Future of Operations & Maintenance 2026

During rail grinding, so-called slag is generated – metallic residues from the rails that accumulate on scrapers, cool down, and harden into lumps weighing up to 10 kg. These lumps fall uncontrolled onto the track and must be laboriously removed at the end of each shift. On average, around 200 kg of slag accumulates per shift along several kilometers of grinding track. Today, this material is collected manually by two employees per shift – a process that is time-consuming, physically strenuous, and increasingly unattractive given the shortage of skilled labor. We are therefore seeking technical solutions that can replace or significantly reduce manual slag collection. The objective is clear: lower workforce requirements, reduce physical strain on employees, and at the same time improve efficiency, quality, and overall safety while cutting costs.

Train operating companies use DB InfraGO’s Anlagenportal Netz (APN) to book tracks in service facilities across Germany. DB InfraGO’s service facilities serve as parking areas for trains, wagons, and locomotives used by both passenger and freight operators. Depending on location and purpose, the tracks vary in length, layout, and equipment — such as power connections, water supply, or toilet servicing facilities. While bookings can already be made well in advance and even on an hourly basis, there is currently no technical solution to send clearance notice. If demand changes, clearance notice are made manually via regional teams. 
As the capacity is limited in service facilities and DB InfraGO AG is facing growing demand, we are looking for digital solutions that make capacity management smarter. The goal: allow customers to release unused storage siding capacity themselves, making it visible and bookable as secondary use.

Every switch in the DB InfraGO production process is unique – with individual geometries, components, and fastenings. Depending on the installation situation, wooden, plastic, or concrete sleepers are used, and each individual sleeper in the switch area is manufactured individually. Today, the bill of materials – i.e., the list of all required sleepers and materials – is still derived manually from installation plans based on a standardized, drawn plan showing the switch position including the sleepers to be used. This process is time-consuming and prone to errors: even one incorrectly ordered sleeper can lead to delays on the construction site.

We are looking for digital solutions that automatically identify the required sleepers from the installation plans and transfer them to a parts list. The aim is to reduce manual checking and transfer work and minimize errors in the current manual process. Of particular interest are approaches based on computer vision, object detection, or OCR+, for example, which can recognize graphic elements, text, and deviations in technical drawings. These systems should independently identify different types of sleepers and special characteristics and derive a digital material list from this information.