DB Cargo has introduced an AI-based spare parts forecasting system for its Class 77 diesel locomotives at DB Cargo Railport Darmstadt.
The system integrates data on mileage, maintenance intervals and workshop context with past consumption patterns to improve demand forecasts.
DB Cargo operates around 60 Class 77 diesel locomotives on non-electrified routes. Built in Canada, the locomotives require spare parts that can have delivery times of several weeks or months, with some components taking significantly longer. According to DB Cargo, conventional forecasting methods have been less effective due to irregular demand for certain parts.
One example is the oil pump for the Class 77. The previous forecasting method did not indicate any demand, while the AI model predicted five units; actual consumption reached six. With delivery times of around 500 days, forecast accuracy directly affects locomotive availability.
Alongside the AI model, DB Cargo revised its existing Excel-based planning tool. Parameters were systematically tested to balance waiting times against inventory levels. Separate parameter sets were defined for different vehicle types to adapt planning to specific operational profiles.
DB Cargo reports that the new methodology and the updated planning tool were implemented within a few months and are now applied to spare parts planning for the Class 77 fleet.
