Big Data and Artificial Intelligence: ALTEN’s data solution to manage aircraft part breakage

1st December 2019

Two major challenges for aircraft manufacturers are maintenance and issues resulting from the breakage of aircraft parts during assembly. Airbus began digging deeper in 2016 with the goal of reducing costs.
The Big Data lab, at ALTEN Digilab in Toulouse, leveraged data science to come up with a solution to predict the cost of breakage during aircraft cabin assembly.

With regard to the specific case of breakage during aircraft cabin assembly, the investigations ALTEN completed for Airbus revealed that the costs associated with breakage were not weighted. So, a lost screw was deemed to have the same cost as a broken armrest.

ALTEN data scientists recovered the logs for the past five years and correlated problems with a history of up to two years, analyzing actual data, including data from invoices. Based on the data, they created algorithms to predict risk and average costs in real time. With a 90% accuracy rate for the costs predicted, the Big Data team was able to determine the cost of breakage in real time. The predictions include financial weighting, ranking of root causes, and prioritization of problems. Once this step was completed, production scheduling data just needed to be added. It was then possible to make predictions for the next six months.

The BIG DATA lab, a “startup” designed to stay ahead of customer’s needs

ALTEN’s Big Data lab is growing very rapidly. It opened in January 2018 and already has more than 55 data scientists working on three major dataprocessing- related topics: display and visualization; algorithms; and prediction.

Digilab’s Big Data activities are set up like a startup or lab. One of the things that makes Digilab unique is that it invests in research and development on a topic before a specific customer’s request comes in. In other words, Digilab is creating customer’s needs by recommending the solutions it develops to customers who do actually need them. This also means that certain PoCs * go right to scale-up.

The Big Data lab is currently working on more than 400 projects, both in-house and on customer’s topics. The lab’s largest customer is Airbus, which represents 90% of its activity. However, it is diversifying into the automotive, food manufacturing, and defense & space markets. The lab is bringing in engineers from industries other than aeronautics to support its efforts to diversify.

* Proof of Concept: an experiment designed to demonstrate the feasibility of a method or concept.