Operations
Predictive Maintenance
Harness distributed equipment and sensor data to better predict component failure and reduce costly downtimes.
Challenge
- Mean-based schedule maintenance was undermining competitiveness as delivery times increased and margins decreased.
- Management suspected that maintenance schedules were too frequent for some regional routes and not enough for others.
Solution
- Devron collaborated with the business intelligence team to deploy and train machine learning models at the sensors on each rig.
- Encrypted model artifacts from each location are sent back to Devron Central Authority to train the main ML model and develop a more nuanced picture of maintenance needs across routes and truck types.
- Company moved from preventive to predictive maintenance which saw an over 35% reduction in downtime.
Outcomes
↑
Enable Preemptive Maintenance Operations
↑
Increase Prediction Accuracy
↑
↓
Reduce Costly Downtimes
↓
Reduce Time to Insight
↓
