NASA-RUL
Turbofan Prognostics

Analyze Turbofan Engine Remaining Useful Life

Run multivariate degradation inference using the trained 2-layer stacked LSTM model. Requires 30 consecutive operational cycles across the 19 selected sensor channels.

LSTM Model Active

Option A: Instant Sample Evaluation

Samples a real contiguous 30-cycle sequence from the NASA FD001 benchmark and evaluates it with the saved LSTM.

How Sample Generation Works

Picks a random engine from the FD001 dataset and extracts a contiguous 30-cycle operational window containing strictly the 19 sensor features (thermal, spool speed, pressure). Input telemetry is passed to the real backend and scaled with the trained StandardScaler.

* Note: Sampled data is for benchmark evaluation, not measurements from a currently operating engine.