• Indian Railways leverages AI for predictive maintenance in Vande Bharat trains
    Times of India | 3 November 2024
  • KOLKATA: In a remarkable move towards enhancing the reliability and efficiency of Indian Railways, the Howrah Division introduced an advanced predictive maintenance system powered by Artificial Intelligence (AI) and Machine Learning (ML) for the Vande Bharat (VB) train sets. Developed at the Jheel Siding Coaching Depot, this pioneering software leverages real-time Remote Monitoring and Management of Locomotives and Trains (REMMLOT) data to proactively identify and address potential issues before they can disrupt service, setting new standards in train operations.

    Using sophisticated algorithms, the AI-driven system monitors multiple operational metrics of VB train sets, comparing them to theoretical performance models. Using inverted logic trees and predefined threshold values, the software detects faults, alerting the maintenance team to take corrective action. The system’s unique capability to analyse running conditions during fault detection enables it to pinpoint specific faulty components, allowing for swift intervention. Maintenance staff can also input feedback, categorise faults by severity, and log corrective actions, thereby establishing a cycle of continuous improvement.

    The system’s integration with Original Equipment Manufacturer (OEM) manuals, historical fault records, and ML libraries enables it to suggest effective solutions for recurring or critical faults. With every input, the software learns and refines its accuracy, minimising future operational interruptions.

    The predictive maintenance software already proved effective, resolving 22 faults to date. These include issues with Static Inverters (STV), loose connections, Low Tension Control (LTC) modules, relays, and defective speed sensors. By addressing these faults early, the system prevented route delays and ensured uninterrupted services. Daily PDF reports generated by the software document each fault, corrective measure, and detailed code analysis, providing transparency and improving maintenance planning.

    The AI-powered system not only increases the reliability of VB train sets but also demonstrates the potential for broader application across the Indian Railways network. As more data is input into the system, it will continue to enhance its fault detection and analysis capabilities, eventually incorporating time trend analysis for recurring faults and temperature profiling to monitor sub-assembly performance, thus enabling optimised maintenance schedules.

    This initiative is a significant step in Indian Railways’ vision of modernising its operations through advanced technologies to boost efficiency and enhance passenger satisfaction. The Howrah Division’s embrace of AI and ML is not just preventing faults but actively shaping the future of predictive maintenance, establishing a new benchmark in rolling stock management.
  • Link to this news (Times of India)