Work2024infrastructure, completed
KORAIL Platform Screen Door Research
Reliability analysis of 10,078 platform screen door failures across 279 stations, with maintenance strategies compared through simulation.
- Role
- Graduate researcher and data analyst
- Evidence
- Normalized and analyzed 10,078 failure records to support evidence-based comparison of rail maintenance strategies.
- Built with
- Python, Reliability Analysis, Monte Carlo Simulation
Run it
Four ways to maintain the same doors, with the same luck. Break one and see which recovers.
Click a station to break a door. Drag to turn the line.
A small model written for this page. Its failure rates and costs are invented; the study's are real.
Problem
Platform screen door reliability affects passenger safety and railway availability, but operational failure records need substantial normalization before component, seasonal, manufacturer, and maintenance patterns can be compared.
Method
I prepared and analyzed 10,078 failure records covering 279 stations and 18 lines. Monte Carlo simulation compared reactive, preventive, condition-based, and predictive maintenance strategies across cost, failure reduction, safety risk, and availability measures.
Outcome
The work produced a quantitative comparison framework for maintenance planning rather than relying on a single aggregate failure rate.
Claims on this page are limited to work I owned and outcomes I can defend. Private systems are described without proprietary implementation details.
Next on the sheet
2023
Mayawalk Staff Management
Full-stack product engineer