When I hear popular culture terms, I tend to borrow the
lingo so I don’t seem remote and out of touch. A term I’ve heard bandied-about
is ‘TMI’, or Too Much Information. ‘TMI’ might be uttered by you following a
conversation with an acquaintance who shared information that was socially
awkward, over-the-top, or just plain overwhelming. But this is the age of
information, which reportedly is the new gold standard. If so, when is
information too much?
Having information or data does not necessarily make you smart;
any more than having a big collection of books in your home or office library
makes you well-read. Data becomes gold
when we put it to work for us. And so it is with the big data being
generated by today’s aircraft.
So Royboy, TMI? What
TMI? Consider the following:
“An A320 generates information from about 20,000 data sources, Airbus
says, compared to 200,000 for an A380 and more than 400,000 for an A350.” 1
“Take an average jet engine from a commercial airliner, which can be
equipped with over 5,000 individual sensors, as an example. According to SAP
business software, these engines can generate around 20 terabytes of data per hour.”2
And there you have it; that
is big data! Let’s develop this; there are many dots to connect.
Here are the topics to be discussed:
- Possible Effects Upon The Aftermarket
- Broad Acceptance Has Been Slow
SOURCES OF THE DATA:
In order to gain a proper perspective on the evolution and
typical sources of aircraft data, I encourage you to read a previous blog: “AUTONOMIC
LOGISTICS”, here’s the link:
Various OEMs are going to call these uses by any number of
names, but generically we can refer to the following:
· Aircraft/Engine/Helicopter Health Monitoring:
Simply, the data being generated is telling us something has failed.
· Predictive Maintenance: Based on trending data,
something, if left unaddressed, may fail in the near future.
· Analytics: This emerging field is software based
and designed to sift through the reams of data and perform the required
analysis. The desired result is that, in the parlance of the CIA and NSA, you
get ‘actionable intelligence’.
Here are the existing and developing uses of all this data:
resolution of write-ups/pireps/unscheduled, non-routine maintenance. While
in flight, data is transmitted to the ground apprising Maintenance Control that
a failure has occurred. Labor, tools, and parts are dispatched or scheduled,
and ready sooner.
and cancellations: a measureable percentage of these are caused by
maintenance issues. With Predictive Maintenance (driven by the big data), the
theory is that some maintenance events will be addressed prior to the hard
failure; this has great potential to mitigate delays and cancellations due to
maintenance. By the way, the biggest cause of delays and cancellations is
weather related, unless of course there are labor problems occurring, but don’t
get me started on that one!
Monitoring reliability of the various aircraft systems indexed by ATA Chapter
and then the respective LRU is critical to running an efficient, productive, on-time
operation. In addition, airframe and engine OEMs choose suppliers based on
guaranteed on-wing reliability performance. The data to substantiate such
performance comes increasingly from these systems.
Debates about units under warranty going to the shop and resultant NFFs, No
Fault Found (meaning the airline is going to be charged for the service which,
if there had been a failure, would have been covered by the warranty agreement)
are rife in the industry. But what if collected aircraft data clearly pointed
to a certain LRU failure (which is under warranty), which then went to the shop
but the shop states its NFF? Will the shop still honor the no charge warranty? Hmmm…fluttering of the eyebrows please…
analysis of Rogue or Chronic components and/or aircraft. For an in-depth
introduction to this topic, see my blog on the issue at:
of MSG-3 programs. Many airlines model their maintenance programs on the
MSG-3 standard, which is driven by data. Moves are afoot to make the big data
generated by aircraft systems an authoritative source of such data.
WHO WILL MINE THE
Right now the big airframers offer fee-based programs for
operators. The engine OEMs are very strong in similar programs, and perhaps
history will write that they indeed pioneered the evolving techniques.
Following the airframe and engine OEMs are the OEMs of the systems and LRUs,
particularly since the aforementioned issues of reliability and warranty are in
play; they are highly motivated to mine the data for those reasons, and to perform
what I call ‘Targeted Analytics’.
How about MROs? Many airlines outsource a significant
portion of their maintenance programs to MROs. This ranges from line
maintenance, to component maintenance, to full airframe and engine overhauls. Will
the airline give big data access to the MRO? Will the airline accept the MRO fees
for acting on ‘actionable intelligence’ derived from management and analysis of
POSSIBLE EFFECTS UPON
Currently, other than the aforementioned OEM use of the
data, there does not appear to be direct aftermarket use of the data; indeed
its use is still evolving. In the near future however, there will be a
trickle-down effect upon the accuracy of the following:
parts: The number and types of parts pooled will increasingly be driven by
of spares: This too will be gradually derived from the data.
- Power By
the Hour: The performance data used (for example MTBR or MTBF) as the basis
to calculate PBH will be based upon and adjusted on this new source of data.
programs: The use or non-use of exchange programs will progressively be
based upon and adjusted by the analysis of spares tied to the data.
In the blog “AUTONOMIC LOGISTICS” mentioned earlier, there is a critical
introduction into the Lockheed F-35’s Autonomic Logistics Information System
(ALIS). In Royboy’s opinion:
ALIS is going to be the harbinger
and standard-setting system by which such big data integrates into logistics
and aftermarket support systems. What works and doesn’t work is going to have a
major trickle-down effect on nascent civilian programs.
Algorithms reflected in software offerings promise to systematically sift
through the big data and provide meaningful analysis. These systems continue to
Intelligence, AI: Analytics will help us arrive at a given conclusion
derived from a given set of data. Then what? The conclusion should be
actionable; but who’s going to follow through? Enter AI. We are on the cusp of seeing Analytics coupled to AI. A
simple possible example: Big data is driving analytics to conclude that an
in-service engine’s Exhaust Gas Temperature (EGT) is starting to climb toward
upper limits and may fail in the near future. AI steps in, analyzes the
aircraft routing, availability of tools and skilled labor at the available
maintenance stations, and issues a work order to perform borescope inspections
at a suitable overnight station with a spare engine in stock; all without human
intervention. We’re not there yet.
data: ATA/A4A Spec 2000 chapter 11 already provides a standardized means
for airlines to share and exchange reliability and performance data. According
to their website, the purpose and outcome of this reporting is the following:
Assist operators and manufacturers to attain and
maintain higher reliability through trend monitoring
Evaluate if certain problem areas are unique to
an operator or observable throughout the industry
Monitor which modification has a better payback
by comparing removal and failure rates of operators who have incorporated
various service bulletins/modifications
Determine utilization rates through aircraft
flight hours, flight lengths and number of landing cycles
Assist operators and manufacturers in
determining the effectiveness of aircraft maintenance programs
There is no doubt that the new big
data will progressively become the source of such information sharing systems.
BROAD ACCEPTANCE HAS
Finally, it seems that broad acceptance and implementation
of the benefits of aircraft big data has been slower than anticipated1.
This is curious since OEMs continue to market their data-driven support systems
as producing maintenance savings of approximately 10-15%. How about:
“…with 10 times return on investment for aviation companies and a
potential 70 to 75 percent reduction in airplane breakdowns, there is a clear
case for automated predictive maintenance in the aviation industry.”2
but reluctance may be driven by the following:
- OEM Support Systems driven by this big data are
fee based. I suspect current cost-benefit math is not a slam-dunk selling
point…yet. After all, there is the
cost-benefit analysis performed by the seller, and one performed by the buyer. Is there a gap to be closed here?
- It seems there is lingering suspicion that using
the data will require a small army of new employee/analysts to monitor the
data…yet another system to monitor and
dedicate labor to; more overhead (perish
- Lingering suspicion of ‘false positives’. For
example, a system used by flight crews (for example, a navigation system) is
reported by big data as having failed in flight, but the aircrew did not write
it up. Is it actionable? Did I just increase my mx costs by servicing “nuisance
Over ‘n out
but Slowly; Aviation Week & Space Technology; April 11-24, 2016; Page