基于差分进化算法的转子有限元模型修正

打开文本图片集

中图分类号：TB9；TH133 文献标志码：A文章编号：1674-5124（2025）06-0141-09

Rotor finite element model modification based on differential evolution algorithm

HUANG Tianyi， LIANG Jie， YAO Shuang， ZHANG Hao， SHI Zhanqun （School of Mechanical Engineering，Hebei University of Technology，Tianjin 30o130， China）

Abstract： In the modeling process，constraints imposed by measurement environments and precision often result insignificant discrepancies between specified parameters and actual values of rotor structures，leading to mismatches between theoretical calculations based on the model and measured values. To address this issue， a model correction method has been proposed，combining experimental modalities with the diferential evolution algorithm. Sensitivity analysis of modal frequencies to rotor material parameters have been initially conducted， establishing an objective function based on the confidence levels of modal frequencies and modal shapes.The diferential evolution algorithm has been applied to correct the model's mass and stiffness matrices，followed by the determination of the Rayleigh damping matrix based on experimental modal frequencies and damping ratios.To validate the applicability of the method，the finite element model of a rotor on a sliding bearing experimental setup has been corrected. Comparative analysis with the initial model reveals that the corrected model exhibits modal frequency errors below 0.02% ，and the modal confidence intervals between computed and experimental modal shapes consistently exceed 0.9. To verify the accuracy of the method， frequency response function and unbalanced response simulations have been conducted on the revised model，and the error between the simulation results and experimental results were both less than 2.5% . This method effectively enhances the precision of the rotor system model， providing a practical reference for dynamic analysis and engineering applications.

Keywords： diffrential evolution algorithm; experimental mode; rotor dynamics; model modification

0引言

在能源、电力、航空航天和水运交通等工业领域中，随着对高功率、高转速、高可靠性的大型旋转设备需求日益增长，准确预测和实时监测转子系统的动态行为变得至关重要，因此需要建立具有高保真度的转子有限元模型[1-3]。（剩余10501字）