Beacon — Industrial Signal Detection Demo

Fault Wave Progress

vibration RMS kurtosis RPM temperature motor current

Inner Race

F1 0.911

PENDING

Outer Race

F1 0.819

PENDING

Ball Fault

F1 0.758

PENDING

Detector Legend

Detector 1 — Structural deviation from healthy baseline

Detector 2 — Fault-pattern prevalence above baseline

Detector 3 — Spectral irregularity (early bearing fault)

Calibration Profile

Status—

Healthy windows used—

Detector 1 sensitivity—

Detector 2 sensitivity—

Detector 3 sensitivity—

Calibration latency—

Fault data usedNONE

CWRU Bearing Benchmark

Fault Type	F1 Score
Inner Race	0.911
Outer Race	0.819
Ball Fault	0.758
Overall	0.832

How Beacon works: Beacon calibrates on healthy sensor windows — vibration, RPM, temperature, and motor current. No fault samples, no labels, no model retraining required. Beacon learns the behavioral fingerprint of a normal machine, then watches for departures from it. When a bearing fault develops, three independent detectors fire: Detector 1 flags structural departures from the healthy baseline; Detector 2 flags fault-pattern prevalence above baseline; Detector 3 flags spectral irregularities — the earliest indicator of bearing fatigue. Any one detector triggering produces a fault verdict.

CWRU Bearing Dataset — Inner Race faults are the strongest signal (F1 0.911). Ball faults are the subtlest (F1 0.758) and lean on structural-deviation detection. Patent pending. Implementation details proprietary to Astrognosy AI.