Inspiration
Industrial machinery failures can lead to costly downtime, unexpected breakdowns, and safety hazards. Predictive maintenance powered by machine learning helps optimize operations, reduce costs, and prevent failures before they happen.
What it does
The project predicts time-to-failure (TTF) for industrial belt systems using sensor data. By analyzing real-time data, it provides proactive failure warnings, allowing for preventive maintenance scheduling. The model predicts the exact time a failure is likely to occur, minimizing operational disruptions.
How we built it
Data Processing: Cleaned and preprocessed time-series sensor data. Feature Engineering: Extracted key time-based and machine-related features. Model Selection: Compared multiple ML models (GBM, LSTM, XGBoost). Final Model: Used XGBoost for its high accuracy, robustness,
Challenges we ran into
Data pre processing
Accomplishments that we're proud of
Multiple model training and results
What we learned
Which model is suitable for which situation
What's next for Capgemini project
Integating maybe a chatbot to warn about the failures
Built With
- python
- xgboost
Log in or sign up for Devpost to join the conversation.