SpainManufacturing2020 – ongoingTalgo
TSMART predictive and condition-based train maintenance
Talgo moved from fixed-schedule servicing to condition-based maintenance on about 100 trains
Problem
Talgo serviced its trains on fixed preventive schedules. It wanted to plan work from the actual condition of each train's equipment.
Approach
Talgo's TSMART platform has monitored the fleet since 2020. With Atos (Eviden), Talgo added AI models and rules, run through an MLOps setup, to track the health of main train systems and estimate equipment lifespan. The system takes more than 30,000 signals per second per train and runs automated inspections more often. It flags early failures so work can be planned before the train reaches the depot.
Results
Takeaway
Fleet monitoring came first, in 2020. The predictive layer was built on that stream of live data.
Categories
- Industry
- Manufacturing
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Atos describes how Talgo shifted to predictive, condition-based maintenance for about 100 trains. It reports gains of 15% in reliability, 10% in maintainability and 5% in availability, and 99% fleet availability. The 200 saved inspection hours per train per year are a projection.
Full details are in the original sources.
