Fig 1.
Factors affecting transformer faults.
Fig 2.
Architecture diagram of state monitoring system.
Fig 3.
Flow chart of model training.
Fig 4.
Assessment process.
Fig 5.
Typical current timing curve.
Fig 6.
Vibration signal.
Table 1.
Five typical operation status categories.
Fig 7.
2D.CNN-BiLSTM neural network model.
Fig 8.
(a) Loss function. (b) Training accuracy and test accuracy.
Table 2.
Performance comparison of different models.
Fig 9.
Comparison of Model Performance.
Fig 10.
ROC curve.
Table 3.
Ablation experiment results.
Table 4.
Assessment results of each state category.