Residual network based rolling bearing fault diagnosis of mine ventilator in high noise scenario
Aiming at the difficulty in rolling bearing fault diagnosis for key components of mine ventilator in high noise environment,an end-to-end bearing fault diagnosis method based on residual network is proposed.By introducing the residual learning framework,the robustness of the network to noise is enhanced.The model is systematically verified by using the bearing vibration data set of Baode Coal Mine,and the key hyperparameters affecting the performance of the model are further determined.The optimization of convolution kernel size and batch size provides higher diagnostic accuracy and stronger noise anti-interference ability for the model.The results show that the fault diagnosis rate of the ResNet34 model is close to 90%in a noisy environment with a signal-to-noise ratio of 0,and the accuracy rate can exceed 54%even when the signal-to-noise ratio is reduced to-10 dB.It provides an effective solution for the fault diagnosis of ventilator rolling bearing in complex mine environment,improves the accuracy,timeliness and reliability of bearing fault diagnosis,ensures the reliable operation of equipment and the sustainable production,and is of great significance to ensure the safe production of mines.
NETL
NSTL
万方数据
