This project develops a Machine Learning-based predictive maintenance system that analyzes IoT sensor data (temperature, vibration, and current) to predict machine failures before they occur.
The system helps industries shift from reactive maintenance to predictive maintenance, reducing downtime and improving operational efficiency.
Traditional maintenance systems repair machines only after failure, leading to:
- Unexpected downtime
- High maintenance costs
- Reduced productivity
This project solves the problem by predicting failures in advance using AI.
Predictive maintenance is widely used in:
- Manufacturing plants
- Smart factories
- Automotive industry
- Power generation systems
Leading companies like Siemens, General Electric, and Tesla use similar systems to optimize operations.
- Python
- Pandas
- NumPy
- Scikit-learn
- Matplotlib
- Joblib
A synthetic IoT dataset was created to simulate real-world machine conditions.
- Temperature – Detects overheating
- Vibration – Indicates mechanical stress
- Current – Monitors electrical behavior
-
Failure (0/1)
- 0 → Normal operation
- 1 → Machine failure
- Data collection (simulated IoT data)
- Data preprocessing
- Feature selection
- Model training using Random Forest
- Prediction of machine failure
- Model evaluation
git clone <your_repo_link>
cd AI-Predictive-Maintenance-IoTpip install -r requirements.txtpython src/main.pyAI-Predictive-Maintenance-IoT/
│
├── notebooks/ # Colab notebook
├── data/ # Dataset
├── models/ # Trained model
├── outputs/ # Graphs & results
├── src/ # Main script
├── README.md
├── requirements.txt
- Understanding predictive maintenance systems
- Machine Learning model building
- Model evaluation techniques
- Data visualization
- Real-world IoT simulation
- Real-time IoT sensor integration
- Streamlit dashboard for visualization
- Cloud deployment (AWS / Azure)
Vaishnava Devi
If you found this project useful, consider giving it a star ⭐ on GitHub!