Definition
Hit rate of the initial diagnostic attribution (sensor problem versus asset problem) against the confirmed root cause, on anomalies investigated to resolution. A degrading sensor and a degrading asset look alike until someone goes out and checks.
The Measurement Gap
Renewable energy sites are predominantly unmanned. Sensors are the eyes and ears of remote operations, feeding SCADA data to centralized monitoring centers. When a sensor shows an anomaly, the diagnostic question is: is the asset degrading, or is the sensor degrading? A fouled irradiance sensor reads low and makes healthy solar panels look like they have soiling problems. A drifting current transformer reads high and makes a normal inverter look like it has an electrical fault. A humidity sensor with condensation shows false moisture intrusion alerts.
The measurement gap is that most operations do not track whether the diagnostic process correctly identifies the source. When a technician is dispatched for an inverter fault and arrives to find a drifting CT sensor, that dispatch cost $500 to $2,000 and the tech did nothing productive for the asset. Worse, when a real asset problem is dismissed as a sensor issue because the ops team has been burned by too many false dispatches, the asset degrades undetected until it fails.
SADR measures the hit rate of the initial diagnostic attribution. For anomalies that were investigated and resolved, did the first assessment correctly identify whether the problem was in the sensor or the asset? This is not about whether the right fix was applied; it is about whether the diagnostic triage pointed toward the right system, which determines whether the response is appropriate or wasted.
The decision SADR supports is whether maintenance is being dispatched for phantom problems created by sensor degradation, or conversely, whether real asset problems are being dismissed as sensor noise because the organization has lost confidence in its sensor data.
Formula
For anomalies that were investigated and resolved with a confirmed root cause, compare the initial diagnostic attribution (sensor problem versus asset problem) against the confirmed root cause. Score the hit rate.
Formula (Notation):
SADR = Correctly_Attributed / Total_Investigated
Companion Metric (False Dispatch Rate):
FDR_sensor = Dispatches_for_sensor_issues_diagnosed_as_asset / Total_dispatches_for_anomalies
Variable Definitions:
| Variable | Definition |
|---|---|
| Correctly_Attributed | anomalies where the initial diagnosis (sensor issue or asset issue) matched the confirmed root cause |
| Total_Investigated | all anomalies that reached investigation and root cause determination |
Edge Cases and Measurement Boundaries:
Anomalies that remain unresolved with no confirmed root cause should be excluded from SADR but tracked separately as an "unresolved anomaly" metric. When both the sensor and the asset have issues simultaneously, count as correctly attributed only if both were identified in the initial assessment. For remote sites where investigation is delayed, extend the assessment window to allow for confirmation before scoring.
Proposed formula, pending validation pilots.