Pressure Checking Fuel Cell

Discussion in 'Sheet Metal' started by Tantrum1, Mar 28, 2019.

Help Support HomeBuiltAirplanes Forum by donating:

1. Mar 28, 2019

Tantrum1

Well-Known Member

Joined:
Jul 8, 2016
Messages:
133
320
Location:
Hey all, I've finished my oil tank and my fuel tank, welded 5052 Aluminum. I need to leak test it. I know of people using air pressure, some using water with an additional "header pressure" of water. My question is I've looked in a few places and can't seem to find any specs of how much PSI(I understand that its LOW PSI) or amount of water column to use. Does anyone have any reference material regarding this.
I have a regulator and I can just test them to 2-3 PSI. I'm just wondering if there's anything published.
Thanks
Mark

2. Mar 28, 2019

TFF

Well-Known Member

Joined:
Apr 28, 2010
Messages:
11,411
3,161
Location:
Memphis, TN
2-3 psi sounds right. I would still want it to bleed out at that psi. You don’t want to turn it into a balloon and miss shape or pop it. The nicer setups have a regulator on an out so the tank can’t get to more than what you want.

3. Mar 28, 2019

wsimpso1

Well-Known Member

Joined:
Oct 19, 2003
Messages:
5,712
2,952
Location:
Saline Michigan
I do not know about regs or even standards on this, but this experienced engineer has done some Failure Mode Analysis on the topic. Biggie seems to be "What is the biggest pressure that can be applied to the tank in use?"

If your header tank is your only tank, and you have a forward facing vent, you can apply dynamic pressure from Vdive to the vent opening and pressurize the tank at that level.

If the tank forms the outside skin of the airplane, you will also have air around the tank at reduced pressures due to Vd * V/v for the part of the airplane - air rushing by the outside of the tank.

Remember to change lb/ft^2 that these are normally calculated into down to lb/in^2.

If you are pumping to the tank from somewhere else, and there is no vent or the vent gets blocked, you can get to the maximum of the pump pressure to the tank.

Typical factor of safety for an aluminum structures in airplanes is 1.5, but on welded stuff 2.0 might make a lot more sense.

Might want to think your way through that sort of thing, and maybe come back to us for a sanity check on your numbers.

Bill

Last edited: Mar 28, 2019
4. Mar 28, 2019

rv6ejguy

Well-Known Member

Joined:
Jun 26, 2012
Messages:
3,747
2,767
Location:
Some people do a balloon test. Pump up a balloon and secure it to the vent line. Leave it for a couple days. If it's deflated a lot, then apply soapy water on all the seams and find the leak. Be careful with applying air pressure, even 2-3 psi can bulge a tank with large flat sides in some cases.

5. Mar 28, 2019

kent Ashton

Well-Known Member

Joined:
Aug 15, 2008
Messages:
715
202
Location:
Concord, NC
6. Mar 28, 2019

Marc Zeitlin

Well-Known Member

Joined:
Dec 11, 2015
Messages:
447
613
Location:
Tehachapi, CA
Along with the aero pressure forces you described, you ALSO have to consider the head pressure from the depth of the tank. If your tank is 12" deep, the pressure at 1G at the bottom of the tank from a foot of gasoline will be about 1/3 psi. But if your airplane is aerobatic, and you've designed it with a 6G limit load, the tank will see 2 psi from the head pressure loads alone, and will be added to the dynamic pressure. This head pressure will far overwhelm the dynamic pressure from the vent pointing forward, which at SL and 120 mph will be about 1/4 psi.

So if you only test to 1/4 psi, or slightly more than that, you will not have tested to the maximum pressure that your tank could see in flight. Obviously, if you have a plane stressed to 3.8G and your tank is only 6" deep, the "G" load numbers will scale accordingly - in this case, 0.6 psi instead of 2 psi. Which is still far more than the 1/4 psi from the dynamic pressure. For faster planes in the "normal" type category, the dynamic pressure may be more than the "G" pressure - at 240 mph, the vent dynamic pressure will be about 1 psi.

So whatever your plane is capable of, you determine the maximum pressure that it could see in flight when at its limits and then use Bill's 1.5 or 2.0 safety factor, then pressurize to that to test.

wsimpso1 likes this.
7. Mar 28, 2019

Dan Thomas

Well-Known Member

Joined:
Sep 18, 2008
Messages:
4,784
2,013
That's what I've done. If the tank is properly built, we're not looking to test the strength of it. We just want to make sure it doesn't leak. A balloon is the way to do it. Much more pressure than that can damage a perfectly good tank.

proppastie likes this.
8. Mar 28, 2019

Dan Thomas

Well-Known Member

Joined:
Sep 18, 2008
Messages:
4,784
2,013
Not all tanks use full dynamic pressure for venting. Vents on faster airplanes are often designed to provide only ambient or slightly more pressure. The strut=braced Cessnas have the forward-facing vent behind the wing strut in a strictly-defined location (which we often found incorrect, as mechanics ignored the manuals), and others like the 177 or 210 have aft-facing tubes between the outer aft ends of the ailerons and the wingtip ribs. No dynamic pressure at all there.

I think of the typical wing tank. Suppose it's a 20-gallon tank averaging about six inches deep and about 24 x 36 inches horizontally. Those horizontal surfaces, top and bottom, would have over 2100 pounds of force on each surface with 2.5 psi in the tank. Those surfaces had better be very well supported and braced if you want the tank to survive. The front, back and sides would also have a lot of force trying to blow them out.

9. Mar 28, 2019

Chilton

Well-Known Member

Joined:
Jun 28, 2014
Messages:
63
15
Location:
Arusha
I know FAR 23 does not apply to EAB, but the answer could still be usefull,

§ 23.965 Fuel tank tests.
(a) Each fuel tank must be able to withstand the following pressures without failure or leakage:

(1) For each conventional metal tank and nonmetallic tank with walls not supported by the airplane structure, a pressure of 3.5 p.s.i., or that pressure developed during maximum ultimate acceleration with a full tank, whichever is greater.

Marc Zeitlin's answer above covered the acceleration with a full tank, so the 3.5 psi may be the limit, or the head pressure may be.

I have found that Avgas can find leaks which a pressure test with water did not, apparently water molecules are bigger than gas!

proppastie likes this.
10. Mar 28, 2019

BBerson

Well-Known MemberHBA Supporter

Joined:
Dec 16, 2007
Messages:
11,648
2,183
Location:
Port Townsend WA
At the factory, we did the FAR 23 test (3 psi at that time)
The tank is filled with water and a clear plastic hose is connected to 8 feet high above and filled with water. That puts 3 psi pressure head in the tank.

11. Mar 28, 2019

proppastie

Well-Known Member

Joined:
Feb 19, 2012
Messages:
3,651
1,048
Location:
NJ
I like this method better than the balloon, as a balloon can loose pressure through the latex. To just check for leaks you would not have to go 8 ft. high. To find the leaks though the balloon will assure you do not over-pressure the tank.

12. Apr 1, 2019

Dan Thomas

Well-Known Member

Joined:
Sep 18, 2008
Messages:
4,784
2,013
That pressure would test the tank, alright, but would destroy an aluminum tank by turning it into the closest thing to a sphere. It's meant to determine that the tank's design won't rupture in a minor crash, not to test a particular tank for leakage.

The other thing to know: Water is more than 800 times the desnity of air, so it takes a lot longer for any leak to appear. Air and soapy water will find leaks far more reliably.

Last edited: Apr 1, 2019
13. Apr 1, 2019

BJC

Well-Known Member

Joined:
Oct 7, 2013
Messages:
9,289
6,060
Location:
97FL, Florida, USA
Water is useful for high pressure testing of pressure vessels (“hydrostatic pressure testing”) because rupture is not an explosive event with the non-compressible water rather than a gas.

Definitely not needed to test the OP’s equipment.

BJC

14. Apr 1, 2019

pictsidhe

Well-Known Member

Joined:
Jul 15, 2014
Messages:
6,408
1,660
Location:
North Carolina
g loads will multiply the pressure in a tank, that's why several psi are specified in far23.
A tank a foot deep, at 6g will have about 2psi on the bottom. Add in 50% safety factor, we are up to 3psi. Maybe your tank is shallow, but if it's long, then you need to consider end loads in extreme or unplanned manuoevers, crashes, too. A split tank is not going to be welcome. I've found FAR23 to be somewhat conservative, but well thought out. Unless you are doing a very detailed analysis, it's prudent to just follow it or at least take it as highly recommended. You can probably skimp 10, 20%. But IMHO, go past that, you are getting out onto thin ice. Some rough numbers for my project a few weeks back showed that using 0.008" roof flashing and round sides, I can easily meet 3.5psi.

3.5psi isn't going to go boom in a particularly nasty way, but having a simple water column is a very easy way to get an exact low psi. A little fluorescein in the water and you can hunt pinholes too.

15. Apr 1, 2019

proppastie

Well-Known Member

Joined:
Feb 19, 2012
Messages:
3,651
1,048
Location:
NJ
The builder I know checked with a water column....but I neglected to say .....he let the column sit for 3 days to check if it changed height.

cvairwerks likes this.
16. Apr 1, 2019

Himat

Well-Known Member

Joined:
May 5, 2011
Messages:
2,763
636
Location:
Norway
To test for leaks, fill with helium. Blank all inlet/outlets and use a helium gas detector to check for leaks. That is the procedure at work for testing pressure vessels for subsea equipment.

17. Apr 1, 2019

Pops

Well-Known Member

Joined:
Jan 1, 2013
Messages:
6,910
5,823
Location:
USA.
Always hated the Hydrostatic pressure test after replacing tubes on the coal fired power plant boiler.
One time went in the plant for a Hydro test at 11 Pm on Tuesday and got out at 10 PM on Sunday. Pump broke down. All double time, in and out pay, made a good paycheck.

18. Apr 1, 2019

BJC

Well-Known Member

Joined:
Oct 7, 2013
Messages:
9,289
6,060
Location:
97FL, Florida, USA
Yup, BYDT.

BJC

19. Apr 1, 2019

BBerson

Well-Known MemberHBA Supporter

Joined:
Dec 16, 2007
Messages:
11,648
2,183
Location:
Port Townsend WA
3 psi is for the one time certification test. 3 psi might crumple it some. For finding leaks I would pressurize the tank with a household vacuum output. About 1psi, I think.
Soapy water on the outside to check for bubbles.

proppastie likes this.
20. Apr 1, 2019

Joined:
Jan 1, 2013
Messages:
6,910