EN中文
Book a demo
Home/Evidence

Case studies

Real elevators.
Real signals. Real faults.

Four cases from the field, each answering a different question: can it reveal a developing anomaly earlier, help narrow down a hard-to-locate problem, verify a repair outcome, and show whether maintenance changed the measured condition? Every case uses data captured from elevators in actual operation, not a demonstration.

2 monthsof trend that revealed a developing traction-machine fault
90% → 1%abnormal runs per day, after one bearing was replaced
Detectedservice windows recognized from the ride data alone

Fault case studies

Four questions, four answers.

2026-07-10·EARLY WARNING
NORMAL WARNING CRITICAL
A two-month trend revealed a developing traction-machine fault before failure
A 15-year-old unit passed every monthly check while the system was already warning. Three recorded trips show the fault growing — the gap between up and down speeds went from 0.06 to 0.24 m/s — and the traction machine was repaired in a planned window.
Traction machine fault Speed gap ×4 No unplanned downtime from the detected issue
Read the full case
2026-06-20·EFFECTIVENESS VALIDATION
BEFORE >90% ABNORMAL AFTER <1%
From 90% to 1%: one bearing made the value of maintenance visible
An office-building elevator rattled and shook but never stopped, and neither the maintenance team nor management could prove how serious it was. Dr. Elevage found over 90% of daily runs abnormal; after the car-top sheave bearing was replaced, it fell to under 1%.
Abnormal runs 90% → 1% Car-top sheave bearing Vibration >10× normal
Read the full case
2026-07-05·FAULT PINPOINTING
REPEATING IMPULSE · FIXED POSITION
Detected from the car top — a counterweight guide-rail fault
A residential elevator juddered and routine checks found nothing. After 24 hours of car-top monitoring, position, direction and frequency all pointed to the counterweight: slightly twisted hollow rails and over-tight shoes. One adjustment, and the jolt was gone.
Counterweight guide rail 24 h to find the pattern Guide-shoe adjustment
Read the full case
2026-06-28·SERVICE TRACEABILITY
SERVICE WINDOW DETECTED PASSIVELY · NO CHECK-IN
Beyond check-ins and photos.
See what changed in the data.
Check-ins, photos and signatures can all be produced without doing the work. A service visit leaves physical traces in the ride data; the platform recognizes them and writes the work order on its own — and shows whether the visit helped: abnormal runs went from 8–11% before the visit to 2% after.
100% passive Auto work order Abnormal runs 11% → 2%
Read the full case

Cases are drawn from deployed units. Building identities are withheld at customer request; the measurements, timings and outcomes are as recorded.

Behind these cases How the detection actually works → Your fleet Put one elevator online and see its story →

Continue exploring

Traction-machine early warningMaintenance verificationElevator fault localization