DB Cargo has introduced an AI-based spare parts forecasting system to improve maintenance planning for its Class 77 diesel locomotive fleet. The system aims to ensure that critical components are available when required while avoiding excessive inventory levels.
The project, named Spare Parts Forecasting 1.0, is being implemented at the company’s logistics center in Darmstadt.
AI-Supported Maintenance Planning
Locomotive maintenance often depends on the timely availability of spare parts. When required components are unavailable, vehicles may remain out of service for extended periods.
The AI forecasting system analyzes historical spare parts consumption together with operational parameters such as locomotive mileage, maintenance intervals and workshop conditions. Combining these data sources allows the model to estimate future demand more accurately than conventional forecasting methods.
The development team included specialists in material planning, data science and technical maintenance operations working at the Darmstadt logistics center.
Challenges in Spare Parts Availability
The Class 77 locomotive fleet consists of approximately 60 diesel locomotives used on non-electrified freight routes. Because the locomotives were manufactured in Canada, replacement components may require several weeks or months for delivery.
Traditional forecasting approaches often struggle to predict demand for parts that are replaced only occasionally. As a result, planners must balance the risk of stock shortages against the cost of maintaining large inventories.
