© Gallerystock

Big Data

There’s gold in those mountains of data!

11/2015 - Never before has so much information been available as today. Anyone prepared to scale these mountains of data can discover real treasures.

Never before in human history has there been such a vast quantity of accumulated data as is available today. Whereas predictions of the future were once no more accurate than reading tea leaves, they are actually feasible today by making comparisons between vast quantities of data – albeit within certain limits. For commercial enterprises in particular, big data analysis opens up opportunities that could once only be dreamed of – and which now offer real benefits.

A new eye for connections

Analysis of this endless tide of data suddenly reveals connections that would once have been undetectable. Today, for example, it is possible to predict with a high degree of accuracy when certain parts of a machine will wear out and require replacement – without having to inspect them thoroughly. For example, as the mechanical linkage of points begins to wear, the power consumption of the switch motor increases slightly. It is then possible to plan the necessary replacement in advance – rather than having to wait for a fault to occur before being able to react. This really saves money and helps to reduce delays.

The mountain of data keeps growing

Noch nie gab es so viele Informationen wie heute. Wer es wagt, sich den Datenmassen zu stellen, kann Schätze entdecken.
© Fotolia.de / Mikko Lemola

In future such forecasts may become even more accurate. But big data only becomes really exciting as the mountain of data grows. Whereas most data today is still gathered manually, an increasing amount will be acquired automatically in future. From machine to machine. Thanks to continuously falling prices, a growing number of specialised sensors can be installed. These automatically send their recordings back to base over a dedicated mobile radio connection. And so they add to the growing mountain, from which real treasures can then can be recovered. Whereas in the past we relied on a few observations and assumptions, these assumed relationships can actually be proven today. And new ones discovered.

Endless possibilities

The possible applications are extremely varied. Because it goes without saying that big data can be utilized not merely for repairs. It is conceivable, for example, that automatic predictions could be made about the number of passengers expected to visit a station at a particular time, based on the number of tickets booked, the day of the week and the time of day. And taking into account, of course, whether it is a holiday period or public holiday. The information obtained in this way may be relevant for the gastronomic and other retail outlets in the station. If a packed train, due to arrive at lunchtime, is delayed for ten minutes, then the restaurant can wait a little longer before putting the sausages on the grill! For station security on the other hand, a calculation of possible risk periods is of interest, for example, where the arrival of a special train with potentially violent football fans is concerned. Possible applications for big data can also be identified for quite different purposes.

A job for specialists

But this job cannot be done completely automatically. It may be true that modern computers can process mountains of data and detect relationships between the individual components. Whether these correlations make sense, however, can only be determined by a specialist – and this spawns new job profiles. Anyone, for example, who evaluates big data analyses for locomotive maintenance not only requires detailed knowledge about the function of the machines, but also about statistical data acquisition: the “data scientist” has been born.

Not merely a pipe-dream

The evaluation of data from individual sensors located in the field has long been a reality at Deutsche Bahn – DB Netz (network division) with its DIANA system, for example, monitors changes in throwing force in points motors. New rail grinding vehicles operated by DB Netz deliver vast quantities of sensor data that only need to be evaluated in order to prevent failures and optimise their maintenance. DB Schenker Rail is working with the support of Digital Vehicle Solutions on the forecasting of failure for its vehicles. In this respect, DB Systel can pool the new technology and its expertise in existing business processes within DB to practical effect. The solutions developed jointly with its partners, therefore, do not bypass the operational business and are instead integrated into the actual existing processes and networked with the existing IT systems. And it will not stop here: in numerous future projects, the digitisation campaign within the Group is set to open up the goldmines hidden in the mountains of data.