Intelligent Predictive Maintenance Using Sensor Analytics for Remaining Useful Life Estimation and Proactive Maintenance

Intelligent Predictive Maintenance Using Sensor Analytics is a machine learning and deep learning-based predictive maintenance system for estimating the Remaining Useful Life (RUL) of turbofan engines using sensor data from the NASA C-MAPSS FD001 dataset. The system performs data preprocessing, causal feature engineering, and engine-level validation to prevent data leakage. Multiple models, including Ridge Regression, Random Forest, XGBoost, and a PyTorch LSTM, are evaluated using RMSE, MAE, R², and NASA PHM08 metrics. The predicted RUL is further converted into configurable Normal, Warning, and Critical maintenance-risk levels with corresponding maintenance recommendations. A Streamlit dashboard provides fleet monitoring, engine-level analysis, sensor trends, model comparison, and maintenance-priority visualization. The system demonstrates an end-to-end approach from sensor data and RUL prediction to actionable maintenance decision support. If the description box has a short character limit: A predictive maintenance system that uses NASA C-MAPSS sensor data to estimate Remaining Useful Life (RUL), compare multiple ML/DL models, classify maintenance risk, and provide maintenance recommendations through an interactive Streamlit dashboard. This wording stays consistent with the project you actually implemented and avoids claiming unmeasured cost or downtime savings. Your original project material describes the goal as moving from reactive toward proactive maintenance, while the completed implementation adds the validated multi-model and maintenance-decision pipeline.