Optimized model for gearbox fault detection: a Random Forest approach using vibration data

Авторы

Аннотация

This study presents the development and evaluation of a Random Forest classifier for predicting gearbox faults using engineered features from vibration data. The model training process, including data preparation, stratified sampling, and hyperparameter tuning, is detailed, and a comprehensive performance analysis is provided. The results demonstrate the effectiveness of the Random Forest algorithm in distinguishing be- tween normal, cracked, and broken gearbox states with high precision and recall. This approach shows promise for practical implementation in predictive maintenance sys- tems, reducing downtime and maintenance costs in industrial settings.