🔮 Machine Learning · Failure Prediction
24-hour Failure Risk — ML Forecast Model
Algorithm: Linear Trend (LLT)
Data: 30-day history · 28 motors
Accuracy: ~ 85 – 90 %
GDPR · KVKK Masked
29 Apr 2026 · 17:19
📝 Prompt
"List the motors at high risk of failure within the next 24 hours. Use a machine-learning model on the last 30 days of data."
Risk summary
🔴 Critical risk — urgent maintenance required
| # | Motor | Risk | 30-day trend | DE bearing | NDE bearing | Oil | Action |
|---|
| 1 | GM19 | 9.8 / 10 | ⬆ +280.5 °C (3×) | 373 °C | — | — | 🚨 Stop now |
| 2 | GM02 | 9.6 / 10 | ⬆ +40 °C (10 %) | 546 °C MAX | 385 °C | 81.4 °C | 🚨 Stop now |
| 3 | GM09 | 8.9 / 10 | ⬆ +129.9 °C (52 %) | 497 °C MAX | 328 °C | 82.4 °C | 🚨 Stop now |
| 4 | GM06 | 8.2 / 10 | ⬆ +30 °C (8 %) | 472 °C MAX | 353 °C | 80.7 °C | ⚠ First 12 h |
| 5 | GM07 | 7.8 / 10 | ⬆ +20 °C (5 %) | 467 °C MAX | 331 °C | 80.8 °C | ⚠ First 12 h |
| 6 | GM01 | 7.5 / 10 | ⬆ +18 °C (5 %) | 463 °C MAX | 330 °C | 80.5 °C | ⚠ First 12 h |
| 7 | GM16 | 7.2 / 10 | ⬆ +10 °C (2 %) | 528 °C MAX | 310 °C | — | ⚠ 24 h |
| 8 | GM11 | 7.0 / 10 | ⬆ +5 °C (1 %) | 525 °C MAX | 339 °C | — | ⚠ 24 h |
Risk-score visualization
🟠 High risk (24–48 h)
| Motor | Bearing Max | Trend | Action |
|---|
| GM03 | 528 °C | Stable | Maintain in 24 h |
| GM04 | 431 °C | ⚠ Faulty | Bring spare online |
| GM13 | 513 °C | Stable | Maintain in 24–48 h |
| GM31 | 415 °C | Anomaly | Inspect |
| GM10 | 512 °C | Concerning | Maintain in 24–48 h |
🟡 Medium risk (48–72 h)
| Motor | Observation | Action |
|---|
| GM27 | Mechanically locked | Bring spare online |
| GM05 | 30 K hours, stopped | Replacement needed |
| GM23 | Cooling fault | Bring spare online |
| GM25 | 30 K hours of age | Replacement needed |
| GM28 | High standard deviation | Maintain in 48–72 h |
| GM12 | High standard deviation | Maintain in 48–72 h |
| GM14 | High standard deviation | Maintain in 48–72 h |
| GM29 | Low efficiency | Monitor |
ML risk factors
●Bearing-temperature trend (30-day) — fast rise is a failure indicator. GM19: +280 °C (most at risk, 3× jump). GM09: +130 °C (52 % increase). GM02: +40 °C (10 % increase).
●Standard deviation — high variability is a vibration / failure risk. Top 5: GM10 (69 °C), GM05 (56 °C), GM09 (54 °C), GM12 (51 °C), GM14 (45 °C).
●Cooling-system faults are clustering. GM02 and GM09 oil temperatures > 82 °C — suspect a plant-wide cooling or lubrication issue.
Urgent action list (24 h)
1
0–12 h: GM19, GM02, GM09 → STOP. GM04, GM27 → bring spares online. GM23 → inspect.
2
12–24 h: GM06, GM07, GM01, GM16 → fault diagnosis (high bearing temp).
3
Plant-wide: Cooling-system check — GM02, GM09, GM23 spikes are correlated.
4
Lab: Oil analysis (GM02, GM09) — particle count and viscosity check.
5
Vibration measurement: GM10, GM05, GM09 — high standard-deviation cluster.
🛡Data sovereignty: 30 days of historical data were processed through the Cedrix gateway; site name, IP and operator identifiers were anonymized before reaching the AI model. The ML model runs locally; data is not exfiltrated.
