AI Predictive Maintenance Simulator

Industrial machines can fail unexpectedly because of high temperature, excessive vibration, and prolonged operation. This project demonstrates an AI-based Predictive Maintenance System that monitors simulated machine parameters and predicts the machine condition as Normal, Warning, or Critical.

The system is developed using Python and a Random Forest machine-learning model. It generates simulated temperature and vibration readings, analyzes the data, and provides maintenance recommendations.

This project is designed as an educational prototype and does not require Arduino or physical sensors.

1. Pictoblox Software: Download PictoBlox | Windows, MacOS, Linux, Chromebook, Android & iOS
2. Laptop/PC
3. Inbuilt or External Camera Setup
4. Speaker or Headphones

Unexpected machine failures can stop industrial production and increase maintenance costs. Traditional maintenance may depend on fixed schedules or manual inspection.

Our idea is to continuously monitor machine parameters and use AI to identify abnormal conditions early.

The program creates 1000 simulated machine records. Each record contains temperature and vibration values.

The Random Forest Classifier learns the relationship between temperature, vibration, and the simulated machine condition.

The program continuously generates new machine readings to simulate a machine operating in real time.

The latest temperature and vibration readings are passed to the trained AI model. The model predicts whether the machine is Normal, Warning, or Critical.

Depending on the AI prediction, the system provides an appropriate maintenance recommendation.

The prototype successfully monitors simulated machine parameters and uses a Random Forest model to classify the machine condition. It demonstrates how AI-assisted predictive maintenance can help identify abnormal machine conditions and recommend maintenance actions.

• Connect real temperature sensors

• Connect vibration sensors

• Add Arduino/ESP32 in future

• Use real industrial datasets

• Add live dashboard

• Store historical machine data

• Predict remaining useful life (RUL)

• Send automatic maintenance alerts

• Add cloud monitoring

This project demonstrates an AI-based approach to predictive maintenance using Python. Instead of waiting for a machine to fail, the system continuously analyzes machine-condition data and identifies possible abnormal conditions. The prototype provides a simple foundation for developing real-world smart factory maintenance systems.