Re: SpaceX Falcon Ready to Go - Again.

"Tom Cuddihy" <tom.cuddihy@xxxxxxxxx> wrote in message
news:1134070432.895490.130580@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
> Ed Kyle wrote:
> > Tom Cuddihy wrote:
> > > Ed Kyle wrote:
> > > > Ed Kyle wrote:
> > > > BTW, there has been a lot of talk about how wonderful
> > > > a place Kwaj is for launching rockets. The truth is that
> > > > there are so many occupied islands in the region to the
> > > > east that launch azimuths are suprisingly limited.
> > > > Azimuths from 55 degrees to 110 degrees present
> > > > downrange problems, for example.
> > > >
> > > > - Ed Kyle
> > >
> > > How much of a problem is that really though? If a 1 mi^2 island 100
> > > miles to the northwest is downrange, the chances that any kind of
> > > launch would actually hit there are miniscule.
> >
> > The problem is that it isn't one island, it is "approximately"
> > 1,225 islands comprising 29 atolls. Unfortunately for SpaceX,
> > Kwajalein is on the western side of the Republic of the Marshall
> > Islands, which places most of the "approximately" 1,225 islands,
> > and 60,000 islanders, on the downrange side, space launch wise.
> >
> > See, for example: "http://www.rmiembassyus.org/General%20Info.htm";
> >
> > - Ed Kyle
>
> First of all, pretty much all useful launch azimuths vary from true
> north to east, and the single alternate launch azimuth to the southwest
> for polar sun synchronous orbits. (though there's no real reason not
> launch with an azimuth of 352T)

Your impression of what constitutes "useful" launch azimuths is purely a
result of American geography. Launching to the southeast (on the appropriate
azimuth) results in the same orbit and the same performance (to first and
second order) as launching on the equivalent azimuth to the northeast. Both
quadrants are equally useful. The only reason Americans launch predominantly
northeast instead of southeast is that the Bahamas and Los Angeles are
southeast of our major launch sites. Launches from Wallops *do* commonly go
southeast. Likewise, we launch southwest out of Vandenberg only because
there is land to the northwest. People at launch sites on the northern edge
of continents - notably Kourou and Plesetsk - consider the northerly
direction to be the normal one for sunsynchronous missions.

> Second, 100% of those people are located on less than 50 sq. miles of
> land (obtained by subtracting some 20 sq. miles of the bigger
> uninhabited atolls from the total 70 sq. miles.)
> spread out over 750,000 sq. miles of ocean. Let's look at probability
> here: assume a 100% chance a rocket launched from kwajalein is going to
>
> come down somewhere in the Marshall Islands (which is rediculous) that
> still gives it about a 50/750000 = 0.007% chance of even hitting a
> populated island--that's not populated area, but populated island. And
> this assumes a uniform probability of hitting a populated island.

I don't know the details of how the Army does range safety in the Marshall
islands, but on the Eastern and Western Test Ranges (i.e. the Cape and
Vandenberg) the rule is that overflight is allowed if, and only if, you can
show that the casualty expection is less than 30x10e-6, or thirty fatalities
per million launches - a few orders of magnitude lower than your figure.

In order to compute this, you examine the planned flight corridor (or more
precisely, the corridor of the potential impact points, which isn't
necessarily underneath the flight corridor), the probability of failure at
each point in time, how far off the planned track the rocket could get, the
size and quantity of the debris that would be created by the failure, the
population living in the impact zone at each of those points in time, and
estimate what the odds are that anyone would actually get hit. It is routine
for the impact point track to safely cross continents late in flight, in
part because the large population exposed is negated by the very brief
duration they are at risk. (Launches out of the Cape almost always have IIP
tracks over Europe or Africa, and polar launches out of Kourou cross
Canada). But the risk within the first thousand or so miles of the launch
site is much higher because the rocket is moving more slowly and spends much
more time there, and therefore it's very difficult to meet the safety
threshold over land anywhere near the launch site.

One important point here is that a rocket launched from Kwajelein isn't
going to fly back and forth over the entire 750,000 sq. mi. ocean range of
the Marshal Islands. It's going to follow a pretty narrow path, and most of
the risk is going to be concentrated along and adjacent to that path. If
that path happens not to cross any populated islands, then it's probably
safe, and probably going to be allowed. If it does cross near an island, the
probability may be much higher than the 0.007 you computed, and it's very
likely not safe.

> In fact, the majority of RMI's population is in unuseful azimuths to
> the southeast, like the only 'city' on Majuro, so a realistic
> probablity would give an even lower number. Add in probablities of
> failure to begin with, realities of ballistic physics, and the actual
> probability of death or injury to a person and it's just rediculously
> silly as a concern.

It's very easy for people who really want to launch rockets to wave their
hands and claim that the probabilities are rediculously low. That's why
there are separate groups of people who are responsible for safety, and not
for rocket launching. Most of them are really pretty reasonable. If you do
the math and prove to them that it is safe, they'll let you launch.

> Our whole launch safety process needs an entire rethinking. If the same
> principles currently used with likelihood of failure were applied to
> aviation, no major city in America would qualify for an airport.

Funny you should bring that up. The 1950s US law that established the range
safety authority used as its primary benchmark the objective that third
party civilians should be exposed to no more risk from rocket launches than
they already were from aircraft spontaneously crashing on them. As I
understand it, that's where the 30x10e-6 value comes from. Despite this
week's incident at Midway, and the occasional Cessna or F-16 crashing in an
urban area, aircraft are allowed to operate over cities because they are
demonstrably much, much less likely to fail than rockets have been to date.
Therefore, many more people can be in the threatened zone without exceeding
a reasonable casualty expectation threshold.


Josh Hopkins


.

Relevant Pages

  • Re: SpaceX Falcon Ready to Go - Again.
    ... and the single alternate launch azimuth to the southwest ... >> this assumes a uniform probability of hitting a populated island. ... But the risk within the first thousand or so miles of the launch ... > for rocket launching. ...
    (sci.space.policy)
  • Re: SpaceX Falcon Ready to Go - Again.
    ... >>> east that launch azimuths are suprisingly limited. ... >> miles to the northwest is downrange, the chances that any kind of ... > 1,225 islands comprising 29 atolls. ... this assumes a uniform probability of hitting a populated island. ...
    (sci.space.policy)
  • Re: SpaceX Falcon Ready to Go - Again.
    ... >>> east that launch azimuths are suprisingly limited. ... >> miles to the northwest is downrange, the chances that any kind of ... > 1,225 islands comprising 29 atolls. ... this assumes a uniform probability of hitting a populated island. ...
    (sci.space.policy)
  • Re: Homeopathy = extreme dishwashing?
    ... Or perhaps they are not alien spacecraft at all, ... I can make a laser spot fly at thousands of miles per ... Secret government spacecraft, trying to remain ... Whenever a launch took place just after ...
    (misc.health.alternative)
  • Re: beanstalks (was Re: Metallic hydrogen ...)
    ... > Compared to EM launch guns? ... > Several hundred miles to a thousand miles of evacuated track, ... point made was if one used a lighter than air gas to fill the tanks and if they were to ...
    (sci.materials)