Multi-National glass manufacturer: Predictive maintenance for manufacturing equipment

Multi-National glass manufacturer: Predictive maintenance for manufacturing equipment

Challenge

  • Experiencing equipment failures unexpected costs in the production process.
  • High machine downtime and low production availability
  • Managers are challenged to see through all of the rapidly growing volumes of machine data captured throughout the process.

 

Solution for predictive maintenance

Solution

  • Using Azure Machine Learning, Neal Analytics identified key variables that influence the failure of equipment such as pressure, current, and duty cycle.
  • Used a classification model to predict future failures in equipment based on known failure events
  • Used an anomaly detection to identify outliers in sensor data that could lead to failures throughout the production process

 

Predictive maintenance results

Result

  • Reduce total maintenance costs through better planning of predictive maintenance programs
  • Able to identify potential breakdowns with 85% Precision up to two days before breakdown.
  • Reduce production downtime and asset utilization
  • Improve spare part supply chain planning