Shuttle Disaster 1

[KLH]

Is it true that the foam insulation that struck the shuttle wing was going +500 mph?? (According to the reports on CNN anyway.)

At the time of the foam separation, both the shuttle and the insulation were moving at the same velocity.

When the foam insulation separated, did its velocity slow enough in that short time (between separation and impact) so that the relative velocities of the shuttle and the insulation exceeded 500 mph?

Can someone explain this to me?

 

[astroclone]

I kind of felt that the report was almost of a religious nature and he's gonna start telling people that Dianetics is the way and so on so forth. I just don't see how if this is such a HUGE problem then it isn't treated in practice or textbooks.
T

 

[IRstuff]

That's easy to answer:

Because, He is the ONLY person with the correct insight.

TTFN

 

[JimMetalsCeramics]

Read sections 7 and 8.2 to 8.5. The phenomenon is real, it really did occur and it was neglected in shuttle design, despite your protestations.

AbuTaha has been an aerospace mechanics expert for over 30 years. The documentation shows that the phenomenon really occured. In fact, this is the easiest argument to make. If you don't think the phenomenon exists and is applied to the shuttle, then how does 1.1 million pounds of thrust from the main engines acting upward, and 200,000 lbs (the weight of the orbiter) acting downward, at a distance of 400 inches from the SRB centerline, generate a maximum measured base bending moment of 700 million inch-pounds?

Why does the vehicle lurch forward substantially past its steady-state defelection point?

 

[IRstuff]

One issue I have is that he's using model rocket thrust characteristics as "proof" of dynamic overshoot, while NASA's page

http://www.grc.nasa.gov/

http://WWW/K-12/airplane/rktengperf.html

on model rocketry clearly shows the same data, but alludes to flame front characteristics. which is a design parameter:

"Designers of solid rockets can produce the given thrust curves by changing the total amount of propellant placed in the engine, by varying the the angle of the cone in the propellant, and by varying the diameter of the propellant (and casing)."

He then compares the model rocket test data with Thiokol test data and says that NASA "missed" the dynamic overshoot. BUT, since Thiokol's data is also measured data, he fails to explain why the test data does not show dynamic overshoot.

The simple interpretation is that Thiokol probably designed the SRB specifically to eliminate the thrust overshoot, since that appears to be a very well known phenomenon of solid rockets.

This is particularly relevant since the SRB does not use the same physical design as a model rocket. A model rocket simply burns from the nozzle forward, while the SRB is a hollow core motor that is ignited along its entire length and actually burns radially, thus signficantly altering the flame front and the thrust characteristics.

TTFN

 

[JimMetalsCeramics]

The excess thrust comes overwhelmingly from transient response and not propellant geometry. Insofar as overshoot being well known; this information, and the acknowledgment of the initial mistake have never been disseminated. Thiokol signed off on the o-ring joint failing on Challenger.

In short, the key areas of stress on the overall launch vehicle are those directly impacted by main engine and SRB thrust. Overshoot occurs during start-up and during engine throttling (during flight). The structural design of the shuttle not only was void of proper understanding of excess effective structural forces due to transient response(ranging from axial tension on the SRB housing to various truss and strut components between the SRBs and the orbiter), but this omission perfectly explains numerous cases of vehicle and payload damage going back to the beginning of the shuttle test and flight history.

 

[GregLocock]

Jim, his misapplication of the quote from Shigley is just plain wrong. Force doubling is NOT a general property of dynamic systems.

I have no problem with high forces being generated as a result of excitation of resonances... it has been my bread and butter for 25 years. But given that NASA had performed a full modal analysis on the shuttle 27 years ago I am reasonably happy to assume that they have a dynamic model that will model 'overshoots' (transients would be a more usual terminology). It is not, if you will forgive the phrase, rocket science.



Cheers

Greg Locock

 

[JimMetalsCeramics]

Greg,

NASTRAN has no transient response module. I have found this matter to be puzzling to engineers, even those working in aerospace(I am familiar with this matter for 15 years now). Certain conditions must be satisfied in order to have dynamic overshoot. The key one is that the system be under-damped.

Is it your contention that there is no first order overshoot of the system response to the SSME and SRB thrust?

 

[JimMetalsCeramics]

I'd also like to add that the manner over the years in which extensive structural damage to the launch vehicle and failure of payloads taken into orbit, has been sloughed over and rationalized is disgusting.

 

[GregLocock]

"NASTRAN has no transient response module."

Yes it does. Time domain modelling based on known force signals is routinely performed all over the world by thousands, if not millions, of people.

I'm not an FEA dude, so I can't tell you the exact commands but, roughly, build FE model, do a static analysis, correlate, do a dynamic analysis (to get the frequency response functions), correlate, build time based force signals, apply to dynamic model, look at pretty pictures on screen.

We then take those time varying stresses and do a fatigue analysis using them.

This is so fundamental that I think I must be misunderstanding what you are talking about.

"Is it your contention that there is no first order overshoot of the system response to the SSME and SRB thrust? "

No. I am saying that the tools to analyse this have existed for decades and that it is a well known property of systems.



Cheers

Greg Locock

 

[JimMetalsCeramics]

If "proper" analysis exists in NASTRAN then this is new. I've had dynamics experts in industry admit that this topic was not properly taught until the early 70s. I had a former undersecretary of Defense tell me that it was clear to him when he attended early shuttle tests that NASA was lost on the mechanical behavior of the system.

The omission of correct transient response from the shuttle design is clear. Let's look at two simple examples from the paper.

Section 7, pg. 23, Fig. 15

This represents data from a flight readiness firing test (FRF). The main engines are ignited and the SRBs are held down. The holddown posts have strain gauges on them to record the force from the main engines.

Using three different calibration methods, the test engineers measure "EXCESS UPWARD FORCE" in excess of 500,000 lbs.

Pg. 24, Fig. 16, "MEASURED DATA QUESTIONABLE -- TOTAL (DELTA) LOAD > SSME THRUST"

Pg. 28, Fig. 17, "REQUESTED OF LEVEL II A VERIFICATION OF THE PRE-IGNITION 1-G CALCULATED LOADS

They are clearly expecting to measure, according to their specification books, a force equivalent to the maximum applied steady state force from the main engine thrust. They see the correct experimental force which includes a roughly 50% overshoot (not the applied force but the system response). They are bewildered.

Section 8.4. Figures 27 and 30. When comparing the two plots for the same measurement, both the max and min numbers for bending moment vs. time are irreconcilable.

Throughout the published shuttle specifications the values you will find are for maximum applied force, sometimes with up to a 14% dynamic factor added.

 

[scotland6]

I don't know exactly how or if they relate to each other but when designing for a shock load, specifically a half sine pulse, the safest approach for a dynamic system is to double the static load. This is a cover all and the actual amplification factor depends on the system natural frequency and damping coefficient but it can never be more than approx. 1.8 times the static load. This allows for an overshoot (apparent increase in force) just after it is applied.
I can't believe that engineers at NASA don't work with something along these lines when doing the mechanical design of the rockets on the shuttle, it is fundamental.

Oh and the report makes interesting reading. I am working my way through it!!

 

[ethical]

JimMetalsCeramics,

Please send a copy of the report to:

briandolezal@mail2engineer.com

 

[astrosag06]

May I have a copy of this report as well?

sts_1OO@yahoo.com the "OO" are letters O's not number zeros. thank you, I appreciate it.

"Man is the best computer we can put aboard a spacecraft... and the only one that can be mass-produced with unskilled labor. "
-Wernher von Braun

 

[dpadler1]

JimMetalsCeramics,

Could I still get a copy? Sorry so late, forgot to mark the thread.

dpadler1@yahoo.com

 

[GraviMan]

I have used Natran for dynamic analysis for fatigue proofing. It is capable of dynamic modelling, but you have to know the velocity as a function of time (seismic mass method). This is difficult if the system is an SRB in the process of being lit-up.

I used to be a non-linear (explicit) FEA guy, so I always mistrust linear results. The norm is to use a half sine pulse, which assumes an upper frequency limit to the dirac (shock) function. This is a fair assumption, since higher frequencies tend to get "absorbed" by any structure.

The big problem comes from the assumption of linearity. The Poisson distribution has already proven to be a mislead to the shipping industry in statistical expectation of ocean wave sizes (turns out to need Schrodinger wave equation to model accurately). Add to that the fact that the structure will deform as a result of vibrations, likely to be large with high tensile materials, and the assumptions soon add up.

I have no read this report, but a statisticion friend of mine used to have grave doubts about the validity of the 6-sigma approach for ensuring that gas turbine white noise vibrations actually gave as little as 0.04% (from memory) that the wings would fall off!

In practice what this really means is that the material ages through fatigue sooner than the calculations would suggest. Basically as the aircraft ages, structural problems will appear sooner than they ought. This tends to affect panels more than say structural members, and could explain why occasionally panels fall off!

Just my thre'pence ha'penny...

Mart

 

[JimMetalsCeramics]

GraviMan has got it. This is a case of low cycle fatigue. The history of the shuttle launch vehicles has been riddled with both major, single launch component failures which are forestalled through empirical strengthening of the system, and low cycle fatigue of components far earlier than initial calculations would have predicted. Everyone asks, if this mistake was made, why doesn't the shuttle explode on every launch. This is because:

1) Systems are spec'ed to minimal material property values,

2) There are finite safety factors, albeit inadequate for long term use, built into the shuttle vehicle structure, and

3) The system has been empirically strengthened over the years.

In the latter case, the SRB segments, for example, are held together with 1" diameter bolts. This necessitates that hole spacing be at least 1.5" (it was 1.51&quot from edge-to-edge in order to avoid one tear out event leading to the adjacent hole and so on. After initial failures of these joints (no doubt due to excessive axial force transients) extra 1" bolts were added between each initial one. Now the spacing is 0.25". OSHA, let alone a pressure vessel designer, would jam on the brakes.

 

[GraviMan]

Thanks JimMetalsCeramics.

Any chance you could bung the report to
martingarrish@aol.com ?

I'm starting to feel left out not having bitten yet!

Mart

 

[JimMetalsCeramics]

Any more responses to the report?

 

[CusterWilson]

In May 1988 Ali AbuTaha's Spaceflight Cover article forecast the Columbia Loss.
Personally, I believe Challenger died for a different reason, however, as his calculations showed areas of the vehicle would have pre-existing metal fatigue my own mechanism -- small explosions in the fuel prompted by over-fast5 drying -- would result in tongues of flame emerging, preferentially, just above the struts that transmit the sideways overstrain (vertical strains are all higher by 40% later in flight because "vacuum thrust" is higher and this, plus 10%, was assumed to cancel the transients but a 1981 launch timing change was not a full redesign and missed that sideways stresses have no such extra 40% -- in 1970 they Launched after 1 second of SSME testing not 6.6, so the vehicle never bent all the way over --Ali wanted to change back but a launch partly bent over might jam one of the SRB clamps; the Companies adopted new Requirements for higher structural strength in effect ENDORSING his claims in 1989 -- Congress refused to spend money to change anything but the new AL-Li ET must meet them as it has only half the expected cargo gain) ...
Now image YOU are a Shuttle --orbiter on your back, Solids your arms: the easily seen flame comes out of your wrist about half forward (35 and 37 degrees) hitting the ET just right of the belly button --- but there are two struts, and a flame came out angled BACK about 37 degrees from directly facing the ET, hitting the Right wing Aileron on the underside of the Right wing (note similar damage on both wings on STS-56, -90, and -107, -90 considered Most Similar Mission to -107 as they both had the same weird 1-push Wind Shears) ... now this impact is AT THE SAME MOMENT that Challenger responded to something the Rogers Commission VOTED was a Wind Shear, however, when they finally figured out Wind Shears at the Weather Bureau 2 years later: all such were circular: there should be a Paired Shear in the opposite direction.
There isn't.
Instead, on fetching Challenger pieces from the missile silos out West where they were stored, they found a fist-sized burn through EXACTLY at the point of impact expected from the PICTURES of the flight -- which Ali was able to find for the furious Weather Services (cf Young's tirades against them for killing his friends by not forecasting Challenger's Wind Shear & CHUCK YEAGHER'S BOYCOTT OF THE CHALLENGER MEETINGS AFTER THE COMMISSION ORDERED NASA TO STOP LOOKING AT PARTS OF THE FLIGHT THAT DID NOT ADVANCE THE O-RING THEORY).
Now as Columbia was also launched in a Florida Drought (with SUVs falling into sinkholes) AND just after a Deluge --just like Challenger -- one should expect a similar hole &
so we see the hole in Columbia's elevon actuator -- the Only piece where the new commission did NOT tell reporters "we have to wait for the chemical and other analyses" but dismissed it as occurring AFTER the breakup (Challenger's hole did NOT match other post-breakup heating and anyway from the moment of impact that Ali discovered her motions clearly matched the the way a computer would double-correct for motions with a control surface damaged -- not applicable here).
In short, had Challenger not had a hole directly facing the ET, and so reached orbit (yeah, the ET did not even break up from opposite the "joint that failed" but higher up the side where a welded seam on the Solid obviously busted)....
... its reentry would have seen the wide area around the burn-through, where heating makes the aluminum less stiff, flex the wing resulting in breaks where it adjoints the almost unbendable RCC panels at the front of the wing, ON THE TOP of the wing.
Now the insane parts of the Columbia Scenario are explained: the STRESS sensor goes first because STRESS breaks it, the TOP of the wing loses pieces at 340 seconds changing the airflow (yet all this is "caused" by a breach at 487 seconds over TWO MINUTES later (!)), the wires all die from the TOP down, the farthest west piece is from the TOP of the wing, the carrier panel directly behind and adjoining #8 lower (the foam impact point) is the LEAST damaged piece in that zone of the wing but #8 carrier UPPER is one of the MOST , Carrier Panels #9 Upper #10 upper, #11 upper are all missing but ALL of the Lower carriers were found by the Debris sorters at a time when it was admitted the orbit object HAD to be a carrier panel (after which we get excuses)... etc etc etc.
Ali's strut-transmitted overthrust causes overpressures in the SRBs to squirt out in a hideously dangerous direction. After patches of insulation melted on STS-56 indicating 1000 degrees (despite never exceeding 200 in either models or on a similar inclination flight) -- at 1100 degrees they die -- the other 3 Shuttles had heavier insulation put on (not Columbia) (not on a wide enough area for so severe a heating incident, probably).
I would: (1) not launch between 1 and 30 hours after a rain (2) start 1 engine 8 seconds before the other SSMEs and Launch 1 second later (+800lb cargo) (3) Provide Intact escape off the Solids using an Idle mode (+22,000 pounds with Centaur OMS) (4) Rehire the fired Scientists (why do you think they had to contract for Boeing to look at the Foam Problem?) (The OMB took over NASA recently (Sean "NASA doesn't need any new starts" O'Keefe), just like the NRC in 1985/6: wipe out the good parts, a shuttle blows, convene a panel containing ZERO Rocket Scientists --- and the next time, the SAME mechanism could kill via a fuel leak -- which clearly happens here cf the sensors that failed after FALLING temperatures showing LH2 pooling in the back of the wing, or aero-force breakup early due to the associated MISMATCH of SRB thrusts (STS-25 -107). Consider the Universal Electrical Failure 2 seconds after the epocal restabilization STS-107s Pilots managed 30 seconds after half the left wing broke off -- LH2 explosion in fuel cell area??? -- other wise they live to bail-out altitude (?).
Finally the economy responds to INFRASTRUCTURE -- when Clinton cut NASA's Space Science by a factor of 2.5 in 2 years, 5 years later all (major) new developments in Computer stopped & the Market crashed....
So for goodness sakes keep NASA at at least HALF JFK's level -- it means $220 Billion a year in the 6th year after, 444 the next 650 the next etc. Comments?

 

[JimMetalsCeramics]

CusterWillson is correct. Challenger exploded because the flame from the right solid rocket booster was aimed at the external tank. The hole in the right SRB was certainly not caused by o-ring blow by (NASA's story) but by a gaping hole torn in the right SRB aft joint due to bending moment stresses exceeding design values (the previous flight damage to the right SRB set up STS-51L to be the mission where the full-blown failure occurred).

Even without the extension of time between main engine ignition and lift-off, there are excessive forces all over the launch vehicle. These include SRB ignition thrust and g-forces acting on the payloads. In the latter case, the excess g-forces would exceed design values that the payload designers were given to design to.

Transient response creates over-stress conditions all-over the launch vehicle. The overstress is generally 60 to 100% greater than the steady-state forces. If you can believe it (and even if you cannot) the launch vehicle was designed to maximum steady state loads with a dynamic loading factor of 14 to 40% used. Given the fact that the transient overshoot exceeds these values, the Challenger accident was a metallurgical/structural catastrophe in the making. Since the problem was never fully recognized and corrected, it must be presumed that it was a primary source of structural damage (including insulation break-up --- if that's was really the ultimate culprit) on Columbia.