- Core codes for the paper "FGDAE: A new machinery anomaly detection method towards complex operating conditions"
- Created by Shen Yan, Haidong Shao, Zhishan Min, Jiangji Peng, Baoping Cai, Bin Liu.
- Journal: Reliability Engineering and System Safety
- Python 3.8
- pytorch 1.10.1
- and other necessary libs
- This repository provides a concise framework for machinery anomaly detection.
- It includes the pre-processing and graph composition process for the data and the model proposed in the paper.
- We have also integrated 4 baseline methods for comparison.
Graph_train_val_test.py
is the train&val&test process of our proposed method;Base_train_val_test.py
is the train&val&test process of base methods.- You need to load the data in following Datasets link at first, and put them in the
data
folder. Then run inGeneral_procedure.py
- You can also adjust the structure and parameters of the model to suit your needs.
General_procedure.py
--data_dir "./data/Case1"; --data_num ['200Hz_0N', '300Hz_1000N', '400Hz_1400N'];
--sensor_number 6; --fault_num 7; --unbalance_train [200, 100, 10]
General_procedure.py
--data_dir "./data/Case2"; --data_num ['G_20_0', 'G_30_2'];
--sensor_number 8; --fault_num 5; --unbalance_train [200, 10]
data
needs loading the Datasets in above linksdatasets
contians the pre-processing and graph composition process for the datamodels
contians the proposed model and 4 base modelsutils
contians two types of train&val&test processes
If our work is useful to you, please cite the following paper, it is the greatest encouragement to our open source work, thank you very much!
@paper{FGDAE,
title = {FGDAE: A new machinery anomaly detection method towards complex operating conditions},
author = {Shen Yan, Haidong Shao, Zhishan Min, Jiangji Peng, Baoping Cai, Bin Liu},
journal = {Reliability Engineering and System Safety},
volume = {236},
pages = {109319},
year = {2023},
doi = {https://doi.org/10.1016/j.ress.2023.109319},
url = {https://www.sciencedirect.com/science/article/abs/pii/S0951832023002338},
}