Wales.Larrison@ofa123.fidonet.org (Wales Larrison) (11/13/90)
Allen, you raised some interesting questions... >Of course, the government usually doesn't buy insurance for the >payloads it puts into orbit. Actually they do. There are differing ways of getting launch insurance (or more properly, assurance). One way is placing a bet with a broker on the probability of success, which is the normal commercial insurance method. This the government does not do. However, another form of insurance is "self-insurance" where the payload originator typically buys an additional satellite. Government program managers for programs with greater than 1 satellite constellation will typically buy an additional satellite and keep it on the ground in storage against a possible launch failure. This was done for DSCS, DSP, Fleetsatcom, TDRSS, GPS, and other government programs. Some commercial firms, Intelsat, for example, have elected to also do this due to large block buys of satellites. Another way of launch "assurance" is to very intensively control and monitor the launch operations. This is what is done for some classified programs which launch one of a kind superbly expensive satellites, and is the current way of doing things on Shuttle launches. >First of all, let's look at the cost of doing the same thing with >the Shuttle compared to the HLV: > >Item HLV Shuttle >Payload $22.5 B $22.5 B >Insurance[1] $ 1.125B $ 1.125B >Launch cost[2] $ 0.150 $ 3.5 B >EVA[3] $ 0.0 $ 0.002B >Engineering[4] $ 0.0 $ 2.25 B > -------- --------- > $23.775B $29.377 > [...Notes left out..] Hmmm... interesting point. However, I think you've been very optimistic about the HLV, and very pessimistic about the Shuttle. Let me do a sensitivity analysis, and play at being optimistic about the shuttle, and pessimistic about the HLV. The difference in cost given here is driven by three two factors: 1) an assumption the long-term reliability of the two systems - a yet undeveloped, "paper" HLV will have the reliability of the Shuttle, 2) the difference in assumed launch costs, and 3) the difference for engineering costs for assembly in orbit. 1). Reliability - the shuttle is, I believe, currently 44/45 (.978). Looking at the equivalent past history of ELVs, the Scout failed 11 of the first 49 launches (1960-66, r=.78), Delta had 3 of the first 43 fail (1960-1966, r=.93), and Atlas had 9 of the first 42 spacelaunches fail (1962-1977, r=.79) [Sorry, couldn't quickly find the data for the Titan program history]. Since each system is now showing reliabilities in the range of .94-.98, I would conclude the shuttle system is going through a common trend in space launchers, known as "infant mortality" where design flaws are found and fixed, and the system reliability increases. Rather than throw in some mathmatical mumbo-jumbo trying to project an increased reliability for the shuttle, I think we can say the Shuttle system reliability should increase with time, if the pattern found in historical launch programs is followed. (cont) -- Wales Larrison Internet: Wales.Larrison@ofa123.fidonet.org Compuserve: >internet:Wales.Larrison@ofa123.fidonet.org --------------------------------------------------------------------------
Wales.Larrison@ofa123.fidonet.org (Wales Larrison) (11/13/90)
However, this does pose some interesting questions. If, as you
say, the Delta HLV design is 7 Deltas strapped together, does this
not imply the reliability to be .98^7 = .86? Also, looking at the
historical data shows that a significant design change to a launcher
typically results in a pattern of failure for a few years (shown in
Delta and Scout upgrades, and in the little bit of data I have on
Titan history). How do we avoid the "infant morality" for a major
design change to a HLV?
For discussion sakes, lets say .90 for the new, untested HLV,
and .98 from the shuttle.
2) The difference in launch costs. I will accept the Shuttle is
never get the economies of scale per individual launch available in
a HLV. So the price of an individual shuttle launch will be always
higher ($/lb) than a HLV. However, if we can increase the shuttle
annual flight rate from 4 to the planned 12, then the cost is
reduced by $2.3 B.
Similarly, your HLV costs can be claimed as very opimistic. If,
LLNL is going to pay for a Space Station to completely recover the
development and financing costs for a HLV, and there is no long-
term recurring market, I think the cost could be increased. Any
problems with early system reliability, delivery schedule,
government funding lines (leading to program stretch-out), technical
problems with the LLNL station, or the HLV delivery schedules would
increase its launch costs.
Again, for discussion purposes, let us propose increasing the
HLV cost by 30% (at least!) for these considerations, and showing
the shuttle costs at a 12 per year flight rate. $0.195B vs $1.170B.
3) Engineering for assembly in orbit. I would assume the
provisions needed to assemble something on orbit are the connections
be simple and very easy to do/undo via EVA. However, aren't these
the same connections needed if the Space Station can be maintained
on orbit by EVA? For example, if one of the LLNL modules were to be
damaged beyond repair by orbital debris, it would necessarily have
to have these EVA dis-assembly provisions, so the module could be
unhooked and a new one hooked in. I would rate this as a wash,
between the two.
Combining all this, gives a set of revised numbers...
Item HLV Shuttle
Payload $22.5 B $22.5 B
Insurance $ 2.250 $ 0.450
Launch cost $ 0.195 $ 1.175
EVA $ 0.0 $ 0.002
Engineering $ 0.0 $ 0.0 Same between the two
-------- ---------
24.945 24.017
Which has now reversed.... I would say the most sensitive part
of this calculation is reliability since Shuttle total costs come
out cheaper is HLV reliability is .94 or less (shuttle at .98).
Based upon the historical patterns for launch vehicle development,
this might be a reasonable bet.
Anyway, I think we've beat this subject to death. Can we agree
to disagree on this? My opinion is that we need a HLV capability,
and I think it would make sense to apply such to any Space Station
program - after a good test period (10-20 flights). I'm just not as
hard over on killing our current programs in favor of untested new
programs - without a much better examination of what this entails.
But I support activities to examine this option.
(cont)
--
Wales Larrison
Internet: Wales.Larrison@ofa123.fidonet.org
Compuserve: >internet:Wales.Larrison@ofa123.fidonet.org
--------------------------------------------------------------------------Wales.Larrison@ofa123.fidonet.org (Wales Larrison) (11/13/90)
>>>[Previous message about future space transportation system mix...] >>I would suggest a shuttle-like vehicle with a on-pad abort >>system, and Liquid Rocket Boosters. I also think it might be >>possible to retro-fit such a capability onto the shuttle a lot >>cheaper than designing a new system. > >Perhaps. But the resulting system would still be far too expensive >to operate in a free market. True, if run by the government for the bureaucracy. But if you are in favor of turning ELV launch operations over to private firms, why not turn the Shuttle over to a private operation? If you could strip the system out of the government and turn it over to an incentive-based private organization, I think it might be worth another look. Based upon my experience and analyses, I think you could easily eliminate the top three levels of shuttle management, and at least 30-50% of remaining personnel, which are most of the "fixed" annual costs. (There are 20 "Shuttle" people at KSC for each person who touches the shuttle!) In my estimation, this would reduce the annual cost of operations from the current $3.5 billion to $2.0B within 5 years. Then, run the system like a trucking company - not a limo. That would allow smooth and standardized flows, and fixes several of the recurrent bottlenecks with specialized missions including special software loads and special training. Among other things, it would also remove the desire to get every last pound of performance, which would increase operational margin, and in turn, eliminate 25-50% of the launch delays. Adding LRBs would attack costs as well as operations. According to the MMC and GD preliminary design studies for LRBs, the estimated recurring costs for LRBs are about $5 M per flight, with about a $2 B development cost. LRBs increase reliability through engine-out on liftoff capability and more benign failure modes. They increase performance by anywhere up to an added 30-50,000 pounds. At current SRB costs of about $30 M per flight (at 8 flt/yr rate), this pays back at 39 flights, or between 3 and 5 years of operations. If we can clean house in the operations, we should be able to get to above 12 flights per year. Talking this over with some of the operations research gurus at work, their numbers show you can get up to 24 flights per year out of KSC with EXISTING facilities and orbiters (Note: this is merely through reducing the OPF flow time to 40 days from the current 95 days, and eliminating the SRB stacking bottleneck in the VAB. The LRBs eliminate the SRB stacking bottleneck by allowing the stacking of non-fueled LRB boosters, and the OPF bottleneck is reduced by merely eliminating the redundant system checks to ensure the first checkout was correct.) Cranking all these numbers in gives $2000 M for 16 flights per year at 65 Klbs (note: I'm allowing LOTS of margin...). This is about $1900/lb (annual cost), without a lot of technical, schedule, or financial risk. Taking more optimistic numbers, $2000 M for 24 flights/yr at 75Klbs gives $1100 /lb. Which is getting pretty competitive to a $1000/lb untested HLV - and is available at about the same technical risk and cost. Whether or not this is the best way to go, I'm not sure. But I think it deserves a good look along with looks at commercializing other launch operations. ------------------------------------------------------------------- Wales Larrison Space technology Investor -- Wales Larrison Internet: Wales.Larrison@ofa123.fidonet.org Compuserve: >internet:Wales.Larrison@ofa123.fidonet.org --------------------------------------------------------------------------
aws@ITI.ORG ("Allen W. Sherzer") (11/15/90)
In article <1565.27401C39@ofa123.fidonet.org> Wales Larrison writes: > However, this does pose some interesting questions. If, as you >say, the Delta HLV design is 7 Deltas strapped together, does this >not imply the reliability to be .98^7 = .86? No the claimed reliability is much higher. A failure which is catostrophic with one Delta may not be as bad if there are six others to take up the slack. I have seen figures on HL Delta reliability but cannot state them here. The reliability is near Shuttle levels although I do not know how the figure was arrived at. >Also, looking at the >historical data shows that a significant design change to a launcher >typically results in a pattern of failure for a few years (shown in >Delta and Scout upgrades, and in the little bit of data I have on >Titan history). How do we avoid the "infant morality" for a major >design change to a HLV? Maybe we don't. I don't know the answer. I do know that: 1. It will cost less than one Shuttle flight to develop the hardware and find out. 2. For HL Delta we aren't making big changes but using well established engineering methods. Maybe you're right. Maybe all we get out of this is a way to send up bulk supplies for a quater the cost of a Shuttle flight. On the other hand, for a very small investment we just might reduce the cost to orbit for everybody by an order of magnitude. Isn't it worth the risk? > For discussion sakes, lets say .90 for the new, untested HLV, >and .98 from the shuttle. Again, this does not agree with the MDAC figures. > 2) The difference in launch costs. I will accept the Shuttle is >never get the economies of scale per individual launch available in >a HLV. So the price of an individual shuttle launch will be always >higher ($/lb) than a HLV. However, if we can increase the shuttle >annual flight rate from 4 to the planned 12, then the cost is >reduced by $2.3 B. This is *VERY* unrealistic. This is three times the average launch rate for the last *TEN YEARS*. There is no reason to think this can be achieved. BTW, a report in this weeks Avation week said NASA is backing off from claiming high launch rates. There is a good graph of NASA launch rate claims and actual for the last ten years. > Similarly, your HLV costs can be claimed as very opimistic. If, >LLNL is going to pay for a Space Station to completely recover the >development and financing costs for a HLV, and there is no long- >term recurring market, I think the cost could be increased. I'll accept your HLV launch cost for the sake of arguement. However, I point out that LLNL will be buying ~24 of these launches plus more for later phases. One byproduct of this may well be enough of an infastructure to get a real space industry started. > 3) Engineering for assembly in orbit. I would assume the >provisions needed to assemble something on orbit are the connections >be simple and very easy to do/undo via EVA. I don't think so. Just because something has field replaceable units doen't make construction easy. For example, it may be possible to fix the wing of a 747 at Podunk Regional Airport but that doens't mean they could build one if given the parts. The Zenith Star engineers considered sending the payload up on a Titan and the Shuttle and doing assembly in orbit. This was considered too expensive because of design changes needed. That is why they went with the HL Delta and Titan V. >between the two. > Combining all this, gives a set of revised numbers... > >Item HLV Shuttle >Payload $22.5 B $22.5 B >Insurance $ 2.250 $ 0.450 >Launch cost $ 0.195 $ 1.175 >EVA $ 0.0 $ 0.002 >Engineering $ 0.0 $ 0.0 Same between the two > -------- --------- > 24.945 24.017 > Which has now reversed.... These two numbers are within 4% of each other. I consider that a wash. So if we use pesamistic numbers on the HLV and optimistic numbers on Shuttle they look the same. On the other hand, if we can't tripple the number of Shuttle flights and MDAC is correct on HL Delta reliability, the HLV will be the clear winner. It will only cost us one Shuttle flight to find out. >I would say the most sensitive part >of this calculation is reliability since Shuttle total costs come >out cheaper is HLV reliability is .94 or less (shuttle at .98). >Based upon the historical patterns for launch vehicle development, >this might be a reasonable bet. Agreed. > Anyway, I think we've beat this subject to death. Can we agree >to disagree on this? My opinion is that we need a HLV capability, >and I think it would make sense to apply such to any Space Station >program - after a good test period (10-20 flights). I'm just not as >hard over on killing our current programs in favor of untested new >programs - without a much better examination of what this entails. >But I support activities to examine this option. I can live with this. I wouldn't kill the Shuttle without having an option in place. I just point out that there is a good chance that that option is here. Allen -- +---------------------------------------------------------------------------+ |Allen W. Sherzer| I had a guaranteed military sale with ED-209. Renovation | | aws@iti.org | programs, spare parts for 25 years. Who cares if it | | | works or not? - Dick Jones, VP OCP Security Concepts |