Chair Force Engineer

Monday, March 31, 2008

Considerations on Mars Sample Return

Taylor Dinerman notes that NASA will attempt a robotic Mars Sample Return (MSR) by the end of the next decade (presumably between 2018 and 2020, depending on when the launch window opens and whether opposition-class or conjunction-class missions are chosen.)

Returning Mars rocks to Earth has always been viewed as the "holy grail" of unmanned, planetary science missions. But the challenges involved in doing so have necessitated a big budget and lengthy schedule for the mission, in comparison with other probes. Before the failures of Mars Climate Orbiter and Mars Polar Lander in 1999, a Sample Return was tentatively scheduled for 2011.

The number of design trades in the Mars Sample Return mission is astounding. For instance, consider the following:
--Do you have a separate lander and orbiter, or will the entire craft descend to the surface of Mars?
--If the craft separates into a lander and orbiter, will the two travel to Mars mated to each other, or will they be delivered by two separate launchers?
--Will the ascent stage of the lander be launched from earth fully-fueled, carrying hypergolic propellants? Or will it produce some or all of its fuel from the Martian atmosphere, as Bob Zubrin has suggested?
--Will the lander collect samples from just its landing site, or will it have at least one rover for collecting samples at different locations?
--If the ascent stage has to rendezvous with an orbiting stage that will return the samples to earth, how will the samples be transferred autonomously from the ascent stage to the orbiter?
--Will the Sample Return Capsule be recovered at sea, on land, or snatched from the air? What measures will be taken to quarantine the samples?

The functionality and complexity of the Mars Sample Return mission will be dictated by the capabilities of the launcher that is baselined for the mission. While Ares I or Ares V might be available by the time MSR launches, the conservative engineer and program manager will baseline a vehicle that exists in hardware form today, with a well-characterized vibrational profile and established performance levels. The only vehicle up to the task is Delta IV Heavy. And while it's tempting to launch two 22-tonne spacecraft and have them rendezvous after reaching Mars (such as an orbiter and lander pair,) conservative engineering practices would frown on making such a rendezvous so far away from earth (at least for initial missions.) Thus, I would not budget any more than 22 metric tons for the spacecraft and its earth departure stage on an initial Mars Sample Return mission. This would appear to rule out a separate rover on the first MSR.

Because MSR is so far into the future, NASA needs a proactive strategy for keeping Delta IV Heavy in production by the time MSR reaches hardware stage. Because Delta IV and Atlas V have common spacecraft interfaces, it might be possible to switch to an Atlas variant (either Atlas V Heavy or wide-bodied Atlas) if United Launch Alliance switches to an all-Atlas fleet. But doing so will also change the vibe profile, potentially affecting the spacecraft.

Beyond the fundamental challenges of mission architecture, there is the issue of how much MSR will demonstrate technologies essential to a human Mars mission. I see great potential for using the Sabatier reaction to produce ascent and Mars-departure propellant. But until the Sabatier reactor is proven in a relevant environment, there's no way the taxpayers will invest in a Mars mission (even an unmanned one) that relies on an unproven technology for mission success. In order for the Sabatier reactor to fly on MSR, it must be tested as an experiment on a near-term Mars mission. This was the plan for the Mars Surveyor 2001, before it was canceled and re-scoped as Phoenix. But here's a novel idea: why not test a Sabatier reactor as part of a Mars rover that could be put into hibernation and then re-activated to collect samples for MSR? It's a possibility, if the rover can be built robustly enough to survive an extended hibernation.

MSR is going to be a costly and expensive mission to pull off. It will be virtually impossible to pull off (and thus a waste of taxpayer money) if NASA takes risky gambles with immature systems. The extreme conservatism in design that the mission demands will probably result in an underwhelming science return for the cost of the mission. From a scientific standpoint, it might be more worthwhile to perform experiments in-situ on Martian soil. From an engineering standpoint, MSR will be a valuable step towards sending real scientists to Mars in the future.

Saturday, March 29, 2008

Broken Acquisition

I was reading a blurb on StrategyPage the other day which claimed that the "B-3 bomber" was being developed in secret, much like how the B-2's development was highly classified. While I've seen no evidence to corroborate the claims of StrategyPage, it did make me think about the words of an AFRL Captain during a recent meeting: the acquisition system is broken. It can be said, somewhat facetiously, that the next bomber will be obsolete by the time its protracted development is completed, and it will take the entire Gross Domestic Product of this great nation to afford just one copy of the next bomber.

The stark reality of the situation is that any new-start program is going to be a protracted and expensive development. In order to ensure that the program achieves an Initial Operational Capability at a reasonable date, the people determining the mission requirements will need incredible foresight to determine relevant mission requirements for a program that may be two decades away from seeing service.

Gone are the glory days when a fighter plane like the P-51 Mustang could move from design stage to first flight in 120 days. The current example is the F-22 Raptor, which took six years to move from source selection to first flight, and another eight years between first flight and IOC. There can be no doubt that the complexity of combat systems has grown exponentially. As long as the US Military insists on purchasing the most complex combat systems that money can buy, every new development will be lengthy and expensive.

There is a good argument to be made in favor of bucking the complexity trend. For example, the F-22 Raptor, at a cost of >$120 million per copy, has enough missiles to engage eight enemy fighters. It doesn't take a genius to see that an enemy can find nine cheap fighters and nine suicidal pilots (for less than the cost of a single F-22) and win the dogfight through lopsided numerical superiority.

The argument against complexity is the reason why the F-16 was conceived in the first place. Because the F-15 Eagle was so expensive (for its time,) it was conceived that a smaller and simpler fighter could complement the expensive F-15. The F-16, by contrast, was envisioned as a no-frills, lightweight fighter that would drop unguided bombs in the daytime and defend itself with short-range missiles. The problem is that the Air Force did not stay true to the original F-16 concept. The airframe was burdened with long-range missiles like AMRAAM, and avionics that enabled the F-16 to drop guided weapons during all weather types and all times of day. The avionics alone account for over 90% of the cost to manufacture a modern fighter aircraft, and they also consume a significant portion of the development cost and schedule.

As long as we insist on purchasing complex combat systems, Congress will have to realize that new-start developments can only be justified if they are amortized over a long production run. Naval acquisition, long held up as the poster-child for the broken DoD acquisition system, has several recent examples of uneconomical developments.

The Seawolf-class submarine was conceived in the 1980's as the ultimate boat for autonomously tracking and killing Soviet subs. When the Cold War ended, the Seawolf class was capped at just three boats. The Virginia class submarine was then launched as a smaller, stealthier boat that was more relevant to modern warfare. The problem is that the Virginia class consumed even more development dollars, and the cost per boat is actually more expensive than the Seawolf-class sub.

The DDG-1000 Zumwalt-class destroyer is another example. It's a hugely-expensive program that may be capped at just two ships, in favor of continued production of the Arleigh Burke-class destroyers.

The US Air Force recently suffered the most expensive plane crash in history, losing one of its 21 B-2 bombers after an engine failure on takeoff. The B-2 was so expensive because its development budget was justified by a planned purchase of 132 aircraft. When the total buy was reduced to just 21 airframes, the development costs made up a large fraction of each example's total cost. While the Soviet threat evaporated, continued production of the B-2 still made sense, for no other reason than to amortize development costs while creating replacements for elderly B-52's.

Because new-start programs are so slow and expensive, there's been a recent trend towards modifying existing airframes to meet the needs of new missions. The KC-45 tanker is a good example of this, relying on the Airbus A330 as a starting point. While a point-design tanker (likely based on the Boeing Blended Wing Body concept) would be best for the warfighter, the KC-45 promises to get a mission-capable product to the field in a shorter amount of time for a lot less money. But there are other times when modifying existing airframes isn't such a good idea, especially when requirements creep threatens to make the program much more expensive (the VH-71 presidential helicopter immediately comes to mind.)

Unless the military acquisition community takes an active approach to fighting the growing complexity of combat systems, the acquisition business will continue to be broken. And unless Congress and the budgeteers start to realize that massive development budgets for new-start programs can only be justified by lengthy production runs, every copy of a major combat system will be a hugely-expensive machine.

Thursday, March 20, 2008

It Could Always Be Worse

For ESAS critics such as myself, it's easy to lose perspective on things. While NASA's plan for returning to the moon is far from perfect, it could always be worse. Much worse, in fact.

Our biggest gripe lies not with the fact that NASA wants to put humans on the moon. That fact alone should excite us. But we've been here before with Space Exploration Initiative and other stillborn dreams of human spaceflight that NASA has peddled. The ESAS critics are afraid that the Vision for Space Exploration is unaffordable, and will soon be confined to the dustbin of history.

It would seem that, in some sectors, the response to Ares I vibration issues is almost gleeful. While I've never bought into Mike Griffin's absurd claims of the problem being a "mountain made out of a molehill," I could never accept that this problem would kill Ares I's development. It now appears that solutions have been identified, although they will cut into the vehicle's tight performance margins.

In the past I have been critical of forcing the Altair lander to perform the Lunar Orbit Insertion burn, and I have questioned the applicability of the ESAS architecture to a future lunar base. But now I stand with egg on my face, as I realize that the baseline Altair descent stage is totally appropriate for unmanned cargo landers to support a moonbase.

The Altair lander is different from Apollo, but Apollo wasn't perfect either. Contrast this with plans for an Apollo LM Shelter, LM Truck, and LM Taxi, which would require a redundant Apollo spacecraft to fly along and provide the LOI burn. The NASA team might not have a good way of explaining their plans to the public, but they have a plan nonetheless.

There can be no doubt that Ares I will be a lengthy and expensive development which offers little advantage over Heavy EELV's. But there are political reasons why government-funded manned spaceflight will not utilize them. As a libertarian believer in fiscal conservatism, I have no problem with laying off the majority of the shuttle employees as that program winds down. They are talented people who will have no problem finding work in space-related fields.

But in the real world, Ayn Rand's heroes could never win a popular election; the real world is run by incorrigible characters like Jim Taggart and Wesley Mouch. The thought of cutting NASA's overhead (the shuttle standing army) is political suicide for elected officials in Florida, Texas and Louisiana. As a federal jobs program, Ares I fills its role very nicely.

In theory, Ares I should be safer than a Heavy EELV. That could play a critical role in keeping Project Constellation going. Can you imagine the Congressional pressure to kill the program if a crew is lost? While the crew launch segment is but a small part of the total mission risk, every reduction in risk does help to forestall the day when Congress pulls the plug.

While I like the idea of coming up with a wide-bodied Atlas for human spaceflight, the idea will only come to fruition when there is a market demand for it. United Launch Alliance will have to convince investors like Bob Bigelow that money can be made off the relatively-modest investment that an evolved Atlas would require. If ULA goes to the government with their hats-in-hand asking for Project Constellation money, they will have become little better than the scoundrels at ATK.

ESAS is Mike Griffin, and ESAS will be the law of the land as long as Mike Griffin is calling the shots at NASA. While there are plenty of alternatives that are better than ESAS on both technical and budgetary grounds, I think the politics of the situation will force us into a choice between ESAS and Nothing At All.

Tuesday, March 18, 2008

Folded Hands for Folded Wings

During the course of America's battles for Afghanistan and Iraq, I've had friends deploy to the Area of Responsibility, friends who are currently deployed there (one of whom just volunteered for a 6-month extension,) and friends who are preparing to deploy. I have been fortunate in that, until today, I have never had to attend a memorial service.

Two weeks ago, Team Kirtland suffered its first loss during the terror wars. Staff Sergeant Christopher Frost, a public affairs NCO, perished along with seven airmen of the Iraqi Air Force when their helicopter crashed in a dust storm. I did not know SSgt Frost personally, but I wanted to make an effort to know a man who wore his uniform with pride and went about his job with great zeal, touching lots of people in the process.

Chris Frost cared about his mission and believed in his purpose when he was deployed to Iraq. He recognized the sacrifices being made by the Iraqi people in rebuilding a nation that has known nothing but continual warfare for longer than he or I have been alive. It is telling that on his final mission, he was the sole American traveling aboard the Iraqi Air Force Mi-17 helicopter, trying to tell the story of Iraq's fledgling Air Force.

Losing a young man in a freak accident of that nature is never easy to accept. The people who wear the uniform put themselves at great risk, even if they aren't dodging bullets on a daily basis, and even if they aren't deployed to the Area of Responsibility. We must go through life with the goal of making the most of each day we are given. We never know when our number is going to be called.

My heart is a bit heavier knowing that SSgt Frost is no longer with us to share his humor and zeal. We can only pray for the family he leaves behind, and hope that his children will be able to grow up proud of all that their father accomplished during a life that was too short but lived to the utmost.

Thursday, March 06, 2008

Was X-33 Really Such a Bad Idea?

I really don't even know what motivates me to post this, as X-33 is a long-dead program that hasn't held the industry on the edge of its seat in almost a decade. Yet the fact that the X-33 hardware is allegedly still in storage shows that, to some people, the concept still carries great hope.

Most observers of X-33 realize that the program was trying to accomplish too many test objectives, including:
--Metallic thermal protection systems
--Lifting Body aerodynamics and control
--Linear Aerospike engines for propulsion and steering
--Composite, multi-lobed cryogenic pressure vessels
--Rapid turnaround
--Autonomous runway landing
--On-board health management avionics

In hindsight, not all of the X-33 technologies turned out well. The aerospikes came in too heavy, and supposedly didn't meet their performance objectives during test firings. The composite hydrogen tanks proved difficult to construct without flaws, and failed their testing.

X-33's most glaring problem was the way it was grossly oversold. Lockheed Martin pitched the fantasy of VentureStar, twice as long as X-33 and over twice as fast--enough to achieve orbit without staging. Or so the fantasy went. Any aeronautical engineer could see through the ruse of the LockMart Public Relations machine and its blatant disregard for the laws of physics. But apparently it was enough to sway NASA into staking its hopes for replacing the shuttle on the mantle of X-33.

It's interesting to think about what would happen if X-33 was de-scoped early in the program. What if metallic propellant tanks were baselined from the beginning, rather than after the hydrogen tank failed? What if a conventional, off-the-shelf rocket engine were used? If these decisions had been made differently, it's possible that X-33 could have flown, and achieved a velocity sufficient to test the innovative Goodrich-developed heat shield (around Mach 12.) It's also possible that the repeated flights of X-33 would have taught us a lot about what is required to rapidly turn around and re-fly a spacecraft.

To be fair, there weren't a lot of choices for alternate engines on X-33. The Space Shuttle Main Engine was probably the best choice. Its vacuum performance was higher than that of the aerospike, but it's a finicky engine due to the complex staged-cycle combustion. Another option would have been the Ariane 5's Vulcain. Similar to the old J-2 in terms of thrust and specific impulse, it would be perfect for Ares I's upper stage (if only it was capable of air-start.) But the Vulcain wasn't designed for reusability, so it's uncertain if the engine would have been useful for the X-33.

So let's imagine for a minute that the X-33, after some moderate descoping, worked as advertised. It achieved a reasonable flight rate and performed successfully. What would be the next step towards replacing the shuttle and having a truly reusable launch vehicle? Forget about the VentureStar delusion and think about the possibility of strapping two or three X-33's together in a parallel configuration. Such a vehicle could probably make it to orbit with a usable payload. It probably wouldn't come close to being a shuttle replacement, but it would allow NASA (likely using the Air Force as a partner and end-user in development of this vehicle) to gain valuable experience with development and operations of a truly reusable spacecraft.

In a situation where two or three X-33's (or any bimese/trimese boosters) are flown in parallel, it is likely that the stack would lift off with all engines at max thrust, followed by throttle-down on the engines of the Stage 2 and Stage 3 boosters. A propellant cross-feed system might be employed, which would definitely add complexity to the system. When a booster would deplete its fuel, it could glide to a landing at a downrange airfield (depending on the chosen trajectory, burnout speed/altitude, and booster Lift/Drag ratio.) Because the X-33 was designed to be transported on a modified 747 (likely one of the two shuttle carriers,) recovering the boosters shouldn't be a problem.

But the "what if" questions about yesterday's choices fade away into the wispy clouds of memory. Instead, the Space Shuttle still flies; National Aerospace Plane, Shuttle II, DC-X, X-33, and Orbital Space Plane remain ideas whose promise was long ago discarded. And while the teams working on Ares and Orion might pride themselves on how far they've gotten, they need look no farther than the partially-assembled X-33 to realize how far along in the development cycle a program can go before it gets killed. The lesson is to determine what the requirements are, scope your program so it meets the thresholds with little risk, and never oversell what you're trying to do.