Error localization of finite element updating model based on element strain energy

13th International Conference on Motion and Vibration Control; 12th International Conference on Recent Advances in Structural Dynamics

Huang, Zi^1 ; Zang, Chaoping^1 ; Wang, Xiaowei^1 ; Jiang, Yuying^1

Nanjing University of Aeronautics and Astronautics, Jiangsu Province Key Laboratory of Aerospace Power System, Nanjing

210016, China^1

关键词: Aeroengine combustors;    Design parameters;    Error localization;    Finite element updating;    Maximum frequency;    Modal strain energy;    Object functions;    Parameter selection;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/744/1/012113/pdf
DOI  :  10.1088/1742-6596/744/1/012113

来源: IOP

【 摘 要 】

An error localization indicator based on modal strain energy changes is proposed and used for selecting design parameters to be updated in model updating process. Taking an aeroengine combustor casing structure as an example, the 'supermodel' of combustor casing was established and validated with the test data and the reduced model (also called the design model) was built with the simplification of modelling. By comparing the modal strain energy changes between 'supermodel' and design model of combustor casing, the error locations of the reduced combustor casing modelling was highlighted by the error localization indicator. Then, the updating parameters of the design model were selected as the areas with significant variations of modal strain energy changes based on the error localization indicator. Defining the updating object function with the minimum of natural frequency errors between the FE model prediction and the modal test data, model updating of the design combustor casing model based sensitivity analysis method was carried out using the experimental modal data. After model updating, the maximum frequency error of the first ten modes was decreased from 27.1% to 1.2%, compared with the test data. The result shows the effectiveness of the proposed method and certain significance in parameter selection for model updating.