Abstract:
Objectives To address the difficulty in accurately describing centering states under nonlinear and time-varying conditions in existing shaft alignment state monitoring models, a neural network-based method for evaluating centering states is proposed.
Method A BP neural network-based alignment state prediction model was developed. Typical working conditions for acquiring training and testing data were defined, and the data was denoised using a moving average method. The rules for adjusting the model's hyperparameters were summarized. Experimental studies were then carried out on both small and large air spring isolation devices.
Results The results demonstrate that the neural network model can accurately predict the alignment state of isolation devices using only air spring pressure data. It exhibits strong generalizability across different device types, with a prediction error of less than 0.5 and an alignment prediction accuracy of 96.29%.
Conclusion The proposed model does not rely on system parameters and performs well in predicting alignment states for both small and large devices. This provides theoretical support for dynamic alignment state prediction and shaft alignment control of power equipment after startup.
下载: