Self-directed project
Predictive Maintenance: Industrial ML for Fault Detection
Condition-Monitoring Engineer · AI / ML · July 2026
A ProoV case study · educational project, not employment
Built an end-to-end rolling-bearing fault detector using vibration snapshots, time-domain statistics, envelope-spectrum defect-band analysis, and a compact decision tree trained only on the training bearings. Evaluated it on a held-out bearing set with reported catch-rate and false-alarm rate, then defended a 0.53 alarm threshold in a maintenance memo that named the outer-race fault from the dominant BPFO frequency.
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
- Correctly identified kurtosis as the strongest time-domain fault indicator
- Trained a shallow decision tree on training bearings only and stated depth/node count against the embedded budget
- Reported held-out catch rate and false-alarm rate rather than evaluating on training data
- Wrote a memo that ties the threshold choice to concrete catch-rate, false-alarm, and cost figures
- Leakage-Free Feature Engineering
- Honest Model Evaluation
The work I submitted19 tasks
11 tasks · 7.7k characters · Python · 8 written answers · 487 words
- Initial Prediction62 characters
- Example Load Plot HealthyPython · 859 characters
- Load Plot Signal CodePython · 1.8k characters
- Load Plot Signal205 characters
- Example Rms WorkedPython · 698 characters
- Time Features143 characters
- Freq FeaturesPython · 1.1k characters
- Train Classifier363 characters
- Evaluation Heldout288 characters
- Threshold Decision334 characters
- Decision Memo1.9k characters
- Initial Prediction9 words
- Example Load Plot Healthy81 words
- Load Plot Signal Code163 words
- Load Plot Signal15 words
- Example Rms Worked72 words
- Time Features10 words
- Freq Features106 words
- Train Classifier31 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.