Strategies for Validating Asset Health Data Integrity


In the modern mining landscape, leveraging sensor data for asset health monitoring and predictive analytics is essential for reducing operating costs. However, the value of advanced visualizations and AI-driven insights is entirely dependent on the quality of the underlying data. Before integrating asset health data into critical decision-making processes, organizations should perform rigorous validation to ensure data integrity.

Follow these five essential steps to identify and rectify "dodgy" or inaccurate asset health data:

1. Analyze Raw Data Streams

Avoid relying solely on aggregated data or high-level visualizations like pie charts, which can mask underlying technical issues.

  • Identify Granularity: Request raw data exports (e.g., XLS formats) to inspect individual data points.

  • Detect Noise: Look for "multiple events" per actual occurrence or false alarms triggered by power-cycling tests, which are often hidden in summarized reports. This can often be accomplished by pulling the data into appropriate charting tools.

2. Conduct Subject Matter Expert (SME) Sanity Checks

Collaborate with technical SMEs to verify that the numerical values align with mechanical and electrical realities. Key questions include:

  • Voltage Levels: Are the recorded voltages appropriate for the specific wheel motors in operation?

  • Thermal Readings: Do brake differential and exhaust temperatures remain within sensible operational ranges?

  • Alarm Frequency: Is the volume of alarms generated within a specific timeframe reasonable for the asset’s condition?

3. Audit Parameter Completeness

Ensure that all available data parameters are being successfully transmitted from the asset to your monitoring system.

  • OEM Benchmarking: Consult with the Original Equipment Manufacturer (OEM) to define the full spectrum of data available via J1939 or CAN bus connections.

  • Gap Analysis: Compare the OEM's technical specifications against your actual data intake to identify missing parameters.

  • Hardware/Firmware Alignment: Address discrepancies by upgrading machine hardware or firmware, ensuring your local support teams are equipped to handle these technical requirements.

Finally, make common sense checks such as “this piece of equipment has 12 cylinders, are there 12 exhaust temperature sensor readings?”

4. Validate Units, Synchronization, and Time Zones

Minor configuration errors can lead to significant reporting discrepancies.

  • Standardize Units: Verify that all units of measure (UOM) are correct and consistent across the platform.

  • Synchronize Timestamps: Ensure all onboard systems are time-synchronized to prevent chronological errors in reporting.

  • Configure Time Zones: Confirm that the correct time zones and Daylight Saving Time (DST) settings are applied to the onboard data loggers.

5. Cross-Reference with Comparative Data Sources

Whenever possible, perform a cross-system validation to confirm data accuracy.

  • Cloud Comparison: If the OEM provides a parallel cloud-based monitoring service, compare those figures against your internal systems.

  • System Reconciliation: Use these secondary sources as a crude but effective baseline for final validation.

Conclusion

Harnessing high-quality asset health data is a primary driver for operational efficiency in mining. By implementing these validation steps, your organization can move beyond "shiny" graphs toward actionable, reliable intelligence.

Need assistance with your asset health program? Contact the experts at MTS to ensure your data is working for you: info@minetechservices.com

Book a 1:1 session with our mining experts