19-year-old Engineer and Entrepreneur Invents New Technologies That Change the World
That headline could be about any 21st century tech startup founder, except it’s actually referring to William Siemens, a German wunderkind who launched his career in the mid-1800s. He started off by creating a new electroplating process, but his innovations eventually included the first automatic dial telegraph, the first alarm bell system to warn railway workers of approaching trains, devices for measuring voltage and resistance, and building the world’s first electric train.
More than 140 years later, Siemens’s legacy is the Siemens corporation, a global leader in engineering and manufacturing, with a rail transportation company — Siemens Mobility — that is dedicated to delivering sustainable, smart rail transport. The company works with more than 3,000 partners, and its customers include Eurostar and Trans Pennine Express. In 2025, Siemens hi-tech trains will begin running on the Piccadilly line in the London Underground.
In the U.K., Siemens Mobility maintains about 500 train units and logs 65 million passenger miles per year. To keep that up, its rolling stock (train cars) and the services the company provides not only need to be available for customers (the operators), but also work reliably and efficiently for riders as well. Martin Klimmek, Head of Digital Development and Operations at Siemens Mobility UK, runs a team that supports about 10 sub-businesses that maintain and service rolling stock fleets for their customers.
“Our mission,” says Klimmek, “is to make our rolling stock data and our maintenance data, maintenance logs, and data from trackside monitoring equipment as telling and as powerful as possible for colleagues who are helping to deliver those fleets every day — but also to create opportunities for external customers.”
They focus on building, integrating, and maintaining digital services that enable and support the required availability, reliability, lifecycle cost, and passenger experience needs of its customers. “Our ultimate aim,” he says, “is for the best possible passenger journey experience.”
In July 2022, Siemens announced the launch of its company-wide Xcelerator, “an open digital business platform featuring a curated portfolio of IoT-enabled hardware and software, a powerful ecosystem of partners, and a marketplace.” It combines open data and portfolio-based applications and is a place where customers’ development teams can collaborate with third-party partners and suppliers. According to Simon Rennie, head of digital at Siemens Mobility, the company’s overall ambition is to develop new cost-effective and scalable tool sets and partners “that can bring in data, then unlock its usefulness and richness to make it accessible to a wide cross-section of our own teams — many of whom have not worked with this data before.”
Data, after all, is at the core of Siemens Mobility, where IoT is often considered a shorthand for “the Internet of Trains.” The company considers its trains “data centers on wheels,” and as a result, believes passionately about observability. It is not only the key to keeping older trains up-to-date and operable for as long as possible, but it is also the basis for creating hi-tech rolling stock of the future.
The tool that has put Klimmek’s team on the right observability track: Grafana Enterprise. With it, Siemens can bring together millions of data points from active train sensors in under three seconds. The company can then take that information and use it to help its customers identify potential faults.
Siemens Mobility’s work in this area was the topic of the GrafanaCONline 2021 presentation “Creating value through IoT train sensor data with Grafana.”
Grafana Enterprise isn’t just a great fit for Siemens Mobility’s most hi-tech rolling stock, it also has helped the company give older trains a “digital makeover.” As Klimmek explains, “The current thrust of what we’re doing is taking dumb trains and making them smart.”
Grafana has been especially valuable for mid-life trains, which are more expensive to operate and maintain than newer models. Being able to monitor data and receive alerts when something needs attention can add another 10 to 15 years of life to a train, he says. “When you have a smart asset that tells you when it needs attention, it can perform markedly better than one that doesn’t.”
With Grafana as a tool in the Xcelerator, Siemens’s maintenance teams will be able to take on business intelligence development capabilities. So rather than directing issues to a specialist, team members will be empowered to observe what is happening through the platform. “Grafana is really helpful in reaching our objective,” Klimmek says.
As he sees it, using Grafana — and having real-time representation of the state of current physical service — has provided Siemens Mobility with “a new vision of ’the art of the possible.’ "
Point of origin
So how did Siemens Mobility arrive at this place? Before Klimmek’s team began using Grafana in 2020, they were trying to figure out how to address several high-level problems.
One major issue was that they had trains producing complex, high frequency data, and they needed a way to help customers that had to make quick decisions based on the relevant information they had available. Klimmek was looking for an open source solution that was flexible and easy to use so that domain specialists could apply their knowledge and expertise on top of the data that was being processed. It needed to be something that they could scale in a cost-effective way that would also provide them with access to a support team.
Around 2019, Klimmek began working with data platform engineer Haluk Tutuk from Periscube, Ltd, a Siemens Mobility UK cloud solutions partner. Together, they had a vision of building accessible self-service portals and workbenches they could use and share to observe all of their important train sensor data. Grafana was already being used to monitor the company’s cloud infrastructure, which made them realize it could do more. “There’s no reason why the same principles that apply to make cloud infrastructure observable shouldn’t also apply to trains,” Klimmek says.
Grafana also was focused on visualizing high frequency time series data and had the ability to integrate with a wide range of tools, technologies, and data sources, such as InfluxDB.
One of the biggest benefits to using Grafana, Klimmek added, is that it is simple to set up and the learning curve is quicker than with other products: “It’s easy to say to a domain expert, ‘This is how you drag-and-drop different elements.’ "
When it came to selecting the best type of Grafana account for their needs, Siemens Mobility chose Grafana Enterprise for three reasons. “One of the deciding factors was rebranding,” Tutuk says. “We had the vision to make our dashboards available to external customers, and we wanted Grafana to look and feel like a Siemens native application.” With Grafana, they could build dashboards with the Siemens colors and its logo, which is what their customers were used to seeing. “One of the main goals,” he says, “was making the dashboards indistinguishable from the rest of Siemens software products” — something that became possible with the release of Grafana 9.
The second goal was being able to have an Okta integration, which came natively with a Grafana Enterprise account. Finally, the third goal was to have a solution that allowed for premium data source connectivity with Oracle and ServiceNow.
For Siemens Mobility, investing in an enterprise account was also like a down payment on future expansion. It gives the company’s users a direct line to Grafana Labs’ Customer Success, Support and Enablement teams and creates a type of partnership. “We want to be at the leading edge of whatever is being developed,” Klimmek says, “so we like knowing that there’s a support team behind the scenes that we can get in contact with and say, ‘Okay, what’s cooking? What’s next? Here are three things that we’re trying to do.’ "
One time, Klimmek needed technical support for the Okta integration, and he says the help provided by the Grafana Labs team “was at surgical precision.” For his part, Tutuk found the support team helpful when a security blocker vulnerability was exposed. They received a quick initial response, and within a week they had a solution. “That was our joint contribution as a paying customer to the open source community,” he notes.
That collaborative spirit fits in perfectly with the mission of the Xcelerator, too. “We want to invest in having as open and as powerful a relationship with the core technology that we use day-to-day,” Klimmek says. “There may be a forum for exchange available in the open source community, but we want to brainstorm openly, and an enterprise setup gives us a closer relationship.”
He adds, “Having a flexible, easy-to-work-with tool like Grafana really makes it very clear what it means to have a ‘self-service’ portal — plus, it looks good, which makes domain experts interested in using it. We see so much potential.”
With Grafana, Klimmek’s team has been able to generate data in real time from the trains to optimize maintenance sequencing — work that has been designed to support performance managers and technical engineers.
One exercise they did examined how they could use recharge data. In that case, a crew on a train in the Thameslink depot ran experiments around a train’s battery, the conditions of the battery, and how quickly it depleted and recharged in different circumstances. There was a Grafana dashboard running on a laptop in the driver cab, which showed the experiments running in real time. “We’re creating lab-like conditions — analyzing data, changing things, and re-analyzing them with Grafana,” Klimmek explains.
Another area they’ve been focusing on is alerting. Rather than generating email alerts on AWS and sending out email-based reports three times a day, the team has tested bringing the alerts into Grafana instead and creating a general status overview of the fleet.
Having proven it works, they’re now thinking about how to scale that for their fleets, including the older trains. “The next step,” Klimmek says, “is to work in the context of our digital retrofit project” (a plan to connect Siemens Mobility’s Class 185 fleet to the cloud to access real time information). They also plan to move away from proprietary desktop-based software, which uses different tools to analyze different kinds of data, and bring it all together in an environment that users can adapt to their needs.
With the new system, Siemens Mobility had to accomodate more data than ever before because the digital retrofit with Grafana allowed them to retrieve data from the “classic” class of rolling stock fleets. One use case involved infrared sensors counting passengers going through doors on a fleet of 80 trains to reveal the current occupancy capacity. Each car had eight to 10 sensors, and the device vendor on the train created the same csv file thousands of times over, which meant each one had to be processed as xml, json, and multi-line json. “We stress tested millions of data points every second,” Tutuk recalls.
The good news: It still worked. Klimmek credited the success to having their Grafana setup available. “The combination of Grafana being very open and playing very well with InfluxDB and AWS allowed us to bring data in, do that stress testing exercise, visualize it, and also create interaction with the customer and supplier of that counting system. We were able to build the understanding and the knowledge that allowed the supplier to fix the problem, reduce the number of duplicates, then have the confidence to go to the external customer and say, ‘Here’s the kind of potential we can reach with passenger accounting data.’ "
Through the roll-out of the Grafana-based observability solution for rail, the digital team at Siemens Mobility is getting closer to their vision of setting up a full system of self-service portals and a workbench, where staff and customers can work collaboratively and interactively across digital and domain. “That’s been the primary value of Grafana for us so far: building POCs and proof points very quickly, then discussing those with internal and external customers,” Klimmek says.
They have already proven the capabilities of the solution with their fleet of about 500 trains, and can look at everything from track data to energy monitoring data and create a train overview. Serious usage developed with early-adopter champions. They have also implemented an online feedback loop to enrich data with domain knowledge through iterative data labeling and data markups. A dynamic and collaborative training and onboarding process has developed to support domain experts, front-line service managers, and analysts to become self-sufficient and confident in training their colleagues in turn.
In the future, Klimmek hopes they can show what it would look like to have a workbench complementing and interacting with commonly-used software to analyze data from systems that are standardized across the industry. Consider a train’s black box recordings, for example: “How can we take some of that very, very rich, log-like data and start looking at that in a flexible online environment, rather than using proprietary, vendor-specific desktop software?” he wonders.
In the meantime, Siemens Mobility will keep happily chugging ahead.