Predictive Maintenance – Project Details

CompanyL&T Smart Manufacturing
RoleML Engineer / Data Analyst
Duration3 Weeks
Skill LevelAdvanced
Tools UsedPython, Scikit-learn, Power BI, IoT Sensors
DeliverablesFailure prediction model, anomaly alerts, visual dashboards
Objective

Develop a predictive maintenance solution for industrial equipment using sensor data from L&T factories. The goal was to minimize downtime by forecasting potential failures before they occurred. Vibration, temperature, and voltage patterns were processed through ML algorithms to detect anomalies and predict machinery failures. A dashboard was also created to track maintenance schedules, anomaly counts, and system reliability trends in real time.

Key Challenges
  • Cleaning noisy sensor data
  • Labeling failure conditions accurately
  • Low-frequency failure data availability
Features
  • Real-time sensor anomaly tracking
  • Failure risk scores by equipment
  • Interactive downtime forecast charts
Outcome

Prevented unplanned breakdowns in key units, reducing overall maintenance cost by 28% and improving operational efficiency.