Model updating of rotor system based on the adaptive Gaussian process model using unbalance response

For predicting the dynamic responses of rotor system with high accuracy, an adaptive Gaussian process model-based model updating method for rotor system using the unbalance responses is proposed. To utilize the data from multiple sensors, multi-channel data fusion is adopted to construct the objective function of the model updating. The fusion feature is the mean of the cumulative distribution function of the response errors of each channel. The Gaussian process model is used to describe the mapping relationship between the selected parameters and fusion features. An adaptive sampling strategy is developed to reduce the computation cost of repeated finite element calculation. Based on Sobol sequence sampling, new sampling parameters are added near the sample parameter with a larger error in the predicted value. The performance of the proposed method is validated by a gas-generator rotor and a dual-disks rotor. Results show that the unbalance responses of the updated model show good agreement with the exact responses and measured responses. The maximum error of the support parameters for the gasgenerator rotor is 3.56%. For the dual-disks rotor, the discrepancies of the first three modal frequencies between the updated model and the measured model are below 1%, 2.5% and 5%. The proposed method is effective for model updating of the rotor system.

Automated summary

An adaptive Gaussian process model-based model updating method for rotor system is proposed in this study, which utilizes unbalance responses data and multi-channel data fusion to construct the objective function of model updating, effectively reducing the computation cost of repeated finite element calculation. Experimental results show that the proposed method has good accuracy and feasibility.