§9.4

Asset management — rolling stock

Keep trains safe, available and reliable through smarter maintenance.

Rolling stock asset management ensures trains remain safe, available and reliable through maintenance, defect management, fleet planning and long-term lifecycle decisions. AI helps move from reactive and calendar-based approaches toward more predictive and coordinated decision-making.

Priority pathfinders

Predictive degradation modelling

also touchesOperations

Problem

High-impact rolling stock components (doors, engines, HVAC compressors, traction converters) cause cancellations and short-forms when they fail unexpectedly.

Approach

Combine telemetry, defect codes, maintenance history and operational conditions to surface emerging issues before they become service-affecting.

Why now

Scope is deliberately narrow and boundaries clear; relevant data already exists across operators, maintainers and ROSCOs if brought together.

In the wildAmygda ↗

DB Regio: 80% of at-risk trains flagged 60+ minutes before departure, working from existing CMMS logs and fault codes — no new sensors. DB Innovation Award.

● Amygda · proof pointBuilt by the team behind this site

80%

of at-risk trains identified 60+ minutes before scheduled departure

Amygda has built this with DB Regio.

From existing CMMS logs and fault codes — no new sensors. 250M+ maintenance records reduced into risk scores per train. DB Innovation Award winner.

Read the DB Regio case study →

AI-supported depot defect triage

also touchesOrganisational

Problem

Work orders are largely free text, fault codes are inconsistently applied, common unplanned defects lack shared templates — repeat misdiagnosis follows.

Approach

Use AI to identify recurring defect patterns, surface likely causes, highlight dominant fixes seen historically — unsupervised clustering and pattern extraction consolidate free-text records.

Why now

Depots already hold relevant telemetry, inspection notes and repair histories. Workflow is repeatable, well understood, with clear safety boundaries.

In the wildAmygda ↗

Structured 1,600 fault codes from 250M+ records into 57 actionable categories — turning free-text depot logs into risk scores.

● Amygda · proof pointBuilt by the team behind this site

1,600 → 57

raw fault codes structured into actionable categories at DB Regio

Amygda has done this.

Free-text depot logs converted into consistent diagnostic groupings. Reusable across operators with messy, unstructured maintenance records.

Read the DB Regio case study →

Pathfinders that also touch this area

from §9.2

Incident management

from §9.2

Train service recovery

from §9.3

Maintenance planning intelligence

from §9.6

Competency and re-certification planning

Share on LinkedIn

tag: #AIRailPlan

Asset management — rolling stock is one of six opportunity areas in GBRX's AI in Rail Action Plan. Here's a 60-second read on what's possible and who's already doing it in the UK.

Source: @GBRX · Built by @Amygda

https://airailactionplan.amygdalabs.com/opportunities/rolling-stock

#AIRailPlan

Paste, then type @GBRX and @Amygda and pick each from LinkedIn's autocomplete to convert them into live tags.