In the previous introduction, we discussed the fundamental concepts of testing and evaluation in product and machine development that remain unchanged throughout history and across cultures, as well as the single-shot destructive test (limit load test) used during the development of the Comet.
This time, continuing from last time, let's focus on fatigue tests that confirmed the fatigue failure that was the direct cause of the Comet jet crash.
Before getting to the main topic, I have an apology to make to everyone. This time, as it's a culmination of the mechanisms I've introduced so far, I apologize for the large number of links to previous articles and for the length of this explanation, as fatigue testing is one of the most important elements in product and machine development, and I didn't want to cut it short.
HoweverThis book is essential not only for product and machine development, but also for intangible aspects of manufacturing, including electrical and electronic products, system and structure creation, organizational development, and policy-making. I hope many people will read it because it provides a fundamental framework for thinking about these things.
Thank you very much.
First, regarding the fracture of an object, please refer to the previous explanation that fracture caused by a single large load is called a one-shot fracture, and fracture resulting from the application of a certain load multiple times is called a fatigue fracture.
For more detailed information on how to get to know them, please refer to this.
Now that the introduction is over, let's consider the fatigue tests of the pressurized chamber during the development of the Comet.
Fatigue testing of the pressurized chamber of the Comet aircraft.
Now, let's consider the fatigue testing of the Comet's pressurized chamber. Before that, however, please refer to this for the difficulties encountered in using the same parts for both the single-shot destructive test and the divided pressurized chamber for testing.
Before considering the problems with fatigue testing in the pressurized chamber at Comet, let me briefly explain the typical (normal) fatigue testing procedure, regardless of time or place.
General fatigue test
This fatigue test is part of the endurance test, which is the final test in the testing process I explained last time, and it is the most important test that is performed last within that endurance test.
Assumed before conducting the fatigue testThis final test can only be performed after all functions and performance have been verified and the product has passed a simple durability test (limit load test).
Furthermore, the basic idea of fatigue testing is that the part being tested is$10^7 times$ (an important indicator of machine lifespan)This test will confirm that it can withstand the load.
In reality, applying a load to a part 10^7 times = 1000 million times is not possible.From an efficiency and time perspective, this approach is not practical. In reality, the company uses its own know-how and experience, along with reliability engineering, to condense the test time by using complex settings and conditions (such as varying the magnitude and duration of the applied load).
Even so, in my experience, fatigue tests can take a minimum of 3-4 days, and longer tests can sometimes last as long as 3-4 weeks.
MoreoverApplying a simple tensile load multiple times to a simple part, such as a metal piece as shown in the diagram, is not particularly difficult. However, machines are composed of multiple parts, and forces are intricately intertwined (tension, compression, torsion, bending, etc.), making the conditions considerably more complex.
In other words, setting the conditions for fatigue testing is extremely difficult.
Tests created through this processThis is generally called a combined durability test (or simply a durability test).
After this combined durability test"Having a sufficient lifespan for market use, and not experiencing any malfunctions during that lifespan."They are mass-produced under that assumption.
In other words, it's not an exaggeration to say that for any decent industrial product, "no product is sold without undergoing this combined durability test," highlighting its importance.
Furthermore, the test conditions and details for combined durability are as follows:The details vary from company to company and organization to organization, and are considered highly confidential.
In some cases, even within the same company, it may be impossible to know the details of a combined endurance test if the departments are different.
In other words, the content of the combined durability isIt's a treasure trove of know-how, built upon the various failures, market performance, and experiences of companies and organizations to date.
The atmosphere of a combined durability test (durability test of an automobile engine)
I will now explain the atmosphere of combined endurance racing using automotive engines, which are my area of expertise.
Simply put, the work of a car engine consists of two types: acceleration and deceleration (increasing engine speed and decreasing engine speed).
However, when you look at how cars are actually used, gradual acceleration, sudden acceleration, and constant-speed driving are all intricately intertwined.
To put it simply, examples include slow driving in traffic jams, sudden starts, rapid acceleration, rapid deceleration, and sudden stops when you're in a hurry, as well as continuous driving at a constant speed on highways.
Furthermore, considering the situation overseas, we have everything from the extremely congested traffic jams in Indonesia to the ultra-high-speed continuous driving at average speeds exceeding 200 km/h on the European autobahn.
For specialized vehicles, such as sports cars, the types of driving involved also include mountain roads and racetracks.
In addition to the load from the engine itself, external factors such as the impact when the car goes over a bump or lateral G-forces when turning must also be considered (e.g., cobblestone roads in Europe or unpaved roads in developed countries).
The combined durability is determined by considering all possible scenarios for how these features might be used, and then integrating the company's mass production track record, know-how, and real-world driving data.
To digress slightly, in Formula 1, the pinnacle of motor racing that many people are familiar with, a system has been in place for some time now that uses real-world data from each circuit to apply a load to the engines that is almost identical to that of an actual race.
Furthermore, it even replicates the driver's individual habits.
For example, some people have a habit of pressing the brake and accelerator at the same time, or they keep the accelerator fully open when shifting gears in the transmission, or they suddenly release the accelerator in unexpected places.
Even such quirks can sometimes be included as conditions for combined durability.
When I was developing racing engines, there was a time when I would carefully consider and design them while scrutinizing driver data (changes in engine speed, throttle position, acceleration, and G-force changes).
There are many things you can't understand from data alone, so I've also asked drivers directly and gone to race sites to observe the actual vehicle behavior and the data displayed on the monitors while thinking about it.
Even after doing all this...New products and machines can unfortunately develop malfunctions due to unfamiliar phenomena and unexpected uses (new things are difficult).
The importance of actual measurement data and real-world driving data as the basis for combined endurance tests.
There's another important point, which those with keen intuition may have already noticed:The essential conditions for combined endurance racing, as explained above, are real-world driving data, mass production experience, know-how, and other forms of experience.
Naturally, unless you participate in races yourself, it's almost impossible to obtain accurate real-world data, and in most cases, the first year is quite difficult (no matter how much you increase the test mileage, it can't compare to actual racing).
AnydayEven with the best thinking skills, and no matter how advanced simulations become, creating a highly accurate composite endurance race, even one conducted within a limited space, rules, and conditions, is extremely difficult (almost impossible).
much lessPassenger cars, which are frequently and freely used by various people in various environments, are extremely difficult to categorize because the conditions are too complex.
In a sense, and without fear of misunderstanding, machines like rockets, which are used only by a very limited number of professionals and strictly controlled, are considerably easier to manage from the perspective of overall durability.(The machines that people use every day are the most difficult to develop.)
In other words, No matter how advanced theories and simulations become, without actual data and know-how, it's absolutely impossible to create a composite durability system (you can't know until you actually try it).
In fact, when designing mass-produced products, actual measurement data and real-world driving data are incomparably more valuable than any studies or simulations.
There is still much we don't understand about natural laws, including human behavior. Even if something is theoretically possible, there are enormous obstacles to overcome when it comes to actually putting it into practice.
そのEven in 2021, there are virtually no tools that surpass experience and actual measured data when it comes to overcoming obstacles.
To illustrate this difficulty with a familiar example, consider the recent trends of electric vehicles and autonomous driving. While each technological element (motor, battery, radar, AI, etc.) is fairly mature, there is a severe lack of mass-production driving data, real-world measurement data, and know-how for actually combining all of these elements into a vehicle. As a result, it may seem like it could be realized quickly, but it's proving to be quite difficult.
Furthermore, when considering electric vehicles, some might think that an electrical company specializing in motors and batteries could be easily established. However, unfortunately, such a company lacks real-world driving data, measured values, and know-how. Therefore, no matter how knowledgeable they are about each technical element (motor, battery, etc.), it is almost impossible for them to immediately produce a safe car.
Frankly speaking, it's too dangerous to drive an electric car if some unknown manufacturer suddenly releases one out of nowhere.
Just becauseIt's not that all new entrants are bad,For new entrants, the key to success lies in understanding the difficulty of mass production and using all available ingenuity and effort to acquire as much driving data as possible that closely resembles mass production.
That's why companies like Google are working with the California state government to run self-driving cars on public roads and collect real-world data.
no matter howNo matter how advanced equipment, massive experimental facilities, or supercomputers try to replace it, they cannot compete with real-world data (actual measurement data).
In other words, In engineering and manufacturing, approximately $ \frac{2}{3}$ of the work is based on experience, making know-how, mass production track record, and mass production data extremely important.
As I've mentioned many times before, the level of accuracy and depth of this kind of combined durability test, based on real-world data and know-how, is a measure of a company's true capabilities that cannot be measured solely by economics (manufacturers with many defects often lack know-how).
The author's advice to the government for industrial development and innovation
If an engineer of my caliber were to offer advice to the Japanese government on achieving its policy of "electrifying many cars," it would be not just about investing money or cutting taxes,I think it will be quite difficult to realize this unless the Ministry of Land, Infrastructure, Transport and Tourism, in cooperation with the police and public security agencies, creates special zones where many prototype electric vehicles can be driven and environments where many people can ride them, allowing participating manufacturers to accumulate actual measurement data, driving data, and performance data.
In my opinion, the manufacturer's true intentions are:I think actual measurement data, real-world driving data, and performance figures are far more valuable than money or tax cuts.(A form of limited deregulation in a sense).
Even Japan's largest automobile company (arguably the world's largest) faces difficulties obtaining permits from the Japanese government, so they have to take prototype autonomous vehicles all the way to California in the United States for test drives (they would really prefer to do it in Japan).
I also tried to obtain permission to collect driving data for the prototype on public roads, but it required coordination with the Ministry of Land, Infrastructure, Transport and Tourism, the National Police Agency, the Public Safety Commission, the Ministry of Economy, Trade and Industry, as well as local public safety commissioners, police, and administrative bodies, which was too troublesome, so I gave up.
The process was complicated and time-consuming, and the conditions for obtaining a driving permit were far too restrictive, making it seem unlikely that I would gain much from it.
Overseas, in Europe, if you get permission from the EU transport authority, you can drive anywhere within the EU, and in certain states in the US, they'll even try to sell you their services to operate in their state.
In other words, I've experienced firsthand how incredibly difficult it is to foster innovation in modern Japanese policy and its vertically structured administrative system.
This meansI believe that this is absolutely essential for realizing things that will change society, not only in the case of self-driving cars, but also in the case of rockets, military satellites and fighter jets, as well as complex machinery and civilian transportation equipment (even if it's limited, such as deregulation and special technology zones).
I've written a bit too much, but I hope you understand just how important this element is.
Finally, let's start thinking about the fatigue test of the Comet's pressurized chamber.
Fatigue testing of the pressurized chamber of the Comet aircraft.
As explained above, the fatigue tests on the Comet's pressurized chamber were likely conducted based on the development and manufacturing company's existing know-how, actual flight data, and proven values (they had the know-how and data from producing passenger and fighter jets).
If the test simply involved applying a constant pressure of compressed air to a pressurized chamber and then removing it, then that company would be in serious trouble (I want to believe that's not the case).
If fatigue tests were truly conducted solely under the commonly cited conditions of "applying a simulated load of 0.56 [atm] to a pressurized chamber, and then, for safety reasons, applying twice that load (1.12 [atm]) once every 1000 cycles," then one would have to conclude that the development and manufacturing company was not at a level to develop such life-threatening machinery (the actual conditions were likely much more complex).
Now let's consider how to conduct fatigue tests on the Comet.
As explained above, fatigue testing is a very important test and an event that attracts considerable attention from companies and governments.
A view of the fatigue test of the pressurized chamber of the Comet helicopter.
To attract attention means,Companies and organizations typically employ a rigorous checking system for fatigue testing conditions and procedures, similar to that used for design drawings.
I apologize for talking about myself, but the final test of the engine was conducted under a name like "Comprehensive Engine Endurance Test," and even a design of my caliber...During the testing period, every day after arriving at the company, before going to my desk, I would always go to the test room to check if the engine had broken down and stopped the test, and if there were any abnormalities in the data during the test (oil pressure, water temperature, horsepower, etc.). Only after checking these things would I go to my workplace.
While design managers don't typically visit the testing site every day, I always reported the test results to my superiors.
That's itThe testers check on the equipment every one to two hours, and if they notice anything unusual, they may stop the test and check each component and test equipment.
Furthermore, if the engine being developed has challenging elements such as new mechanisms, even if there are no abnormalities, the testing is stopped at regular intervals to thoroughly visually inspect the parts, taking great care not to overlook even the slightest cracks or damage.
In other words, it's essential to be able to immediately detect any abnormalities in either the component being tested or the testing equipment.
HoweverThere is no evidence that such basic checks were performed during the fatigue testing of the Comet's pressurized chamber.
It's a bit of a roundabout way of explaining, but I'll describe the events that occurred during the fatigue testing of the pressurized chamber during the development of the Comet.
As mentioned in the section on the pitfalls of development testing during the Comet crash, please recall that compressive residual stress is one of the pitfalls in development testing.
I believe the fatigue test conditions for the Comet's pressurized chamber involved applying a load twice the expected load in a cycle of expansion and compression, repeated once every 1000 cycles.
They explained that the doubled load generated compressive residual stress, making the strength greater than it actually was.
To explain this in more detail, it is believed that cracks actually appeared in the corners of the windows and around the antenna holes during fatigue testing during development.
However, it appears that the crack was crushed by the compressive force of the test before it could grow and lead to the collapse of the pressurized chamber, thus preventing a complete failure.
In other words, During the fatigue testing of the Comet aircraft, cracks had appeared in the corners of the windows and around the antenna holes (they just didn't lead to complete destruction).
If a standard testing management system is in place, then, as explained, checks are frequently performed even at the staff level, so it shouldn't be difficult to detect cracks during testing.
Either the person in charge honestly reported the crack, but their superiors, due to lack of time or panic, dismissed it, saying it was fine since nothing was actually broken.
Furthermore, if a single-shot destructive test is conducted correctly, it will be clear with 100% certainty that stress will concentrate at the corners of the windows and the antenna holes, causing the pressurized chamber to break in sections. Therefore, during fatigue testing, these areas, such as the corners of the windows and the antenna holes, should be checked very rigorously and frequently by many people.
much lessSince the Comet was the world's first jet passenger aircraft, it involved a level of novelty and challenge far beyond what I've experienced with car engines. Therefore, under normal circumstances, everyone should be taking great care in conducting fatigue tests on the pressurized cabin.
What I'm trying to say here isIt is inconceivable that antenna holes or cracks in window corners that occurred during fatigue testing would go unnoticed or undetected.
Checking the pressurized chamber after the fatigue test is complete.
Even if checks are not performed during the test, once such fatigue tests are complete, the machine is completely disassembled and each component is carefully checked for any abnormalities (particular attention is paid to checking for cracks).
Furthermore, since cracks can be overlooked during human visual inspection, a special solution is used to color the cracks, making them more visible (penetrant testing, crack checking).
I wrote that it's a special solution, but it's not that it's difficult to obtain, expensive, or a recent invention; it's an old solution that you can find at a home improvement store if you look for it.
We'll use this solution to stain all the parts and check for any tiny cracks.
In addition to thisEven if there are no visible signs of stress concentration or other suspicious areas identified during the design phase, internal cracks may be present. Therefore, a materials expert carefully cuts the parts and meticulously checks them under a microscope for any traces of cracks or signs of fracture initiation.
Even after doing all this, there's still a possibility of overlooking something, so in order to let more people see it, we're going to display all the disassembled and dyed parts, as well as the parts that have been cut and examined.We'll hold an exhibition that everyone on the development staff can see, including department heads and section chiefs.
In some cases,This is such a crucial project that it wouldn't be surprising if the technical director came to see it.
Normally, if everything goes smoothly up to this point, the testing is considered complete and the next step towards mass production can be taken.
In other words, even if a crack isn't detected during testing, it can be found with 100% certainty if a standard inspection of the completed test product is conducted.
How muchEven if a crack is crushed by compressive load and is not visible from the outside, if you cut the part, you will always find evidence of it (materials are honest and always leave traces of any changes).
も う 一 つThe fatal mistake I can for Comet at this stage is that, although a crack was found in the parts after the test, they may have passed the test because the entire pressurized chamber did not collapse (which is a dangerous judgment to put it in terms of "there was a crack, but it didn't progress, so it's okay").
This is the worst decision you can make for a machine. If a part has a crack, it can withstand loads in a specific direction to some extent, but it becomes extremely weak against loads in a different direction (which is obvious since there's a crack).
Moreover, once a part develops a crack, it breaks down differently from a normal part; the crack becomes the starting point, and the destruction progresses at an incredibly rapid pace, causing it to shatter into tiny pieces (a phenomenon called disintegration).
In other words, the pressurized chamber was able to withstand the cracks only because, by chance, the load and direction of the load applied during the test were such that cracks appeared.If the aircraft were to actually fly with a crack in it, it would be in a very dangerous state where it could be caught in a gust of wind or turbulence, and even a slight load applied from an unexpected direction could cause it to rapidly break apart (there is a high possibility of it disintegrating in mid-air with even a slight disturbance).
A common example is a glass cup, vase, or pot. If you simply drop it and it breaks, it will usually shatter into two or three large pieces. However, if it already has a crack, dropping it again will cause it to shatter into many pieces with just a slight impact—something you've probably experienced.
For example, if you wash a teacup that has a crack in it without noticing the crack, depending on the direction, it might break in half with just a slight touch.
The same applies to metal materials; even if they don't break during durability tests, if they have cracks, they are in a very dangerous state and absolutely unacceptable (dangerous).
In other words, at this point, it's impossible to determine whether it was intentional, accidental, or due to the engineers' incompetence, but there is physical evidence (cracks) that could lead to the series of crashes, and despite the very high probability that the cracks had been discovered, development proceeded as if there were no problems (judging from the fact that test flights were conducted immediately).
This meansIt's hard to believe that the problem wasn't even about the engineering issues of stress concentration and fatigue failure, but rather that there were major issues with the testing check system, judgment criteria, development system, organization, governance, and ethics.
The problems with the fatigue tests conducted so far are as follows:
The author estimates that, given the scale and conditions of the fatigue test, if it were conducted over a one-month period with daily checks, it would have been possible to check the cracks 30 times, or even 90 times if checked three times a day, yet the cracks were missed or ignored.
- Crack checks were not performed on the parts after testing, cracks were missed during testing, or cracks were noticed but ignored because the parts were not broken.
- The pressurized chamber was deemed acceptable as it was not damaged at the end of the test. The cracks were deemed acceptable as they had not progressed.
If something like this really happened at the Comet development site, then...This is not just a technical issue, but also a moral one (in my experience, conscientious testers and designers on site explained the occurrence of cracks and their dangers, but it is highly likely that managers, who were pressed for time or had management responsibilities, covered it up).
We've considered the conceptual and design issues up to this point, but the consequences of this particular problem are critical, as it essentially guarantees a crash.
このThe shoddy nature of the fatigue tests offers a valuable lesson for us today.
Next time, we'll finally start thinking about the details of the test flight.
To those who found this article helpful in understanding design:
Since we're on the subject, I'd like to recommend a book that's essential for mechanical design.
To be honest, the content is extremely unhelpful, but it can be used like a dictionary when you forget the details. If you read this article, you should be able to understand the content and use it effectively. It also includes commonly used standards, making it quite useful.
If you don't already own one, I highly recommend getting one, even though it's a bit pricey. However, new ones are expensive, so if you're considering buying a used one, I strongly recommend checking that the surface roughness conforms to the new JIS standard.
Comment: