Development Of New Aircraft

[dov]

Hallo

A look at the development of new aircraft. The high costs, overlapping development processes, and then static problems in many places. I consider this in the forum because I want to show the current weaknesses apart from any electronic development. From my own experience in development and design in large machine design and construction and series production, I can bring my experience to bear.

I myself was there when vibration problems destroyed a large hydraulic power plant for the first time. At that time, the line from medium-sized machines to large machines was crossed. This means that the elastic modulus of the material and the spring constant of the foundation play an important role and shift critical speeds. So, you may have the nominal speed exactly on the critical speed. Many major projects in power plant construction failed and had to be improved over many years. The development of crack mechanics on the off shore oil platforms analogous to this.

This development already began in aircraft construction with the F-18. Here the development goals were diffuse and ultimately the range of services behind the previous models. We only consider the aerodynamic and static quantities. No electronics. No weapon systems.

With the A-380, for example, the problem of the wing tips. The static problem could not be fixed because the program was too advanced. This tanks in the wingtips never could be filled. Ultimately, it was stillborn. No success.

The static problems of the spars and ribs on the F-35 show problems with all versions today. Local stress free annealing with laser on the A version, B and C version have to do the fatigue test again. If the basic design of the statics is so poor, such a device remains a coffin nail.

This is far from a good program flow. The combined knowledge is not in the minds of decision-makers from today. They are only specialized. In this way, contradictory decisions are made, with the cost pressure restricting every leeway. At the same time, the absolute belief in the unlimited possibility of electronics blinds rational decisions.

How do you see that? What do you think?

[exdraken]

Very interesting and sound thoughts imho!

 

But hasn't it always been the case that wrong decissions not being cirrected early on un a project get very costly to correct later on  if possible at all! With these projects czrrently goining on over 2 decades, engineers still iron out / fix mistakes being made for what ever reason long time ago...

 

And yes, electronics have come a long way and are really helpfull! But can't as you say overcome physics or major design flaws like exsessive drag, etc...

Edited May 31, 2020 by exdraken

[dov]

Good day

Now I would like to give you another little insight into what happened in the late 1970s and early 1980s. It was then that the era of finite element calculation began. What's this?

Imagine a kit made of wooden blocks. You may have 100 different building blocks, each in large numbers, as many as you need. Now shape or design the machine you want to calculate from these same building blocks. So that the contours are exact. These building blocks have a mathematical inner life. This was shaped in particular by the Polish mathematician Zienkiewicz.

However, in order to be able to model exactly, you can also deform any component. This means that the angles become unequal to 90 ° from a cube, the aspect ratios change etc. This is exactly where things start to get exciting. Each module, each element must be checked by reference calculations to determine whether the results under load after the calculation are still correct or incorrect. Which elements you can combine? Where is the geometric deformation limit without load? That is the work that I had to do among others. Set limits on when and under what conditions an element delivers incorrect results.

This was the basis for the recalculation of the Tornado torsion box and the early Airbus models.

The analog game must be carried out when the force is applied. You need a precise picture of how a force acts on a component. Over what surface, evenly or in what course. As with a huge zoom, you have to look closely at all areas of a component. Welded seams are a special challenge because the material is no longer homogeneous and its microstructure is embrittled due to the high temperature and only becomes approximately homogeneous again after stress-free annealing.

All this common knowledge is a prerequisite for a correct calculation. The correct (correct results delivering) model design is a prerequisite for this.

You can also perform dynamic and thermal analysis using this method.

In the current worldwide education, the knowledge required for this is no longer conveyed. Unfortunately, the creation of determinants to understand the mathematical context is completely missing. I have a lot of large corporation (car, engine and aircraft construction) calculations that have hair-raising errors.

With this background knowledge, I look at today's developments.

How do you see that? What do you think?

 

I will continue with some very basic aspects, so that everybody may get a picture of what is present day devolpment. I was part of this for 4 decades.

[dov]

Good day

Now let's look at the results of this new method, the finite elements.

At each grid point of the model you will receive stresses (normal and main and comparative stresses) as well as torsional stresses and deformations. Many elements also have half-length dots. The calculation process takes a long time to complete the corresponding analysis. Back in the late 1970's it was hours. Compared to today's computers that are faster and with better elements, the computing time has increased dramatically without getting a significant improvement in results. It is an evil development. The causes are that internal systems of equations in the determinants of the finite elements are poorly understood. These are the program developers at work. Thus, many useless mathematical operations are caused because you cannot cancel them, because you simply do not know the relationships from physics. On the other hand, you ignore the factors of computing time and computing capacity.

The evaluation of such an calculation output and its conclusions should flow directly into the design of the components. After a few months of experience, the practiced person can estimate the result in the design phase. This in turn saves unnecessary calculations and enables a quick final result.

To examine the mathematical problems in more detail: The zero search is the most frequently required operation outside of the finite elements. Some function has to be brought to zero. Even if two functions are brought together. (Example F(x) = (1 / X2) - X3 + 10). Various functions make this possible, some are slow, some are faster. The most ingenious functions were developed in the course of astronautics. These have been forgotten. Unfortunately. Think of the early Apollo computers. Maximum computing throughput with minimum capacity. Here a generation learned to program optimally. Unfortunately, this knowledge has not been passed on or is ignored today. The same applies to the initial computers on the Tornado or in the F-111 including the navigation system.

If this knowledge could be reestablished, flight computers and scientific calculations would be significantly accelerated. It is brutal and hard work to create such systems of equations. Hardly anyone can imagine this great requirement. Just a comparison from the book by Franz Kafka: Der Heizer. I was part of such an early R&D team. Myself I stood several times at the abyss because of mental exhaustion.

 

What do you think

[dov]

Hallo

Today we only want to look at the drawing level.

I myself grew up with the drawing board. With all the advantages and disadvantages, I worked productively in the various companies for about 10 years. Then came the change to CAD. Here first 25 years on the two-dimensional CAD and only then in the last 5 years on the three dimensional CAD. What are the differences, advantages and disadvantages of the systems?

On the drawing board, besides mastery of matter and geometry, it was manual drawing. It was really a hard work. A little talented employee was easily unmasked and, since everything was on paper, could cause hardly any visible damage. It was also easy to pause changes and continue. To generate versions. Of course, it was associated with drawing work. This time of drawing was also a good time. Here you could rethink everything that was on paper.

Things were fundamentally different here in two-dimensional CAD. Exactness was brought here to the absolute extreme. Intersections were exact. My personal and operational accuracy was 1E-8 mm. Angularity was absolute.    90 ° and this is not 89.99999999 °. Any line that does not go exactly to the intersection is a horrendous mistake that can have unforeseen negative consequences. This absoluteness naturally takes its toll. Absolute concentration during the entire working time. Here we are in the concentration range analogous to the work on the flight deck during the approach. If you now get a drawing for processing from a colleague who works sloppily with all the errors mentioned earlier, you can destroy the entire design work for an entire device, an entire machine. In addition, a whole year work from a team of 10. Analogy of a total loss of an aircraft. I also had such an experience. Tolerance in the team, regarding accuracy is absolutely zero. Here the human relationship changes completely. The first aircraft in Europe to be designed that way was the Alpha Jet. On hot summer days we had to stop working, because the sweat of the arm and hand caused regular short circuit on the tablet. At the first decade, a tablet was usual.

Now for 3D CAD. There are euphoric positive aspects but just as many negative ones. Anyone who knows the mouse can do something here on the screen. Bluff. A design must first be dismantled to check whether everything is actually correct. Consistency checks! However, this control is almost impossible here. They cannot generate a print from a drawing where they automatically switch off partially invisible lines. There is only yes or no. If so, the drawing is chaotically confusing. If not, too much information is missing. It is the same with center lines. It is an absolute waste on ink. The computing times become unusably long with properly large drawings. You can sit and wait or drink coffee. But at all, you lose productivity in the company and calculated annually a huge amount of money! Submitting drawings, division into a team is very difficult even with the greatest caution. Cooperation Europe and China should be rated the same as in the same office with two workplaces. Even if they are geniuses (developers), they are constantly pushing the limits of the system. The system is full of pitfalls. You cannot represent a framework correctly because the angle profiles with the radii blow up the computer. Welds are a total dilemma. The list is very long. A small detour into model making: The engine cowlings on WW1 and WW2 aircraft form an area that cannot be mastered with 3D CAD. Something is possible with tricks. But not really. The cladding panels rest on the fuselage. They do not form a surface which is based on a sum of surface points of the fuselage. Only the screw connection. The rest is somehow hanging in the air. The same goes for the Bf 109. Overlapping sheet metal joints are a horror! Check our climate debate: The waste heat generated by such good computers in comparison threefold the people working in office! So you need the air-condition!

For the sellers, 3D is a feast. You can show the layman pictures. Yes, but that's where we really ended up at the picture book level. All 3 systems have their authorization. Analogue of glider, engine and jet drive. Only rarely does someone create a good symbiosis here.

So much for this topic.

What do you think?

[malpaso]

Your thoughts on the drawing board vs CAD vs 3D match my experience, though luckily architecture is generally a simpler discipline.  Unfortunately there is a lot of reliance on the computer programmer knowing how a building is put together (obviously they have no idea, they know about computers) and younger colleagues sometimes lack experience to spot an issue arising. 

Of course they just think I'm a dinosaur.  In my world the accuracy of the computer can be awkward from another direction, for example a computer UK brick will be exactly 215 x 102.5 x 65mm, leaving aside the ludicrous .5mm on the width, this is a product that can vary by +/- 3mm.  Or old buildings that are nearly but not exactly square, and walls that aren't mathematical planes.

Then there's the joy of annual revisions to the software which more recently include no backward compatibility.  Wow that's a great improvement - for the software company and their resellers, but hopeless for the users.

[dov]

Good day

After we have a very small basic idea of designing a new engine, many steps are still missing to understand the entire process. It does not matter which machine it is in detail.

We are now going through the communication process between the drawing and the workshop. How should and must a drawing of an entire machine be arranged and structured so that no misunderstandings arise, everything is described so clearly that no rejections are produced and the worker does not take too long to understand the instructions distinctly and clearly.

Here we touch an important part in modeling: the instructions. To be honest, the level of modeling instructions is appallingly low, if not catastrophic. If you had delivered such drawings as a designer, your days in the company would have been numbered. Please imagine the reality: Your drawings go to the production company. Locksmiths, welders and lathe operators work there, etc. All work in piece rate. Scheduled time for an activity. In plain language, this means that if the man at the machine doesn't understand your drawing, he won't treat you with the greatest kindness because you will steal part of his wages! That's the reality. In the event of repetition, dear reader, of course you ask yourself the question:  is the worker stupid? Hardly if that were the case, he would no longer be in the production company in question. You are therefore dependent on listening to messages that the worker wishes for easier understanding and implementing them in the next drawings. If you think I'm the Emperor, the other person should see how he gets on, then your career as a designer will come to an abrupt end. Unfortunately, this learning process is barely noticeable among modeling companies. Yes, they reject such comments! No matter which company, whether praised or not, the instructions are often a real laugh. Examples would fill many pages.

But back to our drawings: Every machine is basically broken down into main assemblies. Including subassemblies. Depending on the complexity, this is further subdivided. As an example, take a welding console with two flanges 12mm thick and a rib 8mm. We use hot rolled simple steel. We'll leave the denomination by standard specification of steel out, otherwise it will be too complicated. The console should be accurate at the end of the manufacturing process. So that the part is also cheap, we will take the components raw for the time being, not process them and only cut them. Components of this type are always welded with jigs (very simple). Double fillet welds on the rib with 6mm are sufficient. The support plates are not directly connected to each other. Only through the rib. When we have finished welding, the work piece is cleaned by brush and then a base plate is milled. Then the second at an angle. Drilling come after that. To process this, the exact dimensioning of the individual parts as well as the exact dimensioning of the finished part is required. Material specifications, tolerances of the individual parts and overall tolerances. Form tolerances and position tolerances are very important here. Each measurement should only be as accurate as necessary. And easy to read. As 130 and not 127.14 for example. Or 130.5 and not 130.434. You also have to measure and set each dimension quickly without any reading errors. Each component also needs a clear position number on the assembly drawing that cannot be misread, for example by the dyslexic. Such a component is shown in several projections. Perspective projections are the least meaningful. Orthogonal projections are easy to read and fast to interpret without error. Perspective projections are a popular option in 3D CAD. All information on perfect production must be clear and unambiguous. In addition, there are of course exact material information, weld seam information, surface processing information, until then whether the part is then ultimately painted and with which paint chain. This is, of course, a simple example. But please imagine an entire device made up of a hundred or a thousand such assemblies. Yes, you need an exact overview. What, how many etc.

We have now understood something about the basic idea, the calculation and the drawing and the production. However, the sum of all components must be produced in many companies all over the world and controlled in a targeted manner using automatic lists.

I wrote such programs myself for 20 years. Next time more of that!

[Giorgio N]

Thanks for your description of these various design stages and the risks of certain technologies, it's all information that I'm sure many here are not aware of and will be of much interest.,

However I'm not sure about what you describe as "lost art" and about your comparison of more modern with older designs, where I seem to understand that you are proposing the idea that today aircraft suffer from problems that were better approached in the past. I may have understood incorrectly here, so I apologise if I did...

My view as an aerospace engineer is that my colleagues today are not doing worse than they did in the previous decades, and are actually doing much better. the static design of today's aircraft is much more advanced than it was in the past and this is proven by the fact that the lifespan of combat aircraft has constantly increased over the years.

At the same time it's clear that any new material or manufacturing technique may give troubles initially as it takes some time to get all parameters right. This has proven particularly true with composites materials, where there's the added complication that the material properties are not as easy to model as they are in metals. Even more important, composites properties generally show a more marked variation due to the manufacturing process and the designer has to take these variations into account. Something that in areas outside aerospace don't always happen as it should...

However even with well known materials and techniques, the history of aviation has shown a lot of problems on aircraft and I'd probably spend a day or two writing about aircraft that needed reinforcements during their career as something was not right from the start. This simply because the interaction of the aerodynamic loads and the structure is something very difficult to model in the design phase. Sure there are today plenty of very advanced simulation tools, but they still can't predict perfectly what will happen to the aircraft once in service. Even more considering that once the aircraft is in service it may well have to meet situations that have not been considered during the design phase. Sure, nobody wants to find out that after a couple of year they have to keep airframes checked constantly, but that's very often just the way it is.

And actually this is sometime a deliberate decision ! This is particularly true in civil engineering, where one of the many approaches is to design and build a structure that features a reduced safety factor but this is compensated by the introduction of constant monitoring. A similar approach has been proposed for aircraft structures through the use of the so called "smart materials", it may sound strange but it's an approach that would result in higher performance at a reduced cost.

 

What I see in my everyday job that agrees in part with your observations is that in certain areas there has been a constant trend toward leaving structural design to technicians armed with FEM software rather than employing a more expensive experienced engineer. It is something that I find very dangeorus as even with the best FEM package, whoever prepares and runs the model should know what they are doing and the best way to achieve this is to let someone who know about structures and materials do the job. Unfortunately it's a trend that has been going on for years and there's no sign of change. However this is something that I tend to see in smaller private enterprises, like in construction. I've never seen this approach at a major aircraft manufacturer...

[dov]

Good day

Today I want to introduce the variant-design to you. What's this? It is basically the way, that every design, every part can be designed for multiple purpose in this way. A basic design is done with parameters controlled so that all variants for which this design should apply are covered. As an example, a rail: minimum length. up to maximum length, in common L_min - L_max. A plate with e.g. 15 holes of different diameters, of which only e.g. 6 are needed. A width and length variable framework with different numbers of fields. Variable truss fields with variable diagonal lengths. Width-dependent parts, which are calculated correctly depending on what width is needed. Just imagine ribs on a wing. You can do this both on the drawing board, with 2D CAD and 3D CAD.

What do you need for this: We have to know the engine we want to build well or be able to assess its future needs well. Then we have to know all components well. And all the subassemblies and individual parts that play along. Yes, before I forget, we can also configure the materials, depending on your choice. So I need a main program as input, where I can configure the device. To do this, you have to know the customers and the market requirements well.

The backbone of all equations behind parameters are algorithm. This word is today often used. This are mathematical formulas, like a poem, they tell you after an input the right story. Some are very complex, some are easy. Some are copies of others. As better you generate them, so better you may use them in other new sequences of a program.

Of course, I can also parameterize a main list of parts including many list of parts in a system, and so I get everything printed out correctly depending on the order. And also drawn or noted in the correct geometric dimensions on the drawing so that the manufacturing process can be carried out automatically. To give you an idea: Such a complete set of fully resolved parts lists of an escalator fills several thick folders!

In case we would have no variant design, you may have drawn every truss once again. You may have filled out each list of parts by pencil. Too much time. The time used for a contract in a company, can be extremely reduced by using this technology.

The creation of such a program is very complex. In my case, we had the number of  constants approx. 10⁸ options when entering. Including correct defaults for all values so that the program runs smoothly and correctly, that was a challenge. It is still very important to continuously improve such a program. Error messages on the level of improvement occur at the beginning of a runtime (the application of my programs were about 15 years). This is the standard. Never, it occurred that something was not possible to produce or that financial damage or penance was caused. The logic behind such a system is very strict. This Logic was sold worldwide, produced worldwide, and everyone benefited. This logic, it must be understood in this way, must never be avoided, never replaced by other logics in the network. Otherwise such a system breaks down. The clarity in the program language (no matter how primitive or sophisticated) must be comprehensible. For them, even after 10 years of application and constant improvement, and for everyone else who works on the development. Finding a harmonious and unambiguous language here and only tolerating it needs creativity on the one hand and hardness and brutality on the other. There is no room for compromise here.

On the other way, you have to consider first every step of a design change, or improvement. How it implement this new feature in the overall design and program. The designer has to talk first to the people responsible for the program. I had several cases, when I had to explain the COE, sorry it is impossible. So, we both had to find a new way. By thinking and not by fighting.

They have a basis in aircraft design. Variants that are planned from the beginning can be created quite easily in terms of the material structure as well as the data structure. If no variants but specifications are required, the same applies. It becomes cumbersome when variants only appear at a later point in time.

[dov]

Good day

Language is the link to all activities up to the finished product. Spoken and written. Too little importance is generally given to this fact. Why? I think it's so natural that hardly anyone thinks about it in detail. I also did not, up to a point in time, an experience.

I was lucky enough to have a friend who was Commander on the 777. With his crew members, I had the opportunity to fly the VIE to NRT (Vienna Tokyo) route several times. (9 hours approx.) And to hear the language, the way in which it was spoken. That was new territory for me. That fascinated me with clarity and exactness. Another friend is flying Airbus 320 and he is teaching CRM.

The step-by-step structure and the exactness and clarity of the statements, I gradually realized in the office, on the drawings and in the program. It was interesting that a lot of things had already been handled by me, intuitively.

I recognized something else as a serious shortcoming. This affects the blurriness of the language. We think we can speak a foreign language, but despite fluent speaking, our vocabulary is too limited to be able to express things in a distinctive way. Let's just take English. Learning English with mother tongue German is a matter of course today. I was lucky enough to live in the United States for a long time. So I learned it differently. But the limited vocabulary I feel like a block on my leg today. It becomes even more drastic in other languages. Take Japanese and Hebrew. I speak both of these languages and I have spent a long time in this countries. Studying and working. From my experience, it sounds like a shame to me in Europe today, the bridge language English has an embarrassing level. Only a few people, even in leading positions, can use and pronounce this language correctly. It is unfortunately a tiny minority.

 

This European language incompetence is fatally reminiscent of the Austro-Hungarian monarchy until 1918. More than ten different languages were spoken, few understood them. During the First World War, every effort in production down to the battlefield failed because of this. People were even less familiar with the peculiarities of the people.

Back to the causal. It's really hard to write instructions on drawings that all workers across the globe understand by the same thing. If the language is too precise, a worker with rudimentary language skills understands nothing at all. If it is too superficial, then they have the wrong result. In order to get the right understanding of the right language level here, they can only have it if their workers and fitters report to them at home. Implement what you hear and thus create a constant improvement of the drawing documents. It is analogous with the program and the comments in it. Here they have a much higher level of language, but the programmer generally doesn't listen very carefully because he is too confident of himself. This is the most dangerous trap.

Operating instructions, here the flight manual is the pinnacle of language development. You will not know, but here is the focal point where all new knowledge of language logic and language philosophy come together.

There is nothing better in the world.

What do you think?

[dov]

Good day

 

Today I want to tell you about an area that is difficult to imagine unless you have experienced it consciously. This area dominates us at all times and in every matter.

Knowledge transformation. You may get knowledge from outside that you have not developed yourselves and therefore you do not understand. Now it may be that you learn to understand this external knowledge over time and then you may develop it further, or never understand it and you can only live from the fruit as long as it is edible.

 

I myself have worked with and in such a system for 20 years. Our system supplier was a multinational corporation with well over 10,000 employees and huge capital. I only fully understood the program system he developed after about 3 years. But understood so that I understood the mindset of this group. I could read their minds and develop further. I knew all the weak points and was able to write better than they themselves. So I was given more responsibility in many small steps. Up to the point when I was able to supervise entire devices myself in terms of programming. After some time, I was responsible for the entire system in development. This burden of absolute responsibility pushed me hard on my shoulders. I wore them and defended them against illogical attacks. Even from the COE. As a small five-strong company, we managed a multinational company. Of course, in an upright and correct sense.

 

We have an analog development in aircraft design and development. Please take only a few countries: Germany, USA, Israel and Japan. We are leaving out China. It does not play a role in this particular view. Here again we focus only on the USA (GB) and Germany.

 

After World War II, many of the leading heads of aeronautical research lived in Germany. Outstanding developments were largely made there. Outstanding criteria were the lack of resources and the ingenuity. Key people from this very field were 'invited' to the USA or asked to come. In order to be understood here, I was lucky enough to be taught by some of these people as students and pupils in Europe. I also had the opportunity to be taught by a teacher who worked as a young technician with exactly these Germans in the same company in the United States. Now comes the important statement:

 

The highest doctrine among these German scientists was to answer their American superiors only those questions which they asked! Roger? Not more. The Americans never found out what they  did not see or ask or have overseen. The Germans did not tell them anything voluntary. The lack of respect for American superiors was the driving force. Victory through mass and mass production.

 

With this very doctrine, you can destroy any system in the world. That is the secret of why development looks like it is today. Just a matter of time.

In contrast, the British version. My best university teacher went to the UK. Working there was exactly the opposite of the US. That was the integration of German scientists and the sharing of common knowledge. Respect played a major role in this. British achievements in World War II were highly recognized by German scientists.

 

Look at today's developments from this point of view.

 

I want to end this series.

I hope that I have not insulted anyone with this article; it corresponds exactly to the oral narratives of the scientists working at the time.

 

Happy modelling