How karabogaerdi/bearing-physics-features works
This repository is a bearing-fault diagnosis research codebase: scripts compare engineered vibration features and a raw-signal CNN across Ottawa and CWRU datasets, using random-window, held-out-bearing, and cross-dataset evaluation. Shared utilities cover record loading, signal processing, and feature extraction; experiment scripts own model fitting and metrics. The README documents dataset acquisition, but the referenced download script and some experiment capabilities are not excerpted, so those relationships are not expanded beyond documented behavior.
Main entry points: ortak.py Data utilities · hafta5_cnn.py CNN evaluation · hafta4_veri_setleri_arasi.py Dataset transfer · hafta5_maliyet.py Feature cost
Components 13
Bearing data
- Ottawa dataset
- CWRU dataset
Signal features
- Data utilities
ortak.py - Ottawa exploration
ottawa_eda.py
Model experiments
- Baseline evaluation
hafta2_temel_model.py - Order features
hafta3_fizik_ozellikleri.py - Dataset transfer
hafta4_veri_setleri_arasi.py - Feature cost
hafta5_maliyet.py - Combined experiments
final_deneyler.py - Result figures
final_sekiller.py
CNN baseline
- CNN evaluation
hafta5_cnn.py - WDCNN model
cnn.py
Other components
- Researcher
Connections 25
- Researcher → Baseline evaluation: runs
hafta2_temel_model.py - Researcher → Order features: runs
hafta3_fizik_ozellikleri.py - Researcher → Dataset transfer: runs
hafta4_veri_setleri_arasi.py - Researcher → Feature cost: runs
hafta5_maliyet.py - Researcher → CNN evaluation: runs
hafta5_cnn.py - Researcher → Combined experiments: runs
final_deneyler.py - Ottawa dataset → Data utilities: loads records
ortak.py - CWRU dataset → Data utilities: loads records
ortak.py - Ottawa dataset → Ottawa exploration: examines signals
ottawa_eda.py - Data utilities → Baseline evaluation: provides features
hafta2_temel_model.py - Data utilities → Order features: provides features
hafta3_fizik_ozellikleri.py - Data utilities → Dataset transfer: provides features
hafta4_veri_setleri_arasi.py - Data utilities → Feature cost: provides features
hafta5_maliyet.py - Data utilities → Combined experiments: provides features
final_deneyler.py - Data utilities → CNN evaluation: loads windows
hafta5_cnn.py - CNN evaluation → WDCNN model: trains and predicts
hafta5_cnn.py - Combined experiments → CNN evaluation: reads predictions
final_deneyler.py - Combined experiments → Result figures: supplies results
- Baseline evaluation → Ottawa dataset: evaluates
- Dataset transfer → CWRU dataset: evaluates transfer
hafta4_veri_setleri_arasi.py - Dataset transfer → Ottawa dataset: evaluates transfer
hafta4_veri_setleri_arasi.py - Feature cost → CWRU dataset: evaluates
hafta5_maliyet.py - Feature cost → Ottawa dataset: evaluates
hafta5_maliyet.py - CNN evaluation → CWRU dataset: evaluates transfer
hafta5_cnn.py - CNN evaluation → Ottawa dataset: evaluates
hafta5_cnn.py