Siemens Mobility Helps Rail Operators Avoid Downtime and Unplanned Maintenance

QuadX

Rail customer satisfaction will always depend mainly on the answer to one question: Are the trains running on time? Factors that could delay a train’s arrival include what the industry calls “rail defects”—warps, stresses, and fractures in track materials and other infrastructure components. Rail system operators are always looking for efficient ways to find and correct such issues.

"Detecting and proactively repairing rail defects—this is exactly the kind of problem we want to help our customers solve," says Friedrich Gloeckner, who leads the data-services architecture team at Siemens Mobility. A leader in transport solutions for more than 160 years, the company constantly seeks opportunities to innovate the offerings of its core lines of business: rolling stock, rail automation and electrification, turnkey systems, intelligent traffic systems, and related services.

Finding rail defects once required manual assessment by inspectors walking the tracks or reviewing video footage shot from inspection trains, but both approaches were costly, time-consuming, and error-prone. Now, there is a new method: Video Track Inspector. This video analytics application is a joint project of Siemens Mobility and Strukton Rail, a Dutch company specializing in the construction and maintenance of rail infrastructure. The new solution still uses train-mounted HD video cameras, but it replaces manual review with machine-learning algorithms that analyze the images, identify and geolocate defects, and issue work orders.

A valuable option for the rail industry, Video Track Inspector is one of hundreds of applications hosted in Siemens Mobility's Railigent Application Suite, an open ecosystem for rail data integration, monitoring, and analytics applications running on Amazon Web Services (AWS). "We built Railigent to help our customers avoid unscheduled maintenance and achieve up to 100 percent availability of their rolling stock and infrastructure components," says Gloeckner. "Without the modern IT infrastructure and cloud services we access on AWS, it would not be possible to come close to that goal."

“Our AWS data lake enables not only data scientists and software developers but also about 250 non-technical employees to create custom applications and reports that help maximize the value of the data.”

– Friedrich Gloeckner, Data Services Architecture Team Leader, Siemens Mobility


  • About Siemens Mobility
  • A leader in transport solutions for more than 160 years, Siemens Mobility—a separately managed company of Siemens AG—is constantly innovating its portfolio in its core areas of rolling stock, rail automation and electrification, turnkey systems, intelligent traffic systems, and related services.

  • Benefits
    • Cuts maintenance costs and energy consumption by 10–15%
    • Decreases unplanned downtime by 30–50%
    • Reduces unnecessary transfers to maintenance by more than 30%
    • Open ecosystem enables third-party applications from leading rail specialists
  • AWS Services Used

Bringing Trains into the Cloud

Gloeckner speaks from experience. Siemens Mobility originally deployed an on-premises solution with analytics capabilities similar to Railigent’s but encountered obstacles related to siloed data, labor-intensive data integration and development processes, and slow time-to-market.

According to Gloeckner, one of the most attractive reasons for developing and implementing Railigent on AWS was the opportunity to centralize data. As part of Railigent's new cloud architecture, Siemens Mobility implemented a data lake that uses Amazon Simple Storage Service (Amazon S3) for persistent dataset catalogs, AWS Glue for data transformation, and Amazon Athena for serverless, interactive querying. It also relies on AWS Lambda for serverless orchestration functions and Amazon Elastic MapReduce (Amazon EMR) for fast, cost-effective processing and analysis of unlimited amounts of data.

"For us, a major appeal of AWS is access to services like Amazon EMR, which gives us the ability to run Hadoop clusters of any kind, at any scale, on demand, with pay-as-you-go pricing," says Gloeckner. "Offloading this kind of operations work was an absolute requirement on our end, because we had experienced how complicated it was to run, update, and scale complex solutions like Hadoop in our own data center."

Running on AWS also helps optimize use of the disparate IoT data that Railigent ingests from hundreds of thousands of sensors and other devices on tens of thousands of rail assets worldwide. "In our AWS data lake, we can store large unstructured datasets in Amazon S3 and use the Amazon Athena schema-on-read capability to create virtual tables for specific new use cases as needed," says Gloeckner. "Cloud services like Amazon EMR, Amazon S3, and Amazon Athena give us much more flexibility in dealing with data than would be possible on premises or even with other public cloud providers."

Democratizing Data to Find New Customer Value on AWS

None of these capabilities would have mattered, of course, if they weren’t also helping Siemens Mobility respond more quickly to customer needs and find more value in ingested data. "When Railigent´s precursor was running on premises, data from different sources was siloed, so building apps to make use of that data required complex custom extract, transform, and load [ETL] jobs and assistance from analytics experts," says Gloeckner. "This made it difficult to take full advantage of all our data, and we couldn't easily offer toolboxes and reusable app components to our developers."

Now, the company can centralize data preparation tasks while enabling a broad range of teams to make use of the data. "On AWS, we use a small, centralized team for data ETL, structuring, and enrichment, and then we make the data available for even non-technical employees to experiment with and build on," says Gloeckner. "Our AWS data lake enables not only data scientists and software developers but also about 250 non-technical employees to create custom applications and reports that help maximize the value of the data. This democratization of data is one of the most important benefits of our AWS data lake."

This data democratization enables faster responses to customer requests, for example, by cutting report-generation time in half. "Before we moved to AWS, we had to rethink authentication, authorization, ingestion, and ETL for each custom BI report—and even then, we could offer only snapshots, not live results," says Gloeckner. "Now, on AWS, these problems have been solved once and checked against global Siemens security and governance rules, so these components can be reused by report developers. With our AWS data lake, we only need two to three weeks to create reports running against live data, as opposed to the month or more we once needed to create even a static report.”

According to Gloeckner, on average, Railigent customers are already seeing 10–15 percent reductions in maintenance costs and energy consumption, 30–50 percent decreases in unplanned maintenance, and a more than 30 percent drop in unnecessary transfers to maintenance. And the company is only beginning to explore what’s possible on AWS. "We’re happy to let AWS handle undifferentiated heavy lifting like operating infrastructure and building services while we focus on what really matters for our business. The beauty of running on AWS is the many possibilities it opens up. Really, we’ve only scratched the surface of what we’re going to be able to achieve in the cloud."