oil and would also wick light woods

 

[jim.howes]

My head hurts. What on earth are all the numbers about? Make a manometer (a 'U' tube about 3' high) from a length of clear plastic hose (1/2" bore) stick to a piece of ply with clips to keep its 'U' shape. Part fill with coloured water (antifreeze is ok) leaving enough room in the tube for the water when pressurised to rise more than 12", when water has found its own level mark a line and then mark a line 12" above the level line. Remove all liquids from tank. Blank all openings in tank bar 2, attach manometer to one and attach compressed air supply to the other putting a ball valve in the line to control the flow of air into the tank , gently introduce air into the tank, as pressure builds in the tank the water in the manometer will rise when it reaches the marked line you should I believe have 2psi in the tank. With a mix of washing up liquid and water and a paint brush go around all fittings and welded seams and anywhere else you fancy, if you have a leak you should have bubbles appear. This method is safe because if you try to put too much pressure into the tank the tube is open ended so water just shoots out of the top of the 'U' tube probably drenching you, refill tube and next time try be more careful with the ball valve introducing the air.

 

[MapisM]

I appreciate that with your home made manometer you don't risk to blow air at 2 bar inside the tank, ending with a loud bang.
But being very careful while blowing air inside the tank is not so complicated - just a matter of inflating slowly and with one eye constantly on the manometer.

And ref. spotting the leak (if any), I'm afraid that what you are suggesting is easier said than done.
The tank is in fact installed in a boat, and the obvious external checks have been done already, but there are many areas of the tank which are not only inaccessible for putting soap on them, but not even visible...

As I said, the idea of the test is meant to (hopefully) rule out the possibility of a tank leakage, because that would be the worst case, pretty hard to find and fix. But at the same time, it isn't worth checking anything else if it's the tank that is leaking.

 

[rafiki_]

P, is the crux of this issue that you suspect a leak, but cannot determine where the leak is from? Pressure/leak or any other test will not really help unless you can ascertain the point of leakage surely? Is your bilge showing evidence of diesel? As others have said, most likely leakage points are hoses or connectors. I'm sure you have checked all of these? Can you smell diesel in the bilges?

 

[TheOrs]

The one litre of air is at 1200 absolute pressure. If you put the same air in a 2 litre space it' pressure is 600 mbar absolute. It's all in the maths that I set out above.

Yes. You are right of course! Post in haste, repent at leisure.

Mapis, apologies for the confusion.

 

[penberth3]

I appreciate that with your home made manometer you don't risk to blow air at 2 bar inside the tank, ending with a loud bang.
But being very careful while blowing air inside the tank is not so complicated - just a matter of inflating slowly and with one eye constantly on the manometer...

Using a compressed air supply as jim.howes suggests you risk putting 6 bar into the tank, so don't do it. A hand pump, foot pump, dinghy pump is much safer, and it won't take long to get a to a suitable test pressure.

 

[guyd]

assuming no temperature change - my favourite equation from Uni was the pervert one - PV=nRt
(pressure x volume = n(umber of moles - ie amount of stuff) x R (a constant - I forget) x t(temperature)

filling said tank with fuel / water / whatever and trying to measure a 1mbar change is a joke - the temperature will affect it far more

 

[MapisM]

P, is the crux of this issue that you suspect a leak, but cannot determine where the leak is from? Pressure/leak or any other test will not really help unless you can ascertain the point of leakage surely? Is your bilge showing evidence of diesel? As others have said, most likely leakage points are hoses or connectors. I'm sure you have checked all of these? Can you smell diesel in the bilges?

It's actually a reasoning we were discussing with some folks dealing with the problem on another boat, but yes, I see your point.
And yes, all obvious/easy possibilities were already checked.
Trouble is, the place where the bit of fuel appears (btw only occasionally), is relatively far from the tank and its connectors, but following its path backward is almost impossible without dismantling half of the boat.
So, in a sense, the idea of pressure testing the tank is in the hope of wasting time, because if it would be the tank itself that is leaking somewhere, finding the exact spot wouldn't be trivial at all...
In other words, the aim is mainly to narrow the field ATM - further investigations will depend on the outcome.

 

[MapisM]

filling said tank with fuel / water / whatever and trying to measure a 1mbar change is a joke

Agreed, when talking of filling the tank with air alone. That's precisely the doubt which popped to my mind when discussing the option of using air to test an almost empty tank - hence this thread.
But don't you think that doing the same test on a 1000L tank filled with 999L of fuel would make any leak much easier to spot, since even a small loss of liquid would create an air pressure loss of MUCH more that a few millibars?

Apropos of testing with an almost full tank, my question in point (5) of post #23 has not been addressed yet:
Am I right in assuming that if the tank rated for 20kPa is 0.8m high, it already gets 80mbar of pressure at its bottom when filled with fuel, so it would be unsafe to pressurize it with air at 200mbar from the top, because the bottom would reach 280mbar (i.e. 28kPa, well above its max rated pressure)?
Or am I missing something again...?

 

[jfm]

Apropos of testing with an almost full tank, my question in point (5) of post #23 has not been addressed yet:
Am I right in assuming that if the tank rated for 20kPa is 0.8m high, it already gets 80mbar of pressure at its bottom when filled with fuel, so it would be unsafe to pressurize it with air at 200mbar from the top, because the bottom would reach 280mbar (i.e. 28kPa, well above its max rated pressure)?
Or am I missing something again...?

Been busy/flying hence slow reply

You are 100% correct. In that example bottom of tank will always have +80 mbar due to the liquid head. Putting 200mbar air pressure at the top will make 280 at bottom.

IF you want to test top and bottom at 200mbar, and you want the tank full, and you don't want 280mbar anywhere, then you have to turn the boat upside down. Real answer is probably that the label on the ferretti tank is idiot proof: they are happy to have 20kpa compressed air at the top knowing that this means 28 at bottom, and the tank is actually designed to have a 28 test pressure. I'm only guessing though...

Alternative is to pressurise an empty tank but then its harder to spot the pressure reduction because it will be so small. Dang.

That tank rating 20kpa strikes me as surprisingly low. I'll check my equivalent data on Saturday when I'm on the boat.