Twine

Switch Predictive Maintenance for Railway

Alireza Ghane
Published

0
  • Project creator

Converted sensor readings into actionable insights for rail infrastructure managers to reduce maintenance costs and improve asset availability. The system reused existing sensor data and demonstrated high prediction precision, reducing false predictions significantly compared to native systems.

We have provided both automatic warning system as well as a visualization framework for the technicians to analyze. Our system could provide up to 2 weeks advance notice for switch failures in the rail network.