블레이드 동특성 근사모델의 확률 분석을 통한 블레이드 구조 설계

Abstract

For a rotor system in which the hub/control design and blade OML were determined, the internal structural shape of the blade was designed using the dynamic characteristic surrogate model to maximize the difference between the rotating natural frequency and the integral multiple of the rotational operating frequency. Based on the general blade structure design procedure, an analysis procedure for generating a dynamic characteristic surrogate model was established, and design variables were selected to define the shapes of the internal components. To generate a dynamic characteristic surrogate model, a space filling design method, a type of design of experiment, was applied to generate design variable input values for database configuration. In this process, a range of design variables and constraint conditions suitable for blade design conditions were considered. Based on the analysis results of the finite element 2D sectional properties for various shape input conditions, a 1D beam model dynamic analysis was performed to build a rotation natural frequency database. Using this database, a dynamic characteristic surrogate model for the blade structure design was generated. A regression model and an artificial neural network model were used as surrogate models, and the 1st ~ 4th modes of the blade rotation natural frequency were predicted with the surrogate model and compared with the actual analysis results. Additionally, the values of design variables that can satisfy the dynamic characteristic requirement were selected through probability analysis of the two dynamic characteristic surrogate models, and the accuracy of the predicted values was also confirmed by comparing the analysis results.