Russia's new heavy-lift rocket has critical defect

[Hamartia Antidote]

https://www.japantimes.co.jp/news/2...avy-lift-rocket-critical-defect/#.XELDe89KiRs

Scientists have discovered a defect in the engines of Russia’s new flagship heavy-lift space rocket that could destroy it in flight, an apparent setback to a project President Vladimir Putin has said is vital for national security.

The Angara A5, which was test-launched in 2014, is being developed to replace the Proton M as Russia’s heavy-lift rocket, capable of carrying payloads bigger than 20 tons into orbit. A launch pad for the new rocket is due to open in 2021.

In July, Putin said the Angara A5 had “huge significance” for the country’s defence and called on space agency Roscosmos to work more actively on it and to meet all its deadlines.

The issue with the Angara A5 was brought to attention by scientists at rocket engine manufacturer Energomash in a paper ahead of a space conference later this month.

The paper, reported by RIA news agency on Friday and published online, said the engines of the Angara A5 could produce low frequency oscillations that could ultimately destroy the rocket.

A special valve had been fitted to mitigate the issue, but in some cases the oscillations continued, it said. Energomash did not immediately reply to a request for comment.

Russia’s space program has been dogged by mishaps in recent years, including failed cargo delivery missions into space and the aborted launch in October of the manned Soyuz mission to the International Space Station. Russia’s current heavy-lift rocket, the Proton M, has had a nearly 10 percent failure rate in more than 100 launches since it entered service in 2001, creating pressure to reorganize and improve the space program.

 

[Dante80]

This is a weird one. Oscillations coming from the combustion chamber is not a new thing, and is considered by far the most prevalent problem to solve in a rocket engine. The bigger the engine, the bigger the problem sometimes.

For historic reference, oscillations and thrust anomalies is what caused Rocketdyne to introduce baffles on the gigantic F-1. Also, oscillation problems is what pushed NPO Energomash to divide the fuel into 4 different combustion chambers in the huge RD-170. Hell, if you follow the history, you can go back to Nikolai Kuznetsov being frustrated with the problem and cursing Gluskho for not giving away the solution due to his enmity with Korolev. This ultimately led to the flawed design of the original NK-33 that - among other problems - destroyed the Soviet chances of winning the race to the Moon.

Now, for the weird part. Angara fields the RD-191, which is a modern single chamber ORSC derivative essentially of the RD-170. It is an extremely efficient Kerolox design, with an impressive TWR and advanced throttling and sensor monitoring features.

At the same time, RD-191 itself has spawned more derivatives, with two of them in particular being very important. The RD-181 intended to drive the American Antares redesigned rocket (with two of them in each LV) and the RD-193 which is supposed to replace the old remaining NK-33s on the Soyuz-2.1v.

Finding an oscillation problem at this point is very weird, since the engine has been thoroughly tested and studied. The amount of outstanding data for its operation is immense. Moreover, the other major RD-170 derivatives (the RD-180 used by the US in Atlas V and the RD-171 that used to drive the Zenit LV) have also been thoroughly tested and studied.

This leaves us with the hypothesis that something particular has been changed on the RD-191 itself, that helps produce said oscillations. I don't find weird the fact that introducing a FADEC valve in the system to monitor flow and adjust the mix entering the chamber does not completely solve the problem.

The weird part is that the problem itself exists in the first place..

 

[Nilgiri]

This is a weird one. Oscillations coming from the combustion chamber is not a new thing, and is considered by far the most prevalent problem to solve in a rocket engine. The bigger the engine, the bigger the problem sometimes.

For historic reference, oscillations and thrust anomalies is what caused Rocketdyne to introduce baffles on the gigantic F-1. Also, oscillation problems is what pushed NPO Energomash to divide the fuel into 4 different combustion chambers in the huge RD-170. Hell, if you follow the history, you can go back to Nikolai Kuznetsov being frustrated with the problem and cursing Gluskho for not giving away the solution due to his enmity with Korolev. This ultimately led to the flawed design of the original NK-33 that - among other problems - destroyed the Soviet chances of winning the race to the Moon.

Now, for the weird part. Angara fields the RD-191, which is a modern single chamber ORSC derivative essentially of the RD-170. It is an extremely efficient Kerolox design, with an impressive TWR and advanced throttling and sensor monitoring features.

At the same time, RD-191 itself has spawned more derivatives, with two of them in particular being very important. The RD-181 intended to drive the American Antares redesigned rocket (with two of them in each LV) and the RD-193 which is supposed to replace the old remaining NK-33s on the Soyuz-2.1v.

Finding an oscillation problem at this point is very weird, since the engine has been thoroughly tested and studied. The amount of outstanding data for its operation is immense. Moreover, the other major RD-170 derivatives (the RD-180 used by the US in Atlas V and the RD-171 that used to drive the Zenit LV) have also been thoroughly tested and studied.

This leaves us with the hypothesis that something particular has been changed on the RD-191 itself, that helps produce said oscillations. I don't find weird the fact that introducing a FADEC valve in the system to monitor flow and adjust the mix entering the chamber does not completely solve the problem.

The weird part is that the problem itself exists in the first place..

Its probably some kind of resonance issue caused by the particular (new) bulk stresses being imposed (during relevant flight stages) on what previously were good enough tolerances for the previous mass configurations....given the early simpler configurations worked just fine (so something most likely has to have changed around it, i.e new boundary conditions of the FBD of sensitive interfaces around the combustion chamber etc.. rather than intrinsically within it, if we rule out manufacturing flaws etc):

This can cause new eigenvalues (with enough of shift from previous) for both new vibration and buckling modes (these are both linked themselves too)...which can lead to failure modes. You are correct to assume quick fix valves wouldn't really fix this kind of stuff if its serious enough.

Its a little more common than one would think when scaling things up (especially for thin walled, impulse loaded systems)....in fact I've run into this issue quite a few times in my own field regarding such things as engine screech/whine/shudder envelopes when we play around with how we move things around/change innards of a jet engine iteratively to try make it better in some way....but say a little too impulsively.

Usually the (final envelope) iterations are simulated/tested during the early stages to better account for them (i.e assumptions/end conditions wont hold for critical when you get there...so design for them in mind first as much you can)...rather than a more additive+assumptions hold process that maybe the Russians went with in interests of time.

@Signalian @Joe Shearer @krash @Vibrio @VCheng @MastanKhan @Vergennes

 

[Dante80]

I appreciate the insight, and you are probably right. The configuration itself should be the issue, the fact that this problem is explicitly called for the A5 variant corroborates that.

The more I delve into the story, the less I think this is simply a structural element or a simple POGO.

As much as I understand, it can be classified as a kind of POGO, although it's not in tanks or pipelines. In Russian nomenclature, parasitic oscillations are divided into two categories: "low frequency" (LF) and "high frequency" (HF). The LF has waves moving alongside the flow, and they are of course lower frequency. The HF has waves moving across the flows and is characterized by higher frequency. In order to combat LF, some sort of fluid dynamics solution is used: dampers, blow-off valves, that sort of thing. To deal with HF, they use curtains, fences, and other disruptors.

In case of Angara, the root of the problem may be the extremely low power setting for RD-191: 30% thrust. The rocket needs it because the lower modules have the same fuel capacity. They could give up on that throttling a little bit and that would resolve the HF cleanly, but that in turn drops the payload because of reduced staging effect. And A5 is already under-performing - sort of, so they can't.

 

[VCheng]

Its probably some kind of resonance issue caused by the particular (new) bulk stresses being imposed (during relevant flight stages) on what previously were good enough tolerances for the previous mass configurations....given the early simpler configurations worked just fine (so something most likely has to have changed around it, i.e new boundary conditions of the FBD of sensitive interfaces around the combustion chamber etc.. rather than intrinsically within it, if we rule out manufacturing flaws etc):

This can cause new eigenvalues (with enough of shift from previous) for both new vibration and buckling modes (these are both linked themselves too)...which can lead to failure modes. You are correct to assume quick fix valves wouldn't really fix this kind of stuff if its serious enough.

Its a little more common than one would think when scaling things up (especially for thin walled, impulse loaded systems)....in fact I've run into this issue quite a few times in my own field regarding such things as engine screech/whine/shudder envelopes when we play around with how we move things around/change innards of a jet engine iteratively to try make it better in some way....but say a little too impulsively.

Usually the (final envelope) iterations are simulated/tested during the early stages to better account for them (i.e assumptions/end conditions wont hold for critical when you get there...so design for them in mind first as much you can)...rather than a more additive+assumptions hold process that maybe the Russians went with in interests of time.

@Signalian @Joe Shearer @krash @Vibrio @VCheng @MastanKhan @Vergennes

When one keeps in mind the magnitude, density and rate of energy released in a rocket engine, it is amazing the darn things work at all. All that energy has to be directed and controlled precisely, and all the vibrations including harmonics have to be considered. This is not my area of expertise, but I understand enough to appreciate the talent of those who make such magnificent machines work.

 

[Nilgiri]

I appreciate the insight, and you are probably right. The configuration itself should be the issue, the fact that this problem is explicitly called for the A5 variant corroborates that.

The more I delve into the story, the less I think this is simply a structural element or a simple POGO.

As much as I understand, it can be classified as a kind of POGO, although it's not in tanks or pipelines. In Russian nomenclature, parasitic oscillations are divided into two categories: "low frequency" (LF) and "high frequency" (HF). The LF has waves moving alongside the flow, and they are of course lower frequency. The HF has waves moving across the flows and is characterized by higher frequency. In order to combat LF, some sort of fluid dynamics solution is used: dampers, blow-off valves, that sort of thing. To deal with HF, they use curtains, fences, and other disruptors.

In case of Angara, the root of the problem may be the extremely low power setting for RD-191: 30% thrust. The rocket needs it because the lower modules have the same fuel capacity. They could give up on that throttling a little bit and that would resolve the HF cleanly, but that in turn drops the payload because of reduced staging effect. And A5 is already under-performing - sort of, so they can't.

That would likely explain it...had no idea they throttled it that low....that is pretty heavy handed early intervention like I was sort of alluding to. You seem well versed in rocket tech, are you a huge enthusiast or you are connected to it professionally too?

When one keeps in mind the magnitude, density and rate of energy released in a rocket engine, it is amazing the darn things work at all. All that energy has to be directed and controlled precisely, and all the vibrations including harmonics have to be considered. This is not my area of expertise, but I understand enough to appreciate the talent of those who make such magnificent machines work.

Indeed...us jet/conventional guys get to cheat with free oxidiser all around us that we merely harness with a little fuel....these rocket fellas have to contain and lug their oxidiser with them. Pros and cons of that.

 

[VCheng]

Indeed...us jet/conventional guys get to cheat with free oxidiser all around us that we merely harness with a little fuel....these rocket fellas have to contain and lug their oxidiser with them. Pros and cons of that.

It is not just the oxidizer, is it. The environment and its rapid changes, the phenomenal rate of power release with the required precision, the reliability required, the metallurgy ....... the list goes on, in my layperson opinion.

 

[Nilgiri]

It is not just the oxidizer, is it. The environment and its rapid changes, the phenomenal rate of power release with the required precision, the reliability required, the metallurgy ....... the list goes on, in my layperson opinion.

Yes but all this ultimately stems from the "local" oxidiser method hehe. Taking all that good concentrated stuff and letting it go to town, than having to rely on its ambient presence

 

[Dante80]

That would likely explain it...had no idea they throttled it that low....that is pretty heavy handed early intervention like I was sort of alluding to. You seem well versed in rocket tech, are you a huge enthusiast or you are connected to it professionally too?

Nothing more than an educated fool amateur with an unhealthy obsession towards - among other things - anything having to do with humankind reaching the stars.

To elaborate a little more, have in mind though that this is based solely on speculation. Angara A5 (in both its variants) utilizes 3 URM-1 stages in a LFB+core sustainer configuration. These boosters are expected to have uprated thrust (10% more for the first 40 seconds of flight). Thus, the first alarm bell ringing is the fact that RD-191 has been modified/pushed to provide more thrust in this variant.

The need for deep throttling is pretty much self-explanatory. For starters, RD-191 is not designed for multiple ignitions. It uses the standard TEA-TEB starter for igniting the engine, but doesn't utilize multiple slugs for it (like the SpaceX Merlin for example), it is a straight one and done affair. This means, that all 3 RD-191s in the variant have to ignite on lift-off. The three modules are identical in fuel capacity, and almost identical in structure too. This is done for a number of reasons, the most notable being ease/cost of production and consistency in handling and operations.

Now, the design of the LV works around maximizing the stage effect. This means that the two LFBs have to empty first, and detach from the rest of the rocket. Then, the (third) core module remaining acts as a sustainer, until the second stage detaches with the payload on board. Since the engine cannot ignite mid-air, and all three modules have identical fuel capacities, the center engine has to throttle a lot for the first part of the flight (right up until max-Q I presume), so that it has enough fuel remaining for sustaining the second stage further up. The more fuel you have remaining, and the higher+faster you stage, the more payload you can lift.

So, we have two identifiers for a possible problem. The first is that the RD-191s used in this variant are modified and uprated by about 10%. The second is that the center engine has to throttle deeply so as to take advantage of the stage effect on performance. Both of these features of the system are unique to this Angara variant, which is the one that is reported to have the oscillation problems.

Now, some more hints. We know that the oscillations are described as LF, and that NPO Energomash used a "special valve" to correct the problem. This points us to some form of early combustion instability OR a flow separation event in the nozzle (during early flight the flow is under-expanded, since the RD-191 de Laval nozzle is designed to operate as efficiently as possible in both ambient pressure and vacuum operation).

Both of these can produce LF oscillations that are catastrophic in scenarios where you have a configuration of multiple modules attached like this. Remember, this setup works more like 3 rockets flying in close formation, not a solid stage running. Any serious oscillation or delta in the flow direction can introduce ridiculous stresses on the core stage attach points. It can literally shake the LV apart.

And both of these are extremely more common to occur in engines that are deep throttling.

At least..that is my take on this. If I am correct, the blow-off valve they introduced can only partly mitigate the problem. To actually fix this, they will either need to throttle the main engine less (reduced payload as a consequence) or brute-force a solution on the engine itself (flow detachment dampers on the nozzle, a faster valve for the pre-burner mix entering the main combustion chamber etc).

This is fixable, but expensive.

 

[Nilgiri]

Both of these can produce LF oscillations that are catastrophic in scenarios where you have a configuration of multiple modules attached like this. Remember, this setup works more like 3 rockets flying in close formation, not a solid stage running. Any serious oscillation or delta in the flow direction can introduce ridiculous stresses on the core stage attach points. It can literally shake the LV apart.

Indeed, this is part of what I suspected to be the problem. Very nice post you have given here (I have saved it externally to have a discussion with a rocket guru later when I next find the opportunity to pick his brains hopefully...I will try get back to you on it...could be lot later tho),

Wish I could + rate your post, a nice treasure of an analysis!