How tavas-softech/apms works
BMP-II Engine Health AI is a predictive-maintenance codebase for monitoring UTD-20 / BMP-II engines. Its tree indicates a workflow spanning sensor and vehicle-data acquisition, signal preparation and synchronization, feature generation, anomaly/fault/health and remaining-life analysis, then decision support and reporting. Training, inference, and evaluation have dedicated pipeline modules. The sampled source excerpts are empty, so component-to-component calls and data transfers cannot be verified; the graph therefore shows capabilities and stages without asserting wiring.
Components 24
Data acquisition
- CAN interface
src/apms_core/acquisition/can_interface.py - DAQ interface
src/apms_core/acquisition/daq_interface.py - Sensor interface
src/apms_core/acquisition/sensor_interface.py - Data logger
src/apms_core/acquisition/data_logger.py
Signal preparation
- Preprocessing
src/apms_core/preprocessing - Time and RPM sync
src/apms_core/synchronization - Operating regime
src/apms_core/operating_regime - Data quality
src/apms_core/data_quality
Engine analytics
- Feature engineering
src/apms_core/feature_engineering - Training pipeline
src/apms_core/pipeline/training_pipeline.py - Inference pipeline
src/apms_core/pipeline/inference_pipeline.py - Evaluation pipeline
src/apms_core/pipeline/evaluation_pipeline.py - Anomaly detection
src/apms_core/anomaly_detection - Fault classification
src/apms_core/fault_classification - Health assessment
src/apms_core/health_assessment - Degradation tracking
src/apms_core/degradation - Remaining useful life
src/apms_core/rul - Model explanations
src/apms_core/explainability
Decision and reporting
- Decision support
src/apms_core/decision_support - Health and maintenance reports
src/apms_core/reporting
Persistence
- Maintenance records
src/apms_core/database/maintenance_repository.py - Prediction records
src/apms_core/database/prediction_repository.py - Engine records
src/apms_core/database/engine_repository.py
Other components
- Engine sensors