The Asset Health Level is an indicator generated by an advanced AI that translates what the system identifies for each failure mode based on vibration and energy consumption analyses. This provides a clear view of the equipment's condition even before Insights are generated, ensuring greater transparency in anomaly detection.
Below, we explain how this calculation works and the factors that influence the Health Level.
How Failure Scores Are Calculated
The model calculates an individual score for each failure mode based on data collected by Smart Trac and Energy Trac sensors. These scores are then translated into asset health categories:
Health Status | Score Range |
Good | 76% - 100% |
Acceptable | 51% - 75% |
Poor | 21% - 50% |
Critical | 0% - 20% |
Unavailable | Insufficient data |
The scores are recalculated daily, enabling continuous monitoring of equipment conditions.
Factors Considered in Score Calculation
Smart Trac Failure Scores assess common mechanical failures, while Energy Trac Failure Scores monitor abnormal energy consumption conditions.
Smart Trac Failure Scores
These scores are based on vibration and temperature analyses, using a specific set of symptoms and data:
Failure | Symptom | Considered Data |
Unbalance | High magnitude in 1st harmonic | Velocity Spectrum + RPM + Orientation |
Angular Misalignment | Elevated harmonics H1, H2, and H3 (H1 dominates) | Velocity Spectrum + RPM + Orientation |
Parallel Misalignment | Elevated harmonics H1, H2, and H3 (H2 dominates) | Velocity Spectrum + RPM + Orientation |
Rotational Looseness | High magnitudes at low-frequency harmonics | Acceleration Spectrum + RPM + Orientation |
Structural Looseness | High magnitude in 1st harmonic on a single axis | Velocity Spectrum + RPM + Orientation |
Lubrication | High background noise magnitude | Acceleration Spectrum |
Vibration by Velocity | High RMS value | RMS |
Temperature | Elevated temperature (excluding climate variation) | Temperature reading |
These data points are compared to reference values from a healthy state. The severity of the failure is measured by the difference between the current state and the reference state, normalized to a score between 0 and 100.
If there are fewer than 10 expert samples in the last 48 hours to calculate a score, the asset's status for that failure mode will be marked as Unavailable.
Energy Trac Failure Scores
Unlike mechanical failures, these scores do not indicate issues that may cause equipment breakdowns but rather anomalies in energy consumption.
Failure | Symptom | Considered Data |
Voltage | Phase voltage above or below the limit | Phase Voltage RMS + Reference Value |
Power Factor | Low power factor under full operation | Power Factor |
Energy Consumption | Consumption above or below expected for the day of the week | Consumption History |
These values are normalized and converted into a score between 0 and 100, allowing the identification of abnormal energy consumption patterns.
How the Overall Asset Health Level is Calculated
The Asset Health Level is calculated based on the severity of different failure modes, considering the type of equipment and the potential impact of each anomaly. The model takes into account:
Future Value: A projection of scores for the next 6 hours, using linear regression.
Historical Value: Within the last 48 hours, the model selects a value that represents the trend of the worst moments in the period, filtering out minor fluctuations.
Final Score: The highest value between the historical data and the future projection.
If an asset presents severe failures, its Health Level will be low, indicating a significant risk of failure. However, you don’t need to constantly monitor the Health Level to know when to take action—whenever a relevant issue arises, the platform will generate an Insight.
The Health Level can support maintenance strategies. For example, if you receive an Insight, you can check the Health Level to understand the severity of the problem and verify other related failure modes. If you're planning a scheduled shutdown, it helps identify which assets need priority inspection. During predictive field inspections, it serves as a reference to guide the team and optimize analysis time. Additionally, if multiple machines require attention, the Health Level allows prioritization of those in the most critical condition, ensuring a more efficient action plan.