How farah-belghith2710/projet_fm works
This repository is a comparative bearing-failure prognostics study: vibration records become signal features, health indicators and remaining-useful-life predictions, then reliability and maintenance economics results. The principal workflow is split between feature extraction and a results pipeline; a Streamlit dashboard presents the precomputed outputs. Major documented capabilities also include complementary statistical/XAI analyses and report figures. Edges follow the shown calls and README workflow; some result-file and dashboard wiring is documented rather than visible in sampled implementation excerpts.
Main entry points: 2_lancer_pipeline.py Results pipeline · commun.py Shared dashboard tools · 4_analyses_complementaires.py Complementary analyses · results Analysis results
Components 27
Signal preparation
- Study configuration
src/config.py - Signal reader
src/data.py - Signal filtering
src/preprocessing.py - Feature extraction
src/features.py
Health and reliability
- Health indicators
src/health.py - Population reliability
src/reliability.py
RUL modeling
- Bayesian Wiener
src/rul_wiener.py - Classical ML
src/rul_ml.py - Deep learning
src/rul_deep.py
Results and decisions
- Results pipeline
2_lancer_pipeline.py - Model evaluation
src/evaluation.py - Analysis results
results - Complementary analyses
4_analyses_complementaires.py - Maintenance economics
src/economie.py - Report figures
3_figures_rapport.py
Interactive dashboard
- Dashboard app
dashboard/app.py - Shared dashboard tools
dashboard/commun.py - Data and reliability
dashboard/pages/donnees.py - Health page
dashboard/pages/sante.py - Prognosis page
dashboard/pages/pronostic.py - Model comparison
dashboard/pages/comparaison.py - Economics page
dashboard/pages/economie.py - Overview page
dashboard/pages/accueil.py - Challenge page
dashboard/pages/tests.py
Other components
- Researcher
- PRONOSTIA signals
- Processed data
data/processed
Connections 33
- Researcher → PRONOSTIA signals: provides records
README.md - PRONOSTIA signals → Signal reader: supplies signals
README.md - Signal reader → Study configuration: uses settings
src/data.py - Signal reader → Signal filtering: filters signals
1_extraire_features.py - Signal filtering → Feature extraction: prepares signals
1_extraire_features.py - Feature extraction → Study configuration: uses settings
src/features.py - Feature extraction → Processed data: writes features
README.md - Researcher → Results pipeline: runs analysis
README.md - Results pipeline → Processed data: reads features
README.md - Results pipeline → Health indicators: computes health
2_lancer_pipeline.py - Results pipeline → Population reliability: computes reliability
2_lancer_pipeline.py - Results pipeline → Bayesian Wiener: runs Wiener model
2_lancer_pipeline.py - Results pipeline → Classical ML: runs ML models
2_lancer_pipeline.py - Results pipeline → Deep learning: runs deep models
2_lancer_pipeline.py - Results pipeline → Model evaluation: evaluates predictions
2_lancer_pipeline.py - Results pipeline → Analysis results: writes outputs
README.md - Researcher → Complementary analyses: runs analyses
README.md - Complementary analyses → Population reliability: calls reliability methods
4_analyses_complementaires.py - Complementary analyses → Health indicators: calls health stage
4_analyses_complementaires.py - Complementary analyses → Maintenance economics: compares strategies
4_analyses_complementaires.py - Complementary analyses → Analysis results: reads and extends
README.md - Analysis results → Report figures: provides data
README.md - Researcher → Dashboard app: opens dashboard
README.md - Dashboard app → Shared dashboard tools: uses shared tools
README.md - Shared dashboard tools → Analysis results: loads outputs
README.md - Data and reliability → Shared dashboard tools: uses shared tools
dashboard/pages/donnees.py - Health page → Shared dashboard tools: uses shared tools
dashboard/pages/sante.py - Prognosis page → Shared dashboard tools: uses shared tools
dashboard/pages/pronostic.py - Model comparison → Shared dashboard tools: uses shared tools
dashboard/pages/comparaison.py - Economics page → Shared dashboard tools: uses shared tools
dashboard/pages/economie.py - Overview page → Shared dashboard tools: uses shared tools
dashboard/pages/accueil.py - Challenge page → Shared dashboard tools: uses shared tools
dashboard/pages/tests.py - Economics page → Maintenance economics: uses cost model
dashboard/pages/economie.py