At KONUX, it is our mission to make rail the mobility choice of tomorrow. With each quarterly release, we bring more innovations to our customers aimed at helping them increase the capacity, reliability, and cost-efficiency of their networks. As we were preparing to announce the Summer ‘20 release of the KONUX Predictive Maintenance System for Rail Switches, we sat down together (virtually) with Dr. Volker Kefer and got his perspective on how he believes rail infrastructure managers can use machine learning and industrial-internet-of-things (IIoT) technology to transform their operations.
Up to the end of 2016, Volker was a Member of the Executive Board of Deutsche Bahn AG, most recently as Vice Chairman of the Executive Board and the Board for the Infrastructure, Services and Technology divisions. In 2018 he joined KONUX in a senior executive advisory role.
KONUX: You have witnessed first hand the impact of continuously increasing traffic. Even with the current corona crisis, overall global railway traffic is expected to double by 2050 according to the International Energy Agency. But already today, overutilization is becoming a problem. DB claims that is one of the key reasons behind the decrease in punctuality, stating that the network is currently running at 140% capacity in some areas (Source DW).
Volker: Every time we talk about capacity, it is important to distinguish between the absolute physical limit of the system and what we define as the maximum utilization it can take to deliver our preferred level of safety and performance. Increasing the physical limit of the network, for example, building new lines, is one way to alleviate that, but it is complex and costly. The other way to allow for higher capacity is to operate the network in a way that is closer to its physical limit. One example here is the LZB signaling systems. Another is minimizing unnecessary downtimes. I truly believe that reducing unnecessary downtimes, by optimizing inspections and maintenance planning is perhaps the easiest and most efficient way to achieve higher network capacity. If you look at the numbers you will see that it pays off. For each 1% improvement in capacity, DB can generate millions in additional revenue and improved performance. So, you really want to gather all the information you can and provide actionable insights to the people in the regions. They are the ones who can really benefit from the transparency, they are the ones who ultimately realize the savings. But that is just the beginning. When you start bringing the latest and most accurate information about the health of the network to the dispatchers and traffic planners, there is so much more that could be achieved.
The Network Map gives you the health overview of all switches in the entire network to ensure your planning has the biggest impact on capacity and line availability. (Asset locations are examples only)
KONUX: Most infrastructure managers will agree that network usage is one of the best indicators of the speed of degradation. This is why they often use load categories, such as the UIC load categories, based on the train tonnage and speed, to segment their S&Cs. They then plan maintenance and inspections with different frequencies. The problem is that there are huge deviations between the planned and the actual usage of a switch. We have also seen examples where all switches within the same load category are inspected with the same frequency as the top 2% most used switches in that group, even though they are dramatically less utilized. This is why we developed the KONUX Load Factor – a customer-independent numeric indicator reflecting the load on a railway switch
With the KONUX Load Factor, you always know your switch usage at a much more granular level and are able to adjust the planning of inspections, maintenance, or replacement cycles. You can see which switches are being over- or under- inspected and take action.
KONUX: What are your thoughts on how the KONUX Load Factor can help improve capacity and reliability?
Volker: I fully agree that asset utilization is the best indicator of the speed of its degradation. And of course, it is important to know what really happens in your network, it is important for everyone – for the planners, the maintainers, and the operators. And no, neither schedules nor historic data are accurate enough. They do not provide the precision and flexibility you need if you really want to make use of the adjustment to your inspection and maintenance program. Of course to really materialize the savings and increased performance you need to also change the processes. In the meantime, if you can tell the asset owners, which of the hundreds of assets they are responsible for are the ones that are “burning” and must be prioritized, they can make informed decisions that help reduce failures and delays. But understanding the traffic in your network has many more implications. Yes, you want to focus inspections and maintenance efforts where they are needed most, but you also must ask yourself, why do these discrepancies exist at all? If assets are remaining idle while we experience 140% capacity, why are they idle? Are there better ways to organize traffic to alleviate the stress in certain bottlenecks? Surely there are many more use cases, and we are barely scratching the surface of what we can do with these insights.