Centaur V

I was very excited when I read that ULA had decided to bring elements of the ACES upper stage forward and begin work on Centaur V. This will be a 5.4 meter diameter stage, likely use the BE-3U engine, and not feature the poorly acronymed but technically wonderful Integrated Vehicle Fluids (IVF).
I was a never a fan of the awkward, iterative development plan that was first shown with the Vulcan announcement. It focused early development on the first stage to remove the dependency on the RD-180 engines, but kept the upper stage at a 4 meter diameter, keeping Centaur unchanged until it was replaced by ACES. Bringing forward a few critical items from ACES to Centaur V is a brilliant move by Tory Bruno. The most important thing it does is increase the lift capacity of the first generation Vulcan, so it can carry all of the national security payloads. This means that as soon as it is qualified and in operation, ULA can immediately retire both the Atlas V and Delta IV families of rockets. Having to support 2 completely different, enormous rockets places a huge strain on ULA's ability to produce rockets in high quantities, and now that SpaceX has ended their monopoly on national security payloads, they can't afford that anymore. This is a good move.

But I wanted to offer a different development path that could potentially get Vulcan flying a little quicker for a lower cost, and still satisfy their need to replace the Delta-IV heavy. Instead of investing in new tooling and re-qualifying the main propellant tanks of centaur to a new diameter, they could instead choose to keep the 3.05 meter diameter and stretch the tanks to length to meet the payload requirements needed to keep the national security payloads flying. Vulcan is expected to have the capability to place about 40,000 kg in LEO in its final version, but to replace the Delta-IV heavy, you only have to be able to lift 28,790 kg. With this assumption in mind, I tried to figure out what a stretch Centaur V would look like. 
Obviously, the engineers at ULA must have considered this, weighed it against their options, and rejected the plan for any number of reasons. They have far more insight into their engineering bandwidth, manufacturing capability, actual performance and mass data, and future business climate than I do. A lot of them are outright better engineers than me too. That said, its fun to engage in a little speculation. 

The Centaur V concept I came up with stretched the total length of the propellant tank from just over 10.34 meters to 16.88 meters in length!
At first, this length stretches credibility, but for an end-of-life evolution, it doesn't seem that ridiculous when given some thought. We see it happen again and again on rockets, where they begin relatively stout, and as tanks get lighter and engines get more powerful they gradually lengthen from one iteration to the next. A look at a picture of the original Falcon 9 looks shockingly stubby compared to pencil-thin rocket we have gotten used to seeing recently. The original, dual-engine centaur was pretty stubby. Iteration D-1T, which first flew on the Titan IIIE, first flew in 1974, was only 7 meters long including engines. That puts the total tank length at only 5-6 meters long. 
The 670 kN BE-3U is also ridiculously overpowered for the stage, but an otherwise unused feature is that it has deep throttling capability - likely through the entire range of thrust, which can be used to control the vehicle acceleration when powered by such a monster engine. 
So, what happens if ULA decides to fly a rocket like this?
Similar to the existing centaur, this version would fly under the fairing. It appears that this would seriously cut down on the length of a payload that could ride with it. ACES is not planned to fly inside the fairing, like a stretched Centaur would have to. The longest fairing ULA currently makes for the Delta-IV heavy to support national security payloads is 19.8 meters. Including space for the engine, an enveloping fairing would need to have a total internal length of almost 40 meters!
40 meters is huge. But it doesn't seem improbable that Vulcan may need a fairing approaching 30 meters in length. It has an estimated lift capacity over 40 tons, 30% higher than the Delta-IV heavy it will be replacing. Future evolution may drive that number even higher, and as ride shares of satellites become more popular, they typically need to take up more space inside the fairing to get the full use of the launch vehicle. While National security payloads seem to be pivoting toward smaller, more survivable payloads, that does not necessarily mean they will be in separate orbits. A group of small satellites clustered near each other in the same orbit is just as survivable from ASAT attack as if they are spread out over the globe.
I was going to do some cool FEA to see what the buckling load of this structure would be (it certainly looks long!), but even under a 4.5 g burnout with a max payload, a modest internal pressure of 26 psig would be enough to prevent that from happening.

To make these estimations I started with some assumptions backed up with some hasty research about Centaur and the BE-3 engine. I plugged some numbers in to a python script and generated some lengths for what a stretched Centaur would look like given a target Delta-V and mass. I put the code up on github, so play around with it and see what you can come up with, or tell me what errors I made.
​https://github.com/Lars-0/StretchedCentaur/