Stochastic Finite Element Analysis for High Speed Rotors

1993 ◽  
Vol 115 (1) ◽  
pp. 59-64 ◽  
Author(s):  
T. S. Sankar ◽  
S. A. Ramu ◽  
R. Ganesan

The general problem of the dynamic response of highspeed rotors is considered in which certain system parameters may have a spatial stochastic variation. In particular the elastic modulus and mass density of a rotating shaft are described through one dimensional stochastic field functions so that the imperfections in manufacture and measurement can be accounted for. The stochastic finite element method is developed so that the variability of the response of the rotor can be interpreted in terms of the variation of the material property. As an illustration the whirl speed analysis is performed to determine the stochastics of whirl speeds and modes through the solution of a random eigenvalue problem associated with a non self-adjoint system. Numerical results are also presented.

Author(s):  
Seyed Hamid Reza Sanei ◽  
Randall Doles

Abstract The aim of this study is to present a representative volume element (RVE) for nanocomposites with different microstructural features using a stochastic finite element approach. To that end, computer-simulated microstructures of nanocomposites were generated to include a variety of uncertainty present in geometry, orientation, and distribution of carbon nanotubes. Microstructures were converted into finite element models based on an image-based approach for the determination of elastic properties. For each microstructure type, 50 realizations of synthetic microstructures were generated to capture the variability as well as the average values. Computer-simulated microstructures were generated at different length scales to determine the change in mechanical properties as a function of length scale. A representative volume element is defined at a length scale beyond which no change in variability is observed. The results show that there is no universal RVE applicable to all properties and microstructures; however, the RVE size is highly dependent on microstructural features. Microstructures with agglomeration tend to require larger RVE. Similarly, random microstructures require larger RVE when compared with aligned microstructures.


Sign in / Sign up

Export Citation Format

Share Document