Definition
The percentage of CBM-driven maintenance recommendations that fall within the degradation model's validated operating range, as opposed to extrapolating beyond where the model has been proven accurate.
The Measurement Gap
The CBM engine says "extend this interval to 210,000 miles." But was the degradation model validated that far? The model may be extrapolating past the evidence. Organizations trust CBM recommendations without checking whether they fall within the range where the model was actually calibrated. No legacy metric distinguishes validated CBM recommendations from extrapolations.
Decision Supported:
When the CBM engine recommends an interval, is that recommendation based on validated data or is the model extrapolating past the evidence?
Formula
Plain language: For each CBM recommendation, compare the recommended interval or condition threshold to the degradation model's validated range. Count how many fall within validated range versus outside.
Notation:
CVRC = N_within / N_total
Variables:
| Variable | Description | Unit |
|---|---|---|
| N_within | CBM recommendations within the degradation model's validated range and training envelope | Count |
| N_total | Total CBM recommendations issued | Count |
| validated_max_interval | Maximum interval the model has been accuracy-validated for | Miles, hours, or condition units |
| training_envelope | Operating conditions under which the model was trained and validated | Condition set |
Sign Convention: Higher is better. 1.0 = all recommendations within validated range.
Proposed formula, pending validation pilots.