Safe Is Not An Option

Donald Robertson wrote a five-star review of the book, but apparently Amazon is getting overly (in my opinion) strict about who is allowed to review books there. So I’m publishing it here:

Rand Simberg’s “Safe is Not an Option” is an absolute must read for anyone interested in space policy, and why our expansion into space has been frozen in place for decades.  The book was first published in late 2013 and the author insisted to me that parts of it are out of date.  He is correct, but in any meaningful sense, it could have been written this afternoon.  I should state up front that, with a few very minor exceptions, I fully agree with his analysis, and came to many of the same conclusions independently.  Mr. Simberg writes well and this is a fun book to read.  
 
Mr. Simberg, an aerospace engineer, argues what should be obvious:  spending the majority of your budget to ensure the safety of astronauts in the most inherently dangerous activity humanity has ever tacked is excellent way to ensure you never accomplish anything – or the way I put it, to price yourself out of the game.  Unfortunately, this is not obvious to most in our government, who insist that safety is their first and last priority.  By extension, this means safety must also be NASA’s highest priority.   Not only does this attitude not make sense, it is unique to spaceflight.  We routinely lose hundreds of people every year in deep sea shipping accidents, and we tolerate all the risk involved in driving a car, to ourselves and to third parties, for no better reason than convenience.  But, we still insist on spending billions in a hopeless endeavor not to lose a single astronaut.  Mr. Simberg argues that this devalues spaceflight – space exploration is not important enough to allow volunteers to take the same risks they take driving to the space port.
 
Less obviously, Mr. Simberg argues convincingly that this attitude actually reduces safety.  Following the decision to move on from the Space Shuttle, then NASA Administrator Dr. Michael Griffin rejected using existing rockets with excellent track records.  He reoriented the constellation project to use almost the entire space exploration budget developing Ares-1, because he claimed it would have been safer,  In fact, it would have been anything but.  According to Mr. Simberg, by the time the project was cancelled as unaffordable, estimated costs had ballooned to $44 Billion.  Since Ares-1 essentially duplicated already existing capabilities, that’s $44 Billion that could not be spent exploring.  Part of that cost was due to many ad hoc systems introduced to “improve safety” – which also increased complexity and introduced new opportunities for things to go wrong.
 
There is one key area where I disagree with Mr. Simberg.  While I think the introduction of a US Space Guard to manage human spaceflight, modeled after the US Coast Guard, is an excellent idea, it makes no sense to put it under Air Force management.  Traditional Air Force operations are mostly, though not exclusively military, and generally involve short sorties supplied from the homeland or a small number of bases.  Any serious attempt to explore the Solar System will involve long travel times through an extraordinarily dangerous medium, civilian as well as military responsibilities, living off the land as much as possible, and the ability to make decisions and act independently forced by long communication times.  These characteristics sound a lot more like traditional naval operations, and any USSG should be under the Navy – or better, an independent organization.  We agree that NASA should return to their research and development roots.
 
Until recently, spaceflight was a bipartisan policy arena, with varying support by both Republicans and Democrats.  Mr. Simberg’s conservative political orientation leaks through in the occasional irritatingly snide remark, but overall, this is a refreshingly neutral book that this liberal Democrat can fully get behind.  Mr. Simberg explicitly criticizes some Republican space policies, and praises Mr. Obama’s efforts to replace what constellation had become with a more affordable, technically diverse, and semi-commercialized space program.  He outlines specific policies that could thaw United States space exploration and stimulate it to life. 
 
Had “safety” been better balanced with accomplishing mission goals and keeping costs low enough to fly often, people still would have been lost – inevitably – but much more could have been accomplished for their lives.  The idea that we can conquer the Solar System without losing lives is patently absurd – yet we insist on managing our space program as if that were an achievable goal.  I have argued that if we are not willing to reconsider that, we might as well stop wasting our money.  Mr. Simberg provides a detailed analysis, from an engineer’s perspective, of what needs to change.  Of course, everything is taking longer and costing more than Mr. Simberg assumed when he wrote the book.  Fortunately, in the increasingly diverse and commercial nature of human spaceflight, there are small, early signs that the long winter may indeed be ending.  

Thank you, Donald. I’m glad my “occasional irritating snide remarks” didn’t cost it a star. 🙂

25 thoughts on “Safe Is Not An Option”

  1. My understanding of Amazon’s recently implemented policy on reviews is that only people who have bought the book in question through Amazon or an affiliate or have spent at least 50 dollars with Amazon in the previous six months can now post reviews.

    The genesis of this policy seems to have been an effort to save the credibility of Amazon reviews from the depredations of SJW types who would gang-post terrible reviews of anything they took exception to for whatever recondite violation of political correctness was alleged. Given Bezos’s own politics, there’s more than a little irony there, but he’s always been operationally capitalist and self-interested anent his own enterprises however many progressive shibboleths he mouths in public.

    1. Aside from book review brigading, they have a huge problem with the phony reviews on normal products, which is probably a bigger problem for them economically, while the brigading is a PR mess.

      1. I don’t doubt that troll reviews of pretty much everything were an issue as well. It’s interesting that a progressive decided to implement a solution that has more than a little in common with the restriction of the franchise to property owners in Colonial times. I don’t think Amazon’s new policy is a bad one, by the way, just that it’s more than a bit hypocritical coming from a progressive corporation. But self-interest often trumps ideology among the denizens of all political camps.

  2. Not only does this attitude not make sense, it is unique to spaceflight. We routinely lose hundreds of people every year in deep sea shipping accidents, and we tolerate all the risk involved in driving a car, to ourselves and to third parties, for no better reason than convenience.

    The problem here is that we are talking about the government vs what companies and individuals choose to do. The competition over who gets budget priority is fierce. The deaths of government workers draws a lot of scrutiny and questions of, “What are they doing with my money?” Anything negative is used by other competitors for taxpayer money in the competition.

    Even when there is something critically important to our nation, like a war, a small number of deaths is used by competitors to justify taking the money and spending it on something else.

    This attitude wont change but it wont apply when commercial activities dominate the realm of space. So, the answer is in reducing the role of government but also in understanding that government playing a role will always mean a society wide interest when things don’t go right.

    Getting the populace to accept risk, failure, or even success is possible but improbable because everyone wants something different. Space nerds can’t even agree among themselves much less persuade the general population to do something which is clearly the wrong way for NASA to do something.

    1. “In a report last week related to the National Defense Authorization Act of fiscal year 2019, lawmakers on the House Armed Service Committee said that last year nearly four times as many military personnel died in training accidents as were killed in combat. In all, by the committee’s accounting, 21 service members died in combat while 80 died as a result of non- combat training-related accidents. And this spring alone, the report added, 25 people were killed in military aviation mishaps.”
      https://www.military.com/daily-news/2018/05/14/training-kills-more-troops-war-heres-whats-being-done-about-it.html

      So people in military die in training.
      But obviously, training is important.
      As Rand Simberg says, [roughly] NASA is not important.
      The fact that it is not important “drives” the obsession of safety.

      Also, NASA should not build rockets.
      It would a bad idea to fly in a plane built by any government.

      The only way NASA can become important is if NASA explores space in order to develop new commercial markets in Space.
      Ie, commerical lunar water mining and Mars settlements.
      Then NASA could become more important than our military, but perhaps not as important as the Space Force.

  3. One aspect of this that I consider especially pertinent is the added complexity of safety systems. These systems, unarguably, do impart additional risk factors. The issue is quantifying them.

    For example, would anyone, ever, favor a launch abort system that, statistically, would save the crew 1 time out of a hundred flights, but due to its own inherent risks on every flight would kill them 1 flight in 50? Operationally, your safety system would be achieving a doubling of the fatality rate vs. not having it.

    I’ve always thought that there are safer, cheaper ways to utilize the mass penalty of a LAS.

    The CRS-7 SpaceX launch failure (2nd stage explosion during a Cargo Dragon launch) ended in destruction of the Dragon and cargo because SpaceX had yet to implement a means to allow parachute deployment in such a situation. Had they done so, there is strong reason to believe the dragon would have survived.

    So, what about a passive launch safety system? Such a system would only need to be more effective than a LAS minus the dangers imparted to every mission by the LAS. Basically, instead of pulling the capsule away from an exploding LV, toughen it so that riding it out is survivable.
    In flight, this isn’t as steep a hill as one might imagine; liquid fueled LVs tend not to explode too energetically, it’s usually more of a deflagration. Like the Dragon on CRX-7, this can be survivable.

    Essentially, some minor toughening of the capsule might be required. The largest risk IMHO might be late in the launch profile, when re-entry is a significant factor – heat shield damage could prove fatal. To mitigate that, I was thinking of a kevlar blanket atop the second stage.

    The huge fly in the ointment would be a pad explosion. The capsule and crew could probably survive the conflagration, only to be killed by the fall. The only idea I could come up with in that regard that would not impart additional risk to a nominal mission was a steel net (suspended via heavy springs) around the middle of the first stage (with enough clearance so the LV would not be at risk from it during nominal launch translations, such as due to wind). Unless the capsule fell straight down after the LV exploded, my hope is that it would be caught survivably.

    This sort of methodology, if viable, IMHO would have been a means to use Cargo Dragon as a crew vehicle. All it would need would be a life support system, plus crew controls, and the kevlar blanket atop stage 2. This would IMHO have been vastly easier, faster, and cheaper to do than developing Dragon 2.

    One of the main features of the above would be that it would not impart added risk, so it would be incapable of blowing up the capsule and crew such as we saw in the recent demise of the Dragon 2. (Going kaboom is not a feature I favor spending years and billions to have.).

    1. The Orion crew capsule weighs 22,397 lbs at liftoff, but the escape tower weighs 16,850 lbs (NASA.gov…quickfacts.pdf

      That’s a ridiculously high mass penalty for something that the Space Shuttle didn’t even have. 43% of the weight sitting on top of the service module at liftoff is the abort system. For an equivalent amount of weight you could have a 16.5 foot diameter armor steel plate that’s two inches thick, which is the same thickness as the frontal armor on a Sherman tank, or a 10-inch thick slab of Kevlar.

      But as you say, surviving an in-flight explosion doesn’t require the big delta V and mass penalty of a pad abort. So what crazy ways can a rocket retain the pad abort capability without incurring a large mass penalty for the payload? I’ve figured out one method of just running a dual turbopump fed/pressure fed SuperDraco LAS system from lift off through max q, but I’m not sure it gains anything over the existing Dragon 2 system. I’ll have to do more math on it.

      1. Hrmmm. I fully agree, the mass penalty is enormous.

        Is there a way around needing that delta/v for a pad abort? The capsule can survive the deflagration fireball, though COPVs cooking off might be problematic (a thin kevlar blanket atop the 2nd stage might help). So why the need to pull up and away? Would sideways do?

        Instead of rockets, then, what about a steel cable? Attach it to a tower, off to the side about 100 ft, and attach the other end to a capsule sidewall hardpoint. If the rocket does a RUD on the pad, the capsule swings away by virtue of gravity once no longer supported by the LV. Or, to deal with the case of a LV about to blow up, the tower end is mounted to a winch. One huge downside; the release mechanism would have to work, every time, in a nominal launch. Also, it would be quite useless if the LV blew right above the pad, like Antares (too low for chute deploy).

        I wonder if it’s possible to rig the ‘chutes so they have a ballistic deployment option?

        I also wonder about SpaceX’s Starship. That won’t have a LAS, though it’s possible its Raptors might be an abort mode.

        1. I think chilling in the Starship Raptors on the pad at the same time as the SH Raptors just before launch would allow the Starship Raptors to fire with sufficient dispatch to constitute a viable LAS system. Separating with a full propellant load would also allow maximum flexibility anent RTLS landing after any abort.

          1. I think you’re right. IMHO it’s a net safety improvement via not adding anywhere near the risk factors the way a dedicated LAS does.

            Prechill takes very little propellant, so my hunch is that the early prechill would not cost much in fuel. Hrmmm, plus, with the umbilicals still attached, that should be replaced anyway.

            I wonder if it could be done with LOX only in a staged combustion engine? I can’t see a way it could prechill the methane pump with LOX, though maybe use methane for that and duct it overboard?

        2. Ballistic deployment of the chutes would be essentially a larger version of the system used on some light aircraft.

    2. Thinking about hardening the capsule and various LAS systems including th ones George describes. I think a tested method from the aviation world might be best

      EXPERIMENTAL SPACE CRAFT
      Volunteers only, fly at your own risk.

      14.2 ounces of placards and paint. Light the candle

      1. Don’t forget “You must be at least this tall to enjoy this ride,” even though that conflicts with my idea of conducting lunar missions with an Orion capsule sized for Peter Dinklage, which would really ease the launch requirements. I mean, nothing says the astronauts have to be very big.

      2. @ john hare

        That’s not viable without a detailed design feasibility study of the placards, materiels certification, structural analysis for every phase of the flight regime, a detailed research end development plan, and an environmental impact report.

        My guess is it would take about three years to do, and cost around 2 billion to implement.

        1. The sad thing is that you are not far off in the paranoia of today.

          It would be quite different if the focus was on getting the job done. Thousands of people years working on systems that might, just might, save a handful of lives. That is also a waste of life.

          In a high passenger traffic scenario, this level of attention would make sense.

  4. I thought about a variety of steel cable methods and ran into the same problems you did. The crazy option in that regard is to launch the rocket with a giant catapult in a deep silo, so that it starts its flight with enough velocity for the capsule to separate, arc over, and deploy a parachute. But NASA would need couple hundred billion for the silo.

    So back to my earlier thought on a different type of LAS system.

    The numbers I ran last night didn’t reach a clear conclusion on the continually-firing LAS, and likely won’t without getting into deep design details like stage mass. But what the heck, so I went there.
    Since I started with numbers on the Orion system, I just stuck an Orion on a Falcon 9 because I had numbers for that, too.

    I was considering two cases, running the LAS all the way through max q, about a minute into the flight, or running it all the way through first-stage separation, at three minutes. For simplicity I only went into the details on the latter.

    So this analysis gets necessarily detailed and a bit dry, as it’s a first-pass design study.

    My initial thought was to run the pad abort system from lift off past max q. If you’re going to carry the weight of extra engines, might as well run them and get some benefits. Additionally, if the pad abort system is already running, aborting becomes a simple matter of just separating from the rest of the stack. I instead switched to the idea of running the system for the entire first-stage burn because it made the numbers simpler, and possibly enhanced the performance.

    So, underneath the aforementioned Orion, place six Super Dracos, (slightly enhanced by a larger bell to bring their seal-level ISP up to 260). They are mounted to the side of the service module shroud like strap on boosters. The larger bell brings their thrust up to 17,700 lbf, but 25-degree cosine losses drop the effective thrust back to a bit above the original 16,000 lbs. Once lit, they provide 100,000 lbsf, enough thrust to accelerate an Orion capsule at about four G’s.

    Each LAS strap on has about 500 lbs of internal pressurized fuel, which it relies on during an actual abort. I also assumed about 100 lbs of dry weight per LAS booster. That would give the capsule a free-space delta V of 300 m/sec (670 mph), or a pad delta V of 500 mph due to gravity losses.

    The LAS engines burn through most of the ascent, and are normally supplied from big low-pressure tanks located at the top of the first stage, and fed via a turbopump that is also part of the first stage. That allows the abort engines to run for as long as needed instead of the seven seconds provided by their internal pressurized tanks. A downside is that two high pressure fuel lines have to run up the side of the second stage and past the service module. Bleh.

    In an abort the LAS boosters automatically switch over to their internal pressurized tanks just as the capsule separates, leaving the service module and upper stage behind. Having the LAS boosters behind and to out to the sides helps aerodynamically stabilize the capsule, much like the trunk on the Dragon 2, though the mounting is going to be weird and kludgy.

    So, instead of the Orion capsule, adapter, fairings, service module, and tractor LAS weighing 76,000 lbs, it would only weigh about 63,000 lbs because we’ve got rid of the big solid motor. Most of that weight is the fueled service module which would be left behind during an abort. But when you light the LAS engines for launch, providing 100,000 lbs of thrust up top, the weight of the whole payload section drops to -37,000 lbs (yes, minus 37,000 because the whole payload section wants to fly on ahead).

    That reduces the structural loads on the whole stack, allowing the booster core and second stage to be structurally lighter. That’s good because lower down we’re also adding about 24,500 lbs of extra hypergolic fuel for the 60 second run-time option, or 73,500 lbs for the 180 second run-time option. With either fuel option, the added mass is less than the added thrust, and the heavier 180-second option is still a slight improvement in thrust to weight ratio at liftoff.

    The LAS system’s sea-level ISP of 260 is lower than the Merlin 1D’s 282 second ISP, so what do they do to the overall performance?

    Lift-off thrust increases 6.5%, from 1,530,000 lbs to 1,630,000 lbs. The LAS engines are burning 175 kg/sec of propellant, whereas the main engines are burning 2466 kg/sec, so the total system’s ISP is 280 seconds, only slightly lower than the Merlin’s own 282 sec SL ISP.

    So it comes down to a question of mass ratios, which is going to get tricky and this analysis is necessarily simplified. I’m going with some posted estimates of on the Falcon 9 FT, so some of these numbers may be slightly off.

    The original Falcon 9 FT first stage carries 398,887 kg of propellant and the rocket has a lift-off mass of 549,054 kg, so the mass ratio is 3.66:1 for calculating first stage performance. It’s free space delta-V, ignoring the increase in ISP at altitude, would be 3,585 m/sec. The lift-off thrust to weight ratio is 1.2667:1.

    The newly enhanced FT, for a 180-sec total LAS burn time, carries 430,355 kg of propellant and has a lift-off mass of 584,928 kg, so it’s mass ratio is 3.81:1. It’s free space delta-V, with the 280 sec ISP, is 3,683 m/sec, an increase of 98 m/sec. The lift-off thrust to weight ratio is unchanged at 1.2667:1. Those numbers assume the total LAS system’s dry mass is 10% of its wet mass, for a LAS mass ratio of 10.

    So instead of a 43% weight penalty, which is severe, the performance actually improves over not having a LAS at all. However, the actual improvement may be larger or become a slight detriment, depending on the fine details of how much the system actually weighs, whether the decrease load on the stack can be converted into structural weight savings, or whether the mass shifts, added drag, or high-pressure fuel lines and disconnects end up actually adding weight that is far more than my first-pass guesses.

    But does the system increase reliability? I think it could. Since the boosters are lit with the main engines, the whole rocket can be shut down at ignition if one of them fails. Since they’re getting a full run every launch, engine reliability data would rapidly accumulate. Since they’re much simpler than a Merlin on the pressure fed side, they will likely have higher reliability than the main engines.

    If a single LAS engine fails, either compensate by gimbaling the main engines or shut down the LAS booster on the opposite side, which will still allow for a 2.6 G escape acceleration. If that’s too low use eight boosters instead of six.

    If the LAS turbopump fails, down between the first and second stages, shut down the LAS engines, flip their valves to the internal pressurized thanks, and the LAS system is still good for a potential abort, though there will be a shortfall in performance because of the unburned LAS fuel down in the unpressurized tanks.

    Since the LAS fuel is hypergolics, I’m not sure if that fuel could be safely dumped or not. If it was dumped out of the base of the rocket then almost all of the LAS mass goes away, and the rocket is almost back to being a regular Falcon 9 FT, so there shouldn’t be much of a performance hit at all on the first stage burn, and the second stage burn doesn’t carry the heavy LAS fuel load because that’s on the first stage, and the second-stage burn doesn’t run the LAS engines except in an abort, assuming the strap on LAS system isn’t discarded along with the first stage. Either way, the LAS system’s internal fuel is burned prior to LAS separation, since there’s no use in wasting it.

    The only failure it couldn’t handle is a COPV failure in a LAS booster, which is definitely an added risk over the tower method, but one shared with the Dragon 2’s system.

    The downside is that the fairly expensive liquid fueled LAS engines would get thrown away with every flight, probably making it more expensive than the Dragon 2 method. The upside is that the capsule doesn’t carry its abort system’s weight penalty all the way to orbit and back, which is a bug of the Dragon 2’s LAS-in-the-capsule system.

    I also skipped the question of whether running Super Dracos on the sides of the service module during the entire first stage ascent was a sane idea, because sometimes common sense boxes in the design space. That’s how we end up with solid motor tractor systems on a space launch system designed for the 2020’s.

    As an afterthought to this analysis, another way to think of the idea is to imagine the capsule as a side-mounted orbital vehicle, much like a mini-shuttle. But instead of having it mount three SSME’s as main engines, we’re having it run a strap-on OMS system on steroids during the launch phase, such that if the mini-shuttle separated from the external tank (which rides on its own SSME’s), it would use the small amount of internal abort fuel to accelerate away at about 4 G’s for seven seconds. We’re just running those OMS abort engines for the entire first three minutes of the ascent.

    Any performance penalty would come from the difference in ISP between the main engines and the OMS engines, since that difference isn’t all that large for the Merlin 1D and SuperDraco with a slightly larger bell (I think N2O4/hydrazine can hit a sea-level ISP of about 280 sec), my analysis is having the slight penalty more than canceled by the increase in lift-off thrust, which matches the increase in launch mass.

    Basically, the LAS functionality in this scheme is a performance freebie, existing as a separate, single-stage launch system operating in parallel with the main rocket, one that merely accelerates itself (and the small amount of internal LAS fuel) along the rocket’s ascent trajectory throughout the first stage burn.

    1. This is a wonderful analysis, and let me say from the start that I prefer your continually-running LAS to the current implementation.

      I do, however, have concerns. # 1 is that redesigning the LV (lightening its structure) gives you a LV that can only be used for these missions, which might be problematic from a cost POV. #2 is, as you say, the issue of high pressure hypergol lines running the length of the rocket, plus integrating a hypergol turbopump into a Merlin sounds like a very difficult task. Also, all of these do add failure modes.

      What I especially like about your concept is that the LAS fuel mass is actually useful for delta/v in a nominal situation.

      The approach I took was that LEO is a very different thing than beyond LEO. For the former, I’m not too worried about launch mass, because so long as it (capsule plus any payload) is within the LVs capacity, it’s not much of an issue. So, in the case of using a modified Dragon Cargo for manned LEO missions, even adding a few tons of launch mass would be okay (provided, of course, it’s done in a way that does not increase reentry mass).

      What I utterly lack is the ability to analyze actual risks, statistically. For example, to compare the overall impact on safety of the tradeoff of removing many failure modes for nominal missions while making some abort scenarios somewhat riskier. To do that, I’d need, at least, a reliability number for the LV (how many times out of 1000 launches would the LV fail, and in what modes and situations). I have no clue how to come up with such numbers. I also wonder if NASA does, given its past history of glaring failures in this regard (their Shuttle risk assessments, vs. those of the Air Force, where were dramatically different and far more accurate).

      Essentially, the concept I came up with entailed, #1, a kevlar blanket atop the second stage. I was using the kevlar blankets around the Merlins as a guide. I have no idea of their mass per square foot, but ballistic blankets for bomb disposal and and also used in high-end body armor, III-A level protection, weight around 1.5 pounds per square foot. I think this might be overkill for atop the 2nd stage (velocities encountered, even from a COPV rupture, should be nowhere as high as rifle fire) but it’s the only number I could find, so, I’m going with it.
      The F9 is a 12′ diameter, which gives me a 113 sq.ft area. The top of the 2nd stage contains a dome, so I’m kicking the area up to 150sq.ft (wold guess). Multiplied by 1.5, that gives me 225 pounds. It’ll need hold-downs, attachment points, etc, so I’ll round up to 300 pounds. But, this stays with the 2nd stage, so won’t add to re-entry mass.

    2. I’d forgotten to add any shielding to my analysis, however the current LAS systems are certified without any form of ballistic protection, I suppose because the explosion’s debris not expected to seriously impact the capsule as it pulls away. But the weight penalty for ballistic protection wouldn’t seem to be very high. The Super Draco mounts on the Dragon 2 already include blast containment in case of engine failure.

      As an aside, going back to my suggested system (which I guess an aerospace company would call the Performance Enhancing Launch Escape System, or PELAS), perhaps a simpler way to analyze it is to treat it as a separate, parallel rocket that merely propels the LAS itself.

      So in the above, going with weight estimates like “500 lbs piping, 100 lb turbopump”, etc. I came up with a dry weight of 4,876 lbs and a wet weight of 37,646 lbs, for a MR of 7.72:1. But they might not want to use the small amount of pressurize fuel so they can still retain an abort capability during part of the 2nd stage burn, so if that fuel is counted toward dry weight, the MR drops to 6.3:1.

      Even with a really low ISP of 200 seconds, and not burning the abort fuel, it would still have a bigger delta V than the first stage it’s attached to because it doesn’t have to carry a second stage and a big payload. With a 260 ISP it had 30% more delta V than the Falcon 9’s first stage, and thus would be enhancing the launch performance. So it’s a ride-along with benefits.

      But the failure statistics would be a complicated affair. If we hold that the pressurized LAS system is immune to a chain failure caused by a first or second stage event, which may take out the piping and turbopumps but leave the strap-ons intact, then the mission failure rate due to main stage failures is unchanged, but now far more survivable, taking into account the success rate of the LAS system in abort modes.

      Say both success rates are 98%. For an actual abort failure, you have to suffer a 1st or 2nd stage failure (2% odds) coincident with a LAS failure, which is only going to occur 0.04% of the time.

      If virtually all LAS failures are either a turbopump failure or a single contained engine failure, and they can’t cause a subsequent failure in the main stages, then if the abort system is still viable, the 2% failure rate of the LAS wouldn’t necessarily translate into a corresponding decrease the mission successes rate, depending on where the delta V ends up without the continual LAS burn. Since the two systems are operating in parallel instead of in series, and a loss of crew failure requires both systems to fail on the same mission, the safety should go up, with the caveat that one failure still might directly cause the other.

      Considering the system as two independent but physically interweaved rocket systems, there are twice as many things to fail, along with any unwanted interactions between the two systems, and the high pressure tanks and associated piping in close proximity to the capsule might actually be increasing the risk of catastrophic failure. In pursuit of safety, the system is still adding more things that can potentially go boom, so it depends if the booms are containable. So I will confidently answer “Maybe” on the safety question.

      But it’s a “maybe” that actually adds a performance enhancement instead of adding a 43% mass penalty to at least most of the ascent phase, and screwing up the astronauts’ view out the window.

      Anyway, I don’t particularly recall encountering a launch abort system that actually added to the performance of a flight vehicle, even in design studies, so this might be a new concept. But there have been so many obscure design studies and Power Point vehicles that it’s bound to have been suggested before.

      1. I did a couple of blog posts with similar ideas several years back, does that count?

        Cue a Geatano the Moron rant that I thought of it first and everybody else stole it. 🙂

        1. That would not surprise me a bit! 😀

          I once thought up a way of levitating a space elevator structure by momentum transfer from small satellites in a football shaped orbit (making perfectly elastic collisions at the pointing ends) only to discover that a noted innovator from the British Interplanetary Society had much earlier thought of something very close to that.

    3. –George Turner
      May 4, 2019 At 10:26 AM
      I thought about a variety of steel cable methods and ran into the same problems you did. The crazy option in that regard is to launch the rocket with a giant catapult in a deep silo, so that it starts its flight with enough velocity for the capsule to separate, arc over, and deploy a parachute. But NASA would need couple hundred billion for the silo.–

      Crazy option- much cheaper than 100 billion.
      Much cheaper than 100 million.
      Even NASA could build it for less than 100 million dollars, but it should cost less than 10 million.
      [[But one could add in a lot stuff related to it and all of that could end up with NASA spending more 100 billion]].

      A pipelauncher:
      200 meter tall 16 meter diameter pipe
      20 meter tall 12 meter diameter pipe
      10 meter tall conic cylinder
      16 meters to 12 meters diameter

      16 meter diameter 200 meter long with 1 cm thick walls. Marine aluminum.

      16 meter diameter has circumference: 50.2654 meters. Times 200 =
      10,053.09 square meters.
      In terms volume of material, divide by 100 {1 cm =1/100th of meter} 100.53088 cubic meters of AL. Times 2.7 density =
      271.4334 metric tons.
      Displacement/volume of 1 meter length of 12 meter diameter or 6 meter radius: 113.097 cubic meters

      If 40,000 yield str and safety factor 1.5
      Outside diameter of 12 meters in inches: 472.441 inches
      44.44 psi working pressure. 66.67 psi burst pressure
      https://tridentsteel.com/barlows-formula/

      If pipelauncher with payload or gross mass of 10 x 113.097 or 1130.97 tons then the 16 meter section would push water 10 meters under the waterline and the air inside pipe would 14.7 psi [ 29.4 psi absolute pressure].
      Correction: that was the 12 meter diameter pipe
      The 16 meter diameter displaces:
      8 x 8 x Pi = 201.06176 cubic meter per 1 meter length
      To push water inside 16 meter diameter pipe so it’s 10 meter below waterline is 2010.6176 tons.
      A fueled falcon-9 : 549,054 kg or 549.054 tons
      The 200 meter long pipe is most massive element and weighs 271.4334 metric tons or entire pipelauncher could weigh less than 500 tons.

      The top of pipelauncher has the 12 meter diameter pipe with pipe wall 1 cm thick.
      That has a cap: 12 meter diameter disk which equal to mass of 3 cm of Marine aluminum:
      area of cap: 6 x 6 x pi = 113.097 square meter.
      At 1 cm thickness it’s 1.13097 times 2.7 = 3.053 tons. 3 cm times it by 3 =
      9.16088 tons

      Above cap or deck of pipelauncher would be launching tower and rocket.
      {going to avoid these details}
      Below cap is the 20 meter tall pipe which 12 meter diameter.
      A part of this pipe section would be emergency floatation.
      If 5 meter of 20 meter in total is sealed bulkhead or has floatation foam. Then pipelauncher can’t sink to ocean floor.
      Another 5 meter of the 20 meter could store liquid air- tens of tons of liquid air.
      Both of these sealed from rest of pipe. Say, two floors and both floors will equal 2 cm thick 12 meter diameter disk or about 6 tons of aluminum [2/3rd of cap].
      And lower third section is the “engine room” which is open to rest of the pipe.
      Engine room is where Liquid air is dumped and/or sprayed. And has burners- which might resemble flamethrowers or jet engines [but they don’t add lift- or not counted on to add lift] or any large burner which is designed to heat air.
      So you have plumbing and electrical and access pathways and all this might total:
      30 tons.
      And 20 meter tall and 12 meter diameter with 1 cm thick walls:
      753.9816 square meters / 100 times 2.7 =
      20.3575 tons
      Then have 10 meter tall conic cylinder going from 12 to 16 meter- a few tons.

      As a note, a person probably could survive being in the engine room while it’s launching a rocket. A problem is the pressure changes, so say someone is in a pressure suit.
      Or engine room is not engulfed in flame or something. Getting as warm as sauna but only for seconds of time. Likewise a person could survive on or in the water below the engine room- and diving about 3 feet under water would safest thing to do.
      And that water is not really moving anywhere. There is no ocean water which is pushed or moved, though any changes of air pressure within pipe alters the level of the water. Oh, there is the turbulence from walls of pipe when going up [or going down] but you could
      be quite far away from the pipe walls.

      Entire length 200 + 20 + 10 = 230 meter.
      If Pipelauncher is going to jump out of the ocean, then you have at most about 220 meter distance to accelerate up.
      Distance = 1/2 acceleration times seconds of acceleration squared

      If 9.8 m/s/s
      220 meter / 4.9 = 44.89 and square root is 6.7 seconds of acceleration.
      6.7 times 9.8 = 65.66 m/s [146 mph]
      Jumping out of the water, might be too exciting/not safe.
      If accelerate at 1 gee for 4 seconds:
      16 x 4.9 = 78.4 meters it acceleration up. And is 39.4 m/s [87.5 mph].
      If want to faster than 146 mph, you have to make pipelaucher taller [and it gets more expensive, exponentially]
      Or you can accelerate at higher gee than 9.8 m/s/s.

      It not clear to me that rockets could take as much 1 gee, though it is not much of problem for pipelauncher to deliver
      2 to 3 gees. If rockets can’t take 1 gee of acceleration from pipelauncher, than it would reach slower speed- or have to make
      pipelauncher taller.

      1. Ah, the thread still lives.

        You could use a pipe launcher or some odd pipe-like tower mounting an electromagnetic system. The EMALs on the Ford class aircraft carriers can accelerate 100,000 lbs to 150 mph in 300 feet, which is 2.5 G’s, so a dozen of them along the sides of the launcher would also work for a Falcon 9. But they’re also a big piece of expensive infrastructure that would tie you to staying with a Falcon 9, or whatever rocket they were designed for.

        But I’m not sure if its worth the cost because if you’re going to those lengths to make the abort system’s job easier, then you’re still trying to use an abort system, but now it has to get you away from an exploding rocket inside the pipe.

        However, if the pipe idea is pursued more fully, it morphs into something like a really long semi-vertical vacuum mag-lev system (like a lunar mass driver) that exits at high Mach out of the top of Mount Everest, or an even taller man-made mountain. The height is to reduce the max-q at the muzzle. But something like that would be ridiculously expensive to build, considering NASA spent a billion modifying a launch tower they already had on hand and will probably use again only once.

        A few years ago I suggest launching a rocket that’s laying on its side in a vertical toss and rotate operation, using small engines distributed along its length. One of the reasons for that is that long structures stay cheap, but tall structures get exponentially expensive because they have to self-support and are really difficult to transport, maintain, and access.

        Dealing with a ship that’s 500 or 1000 feet long is no problem, and the Navy does it daily. Working on a vehicle that’s 500 or 1000 feet tall would be a nightmare.

        I always keep in mind that cost per weight to LEO is everything. Any bad lunar architecture that relies on a launcher that delivers a pound to orbit for $600 (like the Falcon Heavy) or less is going to beat any good architecture that relies on a launcher that delivers payloads for $7000 a pound (like the SLS).

Leave a Reply

Your email address will not be published. Required fields are marked *