Fail-Safe Design and Ultimate Load Factor in Military Aircrafts

[jackmckay]

Hi everybody,

this topic is intended to clear out some misinterpretation regarding structural strength of simulated airframes here in DCS.

During my evaluation and testing of simulated modules I have encountered certain damage model behavior that I found not to be inline with industry specifications and standards, historically and current, respectively in cases of ultimate load excess leading to catastrophic failure on wing structure, partially or in full.

As an examples, I point to the occurrences of catastrophic wing disintegration in case of F14, F5, JAS37, MiG21, Su27, J11, Su33 etc. in cases of exceeding nominal limitations given in pilot manuals that occur instantly without prior warnings.

My professional background covers field of structural evaluations of mechanical structures in high load conditions with application of crucial factor of safety that is applied on high stress elements of design with respect to fatigue impact on strength degradation over the life span of any high risk mechanical construction that aims to preserve design itself as lives of operators involved.

Being involved in such evaluations, I find factor of safety the most important element of design that ensures design itself to be safe even after exceeding defined limitations in user manual in certain conditions. Thus, I will give some brief insights on allowed ranges and consequences of exceeding limitations not stated in manuals.

At first, the observed behavior of named modules do not represent real life performance. The initial state of fresh spawned module is projected to be factory rolled state, meaning that there is a minimal work hours accumulated on airframe, theoretically. That implies that the state of structure is at peak strength without any developed structural imperfections and potential seeds of catastrophic failures. The most important structural degradation comes from fatigue induced cracks that are initiated and developed during life span of airframe within work envelope with higher expectancy towards end of life.

In practice, that means that initial strength of airframe is, based on fatique evaluation, multiplied in such manner that, within nominal operational limits, able to withstand forces and moments induced on critical elements that keep the design safe on point of reaching end of design lifespan. Initial design safety factor is, founded by FAR, in range of 1.5 FOS(Factor of Safety) for civil aviation projects. Keep in mind that civil aviation designs are not expected to withstand battle damage of any kind but only environmental threats that include rough weather, turbulence and stress cycles induced mostly by landing shocks.

Thus, military aviation projects are designed in such way that is mandatory to expect higher values of structural loads and performance limitation impact. As stated by FAR, as historical reference, in practice, that would imply that 9G civilian aircraft is safe to withstand 13.5G ultimate load stress without entering plastic, permanent, deformation of geometry. The ultimate load criteria is to be satisfied in time frame of not more than 3 seconds without plastic deformation assuring the geometry of airframe to return to initial state.

That observation is based by FAR, simplified for initial reference. The MIL standards are product manufacturer's specific and are not standardized globally. Those designs do imply structural strength requirements that that exceed civil aviation in many areas simply because of nature of design and operative environment.

So, this post is mostly intended to Module Developers, booth to ED and subcontractors, to revise the state of catastrophic failure occurrence in their modules as those practices do not reflect the nature of design they try to accurately replicate in DCS. If referring to pilot manuals, those limitations are intended to preserve airframe in full life cycle span, which means that initial state of fresh airframe is, based by fatigue evaluation, at least in 3-5 FOS range. Taking the lowest state into account, fresh 9G airframe should be able to withstand 27Gs without inducing structural catastrophic failure as wing separation is by nature.

Nevertheless, this number might appear overrated but the industry does prove this to be the case as this is basic fatigue requirement and those same fatigue evaluations are expedited in at least the double lifespan working cycles. During that evaluation every weak spot is evaluated and reinforced to compensate possibility of occurrences of cracks that might lead to most undesired outcome. Keep in mind that military airframes, intended to be operational at high mach numbers, do have additional scopes of concern as thermal impact on weakening of structure, density fluctuation in airstream, as much as raw size of airframe aggregates bending moments on wingbox that grow exponentially in respect to length of wing against wing root chord length.

My conclusive standpoint is that all involved should redefine the limitations on airframe structure and allow higher margin of safety that should be applied on simulated modules in addition to safety limitations extracted from pilot manuals. The most ideal presentation of structural failure should be presented in a way that, in cases of exceeding nominal limitations, pilot should be warned of incoming catastrophic failure in progressive manner starting from indicative sound (bumps and twists), rivet/bolt loosing and popup(damage module texture), control surface/flap separation (su25T example), permanent plastic deformation (extracted from wing flex feature) with aerodynamic performance degradation and ultimately wing/control surface separation leading to catastrophic event as loss of wing or portion of wing structure is.

By rule of thumb the minimal FOS should be around 2 meaning that fresh 9G certified airframe should have to be subjected to at least 18G of force for catastrophic failure to occur and progressive cycle of damage should occur at FOS of 1.5 starting form 13.5G on referenced airframe above 3 second load condition.

Thank you for your time. I do expect the corrections of this behavior to be accepted and implemented on future updates based on tools and methods available, but for a start, limitations should be set at FOS of 1.5 without entering permanent deformation on lift surfaces.

Regards, J.




 

[draconus]

51 minutes ago, jackmckay said:

As an examples, I point to the occurrences of catastrophic wing disintegration in case of F14, F5, JAS37, MiG21, Su27, J11, Su33 etc. in cases of exceeding nominal limitations given in pilot manuals that occur instantly without prior warnings.

These break wings at about 1.5x of allowable G limits afaik. As expected from you no hard data has been provided to change the limits to 2.0x.

Most of DCS aircraft are not considered factory fresh. Look at the cockpits for example. They are the average aircraft of the type and version usually.

You should know that there is no advanced wing damage simulated so the bending, cracks and losing bolts belong to the wishlist anyway.

Edited March 22, 2023 by draconus

[jackmckay]

 

1 hour ago, draconus said:

These break wings at about 1.5x of allowable G limits afaik. As expected from you no hard data has been provided to change the limits to 2.0x.

Most of DCS aircraft are not considered factory fresh. Look at the cockpits for example. They are the average aircraft of the type and version usually.

You should know that there is no advanced wing damage simulated so the bending, cracks and losing bolts belong to the wishlist anyway.

 

1.5 FOS for 3 secs WITHOUT FAILURE. Quote: AGARD REPORT 815 Loads and Requirements for Military Aircraft, Page 14-2. So, FOS of 1.5 is just beginning of trouble for conventional airframes. Supersonics do even perform better in subsonic regime. So take 1.2 FOS initial limit load criteria and multiply by 1.5 ultimate load criteria it goes to 1.8 WITHOUT deformation. Entering plastic zone of strength of specific material will give extra range for catastrophic failure. So FOS 2 is bare minimum for wing separation.

What kind of hard data you need to be convinced?

 

Note temperature affected degradation of strength. Keep in mind that supersonic skin is hot and still has to comply strength criteria given.

Edited March 22, 2023 by jackmckay

[jackmckay]

2 hours ago, draconus said:

As expected from you

No answer to that. I'll just keep that aside.

15 minutes ago, jackmckay said:

Most of DCS aircraft are not considered factory fresh.

So where would you place those airframes in time period or working hours. some 50% lifespan? And what IF time set in mission is exact year of production rollout of unit? Does that mean FOS will be higher?

Edited March 22, 2023 by jackmckay

[draconus]

28 minutes ago, jackmckay said:

1.5 FOS for 3 secs WITHOUT FAILURE. Quote: AGARD REPORT 815

Isn't this report from 1996 thus not really applicable to mentioned older aircraft? And surely not for Russian/Soviet ones.

I wouldn't place our aircraft as some lifespan % but they are still modeled after performance data of the tested examples (made for manuals) so they may well represent rather new ones, I give you that.

[jackmckay]

Just now, draconus said:

Isn't this report from 1996 thus not really applicable to mentioned older aircraft? And surely not for Russian/Soviet ones.

I wouldn't place our aircraft as some lifespan % but they are still modeled after performance data of the tested examples (made for manuals) so they may well represent rather new ones, I give you that.

It is report from that time BUT the 1.5 FOS is set early in 30s.

FACTOR OF SAFETY - USAF DESIGN PRACTICE Section I

This is from era when fatigue assessments were not executed.

As much as practice I encounter during my work on structural assessments my limit is set to 2/3 of material Yield strength.

..and this is evolution of fatigue assessments starting from mid 50s.
I also stroke, supersonics do have more rigid criteria.
On upper stress-strain graph, the difference between Yield and Ultimate strength is roughly 20-30% in favor of Ultimate.
That is the way industry deals with this issues and that is the way it is and was for a long time.
 

[jackmckay]

23 minutes ago, draconus said:

And surely not for Russian/Soviet ones.

Well, if you intend to design a vessel for colder areas, -60 deg C etc, you need to take into account that materials in cold weather become brittle so to counter this phenomenon you must make your design more elastic to compensate temperature induced brittleness and usually thicker, heavier overall design. that also increases FOS - a lot. So go from -60 on the deck, climb up to ceeling and go mach 3 to heat skin on 1200 deg and come back and settle in -60, then consider the internal stress material has to handle not to develop fatigue induced micro fractures. That is a one tough job and that is why supersonic designs for colder areas require even bigger FOS. that is why, based by my knowledge, soviet design uses UL FOS of 2.5 and in some cases even 4.0 IIRC. Blended wing design is also another feature that further increases FOS. All material strength specification is tested in room temperature at 20-25deg C. That is the starting point. Designers must choose work envelope and the more extreme envelope is, the more radical approach it requires.

[Dragon1-1]

51 minutes ago, jackmckay said:

 that is why, based by my knowledge, soviet design uses UL FOS of 2.5 and in some cases even 4.0 IIRC. 

Note that, as far as Soviets were concerned, their design philosophy was very different. The aircraft had to be economical to produce and cheap to replace, especially "monkey" export models (which were also intended to backfill loses of good aircraft in case of WWIII draining them). Now, the Soviets did like their hardware to be robust (mostly because of an entrenched assumption that an average Soviet grunt was an idiot, very much a self-fulfilling prophecy), but in many cases they did sacrifice reliability and time between overhauls in order to make the design more economical to produce. It was certainly nowhere near as standardized as it should have been.

[jackmckay]

2 hours ago, Dragon1-1 said:

Note that, as far as Soviets were concerned, their design philosophy was very different. The aircraft had to be economical to produce and cheap to replace, especially "monkey" export models (which were also intended to backfill loses of good aircraft in case of WWIII draining them). Now, the Soviets did like their hardware to be robust (mostly because of an entrenched assumption that an average Soviet grunt was an idiot, very much a self-fulfilling prophecy), but in many cases they did sacrifice reliability and time between overhauls in order to make the design more economical to produce. It was certainly nowhere near as standardized as it should have been.

Well, I can not confirm nor deny that soviet FOS state. That is what I heard from mouth of my material science professor on college who had some insights on that side. We can leave that on speculation level. But there is one historical reference that could be taken into account - the IL2. That plane was well known for rigidness and battle endurance as a proof of structural strength concept and was produced in large numbers IIRC. Also when Yak3 La5 Mig3 are compared to P39 Aircobra those had obvious differences, and based on that time comparison, things on FOS side could be at least equal to western approach, regardless of combat performance and efficiency, cos booth sides exchanged technologies in that time.

How things later developed stays on speculation level but one thing is sure, FOS are taken very seriously into account. Wing flex is derived elastic safety principle and it is very hard or nearly impossible to snap the wings just by overriding G limitations by 50%. Suffering some wing/fuselage deformation prior disintegration occurs above that number. From my knowledge that critical factor should be at least and above FOS 2 and above 3 second interval. At current state in DCS this occurs bellow FOS 1.5 and instantly. This correction should be universal at least as common ground for all modules regardless of origin. Changing that limitation in DCS should be trivial and if more consultation with other structural engineers is needed, it should be also easy to confirm that design rule worldwide as basic principle in aviation.

[draconus]

Still, RL accidents of over-Gs tend to be around 1.5x the limit and I don't see anything conclusive in your data provided so far. This was discussed many times before especially on F-15C forum since the aircraft finally got some over-G damage model of the wings a few years back. Here's one of the threads, even before implementation:

Keep in mind the onset rate and assymmetry is also considered in over-G modeling.

Again, as many things in DCS it's also simplified, so you either break it or not. We all would like to see more realistic model but it'll probably come in some distant future - just don't expect pulling 18Gs and stay in one piece though, ever.

Edited March 23, 2023 by draconus

[Dragon1-1]

8 hours ago, jackmckay said:

But there is one historical reference that could be taken into account - the IL2. That plane was well known for rigidness and battle endurance as a proof of structural strength concept and was produced in large numbers IIRC. Also when Yak3 La5 Mig3 are compared to P39 Aircobra those had obvious differences, and based on that time comparison, things on FOS side could be at least equal to western approach, regardless of combat performance and efficiency, cos booth sides exchanged technologies in that time.

Do note that the way Soviets did FOS compared to Western powers was peculiar, in that, at least in engines, it resulted in inferior performance. This again comes back to "grunts are idiots" thing, Soviet designers didn't trust the pilots to take good care of their engines, and thus physically locked out the "combat power" that was available for a limited time to comparable Western engines. A few Soviet designs had "boost" which could be engaged when needed, but a Yak-1 was usually flown at full power for the whole mission. Of course, when those pilots got their hands on Lend-Lease designs like the P39 or the Spitfire, they tended to fly them the same way. While the P39's Allison was itself overengineered compared to the limits in the manual (which is why it got so popular with the Soviets, who didn't care much for the manual), the Spitfire quickly got a reputation for excessive engine wear.

Also, Il-2 was exceptional (mostly by being heavily armored, which gave it a lot of "free" structural strength), and all of those aircraft were built at the time when aircraft design involved a lot less strict numbers and more guesswork than it does now. Soviets did lag behind the West somewhat when it came to electronics, but they had their own computers, and aircraft design practices evolved a lot since WWII's slide rules and mechanical calculators. Jets like MiG-29 and Su-27 wouldn't have been possible using the old WWII methods. Manufacturing quality was also variable across Soviet designs, and what was specified wasn't always what was built.

[jackmckay]

10 hours ago, draconus said:

Still, RL accidents of over-Gs tend to be around 1.5x the limit and I don't see anything conclusive in your data provided so far. This was discussed many times before especially on F-15C forum since the aircraft finally got some over-G damage model of the wings a few years back. Here's one of the threads, even before implementation:

Keep in mind the onset rate and assymmetry is also considered in over-G modeling.

Again, as many things in DCS it's also simplified, so you either break it or not. We all would like to see more realistic model but it'll probably come in some distant future - just don't expect pulling 18Gs and stay in one piece though, ever.

 

Even if it is 1.5 as bare minimum it still does not come after 3 sec in constant linear peak force applied. And that should be counted on end of airframe lifetime, under the condition that regular maintenance and inspection was applied. New fresh have 2, at least, just for the fatigue compensation reserve calculated prior serial production. also no one would expect any pilot to stay out of LOC on 18G for 3 sec. if you were given a task to profile structural reinforcements in airframe, and if you stick to regulatory body guidelines, you wont be given certification of airworthiness unless you comply to that rules.

I said before, the max stress allowed in reports is 2/3rd of material Yield, 50% reserve to max Yield and 30% on top till ultimate strength point. further then you get excess strain elongation reaching breaking point that depends on applied stress time factor. simple as that. the airframe material has to be highly elastic not brittle and that's why there is an excess wing flex and plastic deformation observed prior catastrophic breaking point. even the choice for materials on critical points is to be selected in high elastic, like steel or titanium if heat is applied.

What else has to be conclusive? I'm revealing you the practice of profession and common material behavior. If you want to confront metallurgy principles and regulative bodies practices I have nothing against that. Except no regulatory body will accept lowest criteria then said especially in aviation since the price of product is VIP factor besides safety concern. I don't understand your point. If you think its Ok as now, that is your standpoint, I respect that, but as a chief of production I wouldn't ever sign that paper with your criteria.

[jackmckay]

8 hours ago, Dragon1-1 said:

Do note that the way Soviets did FOS compared to Western powers was peculiar, in that, at least in engines, it resulted in inferior performance. This again comes back to "grunts are idiots" thing, Soviet designers didn't trust the pilots to take good care of their engines, and thus physically locked out the "combat power" that was available for a limited time to comparable Western engines. A few Soviet designs had "boost" which could be engaged when needed, but a Yak-1 was usually flown at full power for the whole mission. Of course, when those pilots got their hands on Lend-Lease designs like the P39 or the Spitfire, they tended to fly them the same way. While the P39's Allison was itself overengineered compared to the limits in the manual (which is why it got so popular with the Soviets, who didn't care much for the manual), the Spitfire quickly got a reputation for excessive engine wear.

Also, Il-2 was exceptional (mostly by being heavily armored, which gave it a lot of "free" structural strength), and all of those aircraft were built at the time when aircraft design involved a lot less strict numbers and more guesswork than it does now. Soviets did lag behind the West somewhat when it came to electronics, but they had their own computers, and aircraft design practices evolved a lot since WWII's slide rules and mechanical calculators. Jets like MiG-29 and Su-27 wouldn't have been possible using the old WWII methods. Manufacturing quality was also variable across Soviet designs, and what was specified wasn't always what was built.

RR Viper engine. Foam boots? remember that story. that material war started very early after WW2. Space race was ahead. soviets were first in space. there must be some high end material/FOS story behind. it didn't happen by accident. the legacy of that time had huge impact on aviation and defense industry. booth sides had reliable space tech form 60 onward besides third parties.

[Dragon1-1]

Soviets pretty much got the "first in space" trophy because Korolev was told to make a nuclear ICBM before the nuke team knew how heavy the warhead was going to end up. The only way to accomplish his goal was to overbuild the rocket massively, so he did. The R-7 ended up so enduring because it was ridiculously overengineered for its originally intended purpose, not because there was some great materials science behind it. Indeed, I'd say that the lack of any fancy techniques was what made it last. The Americans did a lot of things to keep their rockets properly sized for the job, and the reliability of their early rockets wasn't great. The R-7, meanwhile was big, crude, and it just worked. That's the reason why the design, complete with things like igniters on wooden planks across the engine nozzles, simple conical nozzles and the weird booster arrangement, still flies today, while unusual solutions from early days of the US space program had all been retired. On the other hand, those very same factors made the R-7 rather sub-par as a weapon, it was fairly quickly relegated to a space launcher.

Either way, it's quite different from combat aircraft since missiles like the R-7 weren't meant to be operated by frontline combat troops, but by trained technicians at silo facilities. To some degree, the very early MiGs like the -15 and -17 were of a similar design philosophy to the R-7, but later on Soviet aviation started building increasingly elaborate aircraft, no doubt paying more attention to performance than to making them quite as robust as before.

[draconus]

27 minutes ago, jackmckay said:

If you think its Ok as now, that is your standpoint...

I said I welcome more realistic model but you should have started the thread in the wishlist for that.

If you want to just change the current over-G damage limits you have to provide the proof/docs for the aircraft in question. Metallurgy, aviation safety lectures or general reports won't cut it.

[jackmckay]

1 hour ago, draconus said:

I said I welcome more realistic model but you should have started the thread in the wishlist for that.

A good point. Maybe it would be better to go straight to DCS Pope, Mr Grey. This here is on general awareness level. I hope they read it.

1 hour ago, draconus said:

If you want to just change the current over-G damage limits you have to provide the proof/docs for the aircraft in question.

Well, that is close to impossible. The reasons for that is simple, heads of design or production are not available at this time. those people belong to the history. the only relevant sources are their products and manuals. the relevant data can be reverse engineered or .. by accident stats by type in frame age. fail-safe design had to be implemented. It would be ultimately disrespectful not to implement basic military aviation principles.

1 hour ago, draconus said:

Metallurgy, aviation safety lectures or general reports won't cut it.

I wonder how DCS decision makers had all data and sources available per module. Im not sure if that is the case. My engineering standpoint claims that modules are too brittle. In DCS it is common feature gradually implemented. Its great feature but needs to be tuned up by extra 20-30%, time dependent and combined with control surface gradual separation as already seen on certain modules. Little animation triggered by numbers, re-tweaked, following some fail safe initial conditions implemented in fresh rolled jet. those things are nasty beasts in RL. that is too much engineering not to be safe to return after first weak point collapses.


Initial conditions are way better.

[FlankerKiller]

On 3/22/2023 at 12:35 PM, jackmckay said:

Thank you for your time. I do expect the corrections of this behavior to be accepted and implemented on future updates based on tools and methods available, but for a start, limitations should be set at FOS of 1.5 without entering permanent deformation on lift surfaces.

I don't know who you are. Frankly I don't care. But unless you are personally Signing Eagle Dynamics paychecks I don't see where you get off thinking you have the authority to make such a demand. 

[FlankerKiller]

I'm sorry but this is a DCSism that I feel needs to stay. There are other factors IRL that keep pilots from flying their aircraft to such extreme levels of acceleration. Most notably that G Forces are not an option that can be turned off. These are not present in the sim. So having an artificial wing break at a certain point seems like a decent Concession to gameplay. But basically it forces you not to fly the aircraft in a completely unrealistic manner that is allowed by the simulation. Obviously eagle dynamics is capable of modeling more progressive failures, bending of the wings, flight controls coming off, because it is moduled that way on certain modules. But ita a layer of additional complexity for vary vary little gain and maybe even some loss. Unless you play with G force turned off I really don't see how this affects you at all. I for one think it's a better play experience if you have to think before yanking that stick at high speeds.