Some of the comments to my last post got me thinking about what I’m trying to accomplish with this series. The reality is that each of these approaches that I’m discussing could easily fill a full chapter in a textbook, complete with 20-30 pages of text, tons of graphs, equations, sample designs, detailed discussions of tradeoffs, etc. I’m probably not the guy you would want writing such a textbook–that’s something better left to either a Masters/PhD student looking for a fun dissertation, or someone who has more aerospace engineering experience than myself (say a Mike Kelley, or a Dan DeLong or maybe a group of such people).
This morning, while I thought back to the UND lecture that started this all, I realized that the key goal of this series has always been to show that there are several realistic approaches to doing RLVs, and to try and give a high-level overview of the different approaches and how to get there from here. While there is definitely a lot more detail that I could go into on each of these topics, they’re not the only ones I want to write about, and I just don’t have the time to both go into the level of detail some would prefer while also being able to do much of anything else. If someone is interested in taking what I’ve got here, and fleshing these out into a more formal and detailed form, let me know. Otherwise, I’d like to just continue as I’ve been going with giving a high-level overview of the most promising orbital access techniques I’ve been looking at.
In Part III, we discussed a TSTO approach where the first stage provides only vertical velocity, and the second stage provides all the horizontal velocity. As many have probably notice however, requiring the upper stage to provide all the horizontal velocity makes the upper stage design a lot more challenging, and also tends to drive the overall vehicle size up substantially. The obvious question is, are there ways of having the first stage provide horizontal velocity, while still returning to the launch site? It turns our that there are some ways of doing that, and this post will focus on the first, and by-far easiest of those methods: glideback.
Glideback TSTO: An Introduction
As has been mentioned several times in this series, the rocket equation is an exponential function. As you near the “right-side of the curve”, ie higher velocities, the engineering challenge of building a reusable stage becomes rapidly more difficult. The corollary of this is that by moving the velocity requirement for a stage even slightly lower, the gains can be quite large. For instance, by going with air-launch, I showed that making a functioning “assisted SSTO” may actually be achievable with near-term available technologies, while a ground launch SSTO is still a much harder challenge. Likewise, even though the pop-up TSTO approach only saves the upper stage about 600m/s over the air-launched SSTO approach, it too makes a big difference. So, at least on the “part of the curve” (delta-V versus mass ratio) that we’re looking at for an orbital stage, adding even a small amount of downrange velocity can still have a very large, and beneficial impact. The challenge is doing so while still maintaining the operational advantage of being able to have the first stage return directly to the launch site at the end of its mission.
The easiest way to accomplish this is by using aerodynamic lift. The idea behind glideback is that the first stage takes the upper stage up to a certain altitude, and gives it some downrange velocity, then it stages, decelerates a bit, turns around, and glides back to the landing site. Naturally, the better the L/D of your system, the more delta-V the first stage can impart while still making it back home, so while it may be feasible for a VTVL first stage to take some advantage of this technique, it is more naturally suited to HTHL approaches.
This presentation, done by Barry Hellman of Georgia Tech, provides some more details on this approach (as well as the boostback approach which will be discussed in Part V), and more details can be found by googling “glideback” or by searching for “glideback” on NASA’s NTRS site. The idea was previously investigated as part of the Shuttle II or Future Space Transportation System studies done in the late 80s and early 90s. The basic concept is that the two stages take-off horizontally, accelerating to about Mach 3-3.2 (~1100m/s) at an angle of about 45 degrees (thus providing about 775m/s of horizontal delta-V), and then staging at an altitude of about 32.5km. At that point, the first stage performs a high-alpha reentry to slow down a bit, turns around, and glides back to the launch site for an unpowered horizontal landing. Mach 3.2 was chosen as the optimal point for the Shuttle II analyses (though I’m not sure all of the assumptions going into that number), as going much faster would preclude being able to return to the landing site on gliding alone. There are different variations on the theme that are possible, and different assumptions will yield different burnout velocities, angles, and altitudes (ie Your Mileage Will Vary), but that was the basic idea.
Analysis of the Staging Maneuver and Booster Glideback for a Two-Stage, Winged, Fully Reusable Launch Vehicle
So, what are the benefits of this approach compared to the other ones we’ve discussed so far?
- The first stage in this approach is actually imparting a significant amount of horizontal delta-V (almost 800m/s), thus making the upper stage’s job much easier.
- This approach takes a lot more advantage of the benefits of HTHL approaches, in that it’s using wings to lower the required takeoff T/W ratio, and using the wings to do a lifting ascent.
- The engines on the booster stage can be much simpler than for a VTVL booster stage. You might not need throttling or gimballing, thus allowing for a much simpler propulsion system–if MSS had been doing HTHL, and if it had had access to an airframe, our engines were mature enough two years ago that we probably could’ve had our own EZ-Rocket flying for over a year now.
- Due to the lower delta-V requirements on the upper stage it becomes much easier to make the upper stage use a denser propellant combination, without taking as much of a hit for the choice.
- The reentry velocity for the first stage is even lower than most suborbital vehicles, thus completely eliminating the need for any first-stage TPS.
- The first stage doesn’t require a very high mass ratio, thus making it quite low-tech. While much larger than an XCOR Lynx Mk II, the vehicle would only need technology on-par with the Lynx Mk I to be workable–ie the technology risk is very low.
- HTHL vehicles tend to provide for much more graceful abort modes. For instance, a total propulsion failure of the first stage might not even require stage separation. You might just dump oxidizer, and then glide back to a landing. Fixed engines are much easier to “armor” against hard starts (and are much easier to make more deterministic than a throttling engine, thus making hard starts potentially less likely).
- Due to the low-Mach number, and low required Mass Ratio, the first stage has much more in common with a normal aircraft than a launch vehicle–it can borrow heavily from aircraft construction techniques and some subsystems, thus leveraging a more highly matured transportation industry.
- Depending on the flight trajectory taken, the first stage might not actually meet the AST definition of a suborbital rocket. While it isn’t clear why you’d want to have the first stage regulated by the other part of the FAA, if you wanted to, you probably could force the trajectory either way depending on which you thought was more commercially useful.
- HTHL vehicles like this are more likely to be able to operate out of existing airfields. While operating out of LAX anytime soon is unlikely, there are plenty of large airfields out there that could easily attain the required FAA launch site licenses by leveraging work done by the Oklahoma and Mojave spaceports (not to mention just using Mojave or Oklahoma spaceports). This flexibility makes it easier to operate out of multiple launch sites not necessarily tied to existing (and expensive) launch ranges.
- The first stage operating by itself without a fully-fueled and loaded upper stage on top probably has enough propulsive power to make several hundred miles downrange. It can also probably do so while operating as a rocket powered aircraft, thus making it easier to self-deliver the first stage to a given destination. Once again, how much the FAA would appreciate someone trying to do this is left as an exercise for the sufficiently masochistic reader.
- The first stage has a low enough required MR that you can probably include hardware, such as ramps, that would allow an unfueled upper stage to be remounted to the first stage without the use of a crane. Sure, that goes against standard aerospace weight-minimizing practice, but if it allows cheaper and easier operations, it might be worth it. Any time you can allow for ground level servicing, maintenance, and inspection, it makes operations a lot easier.
Once again, there may be other advantages I’m overlooking, but those were some of the key ones that I could think of.
Drawbacks, Limitations, Constraints, and Challenges
As you probably guessed, there are some drawbacks to this approach in general, and the specific implementation mentioned above. Unlike the Pop-up TSTO approach, there’s a bit more flexibility on the exact trajectory, which means that some of these issues may be resolveable by using clever trajectory planning.
- The staging velocity and altitude result in a fairly severe dynamic pressure environment during stage separation. 800psf to be precise (38.3kPa for our metric-using friends). This makes staging a lot more dicey. The article that I pulled the picture from includes some analyses on how to solve this problem, but it still has a fairly high associated pucker factor. It may very well be worth redoing the analysis with staging dynamic pressure being given a higher weighting factor (ie. at the cost of some performance).
- Staging at 32.5km at that speed and angle also means that your first stage apogee only reaches a little over 60km. That means that you’re going to take some gravity losses with the upper stage unless the T/W ratio is really high. This is especially the case if you coast up to a higher altitude to do staging. This will slightly reduce the benefit of the downrange velocity. It might be possible to change the trajectory such that the first stage apogee is 100km, and the staging point is over say 50km to keep the dynamic pressure down, while still keeping some or all of the downrange velocity, but I’m not in a good position to say what the tradeoffs would be.
- The wings and landing gear for the first stage have to be designed to handle lifting the full stack, and for doing emergency landings. Fortunately the first stage doesn’t need very high MR, so this isn’t as big of a problem as it would be for a ground-launched SSTO for instance.
- If the first stage is running a trajectory that causes it to be classified as a launch vehicle, it’s IIP will stop over some downrange point. Also staging occurs with the IIP at some downrange point as well. It will be important to try and locate this point such that it isn’t over populated areas. This may limit somewhat the available launch azimuths, and may require the first stage to have some extra performance margins in order for different launch locations to shape the trajectories to minimize the E-sub-c for the flight.
- There are also issues with scalability. While the NASA study mentioned previously was for an HLV sized vehicle, realistically, it’s going to be a challenge getting anywhere near that big anytime soon.
- The orbital stage TPS problem. Same as with the other approaches, but as the stage gets lower delta-V, it also becomes slightly less fluffy, which tends to increase the TPS material challenge.
- Glide landings are no fun, but depending on the engine concept, it should be possible to do what XCOR does, and have propellant on-board and the capability to do a “go-around” burn. As it is, it’s been fun over the past few weeks watching a certain “Undisclosed Flying Object” do multiple in-air relights (and some pretty sweet maneuvers) over the Mojave Spaceport. For some reason I think that making the landing not have to be a glide landing wouldn’t be that difficult to design in from the start…
There may be other issues, but the two biggest ones have to do with the trajectory, and it might be possible to design the trajectory to avoid them.
This approach shares many of the enabling technologies with the other two approaches. Reusable TPS, orbital tugs to offload some of the dead-weight on the stage, suborbital vehicles help provide experience with handling similar vehicles, composite tanks always help with HTHL design (since you can now do a cryogenic “wet-wing”, and have more integrated structural tankage/insulation), etc.
There’s another potential non-technological (regulatory) enabler that an affiliate of ours at MSS is working on, but I’m not sure if I can go into it yet. It would also be beneficial to suborbital operations including both HTHL and VTVL operators.
The Path Forward
As you’ll be noticing if you’ve read the previous parts, there’s a common theme for most of these orbital RLV approaches. Almost all of them have big unknowns when it comes to TPS. Almost all of them can benefit from work being done currently for suborbital vehicles. Most of them can benefit from subscale “proof-of-concept” testing using suborbital vehicles in development as “first stages”. This is particularly the case for this approach.
In fact, the HTHL work that XCOR aerospace is doing right now for their Lynx vehicle is directly applicable to what would be needed for a glideback TSTO design. In fact, as they’ve been saying for a long time, they’re planning on using Lynx or Lynx Mk2 as a nanosat launcher. Using a slightly modified Lynx or Lynx Mk2, you could do work on things like staging techniques, trying out various trajectories, abort mode practice and planning, etc. Not to mention that the technologies being developed for Lynx and Lynx Mk2 (especially the cryogenic LOX tanks) are directly relevant to this TSTO approach, for the exact same reasons. I know that the XCOR guys, for good reason, are very quiet about their ideas about how to proceed beyond suborbital, but I’m almost positive that something like this is how they’d go about it if they were ready to take that next step.
But as with the other approaches, while the path ahead is fairly clear, it’s still involved. XCOR’s been doing excellent rocketry work for almost 10 years now, and they’re barely getting enough traction in the funding world to get their suborbital vehicle into full-time development. But once it’s in operations, taking the next logical steps should be relatively quick–provided someone has the funding and the interest. But that’s a post for another day.