Conference 7.286::space

Associated Press Tue 05-AUG-1986 18:51                  Shuttle-NearAccident
    
                        By MICHAEL J. SNIFFEN
                        Associated Press Writer                    
    
   WASHINGTON (AP) - NASA came within 31 seconds of launching the
shuttle Columbia last January without enough fuel to reach its
intended orbit because of human error caused partly by overwork and
fatigue, the Rogers Commission disclosed Tuesday.
   The incident occurred on Jan. 6, just 22 days before the shuttle
Challenger exploded 73 seconds after launch, killing its seven crew
members.
   Columbia was successfully launched on Jan. 12, after a record
seven delays, and carried among its crew, Rep. Bill Nelson, D-Fla.,
the second congressman to fly on the shuttle.
   The Rogers Commission, a presidential panel that investigated the
Challenger explosion, said the Jan. 6 scrub of the Columbia flight
    resulted when 18,000 pounds of liquid oxygen fuel ``were
inadvertently drained from the shuttle external tank due to operator
error.''
   It said an investigative report on the incident by Lockheed Space
Operations Co. ``cites operator fatigue as one of the major factors
contributing to this incident.''
   In an interview, Nelson said Tuesday he was told about the error
and the loss of fuel when it was discovered, ``but I haven't heard
about fatigue related to this incident.''
   Nelson said he had been told ``varying interpretations of what
would have happened if we had launched. One person told me we would
have gone into a lower orbit than intended. Others said we would
have run out of fuel before achieving orbit and would have had to
make an emergency landing on a 10,000-foot strip in Dakar, Senegal,
which you don't want to do in a fully loaded shuttle.''
   Nelson said he had thanked the operator who was alert enough to
discover the error, but ``if there is worker fatigue down there, it
ought to be a concern.''
   The Rogers Commission's disclosure came in a previously
unpublished study it made of workloads at the Kennedy Space Center
                                                            More -->
Associated Press Tue 05-AUG-1986 18:51         Shuttle-Near Accident (cont'd)

in Florida. It concluded ``there is no system at Kennedy for
monitoring overtime from the safety perspective.''
   The study blistered the National Aeronautics and Space
Administration as well as Morton Thiokol Inc. and Lockheed, the
contractors primarily responsible for shuttle launch preparations at
Kennedy, for excessive overtime, multiple shift changes and long
periods without a day off. It said those circumstances ``represent a
potential threat to safety and worker effectiveness.''
   The workload study was contained in four volumes of backup
materials and transcripts released Tuesday, nearly two months after
the commission's final report was presented to President Reagan.
   The final report concluded that launch site preparations were not
a factor in the Challenger explosion and the panel made no prior
public reference to its study of workloads.
   But the commission said NASA had agreed that the study should be
published as an appendix, and the panel recommended that the space
agency develop a regular overtime audit procedure and more
restrictions on overtime to reduce the risk of worker errors,
``especially that engendered by unanticipated launch postponements.''
   At the time, NASA officials gave a different explanation of why
                                                            More -->
Associated Press Tue 05-AUG-1986 18:51         Shuttle-Near Accident (cont'd)

Columbia's launch was scrubbed 31 seconds from liftoff, suggesting
that too much fuel had been put into its engine. This actually meant
that fuel had been vented from the tank through the engine. Those
explanations contained no reference to a fatigue-induced console
operator error.
   NASA spokesman Hugh Harris at Kennedy said, ``We've previously
acknowledged the loss of fuel and the operator error, but the
fatigue as a cause is new.''
   NASA spokesman Shirley Green said, ``We recognize this is one of
our problems and we are studying it very carefully. But it is
interesting to note that the commission did not include this in
their formal recommendations.''
   The commission quoted the Lockheed report as saying, ``Had the
mission not been scrubbed, the ability of the orbiter to reach
defined orbit may have been significantly impacted.''
   Arnold Aldrich, NASA's manager of the shuttle system, said the
loss of liquid oxygen ``could have led to serious safety of flight
consequences had the team elected to ... proceed with the launch.''
   The Lockheed investigation revealed that ``console operators in
the launch control center at Kennedy had misinterpreted system error
                                                            More -->
Associated Press Tue 05-AUG-1986 18:51         Shuttle-Near Accident (cont'd)

messages resulting from a failed microswitch on a replenishment
valve,'' the commission said.
   ``Consequently, instead of manually overriding an automatic
sequencer and closing the next valve in the sequence, they pressed
`continue' and caused the vent and drain valves to open
prematurely,'' the panel said.
   It added that the operators who made the error ``had been on duty
at the console for 11 hours during the third day of working 12-hour
night shifts.''
   At the time, NASA officials said a sensor switch failed on a
valve that was supposed to close on computer command and stop the
flow of liquid oxygen into Columbia's external tank.
   They said launch controllers checked pressures and flow rates and
decided the valve had closed, but a second valve, responding only to
the faulty sensor, refused to shut allowing an oversupply of between
1,400 and 2,900 gallons of liquid oxygen to flow into the tank.
   They said then that this lowered the temperature of the fuel
below specifications, triggering alarms.
   The commission found that both Morton Thiokol and Lockheed had
rules requiring special approval for anyone working more than 20
                                                            More -->
Associated Press Tue 05-AUG-1986 18:51         Shuttle-Near Accident (cont'd)

hours overtime in a week. Thiokol required special approval for
anyone working more than 13 consecutive days; Lockheed, for anyone
working more than six days a week, but no policy against more than
six days straight over two weeks.
   The commission looked at work records for the 5,000 contractor
employees at Kennedy from Oct. 6, 1985 to Feb. 2, 1986. During
January, there were five shuttle aborts and two launches.
   The panel found that Morton Thiokol exceeded the 20-hour overtime
limit 480 times and Lockheed exceeded it 2,512 times.
   ``Some individuals exceeded the limit during as many as 11 of the
14 weeks, while others worked nine consecutive weeks in excess of 20
hours overtime,'' the report said.
   It said overtime was concentrated among critically skilled
workers and cited four case histories. One mechanical team leader
worked 60, 96.5, 94 and 80.8 hours in January.
   An electronic technician, whose shift began at midnight, worked
44 consecutive days including 29 that were 11.5 hours long, seven
12-hour days, and two 15-hour days.
   A manager for technical shops worked 26 consecutive days in
January with 18 days lasting 12 hours or more - 16 of them in
                                                            More -->
Associated Press Tue 05-AUG-1986 18:51         Shuttle-Near Accident (cont'd)

succession.
   A lead electrician worked 50 consecutive days on the midnight
shift, 38 of them 11.5 to 12 hours long.
   ``These four cases are extreme but by no means isolated,'' the
panel said. ``They illustrate the frequent pattern at Kennedy of
combining weeks of consecutive workdays with multiple strings of 11-
or 12-hour days. Research has shown that either of these factors
alone produces worker fatigue, but that together they represent a
potential threat to safety and worker effectiveness.''
   Finally, the panel said, ``irregular working hours and
insufficient sleep'' may have contributed to the atmosphere during a
launch-eve telephone conference at which NASA officials at the
Marshall Space Flight Center convinced Morton Thiokol executives to
overrule their own engineers' objections to launching Challenger in
cold weather.
   Because a launch was scrubbed on Jan. 27, the Marshall managers
who participated in the teleconference had been able to sleep as
little as one hour to as much as seven hours the night before the 7
p.m. conference.

    They did not really say what actually DID cause the scrub, you notice,
    and I don't remember.  Notice that T-31 is when they switch over
    to the orbiter computers to run the countdown.  Presumably the
    GPCs made a check and discovered something amiss.
    
    The question I am bringing up is this:  The whole purpose of a
    countdown is not only to sequence all the operations that must be
    done, but also to check that everything is within tolerences.  It
    is fine to be concerned that an overworked technician made a serious
    mistake, but would it have been likely or even possible to launch
    without enough LOX?  Was it in fact a LOX quantity alarm that stopped
    the count to begin with?  It is scare-mongering to say "Columbia
    came with 31 seconds of launch with x wrong" when in fact x is not
    checked until T-31.
    
    Burns

    My understanding is that it was, in fact, the problem with the fuel
    being bled off that caused the count to be aborted at T-31.  This
    is not to say that at T-31 they check the fuel gauge.  Rather, a
    problem was being looked into (the fuel) and at T-31 it was decided
    to scrub.
    
    I agree, in one aspect, this shows that their system WORKS.  A problem
    developed and was discovered, preventing an unsafe launch.  But,
    likewise, this shows that there are problems with the system.  This
    fuel problem never should have occured.  While one part of NASA
    was functioning well, another was falling apart.
    jim
    

    As I remember the abort was caused by a low pressure reading on
    some sensor somewhere (sorry, that's about all I remember) when
    the internal computer took over the launch sequencing at T-31.
    

    Last night McNeal-Lehrer (sp?) interviewed Bill Nelson and elicited
    the same information shown in the previous responses, plus one
    additional detail: the scrub at T-31 was caused by a low temperature
    reading--a thermometer that should have shown no lower than -291
    degrees showed -296.  Nelson implied that this anomalous reading
    was caused by liquid oxygen having been vented incorrectly.  Thus,
    the scrub was related to the problem.
    
    Nelson also described an incident the following day that I think
    surprised even the interviewer.  On January 7 the launch was scrubbed
    again due to weather at two of the abort landing sites.  When they were
    draining the tanks they discovered that a temperature sensor had
    broken.  The part that broke passed through a liquid oxygen line to the
    engine, where it caught in a valve.  Apparently, if they had launched
    they would have been unable to shut down one of the main engines when
    they reached orbit.  The engine would have overheated and exploaded.
    The blades would have torn out the rear of the shuttle.  Nelson didn't
    say what would have happened then, but I suspect they would have been
    stranded in space, since the main engines could not have been
    restarted.  If they could somehow have reentered the atmosphere landing
    would have been very tricky, since the damage would have changed the
    aerodynamic shape of the vehicle. 
    
    Nevertheless, Nelson climbed back into the shuttle, and eventually
    flew.                            
        John Sauter

    I dug out my January stash of Aviation Week and found that both
    incidents were reported in detail.  The low-oxygen problem was reported
    way back in January to have been because of the orbiter valve not
    being checked after overriding the broken microswitch on the ground
    valve.  Thus the orbiter valve stayed open and out came the LOX.
    The only thing AvWeek did not mention was the fatigue factor.  They
    DID say that it was a human mistake.  The hold at T-31 was caused
    by several things being out of spec, including the LOX temp, LOX
    pressure, and LOX quantity.  They held at T-31 for some time trying
    to fix it, and they might have launched that day after refilling
    the tank, except that RCA got nervous since the delay would have
    caused their satellite's backup deployment period to be a day late.
    So they scrubbed at RCA's request.
    
    The sensor problem was also described.  Apparently the engines run
    into real trouble if they are not shut down properly.  In particular,
    they must not get too much oxygen while they are shutting down.
    (I assume this causes either too-hot burning of the remaining
    hydrogen).  If they get too much oxygen, they overheat and are damaged.
    (It was not clear from the article whether the overheating posses
    immediate danger.)
    
    Note that the main engines are not used to deorbit...the oms engines
    are, although they are located in pods which are not all that far
    from the main engines.

    Burns
    

Re .5:

I was surprised to observe from some diagrams in "The Space Shuttle
Operator's Manual" that the main engines (unlike the OMS engines) do
not have 4 valves per propellant pipeline in a redundant configuaration
which prevents any single valve from sticking in either state from
affecting the "open" or "closed" state of the valve system.  Perhaps
the flow rates are too great to squeeze 8 large valves in the space
alotted in the rear of the orbiter.  (Did the first stage of a Saturn
V have redundant fuel valves?)

I also noted from the diagram that it is entirely LH2 which cools the
main engines; no LOX is used.  (It would be interesting if someone
explained the temperature, heat of vaporization molecular weight and other
factors which cause the decision to work this way).  I wonder if the
issue is overheating of the engine itself, the turbopumps, or what?

On the other hand, I presume that if a LOX valve was stuck open at MECO,
that you could hit "ET SEP" early and have some chance of getting away
without the engines blowing up, if only the situation was diagnosed
correctly in enough time.  Unless they use some kind of "dump" ports
to empty the ET after MECO, there probably isn't any great danger from
separating from the tank with hot engines while LOX is still escaping from
the ET ports in quantity.  You indicated that they normally stop the
engines with propellants still in the tank.

Also, while I agree that having pieces of shrapnel shred the aft fuselage
is not fun, allegedly "the manuvering rockets" can deorbit the shuttle,
although it takes 3.5 times as long as the OMS engines.  While the forward
RCS system only has 3 870 pound thrust forward facing engines, compared
to 4 rearward facing engines in the rear RCS system, I assume a full
load of RCS propellants is enough to do the job.

Because of the redundant thrusters, the only bad part would probably be
damage from an exploded engine.  That could do more than rip away parts
of the skin.  It also has the potential to destroy any or all of the
movable aerodynamic control surfaces, and/or electric or hydraulic lines
which actuate them.  And I don't think the shuttle could land entirely
under RCS control, especially if the aft RCS is destroyed and the forward
RCS has used lots of fuel to retrofire.  Normally the control surfaces
take over for roll control at 25 minutes before touchdown, pitch at
23 minutes and yaw at 3 minutes.
	    			/AHM

    The fuel line connections to the ET have "flapper" valves which
    close when the shuttle separates from the ET.  This would prevent
    LH2/LOX leaking if the tank was dumped while the engines were
    still firing, although I'm not sure if the aerodynamics of the
    system would cause the shuttle and ET to collide in such a manuever.
    

The problem is that if the engines burn too lean  (too much
oxygen) the insides of the engine start to erode.  The problem
is not with the turbo pumps or cooling.  I believe that most of
the problem is that you have hot metal in contact with lots of
oxygen and you start to burn up the engine.  This actually
happened during the Challenger disaster.  The hydrogen tank was
opened up first and hydrogen pressure to the main engines dropped
first.  The engine controllers noticed the problem.  Two of the
engines were in the process of shutting down.  The third was
about to.  It appears that although the combustion chambers were
damaged and the engines could not have been used again the
engines would not have damaged the rest of the shuttle.  Most of
this is from the analysis of the recovered main engines from the
Rodger's Commission report.

The flapper valves could shut off the flow of oxygen to the
shuttle if the main engine valves stuck open.  However there is a
potential problem.  The flapper valves close suddenly.  If the
flow rate of oxygen is high this will produce a "water hammer"
effect which could cause catastrophic damage.  Another
consideration is that the turbopumps are run by the hot gases
produced by preburning some of the fuel.  If the hydrogen supply
is shut off they will stop.  That might reduce the oxygen flow to
a small amount.

To really know what would have happened with the stuck probe
would take much more information than I have available.  To know
for sure would probably take a destructive test of an engine.

	Tony 

    Note that at MECO, the shuttle is not in orbit (or at least, not
    in an orbit where the perigee is above the atmosphere).  Thus if
    the OMS engines were destroyed, they would hit the atmosphere roughly
    half an orbit later.  I don't know if this would be within range
    of a landing site.
    
    Burns