Digital Twin Based Vibration Performance Monitoring Method for Gear Transmission Systems

In order to monitor the vibration states of the gear transmission systems during operation, a digital twin-based method was proposed for monitoring the vibration performances of the gear transmission systems. Considering the physical model, the digital twin model and the interaction between the two models, a digital twin framework of the gear transmission systems was developed. A mechanism model of the gear transmission systems was presented based on the generalized finite element method, and the simulation data produced were used to construct a surrogate model which was constructed using the simulation data generated by this mechanism model, so as to replace the original computationally expensive mechanism model. The model parameters were selected by the Sobol sensitivity analysis method, and a genetic algorithm was used to update the model in combination with the experimental data. The updated model was then used to monitor the vibration performances, thus completing the construction of the digital twin model. The proposed method was applied to monitor the vibration performances of a herringbone gear transmission system. The results show that the digital twin model may reflect the vibration performances of the gear transmission systems accurately during operation. Under different rotational speeds, the maximum relative error among the simulation and experimental results at each measurement point is as 15.21%, which verifies the effectiveness of the proposed method.