AI in Predictive Maintenance Quiz

Answers and explanations appear as you take the quiz.

1. What problem does predictive maintenance primarily solve compared to preventive maintenance?
   • It avoids replacing healthy parts on a fixed schedule by fixing them only when needed
   • It eliminates the need for any sensors
   • It makes machines run faster
   • It removes the need for technicians entirely
2. What does 'Remaining Useful Life' (RUL) estimate?
   • The total cost of a machine
   • How long a component has left before it fails
   • The energy a motor consumes per hour
   • The number of sensors on a device
3. Which signal is collected in the Case Western bearing fault dataset for rotating-machine diagnostics?
   • Patient electrocardiograms
   • Satellite color channels
   • Vibration acceleration
   • Text sentiment scores
4. How does an autoencoder typically support anomaly detection in PdM?
   • It labels every fault type by name automatically
   • It generates new machines
   • It replaces vibration sensors
   • It is trained only on healthy data and flags high reconstruction error as abnormal
5. What edge-computing approach does the cited condition-monitoring paper study?
   • Federated autoencoder training using vibration data at devices near the monitored machines
   • Centralizing all raw data in a public image database
   • Training a language model on maintenance manuals only
   • Replacing sensors with a periodic questionnaire
6. How does NIST define a digital twin?
   • A maintenance calendar with no model of equipment
   • A virtual representation of a physical or perceived real-world entity
   • A physical duplicate machine kept in storage
   • A guarantee that a machine will not fail
7. What is the NASA C-MAPSS dataset used to evaluate?
   • Cancer grade classification from biopsy slides
   • Speech recognition in noisy rooms
   • Turbofan engine degradation and remaining-useful-life prediction
   • Global climate averages without time series
8. How should an anomaly alert from a predictive-maintenance model be interpreted?
   • As a guaranteed failure date
   • As an automatic repair prescription
   • As evidence that no human review is needed
   • As a signal for inspection, not proof of a specific fault or failure time