For a long time in the car world, due to the high setup times inherent in some of their equipment (like the massive multi-ton dies they’d use for panel stamping) and other factors, car manufacturers would use a “batch-and-queue” style manufacturing system. The logic went that since it took a full day sometimes to change from one model to the other (due to for example, needing to say swap dies, remount the new die, and then precision callibrate it) that it made sense to switch models as infrequently as possible. So, if you were producing car models A, B, and C on a given production line, you’d produce all the A’s you think you’d need for a month, then proceed to B’s, then to C’s. The problem is that this is wasteful, takes up a lot of space, increases the odds of damaging goods, and requires a lot of capital to be tied up in intermediate stages of production where it isn’t actually generating revenue. It also depends a lot more on forecasting, computerized inventory, and all sorts of other things that tend not to work very well.
When the Toyota Corporation was getting back on its feet after WWII, they realized that doing things that way was a luxury they couldn’t afford. Most people who know anything about manufacturing have heard the terms Just-in-Time or Lean Manufacturing. With Just-in-Time, the goal is to reduce inventory levels as far as possible, so that production signals from the end customer can tell you when to produce what, instead of trying to guess and forecast, and then end up producing lots of crud that nobody wants to buy. In order to do that though, something had to be done about those huge sheet metal presses. There’s no way you can run on only a few pieces of work in process inventory if it takes a day to switch dies. So Shingo decided to set an audacious goal: he wanted to be able to reduce the setup time for one of those dies from as much as 24-36 hours down to 1 minute. The disturbing thing is that he actually went on to do it. The technique he developed, called Single Minute Exchange of Dies or SMED for short, is one of the key enablers to modern Lean Manufacturing. All of the sudden when swapping from model to model to model only takes about a minute each, you could put that machine back into the main line. You could swap dies between each and every panel if you wanted. You no longer needed to produces weeks and weeks of panels at a time, because you could build precisely what was in demand at that time. In short, it was a really big deal. RLV operators would be well advised to pick up a copy or two of Shingo’s book on the topic and take it to heart. I want to see races in a few years where the turn time for an RLV is dropping below 15 minutes.
So, what the heck does this have to do with space?
The way I think this relates is that it shows that the end goal you are working for can often lead to completely different means. Imagine what had happened if Shingo, like hundreds of other talented industrial engineers at the time had merely settled for the “more realistic” goal of dropping the time of die swaps by 50% or even 1 order of magnitude down to a few hours? He probably would have acheived his goal, but in some ways it might have been just as difficult, and yeilded far less benefit. If it still takes you 2-3 hours to swap dies, you’re still stuck at least producing weeklong batches if not monthlong ones. I think the reason why Shingo was the one who made so much progress so fast in that field wasn’t just because he was a genius, and wasn’t because all the other IE’s in the US were a bunch of knuckle-dragging neanderthals. I think it’s mostly due to the fact that he was actually looking at the problem right. Had any of those other IE’s thought “why couldn’t we switch dies in only a couple of minutes”, I think that they probably would have beat him to the punch. Once you accept the possibility that the goal probably isn’t physically impossible, or even silly, you’re 90% of the way to a solution. The last 10% may be a real bear, but you’re most of the way there once your perspective is right.
Take a look at the EELV program, and even SpaceX. EELV’s goal was to reduce the cost of launching satellites for the military from absolutely obscene to merely ridiculous (ie a 50% drop in price IIRC). So, they tried to make some incremental changes to how they build and operate their vehicles. In some areas they’ve gotten a lot better, but the reality is that they didn’t even acheive the modest goals they set out for themselves. It isn’t that they’re dumb, or malicious, or incompetent. It’s just that they set themselves too easy of a goal, so they didn’t actually have to think outside the same high-cost artillery box that they’ve put themselves in over the years.
SpaceX is doing quite a bit better. They wanted to slash the launch cost by 10x. I think they’ll pull it off. But that’s kind of like cutting time from 24 hours to 2.5 on the sheet metal press. Sure it’s a huge improvement, sure it’ll make a difference, and sure it is possible. But ironically it may actually be more difficult than going even more radical. I think that SpaceX will eventually figure out some recoverability for their launchers. Might even cut the price they charge customers by another 10-20% compared to a fully expendable vehicle. They might even get up to two 9’s of reliability. But if they go for the BFR instead of trying to radically change the Earth-to-Orbit transportation market by going fully reusable…They’re probably going to get their lunch eaten. I mean, they could possibly acquire one of the companies that actually develops a fully reusable, high-flight-rate orbital space transport. But the reality is going to be that if they don’t keep pushing more and more reusability into their Falcon line, it’s going to go obsolete. In fact, I’m not even sure if they can get to there (sufficiently full reusability to maintain competitiveness and marketshare) from the vehicles where they’re starting. Don’t get me wrong. I think in the near term SpaceX is doing something absolutely wonderful. I think it will change a lot of things in the space industry. But since they took the “how much lower cost can we go based on evolutionary improvements on the status quo” approach, they’re already limiting their long-term competitiveness in the ETO transportation market. All it takes is a couple of Shingos to figure out that there’s nothing impossible about making a vehicle that is safe enough, reusable enough, and inexpensive enough to drop the launch price by another order of magnitude from SpaceX in order for that to happen. Perspective can be everything.
Also look at lunar exploration and development. For the longest time, I was thinking about how I could reduce the cost of a lunar mission relative to NASA. SpaceDev provided a good example of this a few months back. They developed a plan where they could get back to the moon for about 1/6th of what NASA is likely to spend, and do so in a manner that is actually more flexible (though not neccessarily more capable). Back when I first did my studies on my Prometheus Downport Project, I was thinking along similar lines. I thought “NASA wants to spend many Dirksens on getting back to the moon and building a base, why couldn’t it be done for a single Dirksen?” So I borrowed some ideas from George Herbert and came up with a kinda crazy scheme involving one-person landers, crashing TLI/Descent stages, some on-orbit assembly, and some lunar surface staging. The problem was the idea wasn’t very practical, and didn’t make any economic sense.
A while back though, I started realizing that the nearest term market for lunar access was probably tourism. And that presents a problem. Futron’s study pointed out that even for really exotic destinations, most people are only willing to fork out a couple of percent of their net worth on a trip. For something like being the next person to walk on the moon, they might even be willing to fork out a fairly substantial portion, but the reality is the number is likely going to be less than 10-20% at the most. The problem is that the number of people with a given level of net worth seems to be exponentially inversely proportional to the level of that worth. While there’s only a handful of people on the earth with a net worth over $10B, there’s a huge number of people with net worths of $100M-1B. So, the reality I started running into is that if you want an individual to be able to buy a lunar ticket, the seat price has to be less than $100M, and probably closer to $20-50M before you’ll even get a single taker. Listen to the sound of crickets chirping over in line to sign up for Space Adventure’s $100M translunar ticket. So the question becomes, how can you get the ticket price for a lunar tourist *in the near-term* down below $50M? Honestly, I’m not 100% sure, but a lot of the ideas I’ve been harping on I think will be part:
- Reusable, high-flight-rate ETO transportation
- On-orbit propellant tranfer and storage
- Reusable translunar transportation with aerobraking
- Lots of intermediate space tourism markets like suborbital, orbital, and translunar
- On-orbit refuelable/reusable lunar landers
- Maybe ISRU
In the long run, ISRU and maybe some form of high Isp transportation like microwave thermal becomes critical if you want to push the price point to the low single-digit millions numbers where you really want to reach if you want to open a large tourism market. But the point I’m trying to make is that if you actually want a lunar transportation system that is economically useful in any sense of the term, building huge welfare queen Shuttle-Derived hardware is never cut it. Not even a low-cost Russian expendable booster infrastructure will work. Only a radical rethink has any chance.
And in the long run, I think it is just such a radical rethink that has the highest possibility of success.