Cyril Verdun, Director of the Equipment Engineering Cluster for SNCF Mobilités, has engaged with his teams on the track (no pun intended), predictive maintenance. He will speak at Big Data Paris, during a conference dedicated to the optimization of maintenance processes within SNCF
Ultimately, it is not expected to find a player such as SNCF at the forefront of data innovation. In what context are they exploited? CV: For about 4 years, we have tried to take advantage of this data that our trains are able to generate. It should be known that a modern train is able to raise nearly 70,000 data per month. So many valuable indications if we know how to analyze them. And that’s the whole point. This information is only valuable if we know how to exploit it! As a result, we have had to reinvent our whole value chain, from the collection of information to the homogenization of data and the decision-making process.
The data must allow us to be more efficient so that our equipment is always more reliable. The data we have come from several sources, in the remote diagnosis or in a CBM dimension (Condition-based maintenance) where weak signals are used to detect anomalies before they have too great consequences. Our 15 data scientists are working hand-in-hand with maintenance operators to understand, decrypt, analyze this information and transform it into an order of interventions transmitted intelligently to the right operators at the right time. Our approach meets a key objective: to reduce the unavailability of trains, to improve reliability and to allow operators to get to the basics.
Value chains, Data Scientists, MRO, are terms that are very far from Zola’s Human Beast. It’s a true start-up spirit that drives you?
CV: It is certain that we are far from the images of Epinal! SNCF maintenance operators are now more often equipped with a touchpad and a laptop than heavy wrenches and oil drums. Recent rains have a very high level of technicality. For information, we find among the Parisians, Regiolis, or TGV including, nearly 1.2 km of electronic cables per meter … A train can be up to 160 meters long … Pure mechanics, if it exists, is less present than in the past. So indeed, this technicality, the valuation of all these data that our trains send us daily, it forces us to constantly rethink our methods, in a perpetual quest for innovation, improvement.
As you observed, there is a true start-up spirit, including among operators, who are the first to get excited about this new way of designing maintenance! What are the reasons that prompted you to speak at the Paris Big Data Paris 2018? CV: We want to clearly show that Big Data is already a reality at SNCF. We are truly at the forefront of predictive maintenance, in Europe and around the world. We just want to let it know. We innovate constantly, in a very pragmatic approach that has a direct influence on the daily lives of travelers. This makes our mission very exciting. Our testimony is also a way of highlighting the commitment of all our teams, from engineers to maintenance operators, in the use of data.