Souhardya Bhattacharjee completed Predictive Maintenance: Industrial ML for Fault Detection

Built an end-to-end rolling-bearing fault detector from vibration data: visualized the faulty waveform, engineered RMS/kurtosis/crest-style signal features plus envelope-spectrum energy at the supplied defect frequencies, and localized the fault to the outer race via BPFO dominance. Trained a compact depth-capped decision tree on training bearings only, then evaluated it on held-out bearings with reported catch-rate and false-alarm rate and defended a cost-based alarm threshold using the candidate’s own numbers.

• Leakage-Free Feature Engineering
• Honest Held-Out Evaluation
• Physics-Based Fault Localization
• Cost-Based Threshold Selection