The perfect aerodynamics - interview with A.Dolotovsky
рейтинг: +6+–x

This is a translation of the original article in Russian

Note: This article requires additional updates and proof-reading. Places where the current translation is most questionable or even not complete are marked in red. If you know the specific terminology and can suggest a better option, please do so either directly in the article or in comments below. Everyone is welcome to make the translation better.

Замечание: Эта статья требует дополнительной правки и проверки. Места, где нет уверенности в правильности текущего перевода, или где перевод отсутствует,выделены красным. Eсли вы владеете соответствующей тематикой и можете предложить более праивильный вариант перевода, не стесняйтесь: правьте прямо в тексте, либо оставьте комментарий под статьей. Любые правки, делающие перевод более адекватным, приветствуются.

superjet.wikidot.com editors have got a unique opportunity to interview one of SSJ-100 creators, Deputy Chief Constructor for Aerodynamics Mr Alexander Dolotovsky. Mr Dolotovsky was interviewed by Sadif.

Note: This interview was taken in June 2012

BigThree.png

Mr. Dolotovsky is on the right, and on the left is the famous constructor of the Su-25 attack aircraft, Mr. Ivashechkin (until recent times he used to be the Superjet's Chief Contructor).

History

Tell us about the project and aircraft's characteristics, what was planned and what is really achieved.

Let us make a short historical introduction. This is the first project in the brand new company. Usually, there are no high expectations in this situation. Fellows from Boeing (as you know, we worked with Boeing) have been saying this is a "big challenge", to create a new market player from scratch. These guys were real mature engineers, who have worked on 737, 767 and 777 programs in the past. In addition it was a challenge not just for us, but also for our French partner for engines — SNECMA hasn't had the experience of engine development for civil aircraft in INTEGRATOR's role. In fact, two new companies were created: Sukhoi Civil Aircraft Company (SCAC) and PowerJet, and two new products had emerged: the Plane and the Engine. It's not just the engine, it is the whole powerplant, because it includes the engine nacelle and all internal sub-assemblies.

Boeing was saying: "Guys, if you manage to build a good plane, it will be a big step forward for you. This is because your main task for now is not to build the perfect plane, but to learn how to build planes." We are following EADS's path now when they launched their A300 program, which was far behind their competitors, economically in the first place.

But here we have a big advantage over other new companies. We live and create in the famous aviation country — Russia. Here we have a solid basement formed by our aviation research centres, which have a lot of unique competencies especially around the aerodynamics, aircraft construction and engines. Of course I mean TsAGI and SibNIA. I can speak only for these centres which are related to my competency, i.e. aerodynamics.

That's why we have got an opportunity to introduce a zest to the project, to utilise the experience of our research centres. The zest is: from the aerodynamics's point of view we have created a WIDE-BODY aircraft with characteristics close to a NARROW-BODY.

Aerodynamics

From now on we will be discussing the plane, as far as I understand.

Let me step aside for a moment. I'd read some magazine or a newspaper some time ago, honestly don't remember which one exactly, there was a report from Farnborough. And one phrase caught my eye, something like "I was staying next to Boeing 787's super high aspect ratio wing, and suddenly I have realised how small the Superjet's wing is." In reality, our wing's geometrical characteristics are very close to ones from Boeing. Our wing is also made of modern materials and has the superconfiguration Boeing has the wing aspect ratio a bit more than 11, we have 10. By the way, both numbers are quite record-breaking in its class. Here we are using almost a centennial experience of plane design and production in Russia and Soviet Union.

Let's get back to the subject. Our task was to design a plane with a level of a passenger comfort close to Airbus 320 series. This implies a wide body, because "comfort = space" in the first place. And the lift-to-drag ratio must be at least not lower than, and preferably even higher than Embraer 190 has, which has a smaller body diameter.

IN GENERAL WE HAVE ACHIEVED IT.

As a result, we have got a plane that has the fuselage width-to-length ratio roughly 30% more than Embraer and difference with Bobardier is even bigger, while our lift/drag ratio is higher than Embraer by 0.5 (we have ~16.5 and Embrier has ~16) and just slightly worse than CRJ has. This is because CRJ has a narrow body and a large wing span.

We were able to achieve all that thanks to our partnership with TsAGI. Which factors allowed us to achieve these results?

aerodinamika.jpg

The wing we have developed together with TsAGI is composed of supercritical profiles optimised in a special way. We have utilised the special software package which was designed in TsAGI in 1980s, this system allows to simulate the airflow along the whole aircraft configuration (i.e. 3D, taking into account all airframe elements). TsAGI started to work on this method in mid-70s and constantly improves it. This package allows to optimise the profile by any given number of section-cuts (опорное сечение) defined by a designer. It takes into account multiple requirements, not just aerodynamics but also technology and layout. Usually, when an aerofoil is designed, the first priority is aerodynamics, second geometrical smoothness and the layout follows first two. The method developed by TsAGI allows to take all three factors into account simultaneously. By varying weights it is possible to give a higher priority to one or another requirement. One more benefit of this method is the optimisation of a supercritical aerofoil in terms of delaying the supercritical shock wave with a given lift coefficient, i.e. reduction based on the local airflow speed control on aerofoil's surface.

As a result, we managed to create a rather "thick" wing, its width along the outermost rib is about 15%, with limited sweep and taper ratio, with aspect ratio about 10 (which is a record breaking value for regional jets). This wing is optimised for flights at M=0.78..0.79 with the Kmax realisation close to 90%(при реализации Кмах около 90%).

Any passenger feels when a plane climbs easily and easily speeds up to M=0.8-0.81. The majority of our cruise flights are performed on these speeds.

In addition, the absence of a shock wave on the wing and on the airframe's surface means the reduction of noise and, of course, the reduction of a shock wave drug. Together with TsAGI we spent a lot of time optimising the local aerodynamics: fairings of flap actuators, fairing of wing-to-fuselage adjustment, shape of the nose and tail of the fuselage, wing tips, pylons, etc. We spent about three years on all that. We introduced many changes into the base geometry: for instance, we completely changed the geometry of the fuselage cowl(in 2006) because we realised that we have a zone of a supersonic airflow, and we have found a very elegant solution. We lowered the local airflow speed by increasing the mid-profile in one place. This made the plane closer to a theoretical maximum, widely known as the "area rule principle".

I.e. you have made the airliner closer to a fighter jet?

Closer to the optimal transonic aircraft, to be more precise. We also have other ideas which we will develop further. To do so we will apply CFD (Computation of Flow Dynamic) method. This method allows visualisation of sonic and supersonic airflows and flow pressure along the airframe surface, all this in 3D on computer monitors. The method allows tenfold increase of volume of research, and most importantly, it brings the feeling of intuitive understanding of what to do with the layout. And you start to FEEL the physics of the process. In the past, we were only visualising the whole airframe airflow in a wind-tunnel and only on limited time intervals or on flat (i.e. 2D) models.

We use different software packages for our calculations: both domestic and foreign (e.g. well-known "Fluent"). For instance, using these techniques we were able to optimize the shape of pylons. It was another «Big challenge».

Just from the layout point of view, due to the long nacelle (on our plane it goes pretty deep under the wing) the area where the pylon, engine nozzle and the wing come together, forms something like a de Laval nozzle. In order to avoid supersonic areas at speeds up to M=0.81 in this place, we have optimized the rear section of the pylon in the area of this narrow "bottleneck" and surroundings. This allowed us to reduce the local airflow speed by ~M=0.2 and thereby completely remove supersonic flow from this area. Also at the beginning we had a shock wave at the front section of the fuselage (if you remember, it was a bit similar to A320). To "clean it up" we have smoothed the plane's "face". As a result of all these activities we don't have supersonic airflows anywhere on our airframe up to M=0.81. Supersonic flows appear only at M=0.86-0.88.

When it comes to aerodynamics, it is necessary to take into account that we have optimised the aerodynamic configuration of the aircraft not only for high speeds (large M numbers), but also for high angles of attack, which are reached at the minimum speeds. This was done in order to increase our aircraft's safety.

This is not so simple from the flight dynamics standpoint. When the flow separates from the wing's surface, the controllability and stability of the flight changes in all channels. In a swept wings aircraft, this process is further exacerbated by the airflow along the wing, which provokes a flow separation at wing tips. This process has became a real pain for the first and second generations of swept wing aircrafts. Another process is the loss of a directional stability and the subsequent movement of the aircraft in the yaw channel, which also provokes the wing stall, especially for an aircraft with swept wings and positive cross V. These are just few samples of problems which do not allow to achieve theoretical maximum angles of attacks calculated basing on wind-tunnel tests.

In order to mitigate all these issues, variety of measures are applied during the layout design and often these measures are mutually contradictory. For example, to improve the lift/drag ratio it is necessary to apply a higher load to the end section of the wing, at the same time the separation flow starts with sections with the highest load. Thus, in order to achieve a better lift/drag ratio, we should apply more load to the outer wing sections, while if we are targeting a better stall characteristics, the extra load should be applied on the innermost sections. It is important to provide a reasonable compromise, not losing in the cruise speed and without losing the stall characteristics. Also, we shouldn't forget that modern aircraft are not allowed to lose stability and control in icing conditions, and this is completely different story, especially taking into account our wide centre of gravity range (range of 24% of the MAC is very large, especially for old soviet aircrafts).

To provide the necessary level of wing lifting capabilities during the take-off and landing we have developed and implemented a relatively simple but highly effective high-lift devices composed of the slat and single-slotted Fowler flap. This has allowed us to provide the approach speed within 140 knots (260 km/h), despite a relatively high wing load. In addition, these high-lift devices with minimised number of slots and low deflection angles at take-off, allowed us to have 15 spare EPNdB in terms of airport noise (based on ICAO Chapter 3) or 5 EPNdB for Chapter 4. Only B717 (formerly MD90) and EMB-190 in our class have similar numbers. Take into account that these aircraft thrust at MTOW is 10% higher than the RRJ-95's and they have a lower wing load.

How did we manage to get the glide/take-off path angle not worse than our competitors have? This is due to our high-lifting devices which provide a very high level of lift/drag ratio in take-off and landing configurations. Our maximum lift coefficient (about 2.6 in the landing configuration) is close to what was previously achieved only with double-slotted flaps only (for a swept-wing). Here we benefited from both factors: our wing aspect ratio as well as high-lift devices configuration.

In order to be able to land on a very short (for aircraft approach category C) 1700m runway (just in case: this means the actual landing roll 1700/1.67-450 = 567m),we use not only very effective carbon brakes, but also extensive spoilers. We have 3 pairs of spoiler sections and 2 pairs sections of speed brakes. Thanks to these elements, during the landing roll we are applying a very decent negative lift force. As a result, our braking coefficient on a dry runway is more than 0.42 on average during the landing roll, from touching down to stop, and this is without the use of the reverse thrust! We've got similar numbers for a rejected take-off. Just to give you an idea: the previous-generation aircraft brakes provided the breaking coefficient =~0.3, i.e. more than 30% less.

Certification

And few more things. While doing certification activities we have discovered that even rather old aircrafts, e.g. the whole Boeing 737 family, have a very good stall characteristics. This means that an average pilot is able to counteract all stall tendencies. It was a small revolution for us, because for us the start of uncontrolled motion around one of the axes was always a sign of stalling. Even now students of aviation universities in Russia are learning it in this way.

We set ourselves the target: our plane should comply with the latest standards of the aircraft behaviour at high angles of attack and stall: if the control lever is pulled towards you (this allows to achieve a high angle of attach and low indicated airspeed), the aircraft should loose speed without any tendency to a wing stall or sudden increase of angle of attack (подхват) (this often a pain-point for aircraft with a T-shaped tail), all plane reactions should remain direct.

WE MANAGED TO ACHIEVE THIS TARGET

How?

In general — by using the specific size and position of the tail fins and by optimisation of the geometrical and aerodynamical wing configuration(крутки крыла), including artificial (e.g. using the flaps and slats). In addition, when we were optimising the profiles (we already discussed this before), we tried to avoid or at least minimise zones with the local static instability in the angle of attack channel. We were doing this in the whole range of angle of attacks, up to the maximum.

What have you achieved?

During the tests in a wind-tunnel it was not easy to put our model into a spin-dive, and it is enough just to put controls into the neutral position for recovery. Model does not enter the flat spin even with the fully deflected elevator. This is a very good result — for the first time in its history TsAGI allowed us to perform high angle of attack and stall test flights without an anti spin parachute or rockets.

birjukov.jpg
Vladimir Victorovitch Biryukov – honoured test pilot

Flight tests have confirmed that we indeed got a very good result. Vladimir Biryukov (a honoured test pilot, the certifier of the Gromov Flight Research Institute, who has a very large experience of high angles of attack test flights, not only in domestic, but also in foreign aircraft) after flying on our plane in November 2008 admitted, that he has never seen before such a good stability and controllability at high angles of attack among any other civil aircraft he has flown.

Behaviour that he was talking about is the ability of the plane to achieve the angle of attack ~30 degrees, than slow down to a speed less than 200 km/h (it's very slow for aircraft with such a high wing load), and still remain under the full control in all three channels, even rudder, without any tendency to stall unexpectedly. So stalling for RRJ is entering the parachute mode when the plane almost falls flat (at this moment there is not enough lift already to maintain the level flight, g-force before recovery drops to 0.6), but the plane falls down under control and can be recovered by simple control actions apparent to an ordinary, average pilot.

And this is additional proof in favour of our aerodynamics.

What did it give us apart from the ability to surprise experienced test pilots?

We have a good take-off and landing performance, while our thrust-to-weight ratio and the wing load are close to the statistical limits. And we made this on purpose.

Questions and Answers

Engine

The thrust-to-weight ratio is driven by the engine.

The engine was designed to produce the required thrust. It was possible to build a more powerful engine with this gas generator.

But it was not built.

No, it wasn't. And here is why. The engine maintenance costs and a number of other parameters that are not directly related to performance, but related to money, are directly dependent on thrust. For example, the engine selling price is based on its thrust in kilograms. So we deliberately limited the engine thrust to reduce the maintenance costs1 of the entire aircraft (and increase the useful life of the engine) .

Honestly, I have never heard of this approach before

Then I just revealed a commercial secret to you. Smiles — it's a joke. Once again, if we have a look at our position among our competitors in terms of the wing loading and thrust-to-weight ratio, then we'll see that we are located at the bottom level of statistics. At the same time our take-off performance — a balanced take-off distance, MTOW limits dictated by the climbing rates at airports with the high elevation and hight temperatures — is not worse, than our competitors.

Close to these characteristics of Yak-42?

Yak-42 is much better than us in terms of thrust-weight ratio. Remember, it has one extra engine and a huge wing. This is a famous story: when they transitioned from the straight to the swept wing, they had to extend the wing area to over 140 squared meters, and that was for the same weight as we have. We have a small wing: ~77 squared meters #red|(по трапеции всего 77 метров)##. But thanks to this we have a very high percentage of implementation: cruising lift/drag ration is 90% of maximum. In other words, we are fully using our wing in all modes: on take-off, cruise and landing. We do not carry extra weight with us.

By the way here we have got a lot of help from Siberians (SibNIA), we appreciate their help very much. They have a very good experience in high-lift devices optimisation. Let's see, we don't have a very extensive hight-lift devices, ours are not as advanced as, for example ,ones used on the Tu-204, which has a magnificent high-lift devices geometry both - by chord and when lowered. But even with our relatively small slat, we have a very good landing lift coefficient = 2.5, which is by far one of the best value for a single-slotted Fowler flap on a swept wing.

Overweight

All subsequent produced planes will have the enforced wing, how is it different from the current one?

The wing enforcement is not my area of the responsibility, sorry. But I can confirm that the geometry of the wing and high-lift devices will remain the same. I think it is obvious why we are going to strengthen the airframe. We are receiving a strong criticism from every possible source due to the fact that our plane is heavier that it was initially planned, and this is true. Indeed, the empty weight of the aircraft has increased by 12%, this is almost 3 tons. In order to keep the announced range of 3000 km for the base model we had to add these 3 tons to the maximum take-off weight. During the initial airframe design some structural reserves were put in place by our chief designer Yuri Viktorovich Ivashechkin. These reserves allowed us to maintain the characteristics of the basic model, despite the increase in the weight of the empty aircraft.

By the way, this extra weight has a minor effect on the operational costs: for an average 500nm leg the effect of increased weight by 12% on fuel burn is less than 4%. This is due to the fact that when flying on short routes the volume of fuel consumed during the cruise flight is close to the volume consumed during the take-off, climb, descend and landing. These components do not change when the basic empty weight is increased. If we only consider the cruise flight mode, the increase in the Basic Empty Weight brings additional fuel burn in proportion of ~0.5% of fuel per 1% of the weight increase. Thus the "11% of extra fuel consumption" which you can often find in media, is not true. In fact, according to our estimates, we have the advantage in fuel burn per flight ~2% for EMB-190 and more than 12% for A319, assuming the same load, of course.

Now we need to built the «Long Range» version which is expected by customers. The take-off weight of the LR is about 49…50 tons. To make this possible we need to strengthen the airframe, but I want to stress that these are local enforcements. We take into account the results of static testing performed at TsAGI and fatigue tests in SibNIA. These local enforcements related to the results of fatigue and static tests is a normal process. Here is an example from our own history. After the WWII, the famous Tupolev Tu-16 was launched, which has became not only our main bomber, but also the base model for Tu-104. It's not a secret that the project was kicked-off twice. For the first time the aircraft was designed using the "old way" setting the structural strength on the first place. As a result, the aircraft turned out to be very heavy and did not meet the specification. After that Mr. A. Arkhangelsky had came up with the brilliant idea: instead of trying to lighten the overweighted airframe by "biting off" unnecessary weight here and there, let's try do design the airframe with a smaller structural strength reserves, test is thoroughly and then determine the places which require local reinforcements. They have done so and were able to remove few tons of weight from the plane. Since that time in order to minimise the weight of the airframe, aircraft are designed with a minimal structural strength reserves. I want to stress that we have achieved the optimal weight. That's why we do strengthen local areas. I am confident to say so because the wing on our static tests plane broke at ~99% of the projected load, this is the amazing result of efforts of structural engineers and constructors. While some people say that our aircraft is overweighted, it's not, in fact IT IS LIGHT, it is lighter than any of its competitors even physically (EMB-190 is 700kg heavier), and given the size of our fuselage, the relative weight is even smaller!

You may have a question: Why we were still announcing the basic weight =24.5 tons even when we already knew that it is heavier? Well, look around, this is a "feature" of modern age, real characteristics are never publicly announced until the certification is achieved. This was the case with A320 and B737 NG, and with our direct competitor EMB, and with Dreamliner B787. This is done for marketing reasons, which are far away from normal engineer's way of thinking …

Useful life

By the way, how many cycles have been passed so far on the fatigue tests plane in Novosibirsk?

I don't even know, most likely Igor Lvovich (Vinogradov) knows this. This parameter is under the constant control from the top, because the useful life — is what can be sold.

Mr. Vinogradov has answered this question later: 21000 cycles (by June 2012).

For IL-96 the useful life is 70 000 flight hours, we would like to see at least the same number for this aircraft.

I want to admit that for a regional jet, the figure 70 000 flight hours is much more "expensive" than for IL-96. Because for the long-haul planes the standard cycle duration is from 6 to 12 hours, and our standard cycle is 1-2 hours, i.e. the number of cycles per the same number of flight hours several times (sometimes 10) higher. Effective airline companies are well aware of this feature when they are leasing aircraft. The cost of leasing depends on the cost of an aircraft and a number of remaining cycles. So while the plane is new, it is put on short routes, "knocking out" remaining cycles. Afterwards it is normally sold to those who utilise her on long routes, working out the remaining "calendar" useful life measured in hours.

Choosing the size

Let me step aside again and tell you few words about the selected size of the plane. The question "why did we go to this niche", was discussed quite frequently, so I will answer another question: Why there is a demand for such planes?

In those days, when the price for a barrel of oil exceeded $30, operation of wide-body aircraft with a less than 75% of loading had became not profitable. While later the price of a barrel began to grow higher and higher, (today, as you know, it fluctuates around $100) the need had emerged to create a plane that would be cost-effective and capable to get a 100% load where B-737 and A-320 are loaded below the profitability level (which is around 80% of the total seat count of the aircraft). The obvious way to achieve this is to reduce the seat count. In addition, for a start-up company this market with the least possible competition is much more attractive than, for instance, bigger market for an aircraft with 200 and more seats, where de-facto government giants Boeing and Airbus play main roles.

But there is a technical dilemma in this approach, because by reducing dimensions we automatically degrade the aerodynamics. The local Reynolds number is reduced, directly affecting the lift/drag ratio because the laminar-flow region is reduced. For example, for an aircraft of our class the maximum achievable lift/drag ratio is 16.5…17.0. But if we stretch such a plane, say, twice, that number would be more than 18 instantly, only because of larger dimensions. Therefore, we had to "hit the eye ball" of the wing optimisation, we have discussed this already.

So we have reduced the dimensions of our plane, as a result we obviously got a lower weight and lower maintenance costs — due to the fact that we are able to use cheaper engines. But our lift/drag ratio is still close to that of A-320. As a result, our plane carries 100% load when even A-319 will carry only 70%. As I have already mentioned, the savings on fuel alone is more than 10%! In addition there are saving on the cost of ownership and airport and air nav services. If you sum it all up, you'll get a large percentage.

In the West instead of our parameter grams/per passenger/per km (which is inapplicable in the economy because this parameter depends on the aircraft's range and the range is different for different aircrafts) another diagram is in use: Fuel per flight/fuel per seat. One axis shows the fuel consumption per flight (what you have to burn), and the other — the consumption of this fuel per seat. This is done for a fixed range. Obviously, the closer to the left and lower we are in this diagram, the better the plane. Similarly, it is obvious when we take a regional plane and compare it to a 150…200 seat plane, we have a worse value for fuel/seat consumption, but we always win in terms of fuel per flight! This is a direct indicator for regular airlines: if the number of passengers drops on a route, it is necessary to shift to a smaller plane in order to reduce costs

Bombardier was the first to smell the money in 90s and they entered this market by stretching their business jet, first to 50 seats and then to 70 and 90 seat planes. They were followed by Embraer, which has created their E-Jet series. Embraer had got the idea right: clients expect a specially developed 70..100 seats aircraft, rather than a stretched business jet (like Bombardier) and not a shrank long-haul version (how it was done by Boeing and Airbus). Shortened versions turned out to be unprofitable.

That's why we have also followed into this niche. Do not forget that our project is commercial from the beginning and the extrabudgetary funding is still prevail. It's only possible to compete with Boeing and Airbus if you are supported by the government, like it is done with Irkut's MS-21.

The only thing Embraer hadn't implemented is the size of the fuselage, psychologically they were coming "from below", from their small regional aircraft family ERJ-135 … 145. For the E-Jet they have chosen the 2+2 scheme, and though it is bigger than CRJ, but nevertheless it is still pretty small. And we used this fact. When we were planning the size of our aircraft, we already knew the new Embraer's dimensions. We decided that our plane should exceed the EMB-190 in the passenger's comfort in the first place, so that a transit passenger would feel the same level of comfort as in A320 or B737.

Level of comfort

What can you feel by your own body, when you enter the plane?

In the first place it is the space in the cabin, similar to one in the cabin of long-haul planes:height in the aisle, aisle width, seat width, pitch between seats, height, width and depth of the luggage rack. I want to stress: not the volume but namely the height, width and depth.

As a result: when you and your colleagues entered the cabin some time ago — were you able to feel the difference with A320? No.

Aeroflot passengers also notice this. More often this is noticed by Europeans and other people who, unlike us, often fly CRJ and EMB. You know, for instance, that if you try to get onboard CRJ with the hand luggage that you just carried with you on the racks of A320 or B737, it would be taken away and put into the front luggage compartment. And this is the only way because there is simply no space for it in the cabin.

SSJ-130

You already have 103..105 seats with 29" pitch in the economy. Why do you want to build the 130-seats version?2. May be it's better to go straight to the 150-seat version?

Why do we need the 130 seats plane? The aviation market has several "centres of gravity". Once oil became expensive and the air transportation market broke-up, two "centres" were formed. First one is located at ~50 seats, and it is firmly occupied by turboprops and small jets like CRJ200, CRJ700; the second is located about the 100-120 seats mark. In addition it is important for air companies to always maintain a high occupancy rate by varying aircraft types. The size of the fuselage we have chosen allows us to develop the 130 seat version relatively quickly and easy. This version will allow us to fill the gap in our model range that has opened after we have cancelled 60 and 75 seat versions. The 130-seat version will be not only comfortable and light-weight plane, but thanks to a longer wing, it will also bring us back to the narrow-body type. In the current layout we plan to have a lift/drag ratio >= 18. This will give us a good economy advantage even without new advanced engines! And though this machine will become a competitor for younger members of the family of B-737 and A-320, it is still not a competitor to Boeing and Airbus in their main segment: 180-200 seat wide-bodied planes.

Fly by wire

A separate subject for the discussion is our Fly By Wire (FBW) system which has been built based on almost whole experience and knowledge that our aviation industry have, including even "Buran".

The wing is changing, tail fins are changing, does avionics also change?

Our avionics is the most interesting part of the aircraft. The «Pipe», i.e. the airframe is the cheapest component the plane has. And it must be cheap because the intellectual component is not so large it in. All methods for an airframe design, including aerodynamics, are already developed in the past. Modern airframes are built using existing high-effective methods. But all controllers in SSJ(there are 15-20 controllers altogether) which control other aircraft's systems, were created based on very basic components, or even from scratch . And even this is not the most important thing.

All software that controls the electronics is actually created by us. We have developed the complete logic of control laws. This intellectual property was created while we were developing our plane.

I am especially proud of the FBW control laws, which are written in collaboration with the Directorate of aerodynamics experts from NIO-15 of TSAGI. We will discuss this topic later. This logic, the principles of control and interaction, are really worth a lot of money and a lot of work. For instance, when we are developing a new aircraft and we change the wing ,tail , fuselage but keep the avionics intact, we are greatly saving money and time.

In fact today the aircraft industry has came to the same principle as called "the platform" in the automotive world. In the last thirty years, car makers change their models every few years. In reality it's a "face lifting" of a core group of components which is called "the platform". The platform changes not so often, once every ten years on average. Now there is also a concept of "the platform" in the aviation. The development of a platform requires maximum effort and a lot of time and money. It is very important to build and establish this platform.

Of course Boeing was the first one with its 737 family. They grasped the idea and rolled it off. Later all manufacturers did the same. Note, that the airframe of planes of the same "family" can differ a lot,but there are some "common points", which are the same for all aircraft family.

This is the platform. Normally the Platform is Avionics + Flight Deck. Plus maintenance principles,logistics and much more …

We are going to follow the same way. We have the platform. By the way - quite a good platform, in many parts even exceeding the the level of long-haul aircraft. This even "stroke us back" during the operation - Aeroflot crews are desperately comparing us with A320, completely forgetting that the RRJ is TWICE smaller and twice cheaper! But crews do not see much difference in general and require the same level in all aspects. That's why, we get their criticism in a constructive way, often with gratitude and even , in a sense , as a compliment.

So, we call avionics and flight deck the platform. And we already have it. A new wing, tail fins and mechanics are now can be developed in a relatively short time and with less risk .

Fly By Wire laws

Now, as it was promised, lets talk a bit about the FBW (Fly By Wire) laws and about the system itself. Probably everyone knows the main "slogan" of our FBW. The world's first fully remote system without a mechanical backup, having the side stick as the main control lever. How have we ended up like this? Back in 2004 we were holing an advisory board with multiple airline's representatives. Such a giants like Air France, KLM (not yet united at that time), Lufthansa, Delta and others has participated. At this event we have presented the ideology of the future cockpit which was almost fully unified with B737 - the etalon and the short-haul market bestseller, as we were considering it at that time. Accordingly, the planned FBW system was a primitive replacement of direct mechanical links with dampers on all three channels and a minimal optimisation for different flight modes. Grandees of world aviation didn't find this idea the best, they told us that the new aircraft should be focused in the future, and the most promising is the ideology of the "protected aircraft" implemented on the Airbus A320 family. We can't say that this idea was quite new for us. Out fighter jets and Tu-204 were already using algorithms for automatic restrictions from going beyond the operational area, i.e. we already had the required skills and knowledge. In addition, the side stick has been already tested in 1980s on a test version of Tu-154 #317 in Gromov Flight Institute, there was even idea to put the side stick on Tu-204. Unfortunately the idea was rejected by someone on the top level of Ministry For Civil Aviation. Our supplier LLI (Liebherr Lindenberg), was ready to provide us with the required operation speeds and backup levels of computers and actuators for the FBW in order to implement the full set of algorithms. So we have a baseline, and it was quite solid.

We decided to take this on our plate and started working. Frankly speaking, the work was challenging but fascinating. We took something from TSAGI(it was the core (Базовая интегральная часть)), took few things from A320 and made a lot ourselves. To achieve the result we actively used various modelling stands, ranging from the simplest PC-based station with a gaming joystick and pedals (i still keep them in my closet) , up to the sophisticated mobile stand PSPK-102 from TSAGI, which was built in the USSR for the "Buran" program and is now used for virtually all tasks which require a high-quality full-scale simulation. Note that even Boeing still uses this stand, despite having its own very solid stand base.

What have we got as the result?

Our FBW performs the following tasks:

Reduces the load on a pilot by:

  • Automatic balancing in pitch and roll channels after controls are released;
  • Optimising of control characteristics for different flight modes, taking into account the high-lift device configuration, speed, Mach number and centre-of-gravity position

Provides protection from entering the aircraft upset by

  • Restricting roll and pitch angles, with the return of the aircraft back to normal operating limits for these parameters when controls are released;

Provides protection from entering the stall mode with a critical loss of speed by

  • Limiting the angle of attack, pitch and roll;
  • Restricting the high-lift devices retraction while flying at unacceptably low speeds

Provides protection against structural destruction by

  • Limiting the indicated airspeed and Mach number, taking into account the high-lift devices position;
  • Limiting the maximum G-load taking into account the high-lift devices position;
  • Automatically adjusting the high-lift devices position;
  • Automatically restricting angles of deflection of control surfaces depending on the airspeed and Mach number;
  • Automatically limiting the pitch angle during the take-off

Provides protection against loss of ability to maneuver at low speeds and low altitudes by:

  • Automatically increasing thrust to the takeoff-level in critical modes when the total energy of the aircraft is lost(the total energy is the sum of kinetic and potential energies);
  • Automatically retracting the high-lift devices into position "1" (actually slats) when the speed is below a predetermined threshold.

All this stuff works even if two hydraulic systems are lost or when electric supply is switched to a backup source (RAT). This set of features is able to prevent crashes such as A320 near Sochi, B737 near Perm, Tu-154 near Irkutsk and Donetsk, ATR42 near Tyumen, and many others where a plane entered the upset, stalled, or critically lost the height during a maneuver.

What happened to the basic algorithms which were developed in the beginning, designed for the control column?

They are now in the minimum, backup mode. Thanks to these algorithms and the good stability of the aircraft, in the «DIRECT MODE» RRJ reminds our old good Tu-134. I won't tell you a secret if I say that we easily lifted the plane in the air and flew more than three months of test flights (if we count the whole program) using the Direct mode. By the way, some of our pilots like «DIRECT MODE» even more than the «NORMAL», because in the Direct mode the plane does not restrict a pilot and control characteristics differ little. This is also a great achievement, because , for instance, entering the «DIRECT MODE» during the education on A320 is only performed on a simulator and during the flight, while we can perform the whole flight in the «DIRECT MODE» and mid-level trained pilots can still control the plane.

By its flying characteristics our "Jet" is really "Super"! It is especially important that this part of its internals is fully owned by the SCAC

Now few words about the hardware. Since the FBW system does not include a mechanical backup, the reliability requirements are very high! For this we have to say "Thank You" to AR MAK, i am talking seriously. Take into account that this plane was designed to fit into a predefined price, so we had to make it cheap and sophisticated, that seems to be impossible. The Russian creative approach in conjunction with German thoroughness is it turned out, are able to create a miracle. I won't go into complete debris, just let you know that taking into account the cost/quality of modern processing units, a complete two level network is deployed on-board, with such a large number of computer nodes that it's not longer possible to talk about a per-channel backup in a traditional way.

In order to "knock out" such a system, it is necessary to destroy more than 70% of computers, which is almost impossible given their heterogeneous hardware and software. At the same time, due to the development of a technology in the microelectronics segment,our system is relatively cheap, and despite the fact that the number of processing units in it is more than in A320, the FBW cost is lower, having a higher reliability.

Furthermore, this structure gives us a very large increase of operational flexibility due to the inclusion of non-performing units in the MMEL. For example, a scheduled flight is allowed3 for any failed top-level computer and three failed computers on the lower level. With this we allow a failure of up to four actuators, provided that each rudder and the elevator still have at least one operational actuator. Departure for ferry flight is allowed when just one top-level computer is operational and even larger number of failed computers on the lower level. Frankly speaking, we have never got a chance to use this ability because we have never had FBW failures due to equipment failure. German quality in its best! By the way our FBW as well as other controllers on board has been tested not only for wire breakage, but also for short-circuit, as well as on HIRF and electromagnetic compatibility. The unprecedented for Soviet and Russian aircraft amount of tests has fully confirmed the correctness of solutions included at the system architecture level. The FBW on RRJ doesn't have a single point of failure. We share this intellectual property with LLI, because the architecture development was carried out jointly.

Epilogue

It was probably more than enough for such a short meeting. Trust me, I have much more to tell about our project and our wonderful team, which has attracted specialists from almost our entire aviation industry. We employ people not only from Sukhoi and mostly not from Sukhoi. People are gathered from Ilyushin, Tupolev, Mikoyan (where i came from) and Yakovlev. For example, our small team of aerodynamics combines specialists from TSAGI ,"Molnia",Ilyushin and even Embraer… Also we have wonderful young people who have become a world-class specialists over the years. Now they are ready to work on equal terms with Boeing and Airbus, confidently prove the correctness of their decisions not only to AR MAK but also to EASA, and, in future,to FAA.

Therefore despite the fact that our work is challenging, I am optimistic about the future and i believe that our first project would be the pioneer of a whole line of planes which in turn will revive our aviation industry and bring it to a new level!

Many thanks for your questions , Vladimir. It was a pleasure to talk to.

Sadif: I want to thank you for the attention you paid while answering the questions of aviation enthusiasts. During the preparation of this interview you have spent a lot of personal time to fix these mistakes and mismatches that i made during the preparation of the material.
That's why i would like to express my gratitude to the designers and test engineers Alexander Dolotovsky , Igor Podorvanov, Igor Sobolev for patience while answering my sometimes silly questions, and for the time that they have spent. I hope i will be allowed to take some more of their time in future.
In short: Thank you to all our engineers who do their job despite of all issues and challenges.

The interview was taken by sadif, we do appreciate it very much!

Discussion

Valery Popov wrote: Here is what i can add: When we were passing the certification in EASA their experts paid a great attention to failures affecting the structural strength. They used quite a simple approach - they were comparing our list of failures with one from A-320. Every time they found a mismatch they asked - why A320, for instance, has a failure for spontaneous deflection(самопроизвольного отклонения) of more than one control, but you don't have it. As a reply we demonstrated them the probability of failure ~1E-20 or 1E-23! Sometimes they just didn't believe, and demanded to conduct analyses, despite the fact that the probability of failure was lower than 1E-9.

Pictures

27 Jan 2014 00:31

Please, rate:

рейтинг: +6+–x


Facebook vk16.png twitter-16x16.png livejournal.gif mailru-share-16.png odklsmall.gif

Добавить новый комментарий
fancy-divider.gif

You may use site content under the terms of Creative Commons Attribution-ShareAlike 3.0 License

Сноски
Пока не указано иное, содержимое этой страницы распространяется по лицензии Creative Commons Attribution-ShareAlike 3.0 License