Shuttle Foam Test Yields Hole in Wing (Produces vs Yields)

AP ^ | 07/07/2003 | MARCIA DUNN, AP Aerospace Writer

[DoughtyOne]

Science - AP

Shuttle Foam Test Yields Hole in Wing

59 minutes ago

By MARCIA DUNN, AP Aerospace Writer

SAN ANTONIO - The team investigating the Columbia disaster fired a chunk of foam insulation at shuttle wing parts Monday and blew open a gaping 2-foot hole, offering dramatic evidence to support the theory of what doomed the spaceship.

The crowd of about 100 gasped and cried, "Wow!" when the foam hit.

The foam struck roughly the same spot where insulation that broke off Columbia's big external fuel tank during launch smashed into the shuttle's wing. Investigators believe the damage led to the ship's destruction during re-entry over Texas in February, killing all seven astronauts.

It was the seventh and final foam-impact test by the Columbia Accident Investigation Board, and it yielded by far the most severe damage.

The 1.67-pound piece of fuel tank foam insulation shot out of a 35-foot nitrogen-pressurized gun and slammed into a carbon-reinforced panel removed from shuttle Atlantis.

The countdown boomed through loudspeakers, and the crack of the foam coming out at more than 500 mph reverberated in the field where the test was conducted.

Twelve high-speed cameras — six inside the wing mock-up and six outside — captured the event. Hundreds of sensors registered movements, stresses and other conditions.

NASA (news - web sites) will continue gathering more information about the poorly understood pieces that line the vulnerable leading edges of shuttle wings, board member Scott Hubbard said.

One month ago, another carbon shuttle wing panel — smaller and farther inboard — was cracked by the impact, in addition to an adjoining seal. This time, the entire 11 1/2-inch width of the foam chunk — rather than just a corner during previous tests — hit the wing, putting maximum stress on the suspect area.

[El Gato]

That's gotta have alot of volume of air/nitrogen or whatever they're using to be able to launch a non-tube conforming 'round'

They use a sabot to seal the "payload" from the propellent gas, and have a "stripper" that stops the sabot, but not the payload, be it chicken or foam. Although in the case of the foam, I think, but am not certain, that they put on a new tube that was the shape they wanted. Pictures show it basically square in cross section. I don't know how they "valve" the compressed gas into the barrel behind the sabot, possibly by puncturing a diaphram, as I've read that's the way some similar "guns" do it.

[fnord]

So the foam travelled a relative distance of 53.4 feet in .252 seconds, right?

Initial velocity of 682.5 m/s
Initial acceleration of 12.1 m/s2

Is this what you started with in your calculations?

[Resolute]

No, those were the approximate initial conditions for launcher & orbiter (cojoined) at the approximate time of foam separation. I derived the initial conditions by applying curve fit to the data presented here:

Observation on TPS damage on Orbiter (Post 2431)

The initial acceleration (actually deceleration) of the foam is 1329.0 m/s2, applied in the opposite direction to the boosting acceleration of the launcher and orbiter. The deceleration due to drag decreases rapidly as the foam slows down: Drag is proportional to the airspeed squared, assuming a constant drag coefficient (in this case 0.65).

[Resolute]

You didn't ask for this, but here is the Fortran source code that I used to estimate the foam separation "dynamics" with respect the axial direction (no lateral or rotational dynamics):

program test

data dx0 / 0.0 / ! m
data rho / 0.08823 / ! kg/m^3
data press / 7345.4 / ! N/m2
data xmass / 1.211 / ! kg
data sref / 0.120492 / ! m^2
data g / 9.80665 / ! m/s^2
data v0 / 682.5 / ! m/s
data deltat / 0.00001 / ! s
data tfinal / 0.5 / ! s
data dtout / 0.030 / ! s
data dteps / 0.000001 / ! s
data cd / 0.65 / ! non-dimensional

vmag = v0
dx = dx0

nsteps = nint ( tfinal / deltat )

do i = 0 , nsteps

time = deltat*float ( i )

! compute drag force

Q = 0.5*rho*vmag**2
drag = cd*Q*sref

! integrate relative deceleration, including drag +
! gravity + powered boost

acc_new = - drag/xmass - g - 12.14
if ( i.ne.0 ) then
vmag = vmag + 0.5*deltat*( acc_new + acc_old )
end if

! integrate relative velocity

vrel = v0 - vmag
if ( i.ne.0 ) then
dx = dx + 0.5*deltat*( vrel + vrel_old )
end if

! save old states for trapezoidal integration

acc_old = acc_new
vrel_old = vrel_new

! check cumulative displacement, exit when threshold is met

if ( dx/0.3048.ge.53.37 ) then
write(*,1) time,vrel/0.447028,dx/0.3048,Cd
write(8,1) time,vrel/0.447028,dx/0.3048,Cd
exit
end if

! output the relative velocity and translation at scheduled intervals

test = time - dtout*anint ( time/dtout )
if ( abs(test).le.dteps ) then
write(*,1) time,vrel/0.447028,dx/0.3048
write(8,1) time,vrel/0.447028,dx/0.3048
1 format(f9.6,3f9.2)
end if

end do

pause 'end of sim'

end

[Darksheare]

Coolbeans.
Being a former artillery soldier, I'm used to the 'old fashioned' way of sending something on it's way.

Thanks for the info.