Bjorn’s Corner: The challenges of airliner development. Part 19. Test rigs for ground and air use

By Bjorn Fehrm, Henry Tam, and Andrew Telesca

September 3, 2021, ©. Leeham News: Last week, we looked at the Certification Compliance Planning we do concurrently with Detailed Design, Figure 1.

It’s now time for us to work on our Test rigs and systems for ground and flight testing. We need to get these defined before we freeze the aircraft’s configuration and start making our flight test aircraft.

Figure 1. The program plan for our project. Source: Leeham Co.

Testing rigs and systems

As we work on our detailed design for the aircraft, all our system suppliers are working on the systems we will integrate into the aircraft.

The suppliers will do system testing of their parts before delivering the system to us. But we must integrate these systems with other systems, and make sure they work together. First on the ground, then in the air.

Ground test rigs

The most spectacular ground test rig we need to plan and produce is the static strength and fatigue test rigs. With the static strength rig, we must verify that our design can withstand the design’s Limit loads (the highest loads expected in service) and the Ultimate load (limit load times 1.5).  

Our rig will test these loads for the wings, fuselage, and empennage structures. As we have an un-pressurized aircraft our fatigues testing (high number of cyclic loads on our structure mimicking flights) focuses on the same areas as our static strength tests. As the static test structure can pass the Ultimate load test with plastic deformations (it only has to stay together and be flyable), we use a second structure for the fatigue tests.

Figure 2 shows the static testing of the Airbus A350 wing, taken from this video that also discusses the ultimate pressurization loads for this pressurized aircraft.

Figure 2. The Ultimate load test of the A350 wing, performed in the A350 static load test rig. Source: Airbus Youtube video.

To understand why we need ground test rigs for our systems, let’s take the example of the flight control system. Our basic control system is a mechanical concept with Yoke and Pedals, pulling wires over bellcranks and pushing pushrods back and forth to move our flight control surfaces.

But we also need a yaw damper, which is part of our autopilot, to make the aircraft fly comfortably for our passengers (the normal compromise between spiral stability and dutch roll for a passenger aircraft is to allow a bit of dutch roll which you then control with a yaw damper).  

The yaw damper is a servo system that is part of our Avionics system. The autopilot/yaw damper must interface with the flight control system, pulling on the wires/pushrods just before our flight control surfaces for pitch, roll, and yaw. 

It does this with electric servos that work in serial for the yaw damper (i.e. the pilot can’t feel its interventions in the Pedals) and in parallel with the mechanical flight control system for the autopilot in pitch and roll (i.e. the pilot can follow the autopilots work by watching/touching the Yoke).

The electrical servos for the Yaw damper/Autopilot use electric power from our electric power generation system. We must make sure the servos don’t have time delays or lower slew rates when our electric system is at its lowest power position, for instance, as the pilot throttles back the engines to flight idle because he’s too high and fast in an approach. Lagging servos can cause dangerous positive feedback situations that can jeopardize the aircraft.

The flight controls also command flaps and spoilers, both operated hydraulically as there we have high forces involved (hydraulics can generate high forces from small physical devices). We must make sure all this works together smoothly even when the hydraulic is at its lowest supply position.

We must create ground test rigs for all critical systems and first test them in isolation and then gradually adding them together. 

We must, as an example, make sure the landing gear is not stealing the hydraulic power from the flaps as the pilot commands gear-up and flaps to start position as he does a go-around after a botched approach.

Avionics/systems integration test rig is another important tool for a development program.  The avionics system communicates with almost every system on the aircraft.  Engineers can use this rig to test out “beta” software to identify and fix issues before loading it onto an aircraft.  Professionals in this industry test as much as possible in a low-risk environment instead of allowing beta software onto flying aircraft.

For all ground test rigs we need to have sensors at the right places to log pressures, flows, slew rates, forces, current, voltages, and temperatures. All these values must be recorded and stored in a well-thought-out manner, so the results can be conveyed to those who need them in the project and to form part of our proof of compliance for the Certification documents we produce.

We need to get these rigs to arrive so we can do tests before we have come too far in our detailed design. This means creating all test plans, work with our manufacturing team, work with suppliers, and of course the FAA as we need test plans with agreed requirements and setups to ensure the data we acquire can be approved.

Flight test rigs

We also need to design the flight test rigs and devices, the ones that shall be mounted in our flight test aircraft. 

Our first aircraft will verify our flight envelope and flying characteristics. It shall be as close to the final production specification for structure, propulsion, flight control systems, etc as possible. The test rigs and systems must log all the relevant parameters for these tests. It must also log all the pressures, flows, movable slew rates and angles, and structural stresses that we measured in our ground rigs. 

An important part of flight tests is flutter tests, a very dangerous checking that we don’t have an aerodynamically caused self-resonance in our structure during flying. For this, we have installed rate gyros, g-meters, and stress gauges on different parts of the aircraft (wings, tail surfaces, engine nacelles…). 

Here we must have a real-time evaluation of results. Shall this be done from the ground? In such a case, we need a reliable data-link system. Or, shall the flight test engineer in the cabin check that we don’t enter a dangerous situation while exploring the envelope?

Unexpected structural flutter is the most common cause of test aircraft break-up in the air. Therefore flutter test and clearing of the flight envelope for flutter is something all experienced OEMs and flight test organizations treat with the utmost respect.

Our second test aircraft is equipped with a full ECS system and cabin. This shall be a production conform aircraft so that we can finish the tests with the required Function & Reliability testing. On this aircraft, we check all aspects of passenger comfort such as airflows, temperatures, vibrations, sound levels, and sound effects from our systems when we deploy flaps, spoilers, and landing gears.

Everything must be measured with sensors, routed via a special flight test EWIS to the test rigs where the data is recorded and concurrently routed to the flight test engineer’s place and there presented on his laptops.

With all the sensors and their data, we check that we get the same results during flight as the values from our ground simulations and test rigs. Flight testing has migrated from finding problems to be a verification of our analysis and simulations, and to gather compliance evidence.

Research, design, and manufacture rigs in time

We need to conclude our research and system work for these rigs and installations in time so we can get the test results we need for the project before the detailed design has come too far. If there are changes needed because of ground testing we want these changes to make it to our test aircraft.

Otherwise, we need expensive modification periods for the test aircraft to final configuration or we might even need to produce a third test aircraft to have a final configuration for our certification flights.

Eventually, we will need a third test aircraft as we plan to offer a cargo variant, but this can be scheduled so this aircraft can be produced on our production tooling (more on this next week) to get the cost of manufacture down.

This is all tricky stuff and now any misses in our work cost major money. This is where project busts their budgets and run into major delays.

39 Comments on “Bjorn’s Corner: The challenges of airliner development. Part 19. Test rigs for ground and air use

    • The illogical and jumping to unsupported conclusions is stunning.

      You feel the FAA and Boeing owe it to you to be fully kept in the loop? Write them and let them know.

      The 777X test result validated why we test. The A380 also broke its wing before it should have (147%). The 787 had a wing join issue.

      Any test failure is addressed. To assert this is going to be or even could be covered up is absurd. Yea I know, the MAX failures prove your point in your mind at least. They don’t but clearly they have become conflated.

      Truly getting into Tin Foil hat area when someone would post that.

      • The aircraft failed the test — period.
        A door blew out, the fuselage ripped open, and there are even reports of structural damage to the floor.
        These failures need to be addressed — period.
        Nobody needs to keep *me* in the loop, but one would expect such a spectacular failure to at least lead to some form of published update. But, then again, the 777X program still hasn’t even managed to secure a green light for TIA…so is anyone actually surprised?

        If this had proceeded in this manner in China, you’d be shouting about low Chinese standards.

        • Boeing tried to game the test with overpressurization.
          ( What they advertised as “beyond requirements” and imu indicating that they knew conformance to be marginal.)
          Still they botched the test.

          Why is it that TW seems to be the condensation of all things “Dunning and Kruger”?

          • Regarding your last paragraph….noticed that too, had you?
            Running around telling others that they’re ignorant and that he’s a beacon of enlightenment…it might be more fitting for him to set up shop at Delphi 😏

          • If you watch the video Bjorn linked to you’ll not AirBus also over pressurized the A350 during the ultimate load test.

            Why this is done I don’t know.

          • Jbeeko: “Why this is done I don’t know.”

            You stiffen the structure of any thin skinned pressure vessel. ( similar domain is failures in silo structures, thesis of a friend long time ago. )
            i.e. you move more of the structure out of compression loading. getting around a bunch of Euler failure modes.

            The Boeing “Pravda” tick off was “beyond requirements”. Usually tends to serve some nefarious purpose.

        • “If this had proceeded in this manner in China, you’d be shouting about low Chinese standards.”

          That is truly a hoot, you have not a clue.

          I would be delighted let alone encouraged to see Chinese test results. So would the world. Haven’t been revealed have they?! Even the FAA despite its low standards at the time could not give Comec a Certification pass. EASA gave up as well

          What level did the ARJ -21 and the C919 wing fail at? What was the failure point prediction (let alone any other tests including life fatigue)

          Funny to see a Chinese cheerleader for a black program with no visibility (Trust us, its fine) and an open one gets dissed. Yep, trust us, please be sure to ignore all those other fiddly issues. Nothing to see here, keep moving folks.

          Selective Tin Foilism at its best.

    • Failures close to the required limit do not necessarily require a retest. What is required is that the failure reason be understood and the design be modified so that modelling predicts a failure beyond the limit. Then the certifying authority will decide if a second test needs to be performed.

      From your linked article it seems Boeing and the FAA are in the process of investigating the reason for the failure and from that will flow a redesign and possibly a second test.

      • IMU any failure that hugs margins is a useful result.
        ( 147 or 155% for a target of 150% is a rather good outcome.
        The estimators actually knew their thing. and IMHO failing slightly below margins is more helpful than going beyond target without failures.

        Things _unexpectedly_ failing in unexpected places
        are a completely different thing. ( again see the 787 )

      • @ jbeeko
        The catastrophic failure was two years ago at this stage (Sep. 2019)…surely we should have had some form of update / white smoke by now?
        Boeing doesn’t seem to be very good at things involving software, so that doesn’t inspire much confidence in its modeling ability.

        When a cargo door blows out and a gaping rupture appears in the fuselage (not along a joint), one can argue that a physical re-test is very much required.

        • Why would you expect Boing to publicly publish anything about their redesign? What would they gain from putting anything out there? I’d expect them to just get on with it submit designs to the FAA and that would be that.

          As for the extent of the damage, when pressure vessels burst it is usually violently with extensive damage extending well beyond the location of the original failure. Hard to pop a balloon without drama.

          • jbeeko:

            You are spot on in regards to failure and the need for retest.

            In the case of the A380 wing, it failed exactly where they predicted it would, but at 147%. That is some extremely good spot on engineering work.

            Airbus was allowed to beef up that spot and NO retest due to their prediction and how close it was. Right call.

            I believe Boeing will have to retest as they did not predict the failure let alone where and the type.

            But other than someone screaming the sky is falling, as noted, its why we test and no failure in this type of testings has ever passed.

            Anyone who thinks this could be let alone would be covered up and slathered over is living in an alternative facts universe.

            But then we have people who take Horse Parasite medication rather than a proven vaccine.

          • @ jbeeko

            Boeing doesn’t necessarily have to publish details of the re-design.
            But it could certainly inform the public that a re-design had been made, vetted by the FAA, and passed — assuming, of course, that any of that had actually occurred.

            I’d remind you of that Calhoun memo to Boeing employees telling them to be more transparent. A quote from the link below:

            “I see greatness in this company but I also see opportunities to be better. Much better,” Calhoun said.
            “That includes engaging one another and our stakeholders with greater transparency, holding ourselves accountable to the highest standards of safety and quality and incorporating outside-in perspective on what we do and how we do it.”


          • @ jbeeko
            As TW has kindly pointed out, the A380 wing failed “exactly where they predicted it would”.

            In stark contrast, that’s not the case for the 777X: nobody foresaw that the fuselage would be ripped open during a wing load test. Technically, the test was interrupted before completion, simply because the ripped fuselage prevented the test from proceeding along its normal course.

            When a wing load test reveals an unforeseen fuselage weakness, a major engineering problem needs to be addressed.

          • @TW
            It looks like this topic has really touched an open nerve with you.
            You’ve now gone beyond your normal tantrums and are throwing out insults left, right and center. Just in 3 of your remarks we’ve already had “tin foil hat”, “Chinese cheerleader”, “screaming the sky is falling”, “living in an alternative facts universe” and “people who take Horse Parasite medication rather than a proven vaccine”.

            One can only wonder what’s next…

          • TransWorld has been suspended for a week.


    • Bryce,
      All your words and posted links in this thread contain no evidence to support your opinion that Boeing ignored/fudged over the 777X ultimate load test failure. You are making serious accusations here, albeit in a weaselly way. I would think that one would follow accusations of this nature with something substantive, but instead you proffer emotional arguments about the violence associated with the test result, and your perception of broken corporate transparency promises. Pretty flimsy.

      Some facts about what happened.
      The failure occurred only 1% shy of the ultimate load.

      “But as Boeing personnel along with six FAA observers watched from the windows of a control room, at 1.48 times limit load — 99% of ultimate load — the structure gave way. Under the center fuselage, just aft of the wing and the well where the landing gear wheels are stowed, the extreme compression load caused the plane’s aluminum skin to buckle and rupture, according to the person familiar with the details.”

      The failure occurred in a location where it is structurally complicated. It is not just a simple case of over-pressurizing the fuselage to cheat on the buckling results, as Uwe keeps on claiming.

      According to the FAA, Boeing will likely not have to redo the test.

      “A safety engineer at the Federal Aviation Administration (FAA), speaking anonymously without permission from the agency, said that because the blowout happened so close to the target load, it barely counts as a failure.
      Boeing will have so much data gathered on the way to the 99% stage that it can now compare with its computer models to analyze the failure precisely, the FAA engineer said. It can then reinforce the weak area, and prove by analysis that that’s sufficient to cover the extra 1%.
      The engineer said it’s not that unusual to find a vulnerability when taking an airplane structure to the edge of destruction.
      “The good news is they found it and can address it,” the FAA engineer said. “They found a problem they can fix. They can beef up the structure based on analysis.””

      • If you re-read my posts, you’ll see plenty of support for my stance.
        – Airbus issued a public description of its wing fix just 5 months after the failed load test. It’s now 2 years since the failed Boeing test and we’ve heard no update. This is at variance with the Calhoun memo that I cited.
        – I and other commenters have pointed out that the Boeing failure is a serious matter because it occurred at an unexpected location. A wing load test that exposes a serious fuselage weakness issue is not to be taken lightly.
        – I also pointed out that, technically, the test was not performed to completion because the unexpected fuselage rupture interrupted the test.

        I not that you’re interchanging the concept of “opinion” and “accusation” in a convoluted manner. You’re also confusing generic with specific.

        • Bryce,
          I read your posts the first time, and there is nothing in them that supports the notion that the 777X ultimate load test results are being “ignored / fudged over” as you claim.

          – How Airbus chose to deal with their own ultimate load test failure with the A380 has nothing to do with the 777X. These are different companies in different times, in different political situations, running programs in different ways, with different PR departments. Airbus was running a series of load tests on the A380, starting low and gradually ramping up to the ultimate load test. They changed the production design along the way. The quick fix you crow on about was a temporary one, not needed in the production design because the production design was already different than the design that was tested to failure.

          – Just because it hasn’t been reported on, doesn’t mean Boeing is ignoring it, or isn’t working on it. Claiming that Boeing is ignoring or minimizing the problem, or worse, fudging over the the problem is a gross overreaction that is not supported by lack of information.

          – Any test failure during a large commercial aircraft development program is a serious matter. I don’t think anyone here disputes that. The FAA engineer quoted in the article I linked also considered the test a failure, albeit barely. This is because it was very close, 1%, to being a success.

          – How do you, or the other commenters for that matter, know that the failure occurred in an unexpected region? The wing and fuselage were tested together with the nose and tail bent down as the wings were pulled up. It wasn’t just a wing test.

          – Exposing potential weaknesses is the whole point of flight test rigs, or did Bjorn not make that clear in his article.

          – The test was performed to 99% of completion, more than the 97% completion of the A380 ultimate load test. Extrapolating 1% is most likely easier than extrapolating 3%.

          Your opinion that Boeing is ignoring / fudging over the 777X ultimate load test results seems to me to be based more on how you feel about Boeing, and how your feelings color your interpretation of some select, minimally relevant facts, than on the more relevant facts themselves.

          Opinions can contain accusations, by the way.

      • And as another poster has mentioned, a failure at 1% shy of the targeted ultimate load is better than one at 10% over. Failure that close to the target (over or under) implies good modelling and a close to optimal structure.

        Several have pointed out that the failure was at an unexpected location. But that is not a condemnation of the complete design, wing, wing-fuselage interface and fuselage. Let’s assume there are 10 load paths each with a near optimal design. It is going to be very difficult to predict which fails first if they are all designed to fail at 1% over ultimate load. In fact I’m not sure I’ve seen anything indicating a prediction of where it would fail.

        • If the matter is as straightforward and innocuous as some commenters here are suggesting, then the FAA or Boeing should have no problem issuing a formal clarification of the matter…right?

          Some people may be content with a sweeping, generic comment from a single, anonymous FAA employee — formulated in conditional language — but the flying public deserves better than that. After all, we don’t want any suspicion that MAX shortcuts are being repeated, do we?

          • Right, but that doesn’t mean they have to, or even that they should.

            I’ll gladly take the anonymous FAA employee reported on by a reputable, fair reporter with an excellent track record of accuracy over a non-reporter, non-industry insider, non-expert blog commenter any day.

          • Since you attach such weight to “a reputable, fair reporter with an excellent track record of accuracy”, you’ll be delighted to read this article by him (see link below), in which he answers one of your questions above:

            “This time, however, though the wings did not give way; it was one of the doors that failed — an outcome that is definitely not supposed to happen.”

            He also confirms a point that I made, namely that the test was suspended (emphasis added), and thus never completed. Legally, the FAA is treading in a potential mine field if it allows certification despite a suspended test:

            “After the incident was first reported Friday by KOMO News, Boeing confirmed that a serious incident had occurred but offered few details.
            “During final load testing on the 777X static test airplane, the team encountered an issue that required SUSPENSION of the test,” it said in a statement. “The event is under review and the team is working to understand root cause.””


  1. My take is its stunning anyone builds a new aircraft, even a simple one at all.

    The incredible amount of work and detail that goes into getting it aloft is mind boggling.

  2. With current technology, systems integrations rigs can become a full test environment, capable of performing simulated flights, and even be possible to have Level D qualification to be used in certification tests in lieu to flight tests, specially for failures testing. A well conceived and integrated rig can offload lots of flight test hours during development and certification. And guarantees that a very good maturity is achieved for software development.

    • Interesting, ever notice how across all OEMs derivatives have ever more bin space inside the same envelope? For example Airspace has 40% more, over plain neo which had more than ceo.

      Like they are tapping into alternate dimensions.

      • Yes, indeed: it’s absolutely *amazing* what has been done with overhead bin volume. In the process, the shape and hinging structure have changed markedly. Presumably, certain utility conduits have been re-routed in order to accommodate this.

        I have literally seen 6 carry-ons side-by-side in an A350 bin.
        I have also seen passengers bickering over (severe) lack of bin space in an old 767.
        This is one innovation that was badly needed.

        • With all that mass storeable overhead crash worthiness should turn into an intriguing job?

          Q: was there ever a difference in the certification environment between FAA and EASA.
          IMU hinged bin arrangements were mostly found under the FAA regime?

          • Regarding crash-worthiness: now that seat pitches on many airlines have reduced to 28-29 inches, some consumer organizations are lobbying for re-evaluation/testing of evacuation processes.
            Overhead bins fit into that picture: what happens if all that extra overhead baggage falls out onto people and/or blocks the aisles?

  3. Scott’s new book is out…loot at the LNA Twitter feed!

    All regular commenters here get a free copy…just dial 1-800-LEEHAM 😉 🙂

    Presumably, John Leahy (ex-Airbus) will be getting some portrait royalties for that big picture of him on the front cover 😉

    • @Bryce: All regular commentators get charged treble for the book. I need the extra dough to pay for those “Leahy royalties.” (For the literal-minded, I’m just kidding on all counts, of course.)

      • What luck that i did the stealthy “Prime” click on Amazon yesterday.
        I’ll foist my opinion after reception 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *