Inflight Dual IRU Failure on Boeing 787

[MolaKule]

Apparently, in January of this year, there were dual Inertial Reference Unit (IRU) failures on a Boeing 787 (UA613) that disconnected the autopilot, and for 12 minutes a series of flight "excursions" occurred.

https://aerospaceglobalnews.com/news/iru-failures-led-to-united-787-upset-over-cote-divoire-ntsb/

A dual failure of this nature is unusual. The NTSB must still be investigating since a cursory search of the NTSB Aviation database didn't turn up any reports.

 

[Astro14]

It is still under investigation.

 

[AZjeff]

What is a flight excursion? Roller coaster ride?

@Astro14, the article mentions loss of voice recording because of 2 hour recording capacity. This is mentioned regularly it seems. Is the voice recording critical or are the black boxes enough data? How hard would it be to increase recording memory?

 

[Astro14]

What is a flight excursion? Roller coaster ride?

@Astro14, the article mentions loss of voice recording because of 2 hour recording capacity. This is mentioned regularly it seems. Is the voice recording critical or are the black boxes enough data? How hard would it be to increase recording memory?

Two hours is the recording capacity on newer aircraft. It used to be 30 minutes. I don’t know what the requirement is, but if you want airlines and manufacturers to spend significantly more money on something, you kinda have to make it a requirement.

The real question is what value would the voice recorder add?

This is a crew that came back, you can debrief them, on their actions, their thought process, so what they said, is frankly insignificant compared with that data.

The voice recorder is really there to try and gain insight into the crew’s thought process and actions, when the outcome results in a fatal crash.

So, sure, we could mandate improvements in black boxes, and force airlines to spend millions and millions of dollars to upgrade their fleets, but to what end? Would it improve the accident investigation in this case?

There are better things to spend money on.

 

[mk378]

Despite attempts to re-engage the autopilot, the aircraft continued to experience altitude deviations,

That sounds more like "due to attempts to re-engage the autopilot," If the pilots had decided to fly it manually would there have been a problem staying straight and level with the remaining instruments?

 

[Astro14]

That sounds more like "due to attempts to re-engage the autopilot," If the pilots had decided to fly it manually would there have been a problem staying straight and level with the remaining instruments?

There is a lot more to flying than “straight and level” - most of the world is now RVSM, which mandates the use of autopilot and has very strict compliance requirements for altitude deviation. An autopilot failure requires leaving RVSM, which would greatly slow the flight, increasing fuel burn and requiring a landing short of destination.

So, the crew needed to attempt the autopilot re-engagement. It was required. To fly for a dozen hours, to that RVSM standard, by hand, is a lot like saying - “why do you need a circular saw? You could frame that house with a hand saw.”

https://www.faa.gov/air_traffic/separation_standards/rvsm

 

[IndyIan]

So the plane on autopilot started going up and down violently, and then the pilots try to re-engage it? Which caused more "deviations in altitude"? Seems odd to me?
I do know I'd rather use a handsaw for a day, than use a circular saw that was giving my hand the odd 110V tingle up while I was up on a ladder...

 

[Astro14]

So the plane on autopilot started going up and down violently, and then the pilots try to re-engage it? Which caused more "deviations in altitude"? Seems odd to me?
I do know I'd rather use a handsaw for a day, than use a circular saw that was giving my hand the odd 110V tingle up while I was up on a ladder...

I wasn't there - so, I don't know. I don't know what the Boeing procedures are - they may require an autopilot re-engagement attempt. They may not. Maintenance may have said it was good - until it wasn't. But re-engaging it may have been a matter of requirement, not just convenience. And "a day" is a lot shorter than most 787 flights.

 

[SC Maintenance]

Back when I did more process controls we did a lot more redundant systems. Everything can fail. Having two units fail within an hour is highly unusual but not impossible. Given the number of these systems in existence might very well just be random occurrence you never see again?

 

[Exhaustgases]

What is a flight excursion? Roller coaster ride?

@Astro14, the article mentions loss of voice recording because of 2 hour recording capacity. This is mentioned regularly it seems. Is the voice recording critical or are the black boxes enough data? How hard would it be to increase recording memory?

Maybe $200. per plane, memory is dirt cheap now as compared to say the early to mid 90's. No excuses not to.

 

[goblin]

Maybe $200. per plane, memory is dirt cheap now...

Unless there's hard data on part numbers and supply chains - I call utter BS on that.

Even if it's identical to off the shelf memory (which it probably won't be) - it will have one or two zeroes added to the off the shelf cost because of the testing and certification required.

If anything - the next generation of black boxes could have x10 or x100 the recording capacity of the current gen and not be more expensive. It's just that they are not certified yet. And the day they are - they'll go in newly built planes only, and will still be electronically generations behind whatever memory will be new then. These things always are.

Decades ago NASA had published cost lists of equipment used in their latest probes vs the cost during the moonshot. The communications part was epic: Dozens of millions when it was bespoke-built equipment during the moonshot, then modified Motorola Marine modelms for these probes.

The Motorola Marine modems were $1k-ish versions of a $12 modem, certified for marine use. The ones used in the probes were Marine versions "adapted for use in space". At a cost of several hundred thousands.

Still orders of magnitude less expensive, but by no means comparable to off the shelf units. I could be wrong on the exact numbers, t'was a while back. But the orders of magnitude I believe I remember correctly.

 

[SC Maintenance]

Unless there's hard data on part numbers and supply chains - I call utter BS on that.

Even if it's identical to off the shelf memory (which it probably won't) - it will have one or two zeroes added to the off the shelf cost because of the testing and certification required.

If anything - the next generation of black boxes could have x10 or x100 the recording capacity of the current gen and not be more expensive. It's just that they are not certified yet. And the day they are - they'll go in newly built planes only, and will still be electronically generations behind whatever memory will be new then. These things always are.

Decades ago NASA had published cost lists of equipment used in their latest probes vs the cost during the moonshot. The communications part was epic: Dozens of millions when it was bespoke-built equipment during the moonshot, then modified Motorola Marine modelms for these probes.

The Motorola Marine modems were $1k-ish versions of a $12 modem, certified for marine use. The ones used in the probes were Marine versions "adapted for use in space". At a cost of several hundred thousands.

Still orders of magnitude less expensive, but by no means comparable to off the shelf units. I could be wrong on the exact numbers, t'was a while back. But the orders of magnitude I believe I remember correctly.

Your not wrong in general, but space (complete vacuum) is quite different than even an airplane, primarily due to off-gas of electronics and heat dissipation (there is no medium to move the heat away). I doubt it was a Motorola Marine Modem. It was likely a Motorola Marine Modem made completely of special components that don't off-gas and produce less heat.

You are right however - its typically the testing and standards are the expense, then of course you have to do all the qualifications up front then wait years for an actual order.

 

[goblin]

Indeed, I was insisting on the testing and certification.
The Motorola thing - I'm trying to find that article, as it was quite interesting. It was definitely Motorola Marine. The details however I can't remember. All I remember was - an off the shelf (if quite specific) part that cost X, and a "adapt it for space use" cost that was possibly 50 or 100x.

 

[SC Maintenance]

Indeed, I was insisting on the testing and certification.
The Motorola thing - I'm trying to find that article, as it was quite interesting. It was definitely Motorola Marine. The details however I can't remember. All I remember was - an off the shelf (if quite specific) part that cost X, and a "adapt it for space use" cost that was possibly 50 or 100x.

100X is cheap. The company I work for redesigned some instrumentation decades ago for the military - high heat, high shock, high vibration. Its basically a complete redesign, with very special components, and a lot of testing (you have to re-think everything), and then when your done you have to re-coop your cost on a very low volume. So if it costs $1M to make a marine radio, and you sell 1 million, then each radio can contribute $1 to the R&D offset.

If you then make one for NASA, or Boeing, or whomever, and it costs $1M in R&D, but they only need 10 units.....

Anyway was before my time and I was told we will never do that again. We also did something for vacuum. Have been told we will never do that again either.

 

[Exhaustgases]

Unless there's hard data on part numbers and supply chains - I call utter BS on that.

Even if it's identical to off the shelf memory (which it probably won't be) - it will have one or two zeroes added to the off the shelf cost because of the testing and certification required.

If anything - the next generation of black boxes could have x10 or x100 the recording capacity of the current gen and not be more expensive. It's just that they are not certified yet. And the day they are - they'll go in newly built planes only, and will still be electronically generations behind whatever memory will be new then. These things always are.

Decades ago NASA had published cost lists of equipment used in their latest probes vs the cost during the moonshot. The communications part was epic: Dozens of millions when it was bespoke-built equipment during the moonshot, then modified Motorola Marine modelms for these probes.

The Motorola Marine modems were $1k-ish versions of a $12 modem, certified for marine use. The ones used in the probes were Marine versions "adapted for use in space". At a cost of several hundred thousands.

Still orders of magnitude less expensive, but by no means comparable to off the shelf units. I could be wrong on the exact numbers, t'was a while back. But the orders of magnitude I believe I remember correctly.

Then don't integrate it into the black box, if as is the black box is sufficient leave it alone. Build a cheap inexpensive cockpit voice recorder that records say 24 or more hours that is meant for incidents that it can survive. Many times the lines to the black box is severed from things like fires etc. anyway. And since its not of huge importance then low cost components can be used, and it should not have an on and off switch, so nothing can be hidden from it.
Since its of low importance no need for critical certifications or anything. They could probably get them cheap from amazon.

 

[goblin]

...
Since its of low importance no need for critical certifications or anything...

I don't think a pilot would be prohibited from carrying a personal voice recorder that could record for weeks. But if it's part of the plane - there's no such thing as no need for certification, or low importance.
In aviation more than elsewhere, all rules and regulations are written in blood.