Fault diagnosis of ship motor bearings based on multi domain information fusion and improved ELM

Abstract：Objectives Aiming at the problems that the symptom parameters of the monitoring signal in a single analysis domain are difficult to completely characterize the running state of the monitoring object, and the model parameters of the Extreme Learning Machine (ELM) network are difficult to achieve the optimization, a fault diagnosis method of ship motor bearing is proposed based on multi-domain information fusion and improved ELM. Methods Firstly, based on the feature information of ship motor bearing vibration signal in time domain, frequency domain and time-frequency domain, a multi-domain feature parameter set was constructed as the input of the fault diagnosis model. Then, the sparrow search algorithm was used to improve the model parameter optimization method of ELM network, determine the optimal weights and thresholds, and improve the recognition accuracy of ELM model for fault diagnosis. Finally, the fault state of motor bearing was identified through the experimental data of self-made test bench and open source experimental data. Firstly, based on the feature information of ship motor bearing vibration signal in time domain, frequency domain and time-frequency domain, a multi-domain feature parameter set was constructed as the input of the fault diagnosis model. Then, the sparrow search algorithm was used to improve the model parameter optimization method of ELM network, determine the optimal weights and thresholds, and improve the recognition accuracy of ELM model for fault diagnosis. Finally, the fault state of motor bearing was identified through the experimental data of Marine motor test bench and open source experimental data. Results The experimental data verification based on the Marine motor test bench shows that the recognition accuracy of the fault diagnosis model using multi-domain feature parameter sets is 100% on the training set and the test set. The verification based on open source experimental data shows that the test set recognition accuracy of the improved ELM model is 90.5%, which is 12.7% higher than that of the original ELM model, and the training set recognition accuracy and test set recognition accuracy are higher than other diagnostic models. Conclusions This study has improved the input symptom parameter set and diagnosis model. The proposed method can effectively identify the fault state of motor bearing, and the model has good stability, which provides reference for the fault diagnosis of ship motor bearing.