Subscribe to DSC Newsletter

That’s Data Science: Airbus Puts 10,000 Sensors in Every Single Wing!

In a meeting with Airbus last week I found out that their forthcoming A380-1000 – the supersized airliner capable of carrying up to 1,000 passengers – will be equipped with 10,000 sensors in each wing.

The current A350 model has a total of close to 6,000 sensors across the entire plane and generates 2.5 Tb of data per day, while the newer model – expected to take to the skies in 2020 – will capture more than triple that amount.

In an industry as driven by technology as the aviation industry, it’s hardly surprising that every element of an aircraft’s performance is being monitored for the potential to make adjustments which could save millions on fuel bills and, more importantly, save lives by improving safety.

So I thought this would be a good opportunity to explore how the aviation industry, just like every other industry, is putting data science to work.

There are 5,000 commercial aircraft in the sky at any one time over the US alone, and 35 million departures each year. In other words the aviation industry is big. And given that every single passenger on each of those flights is putting their life in the hands of not just the pilot, but the technology, the safety measures and regulations in place are extremely complex.

This means that the data it generates is big, and complex too. But airlines have discovered that with the right analytical systems, it can be used to eliminate inefficiencies due to redundancy, predict routes their passengers are likely to need, and improve safety.

Engines are equipped with sensors capturing details of every aspect of their operation, meaning that the impact of humidity, air pressure and temperature can be assessed more accurately. It is far cheaper for a company to be able to predict when a part will fail and have a replacement ready, than to wait for it to fail and take the equipment offline until repairs can be completed.

In fact, Aviation Today reported that it can often take airlines up to six months to source a replacement part, due to inefficient prediction of failures leading to a massive backlog with manufacturers.

On top of this fuel usage can be economized by ensuring engines are always running at optimal efficiency. This not only cuts fuel costs but minimizes environmentally damaging emissions.

In the case of Airbus, they partnered with IBM to develop their own Smarter Fuel system, specifically to target this area of their operation with Big Data and analytics.

Additionally, airlines closely monitor arrival and departure data, correlating it with weather and related data to predict when delays or cancellations are likely – meaning alternative arrangements can be made to get their passengers where they need to be.

Before they even take off, taxi times between the departure gates and runways is also recorded and analyzed, allowing airlines and airport operators to further optimize operational efficiency – meaning less delays and less unhappy passengers.

This sort of predictive analysis is common across all areas of industry but is particularly valuable in commercial aviation, where delays of a few hours can cost companies millions in rearrangements, backup services and lost business (The FAA estimates that delayed flights cost the US aviation industry $22 million per year).  

Specialist service providers have already cropped up – masFlight is one – aiming to help airlines and airports make the most of the data they have available to them.

They aggregate data sets including weather information, departure times, radar flight data and submitted flight plans, monitoring 100,000 flights every day, to enable operators to more efficiently plan and deliver their services.

In marketing, too, airlines are beginning to follow the lead of companies such as Amazon by collecting data on their customers, monitoring everything from customer feedback to how they behave when visiting websites to make bookings.

Now we are used to generating and presenting tickets and boarding cards through our smartphones, more information about our journey through the airport, from the time we enter to the time we board our flight can also be tracked. This is useful both to airport operators, managing the flow of people through their facilities, and to airlines who will gather more information on who we are and how we behave.

So businesses in the aviation industry, including Airbus, are making significant steps towards using data to cut waste, improve safety and enhance the customer experience. 10,000 sensors in one wing may sound excessive but with so much at stake – both in terms of profits and human lives – it’s reassuring that nothing will be overlooked.

I hope you found this post interesting. I am always keen to hear your views on the topic and invite you to comment with any thoughts you might have.

About : Bernard Marr is a globally recognized expert in analytics and big data. He helps companies manage, measure, analyze and improve performance using data.

His new book is: Big Data: Using Smart Big Data, Analytics and Metrics To Make Bette... You can read a free sample chapter here.

Views: 15957

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Michael Wm. Denis on August 10, 2015 at 7:39am

Directionally correct and interesting article, although there are a few inaccuracies that #AvGeeks would point out.

Aviation has been doing a lot of what other industries call Industrial IoT - and melding OT/OD with IT/ID (traditional enterprise information data) has been a main stay of our industry for decades.

Aviation data fits all four Vs of BigData - and especially making things challenging is not as much around Volume or Velocity but Variety (lots of documentation content in multiple formats from paper, raster files, SGML, XML DTD, XML XSD, ...) which leads into a Veracity challenge.

The new area of investment for aviation is predictive maintenance to include predictive analytics, case based reasoning diagnostics, Bayesian prescription, prognostics, health management and autonomics. 

Autonomics = Sense & Respond in a automated and autonomous optimized system.

Comment by Anto Franklin Christuraj on May 6, 2015 at 12:32am

Very insightful article on increased use of big data in aviation industry. Although I am not from Aviation industry background, I found this article simple and very explanatory. This would give readers an idea of how data analytics and predictions could be used to optimize both time effeciency and safety.Thanks for sharing it with us Bernard!

Comment by Sione Palu on April 13, 2015 at 2:43am

Realtime sensor monitoring of aircraft performance is not new. It has been going on for years in the last 30 years or so, but  we hear  more often about them now because it has always been a sole domain of engineering, not data science. The purpose of real-time monitoring is mainly for aircraft performance design.  Control System Engineering students have largely done simulations of aircraft flight dynamics by using popular industry Matlab software, with its various toolboxes for Simulink, Control System, Aircraft Control, Signal Processing, Robust Control, Even final year students get to design flight control of an F-14 tomcat fighter.  The simulink model for this design exercise is bundled in the Simulink matlab itself, but students can add-on their own model block (either linear or non-linear) to see if they can improve upon the original design, like the one below.

http://robot2.disp.uniroma2.it/astolfi/CAastolfi_1314/Matlab/sldemo...

Its only a simulation but the data that's generated out of this is quite massive. Students have to retrain their model on this massive dataset to tune the performance of their aircraft controller models be it nonlinear NARX or linear ARMAX. There are a few  textbooks that had been prescribed over 15 years but "Aircraft System Identification: Theory and Practice" has been the most popular in recent years because it comes with its own Matlab toolbox in addition to the MathWorks control system toolboxes (http://au.mathworks.com/support/books/book48721.html?category=13). The generated simulated data is actually what's happen in realtime sensor monitoring of aircraft (not simulation).

Follow Us

Videos

  • Add Videos
  • View All

Resources

© 2017   Data Science Central   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service