An optimized CNN-BiLSTM network for bearing fault diagnosis under multiple working conditions with limited training samples.
Baoye SongYiyan LiuJingzhong FangWeibo LiuMaiying ZhongXiaohui LiuPublished in: Neurocomputing (2024)
Keyphrases
- fault diagnosis
- training samples
- monitoring and fault diagnosis
- operating conditions
- expert systems
- fault detection
- neural network
- fault detection and diagnosis
- chemical process
- high dimensional
- training set
- electronic equipment
- learning algorithm
- test sample
- feature space
- number of training samples
- power transformers
- analog circuits
- fuzzy logic
- supervised learning
- bp neural network
- rbf neural network
- training data
- multi sensor information fusion
- multiple faults
- rotating machinery
- image processing
- gas turbine
- power plant
- multiscale
- tennessee eastman
- feature extraction