Bearing Fault Detection and Classification using Signal Processing and Machine Learning

This project focuses on detecting and classifying faults in ball bearings using a combination of signal processing techniques and machine learning models. The dataset includes vibration signals collected from an induction motor-driven system under different fault conditions. Key methods include pre-processing, feature extraction using techniques like Hilbert Transform, and the application of machine learning algorithms to classify bearing health conditions.

Key Features:
- Signal Pre-processing: Applied filtering and noise reduction techniques to the raw vibration signals for accurate feature extraction.
- Feature Extraction: Utilized advanced methods like Hilbert Transform, Envelope Analysis, and Morphological Analysis to capture key characteristics of bearing vibration signals.
- Machine Learning Models: Employed models such as Support Vector Machines (SVM), K-means clustering, and decision trees to classify various bearing fault conditions, including outer race, inner race, and ball faults.
- Cross-Validation and Performance Metrics: Used cross-validation techniques to evaluate model performance and ensured high classification accuracy through performance metrics like precision, recall, and F1 score.
- Automated Fault Diagnosis: The system is designed for real-time fault detection and classification, making it suitable for predictive maintenance in industrial applications.