Improving on NASA's work with induction motor bearing fault detection using RNN-powered smart sensors.
Improving on NASA's work with induction motor bearing fault detection using RNN-powered smart sensors.
For starters, you'll want to run source setup_venv.sh
to automatically setup a Python virtual environment under bearing_venv
. You may want to experiment with different versions of analysta
(to be found in the anomaly_detection
submodule) to make sure training works properly.
Then:
bearing-fault-detection/data
, then cd bearing-fault-detection
and run python3 preprocess_data.py
.cd bearing-fault-detection
and run analysta -vv model single -c lstm_config.json
.cd bearing-fault-detection
and run python3 spectrogram.py
.