Numerical characterization of turbine blade dynamic properties via proper orthogonal decomposition

Forced vibration analysis is an essential process during the design of a turbine blade and requires enormous computational expense because of its complicated numerical procedures with large size matrices. This paper focuses on the application of the proper orthogonal decomposition (POD). It is under the spotlight for its usefulness on the reduced order model (ROM) construction. POD applications include nonlinear static, modal analysis, and forced vibration (harmonic response) analysis of the turbine blade. Performance of POD-ROM is examined for various operating conditions, and effect of the number of POD modes used is also investigated. With the sufficient number of POD modes, accurate results of turbine blade response are obtained with drastically reduced computational resources.