🚀 go-pugleaf

On Date: Thu, 14 Jan 93 00:57:14 EST, John Roberts <roberts@cmr.ncsl.nist.gov>
writes:

>It turns out that with the measured wind patterns, if one of the SSMEs had
>failed (it didn't) right when the Shuttle was passing Mach 1.55, three of
>the struts in the structure would have had a safety factor of 1.36 (36%),
>while the target safety factor under those conditions is a minimum of
>1.4 (40%). Analysis of the situation and of the particular orbiter indicated
>that this safety factor was acceptable, so permission was granted to proceed
>with the launch.

Strictly speaking, the *safety factor* was not reduced.  The *margins of 
safety* for the struts would have been reduced.  The following 
definitions (not exact spec wording, but close to the point) may help:

Limit Load (or Stress): the maximum predicted structural load which may 
be experienced by the vehicle while performing its mission.  This 
typically defines either a "worst case" or statistically derived (e.g. 
3-sigma) magnitude.  The SSMEs use a 2-sigma operations criteria, and the
rest of the structure may use the same.  Uncertainty factors may be included
in the calculations (varies from program to program but typically 1.5 at PDR,
1.25 at CDR, and 1.0 at FRR).

Allowable Yield Load: the maximum load the structure can withstand 
without unacceptable permanent deformation.

Allowable Ultimate Load: the maximum load the structure can withstand 
without collapse.

Factors of Safety (FS): the values by which the limit load is multiplied 
to obtain the yield and ultimate loads that the structure is designed to.

Margin of Safety (MS) = [Allowable Stress/(Limit Stress x FS)] - 1

The FS is a traditional "fudge factor" used to account for uncertainties 
such as variations in material properties, dimensional tolerances, etc.  
The Shuttle payloads (and I assume the Orbiter as well) use 1.4 for 
ultimate load.  A factor of 1.1 is typically used for yield load, but I 
don't know if the Orbiter structural design uses that.  Recent work has 
been done on SSME and SSF components to replace this traditional method 
with probabilistic structural analysis which uses a statistical treatment 
for uncertainties in both the applied load and material strength.

If a structure designed by this criteria has an Ultimate MS = 0.0, it 
will collapse if the actual imposed loads are 40% higher than the maximum 
expected.

<Pat was concerned by mechanical safety margins, and felt that the 
<Shuttle operates within 5% of destruction for its major mechanical and 
<structural systems.

Based on the information presented, it's unclear but I *think* that in 
the event of a single engine failure, the struts have an Ultimate MS = 
0.0.   It could thus be said that the Orbiter would then be flying within 
40% of destruction (and possibly 10% of being deformed such that it would 
need repairs).  In the case of the flight in question, where predicted 
winds aloft were higher than those which were used to design the Orbiter, 
it appears that the Ultimate MS was = -0.04 and it was flying within 36% 
of destruction (6% of yielding).

<The spokesman (Brewster Shaw, Deputy Director of the Space Shuttle 
<Program) replied that the safety margin stated is not to the point of 
<destruction . . .

Does anyone on the net know what the actual Yield and Ultimate MS were 
for this flight?  Given the fact that payload launch load analyses 
typically take one year to complete, was the analysis for this launch 
done "real-time", or were the results already available from previous 
studies of possible flight envelope expansion?

<One of the questions during the press conference was "what if something
<additional had happened, and the stress was increased another 36%, so 
<that zero safety margin was passed?"

The same thing that happens to bridges which are subjected to loads 
beyond the maximums expected by the engineers, something breaks.  An 
excellent book on the limits of engineering ability to predict the 
behavior of structures and the public's expectations of these predictions 
is "To Engineer is Human, The Role of Failure in Successful Design" by 
Henry Petroski.  Has anyone else read this, and if so what did you think?

Speaking of bridges:
On Date: 20 Jan 93 17:49:34 GMT, Pat <prb@access.digex.com> writes:
> Certain things must expect certain conditions.  Now i dont expect a bridge 
> to withstand the North AMerican Fat ladies precision Hop SCotch squad :-)
> but i do expect it to withstand 3 cement trucks in close order.

Petroski's book describes the inadvertent proof test of the Golden Gate
Bridge performed by 250,000 pedestrians who packed onto the bridge on its
50th anniversary.  The bridge swayed several feet and its arch flattened out
as much as 10 feet.  If the bridge had collapsed under this unanticipated
load, would the engineers have been to blame? 

Related to this subject, I am researching the use of traditional 
uncertainty factors in defining limit loads in the fields of aerospace, 
automotive, nuclear, etc.  It appears that very little has been formally 
published in this area and the most of the information is in the brains 
of the people who have done the work.  Any information in this area would 
be greatly appreciated.

I apologize if this goes beyond the scope of sci.space.  Does anyone know
of any engineering (especially aerospace and/or structural) news groups?

Disclaimer: Opinions stated are solely my own (unless I change my mind).
Ben Muniz     MUNIZB%RWTMS2.decnet@consrt.rockwell.com    w(818)586-3578
Rocketdyne/Rockwell:Space Station Freedom:Structural Loads and Dynamics
   "Man will not fly for fifty years": Wilbur to Orville Wright, 1901