Improved hyper-reduction approach for the forced vibration analysis of rotating components

Abstract

Forced vibration analysis is an indispensable process for the design of a rotating component. However, rather expensive nonlinear static and linear frequency response analyses are usually accompanied by a frequency domain analysis. The traditional mode-superposition method (MSM) effectively reduces the cost of the frequency response analysis. However, the nonlinear static analysis of earlier processes remains as the computational bottleneck. In this paper, the application of the hyper-reduction method will be proposed along with the model order reduction (MOR) framework for rotating component forced vibration analysis. The energy-conserving sampling and weighting (ECSW) method will be employed for the nonlinear iterative computation. The pre-stressed stiffness matrix of the reduced finite elements (FEs) resulting from the ECSW will be used for the post computation stage. Also, a variety of MOR will be attempted for the performance comparison, including MSM, proper orthogonal decomposition (POD)-based reduced order model (ROM), and a hybrid approach. It is found that the present ECSW-combined MOR will significantly relieve the computational bottleneck, provide a minimal loss of accuracy, and be compatible with both nonlinear and linear analyses of the rotating component forced vibration analysis.