Predictive Configuration

Predictive configuration is the control panel for the predictive maintenance engine. It allows each tenant to adjust the parameters that govern how health indexes are calculated, when an asset is considered at risk, and how much minimum confidence is required to automatically generate an intervention.

All changes apply in real time and are tenant-specific, meaning two organizations can use the same engine with completely different behaviors depending on their processes and risk tolerance. A 5-minute cache ensures that frequent reads do not generate unnecessary database load.

When adjustments do not produce the expected results, the reset button restores all values to factory defaults in a single click.

What is it for?

• Personalizes the predictive algorithm without code changes or deployments.
• Adapts intervention thresholds to each organization's risk tolerance level.
• Adjusts AHI (Asset Health Index) weights according to the relative importance of each sensor.
• Defines failure probability breakpoints that trigger alerts and work orders.
• Configures data maturity requirements before the engine issues predictions.
• Sets a confidence ceiling to filter low-certainty predictions.

How does it work?

Configurable parameters

AHI weights

AHI weights determine how much each data source contributes to the asset's overall health index:

{
  "vibration": 0.35,
  "temperature": 0.25,
  "current": 0.20,
  "pressure": 0.15,
  "efficiency": 0.05
}

Weights must sum to 1.0. The system validates this condition on save.

Probability breakpoints

Breakpoints define risk bands:

• Green (normal): probability below low_breakpoint
• Yellow (attention): between low_breakpoint and medium_breakpoint
• Orange (alert): between medium_breakpoint and high_breakpoint
• Red (critical): above high_breakpoint

Configuration cache

Configuration is cached for 5 minutes to avoid repeated reads on every evaluation cycle. Saved changes are reflected in the next evaluation cycle after the cache expires.

Reset to defaults

Reset restores all parameters to the tenant's factory values. This action is irreversible but previous values can be recovered from the change history for 30 days.

Using it from the Dashboard

1. Go to Settings > Predictive Engine.
2. Adjust AHI weights using the sliders per component.
3. Modify the probability breakpoints in the Risk Bands section.
4. Configure the minimum maturity and confidence ceiling.
5. Click Save Configuration.
6. To revert to the original values, use Restore Defaults.