Self-directed project
Predictive Maintenance: Industrial ML for Fault Detection
Condition-Monitoring Engineer · AI / ML · August 2026
A ProoV case study · educational project, not employment
Built an end-to-end rolling-bearing fault detector using vibration features, envelope-spectrum energy at the supplied defect frequencies, and a compact decision tree trained only on the training bearings. The candidate also validated the model on a held-out set with a full confusion matrix, selected a cost-based alarm threshold, and correctly localized the fault to the outer race via BPFO dominance.
Graded against
- Feature engineering (signal to honest numbers)25%
- Held-out evaluation (honest measurement)25%
- Cost-based threshold decision + memo25%
- Compact model + embedded-budget justification15%
- Plain-language explanation10%
Passed · pass mark 60/100
What stood out6
- Computed envelope-spectrum energy at the supplied defect frequencies and correctly identified BPFO as dominant.
- Used a depth-capped decision tree with stated depth 1 and 3 nodes, aligned to the embedded-budget constraint.
- Reported a full held-out confusion breakdown with catch rate and false-alarm rate, including all four cells.
- Justified the alarm threshold using the tradeoff between missed failures and false alarms, not raw accuracy.
- Leakage-Free Feature Engineering
- Honest Held-Out Evaluation
The work I submitted19 tasks
11 tasks · 9.5k characters · Python · 8 written answers · 431 words
- Initial Prediction62 characters
- Load Plot Signal CodePython · 1.8k characters
- Load Plot Signal205 characters
- Example Rms WorkedPython · 697 characters
- Time Features143 characters
- Threshold Decision335 characters
- Freq FeaturesPython · 1.2k characters
- Fault Location239 characters
- Train Classifier363 characters
- Evaluation HeldoutPython · 3.8k characters
- Decision Memo739 characters
- Initial Prediction9 words
- Load Plot Signal Code163 words
- Load Plot Signal15 words
- Example Rms Worked72 words
- Time Features10 words
- Threshold Decision26 words
- Freq Features122 words
- Fault Location14 words
Summarised on purpose — the submitted code and writing stay private so this page cannot be reused as an answer key. The full submission sits behind the verified certificate.