Prototype of a rotating detonation engine, previously thought impossible

I’ve personally worked in rotating detonation engine (RDE) research, so I’d like to add some background on the field as well as my perspective on the novelty of the results in this paper.

But first, I’ll link some cool videos of RDEs firing on Youtube that you might enjoy.

Purdue University firing a RDE using LOX/Natural Gas (successful firing at ~45 second mark): https://www.youtube.com/watch?v=pHcxI-8GtZg

UT Arlington video from 2009 of RDE running H2/O2 (it looks like a successful firing occurs at 3 minutes, 22 seconds). https://www.youtube.com/watch?v=QbJgOk2rDbI

I agree with another comment here that it can be hard to tell when an advancement, such as the one claimed by the authors of the paper, really first occurs. However, the work of the article linked is certainly not the first time that a rotating detonation engine running H2/O2 has been demonstrated successfully or for an extended period of time (read: dozens of minutes). Some short tests using H2/O2 are shown in the UT Arlington video above. A wealth of research on RDEs successfully running H2/O2 can be found on Google Scholar going more than 7 years back.

AFRL East (specifically the Turbines and Combustion Branch at Wright-Patterson AFB) has been running and publishing results of rotating detonation engines using oxidizer/fuel combos of atmospheric air and standard liquid hydrocarbon aviation fuels for several years now. AFRL West (in this case referring to Edwards AFB, which was involved in the research in the linked article) is interested in RDE applications where there is no atmosphere to use as an oxidizer supply, so an oxidizer would need to be held in a tank. AFRL East is instead more interested in terrestrial applications, so RDE research work there has naturally focused on using air as the oxidizer.

There are several aspects of using liquid fuels and regular air that create additional challenge compared to using gaseous fuel/oxidizer combos (including H2/O2), not the least of which is ensuring proper atomization and mixing of the liquid fuel in the vary narrow window of time that it can be fed into the detonation annulus before a detonation front returns to consume them.

For a 6-inch diameter RDE, the detonation wave circles the annulus around 3000 times per second, depending on the fuel and oxidizer used and their equivalence ratio (the ratio of fuel and oxidizer in the annular detonation chamber). This creates an incredibly loud sound that is roughly 3 kHz, which is directly related to the 3000 times/sec the wave circles the annulus. Sometimes, two or more (in the paper they mention five) separate waves (distinct detonation fronts) will emerge, usually distributed equi-radially. This means that a detonation wave will pass a given point in the annulus five times as frequently, so the noise heard by an observer is then 5*3kHz = 15kHz.

When using aviation fuels in air-breathing RDEs, it is typically necessary to preheat the liquid fuel before mechanical atomization via injector, which is easier in H2/02 configurations. Hydrogen and oxyen are stored in tanks as liquid and may be quite cold, but between raising their temperature before injection using waste heat [0] and their higher volatility, getting them to atomize and mix is considerably simpler.

At my former employer, when we tested new parts for our research RDE rig that were fresh from the machine shop, we would run H2/O2 or H2/N2O as an easier cheat test to see if there were any glaring manufacturing issues before moving on to the more involved air and liquid aviation fuels tests.

So yes, I’ve personally seen larger RDEs running on H2/O2 fire literally hundreds of separate times. I can think of five or so other institutions doing RDE research that have already accomplished (and published on) more than what I see in this article. I love the RDE research space, but the linked article has me scratching my head at their claim to be first.

[0] https://en.wikipedia.org/wiki/Regenerative_cooling_(rocket)

How much higher ISP, and how much more engine weight/complexity? If it’s 15-20% higher but requires heavier engines that are harder and more expensive to maintain or use, it’s not going to be that useful or revolutionary. If it causes excess vibrations, it’s unusable. If it doesn’t scale up over a hundred thousand pounds of thrust, it’s a not very useful toy.

The biggest cost in space travel isn’t fuel, fuel is so cheap it’s a rounding error in today’s launch costs. The big costs are the manufacturing cost of making expensive hardware.

We already know the solution to dramatically lowering the cost of space access, it’s reusability. Burning up your entire stack every launch is insanely expensive. SpaceX ‘s reuse of Falcon 9 first stages saves as much as $30M per launch on the already cheapest stack due to mass manufacturing techniques.

The Shuttle cost was around $50k per pound of payload to orbit. Before SpaceX commercial launchers were $5k to $10k per pound of payload. A expendable Falcon 9 sells for around $1,500 per payload pound, a reusable Falcon Heavy launch sells for near $1,000 lb.

A fully reusable launch system would be dramatically cheaper. Each Starship launch is expected to put over 200,000 lbs of payload into orbit for about $500,000 in fuel. Keep refurbishment costs to $10M per launch and amortize the build cost over dozens of launches, and your cost per pound of payload is down to roughly $100. And it would be the largest, most capable lift vehicle ever made.

And it will do it with engines 20% less efficient than the Shuttle. Efficiency is only useful if it’s application cost is low.

Reading news about this research over years, hard to be sure when the "world-first" event actually happens. For example, a couple of years ago (?) similar achievement was claimed by Energomash in Russia.

In any case this seems to be wonderful technology, when made working. Kudos to scientists and engineers involved, it's hard research.

This page [1] provides a useful overview of the fundamentals of detonation engines (pulse and rotating), including some more background on why detonation is more efficient than deflagration (see Figure 1 in the linked document for the relevant pressure-volume diagrams, followed by the expressions for their efficiency).

[1] https://www.intechopen.com/online-first/a-theoretical-review...

Do we have any idea about what kind of ISP these rotating detonation engines can achieve?

And secondly, want to bet that the KSP2 devs are rushing to add this engine ASAP?

Anyone else think about the Aurora spy plane fables from the 1990's?

"strange, loud pulsating roar... unique... a deep pulsating rumble that vibrated the house and made the windows shake... similar to rocket engine noise, but deeper, with evenly timed pulses."

https://en.wikipedia.org/wiki/Aurora_(aircraft)

This is speculation on my part, but it seems like this might be difficult to sustain as the external air pressure changes as a rocket climbs. While the gas supply pressure can be controlled, changes in external pressure seem likely to affect operation, e.g., as air density gets lower it exhibits less resistance to the exhaust, changing the backpressure. Maybe that can be sensed and compensated for? That may add additional complexity and reduce reliability, especially if achieving stability requires very fine tuning. On land or in space there may be no such issues, but air pressure and temperature changing as a rocket climbs may cause issues. I'd want to see it fly before I'd believe it could be used in rockets.

The first question that comes to mind is: can you still have a pre-burner and turbopump? I think the answer may well be "yes", but it could be complicated. What if having any exhaust contents (water, in this case) mixed with the fuel and oxidizer prevents detonation? Then you'd have to exhaust pre-burners separately, but now it's not a pre-burner but a small-burner for running the pump.

Maybe that would be a simplification: you wouldn't need high pressure pipes conducting pre-burner outputs into the main combustion chamber,. Then the only high-pressure components would be the small combustion chamber for the turbopump, the turbine, and the main combustion chamber. And there'd be no fuel rich/lean/full-cycle pre-burners, just a single separate and small burner burning just enough fuel and oxidizer to power the pump into the main chamber.

Another thing is that hydrogen is a terrible fuel because it's so damned light. It damages pipes and pumps and everything by making materials brittle as it invariably seeps through. That makes hydrogen/lox engines less reusable than methane/lox engines. Ideally the detonation engine could work with methane and lox.

This could be used as a combustor in a turbine engine, increasing efficiency by a few percent... which could save huge amounts of fuel.

They specified the engine in terms of lbf. Why don't they use metric? I realize that imperial units are used colloquially in the US but surely that doesn't extend to actual engineering!

> We can even inspect the value of $PATH, here:

> (...)

> Your output is going to be different from mine because of the hashes in the store paths.

Does this mean that a nix install is not reproducible? If the same installation is performed twice, the file names and contents will be different due to some randomness?

>The team has published their results in the journal "Combustion and Flame".

Sounds like a rip-roaring read!