Predictive Maintenance on NASA Turbofan Data

A machine-learning study predicting the remaining useful life (RUL) of turbofan engines from NASA’s run-to-failure degradation dataset, a benchmark problem in predictive maintenance.

What it does

• Models multivariate sensor time-series to estimate how many cycles remain before an engine fails.
• Covers the full workflow: data preparation, feature engineering on the sensor channels, model training, and evaluation against the RUL ground truth.
• Aimed at the practical maintenance question: catch degradation early enough to act, without pulling healthy units offline too soon.

Code on GitHub.