Raptored Vulcan Part 2: Performance
In Part 1, I introduced the idea of Raptor engines on Vulcan and why this has been proposed by space launch fans, and I ended with three questions:
Would Raptors work on Vulcan?
Would ULA buy them?
Would SpaceX sell them?
In this post, I will try to answer the first question, leaving the other two mostly for Part 3.
I will briefly touch on the second question, though. ULA's CEO Tory Bruno has already politely said no to Elon Musk's joking offer and has also publicly stated that the BE-4 is the right engine for them.
And indeed, replacing BE-4 even with a similar methalox engine would be no small feat. Mount points and plumbing, control software and timings, validation and integration procedures would all need to change. The oxidizer to fuel ratio would change and thus the oxygen/methane tank partition would no longer be optimal. Tanking operations might be impacted. And then there's of course all the negotiation and paperwork that would go into ensuring ULA was getting a good deal on reliable engines that would be delivered in a timely manner, not to mention a likely costly cancellation of their contract with Blue Origin.
Given how close BE-4 is to being ready, changing engines now would not make Vulcan fly any sooner. It would add further delay.
But what if Raptor wasn't a last minute change to get Vulcan off the ground as soon as possible, but an upgrade for a future version? Would that make sense? And what would it look like?
One of the common responses to the suggestion of replacing BE-4 with Raptor is that you'd need three of them, because Raptor doesn't have as much thrust as BE-4 (Vulcan uses two BE-4s). This argument often assumes 185tf (ton-force) thrust that matches the initial iteration of Raptor used on Starship prototypes. But Raptor 2 is in late stage development, with over 230tf of thrust planned -- still a bit lower than BE-4, but not by much.
Vulcan comes in four configurations, with 0, 2, 4, and 6 solid rocket boosters. For every configuration with solid rocket boosters, the slightly lower thrust would not be an issue for takeoff, but would just lead to slightly higher gravity losses that could be more than compensated for by the higher isp.
For the "slick" configuration with no side boosters, Vulcan will already have to be underfueled to have high enough liftoff thrust-to-weight ratio. A Raptor-based Vulcan would have to be underfueled by a bit more. We'll take a look at how much this might impact performance down below.
Another objection is the oxidizer to fuel ratio. Because of its higher chamber pressure and better propellant mixing thanks to its full flow staged combustion cycle, Raptor operates at a higher O:F ratio, closer to stoichiometric. This means that Vulcan's tanks won't be proportioned optimally for Raptor, suggesting the common dome location would have to be changed.
While it would be best to make this relatively minor change, it's not strictly necessary. Because more than three quarters of the propellant is oxygen, simply fully filling the oxygen tank and underfilling the methane tank to get the right O:F ratio yields only about a 1-3% drop in total propellant load, depending on how much of a difference there is in the O:F ratio.
One more qualitative question to resolve before we get to a quantitative analysis is which Raptor engine to use. Raptor, aside from its continual technical evolution, has two major variants: a sea level optimized version and a vacuum optimized version, which I'll refer to as RSL and RVac. The difference is in the nozzle -- RVac uses a larger nozzle for a higher expansion ratio, which enables greater efficiency and thrust at high altitude, at the cost of efficiency and thrust at take off and for a little while after.
Most vacuum-optimized nozzles have such high expansion ratios that they can't be safely operated at sea level, with flow separation threatening to destroy the engine. However, likely because Starship needs very compact engines, both RSL and RVac have lower expansion ratio nozzles than what might be expected, and because Raptor operates at very high chamber pressures, RVac is able to have very high efficiency without excessively low exit pressure (comparable to and likely a bit higher than the Space Shuttle and SLS's RS-25s), enabling it to operate at sea level.
When I mention this, I'll sometimes encounter the retort that RS-25 has a special feature to enable it to operate at sea level. The special feature, as best as I can make out, is simply nozzle geometry -- it has a lower than optimal nozzle angle at exit. More importantly, it's been confirmed repeatedly, including a few months ago by Elon Musk that RVac can in fact fire at sea level.
So, we have options.
Since even the first stage of Vulcan will do most of its work at high altitude, it actually makes sense to opt for the vacuum engine, except for the slick (no side booster) configuration, where the higher sea level thrust of RSL allows a greater propellant load. The solution, then, is to use RSLs for the slick configuration, which actually seems unlikely to see much use, and to use RVacs for the configurations with 2, 4, and 6 side boosters.
But two other questions arise: can RVacs gimbal? and will they fit?
The two engines of Vulcan need to be able to gimbal in at least one dimension (orthogonal to the axis connecting their centers) to provide 3-axis attitude control (yaw control can be provided using differential thrust without gimbal, pitch control through gimbaling the engines in the same direction, and roll control by gimbaling them in opposite directions).
It's true on that on Starship, RVacs are currently fixed with no gimbaling mechanism, but since RVac has the same powerhead as RSL, there's no reason to think that RSL's gimbaling mechanism cannot be applied to RVac. On Starship, the RVacs seem to have been fixed to economize on space, mass, complexity, and cost, much like the outer ring of RSLs on the first stage that are planned to be identical to the center RSLs except that they will lack gimbaling.
As for whether two RVacs will fit on Vulcan, the answer again seems to be yes. Their outer diameter seems to be about 2.3m, compared to RSL's 1.3m and BE-4's 1.9m. Vulcan has a core diameter of 5.4m, leaving 80cm of margin, enough space to keep the nozzles separate from each other and within the diameter of the core, even when gimbaling, since they don't need to gimbal towards each other.
So... what would Vulcan with Raptors look like?
You can see that the physical changes are relatively minor in @StarshipFairing's beautiful illustrations above.
For the Raptor-engined Vulcan core, we decided to keep the oxygen tank volume the same but shrink the methane tank to match the higher O:F ratio (3.5 vs 3.2 per Aeneas). I went with a 8.57% reduction in methane tank volume and a 2.04% reduction in overall propellant load for the performance simulations. In @StarshipFairing's illustration, you'll see a somewhat more modest reduction in the methane tank size that's just enough to accommodate the longer RVacs.
The slick configuration is a lot shorter with RSLs, but since no SRBs are used with that, we don't have to worry about RSL exhaust impinging on the SRB nozzles.
The slightly lower maximum propellant load had relatively minimal payload impact on configs with SRBs and no impact on the slick configuration, since it is underfueled for the sake liftoff thrust to weight ratio, anyway.
Here are close-ups of the engine area:
Engine Specs
To evaluate the impact of a potential engine change, we need engine specifications for both the outgoing engine and its replacement. Unfortunately, publicly available data is sparse, and in the case of Raptor, ever changing. But based on what has been shared publicly, we can try to puzzle out other key figures. For BE-4, RSL, and RVac, this is what Aeneas (@Phrankensteyn) has come up with:
Credit: Aeneas @PhrankensteynHere, Raptor is taken to operate at 305bar of chamber pressure, which sounds like a sensible estimate. Apparently it currently operates regularly at 300bar, while SpaceX seems to be working on pushing it higher, possibly as high 330bar. Isp figures, at least vacuum isp for RVac, as well as twr figures, also seem inline with the ranges Elon Musk has indicated.
Published payload capabilities for Vulcan, on the other hand, suggest that the BE-4 isp figures above might be optimistic. In simulation, getting the performance down to match ULA's announced figures with these engine specifications called for both a greater mass at cutoff for the Vulcan core than might be expected and a lower propellant load than the tank sizes suggest.
This could be due, at least in part, to high propellant residual fractions. Another possible explanation is that BE-4 has a substantially higher mass than the estimate in the table. This would make sense when comparing it to other engines. The RD-180 used on Atlas V, for example, uses a similar ORSC cycle with a denser kerolox propellant at a higher chamber pressure, leading to a more compact engine that is also pushed harder, unlike the BE-4 which targets gentle use to ease reuse. RD-180 has a twr of 78, whereas BE-4 here is estimated at 98.
Relative Payload Performance
For the slick configuration, RSL managed to match and even slightly exceed the performance of BE-4, despite a lower propellant load and a smaller isp gap. And as mentioned earlier, this configuration with the least Raptor advantage seems unlikely to get much use given that apparently it was not requested once in the first 30+ orders.
Raptor showed up to ~80m/s higher gravity losses for the slick configuration, and ~60m/s for the 2-srb configuration, but this was more than compensated by its higher isp. The 4- and 6-srb configurations showed no discernible difference in gravity losses, but because of unknowns around exact launch trajectories and optimal Centaur fueling, I added a 40m/s gravity loss for those Raptor variants to give a more conservative estimate.
For the slick configuration I underfueled both vehicles to achieve a liftoff twr of 1.2, with less propellant on the Raptor-engined Vulcan to match its slightly lower thrust.
ULA's Vulcan payload figures from 2019 show a cutoff at 27.2t (vehicles that should exceed that figure to LEO, like Vulcan Centaur w/ 6 SRBs and Vulcan Centaur Heavy, are capped at that level). There is a good chance this is a structural limit on Centaur V. Structures could be upgraded to increase LEO payload or LEO performance could be capped at 27.2t as represented by the dashed line in the chart.
To be clear, the figures in this chart may not be very accurate and should be taken with a big grain of salt. Especially beyond LEO+5km/s, I would not expect the charted figures to be very accurate. The purpose of this analysis was to show the comparative performance of Raptor and BE-4 in a Vulcan application, and the conclusion seems clear enough even with large error bars. Two Raptor 2 engines on Vulcan would match or exceed (by a lot) the performance provided by two BE-4s.
One additional caveat here is that the Raptor performance figures might be baselined on subcooled propellants, which would enable the powerhead to pump a bit more propellant to a bit higher pressures. If they are and if ULA isn't planning to use subcooled propellants for Vulcan (they probably aren't), Raptor thrust would take a hit, and by extension so would Vulcan-Raptor performance, though it would still likely have substantially higher performance than Vulcan-BE-4 in the configurations that use SRBs. Alternately, ULA could do the work of switching to subcooled propellants, but this might not be worth it for them, and it would also improve Vulcan-BE-4 performance, though not as much.
What does this mean?
Raptors could work on Vulcan with modest changes to the vehicle's structure and provide a big performance boost. But performance alone is unlikely to be enough motivation for such a change.
Are there other factors that might make it worth it? Read more in Part 3.
Comments
Post a Comment