How BHP achieved $5.5M in annual savings through predictive maintenance for haul trucks
ML models trained on failure modes and telemetry prioritized risk-based maintenance. That cut unplanned failures and safely extended intervals.
Context
The operation relied on scheduled maintenance that often triggered downtime even when equipment was still healthy. At the same time, unexpected failures continued to occur because maintenance did not reflect true asset condition. This created both lost production and inefficient use of maintenance resources.
Solution
The company implemented risk-based predictive models trained on historical failure modes and real-time telemetry. These models produced health scores and recommended actions that the maintenance COE monitored through dashboards and KPIs. This shift enabled maintenance to align more closely with measured equipment risk.
Results
The approach led to $5.5M in savings at one site by reducing avoidable work and lowering the number of unexpected outages. Maintenance intervals were extended safely, and equipment availability improved through earlier risk detection.
Data Inputs
Telemetry, CMMS/work orders, sensor data (vibration/temp/pressure)