11. March 2026, 09:24 o'Clock
Article: DB Cargo: AI-supported spare parts planning for Class 77
DB Cargo uses AI to reduce bottlenecks in the supply of spare parts for Class 77 locomotives.
When spare parts are missing, operations come to a standstill
When maintaining locomotives, missing spare parts can significantly delay processes. To avoid this, DB Cargo has launched the "Spare Parts Forecasting 1.0" project. The aim is to provide spare parts exactly when they are needed – without unnecessarily increasing inventory levels.
AI-supported forecasting at the Darmstadt logistics centre
Material requirements planner Jörg Soyka, data science specialist Frederic Tausch and project manager Philipp Nowak are jointly developing the AI-supported spare parts forecast.
At DB Cargo Railport Darmstadt, decisions are made every day about whether vehicles will be repaired quickly or whether they will be out of service for longer periods. Project manager Philipp Nowak, material requirements planner Jörg Soyka, data science specialist Frederic Tausch and technical expert Anne Kulinski developed the AI model there. It combines historical consumption data with information on mileage, maintenance intervals and workshop contexts.
The Class 77 fleet comprises around 60 diesel locomotives for non-electrified routes. As the locomotives were built in Canada, spare parts can take weeks or months to deliver. Traditional forecasting methods reach their limits here because many parts are only needed irregularly.
Practical insight: Three key findings from the project
The benefits of AI-supported forecasting became apparent after just a few months:
- Targeted availability instead of full warehouses: Expensive and long-running parts are reliably secured, while quickly available parts are planned more leanly.
- Context beats pure history: Information on mileage and maintenance levels increases forecast reliability, even with irregular demand.
- Practical benefits in a short time: The new methodology and the existing planning tool were successfully implemented within a few months.
This insert shows that AI forecasting delivers concrete practical benefits – not only for individual vehicles, but for all spare parts planning.
Thanks to AI forecasting, spare parts such as oil pumps are available on time - long waiting times are avoided.
A concrete example: Class 77 oil pump
A clear example is the oil pump: while the old method did not identify any demand, the AI model predicted five units – actual consumption was six. With delivery times of around 500 days, this determines whether a vehicle is out of service or remains operational.
Optimised planning tool increases efficiency
In parallel with the forecast, the existing Excel planning tool was improved. Parameters were systematically tested to balance the conflict between waiting time and capital commitment. Separate parameter sets were derived for different vehicle types, making planning even more practical.
With its AI-supported spare parts forecast, DB Cargo demonstrates how data-based planning makes maintenance more efficient and reduces downtime.
Go