导航

ACTA AERONAUTICAET ASTRONAUTICA SINICA ›› 2015, Vol. 36 ›› Issue (10): 3305-3315.doi: 10.7527/S1000-6893.2015.0025

• Solid Mechanics and Vehicle Conceptual Design • Previous Articles     Next Articles

Reduced-order modelling and dynamic response prediction method for mistuned bladed disks

ZANG Chaoping1, DUAN Yongliang1, E. P. PETROV2   

  1. 1. Jiangsu Province Key Laboratory of Aerospace Power System, College of Energy and Power, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China;
    2. School of Engineering and Informatics, University of Sussex, Brighton BN1 9QT, United Kingdom
  • Received:2014-10-20 Revised:2014-11-13 Online:2015-10-15 Published:2015-03-25
  • Supported by:

    National Natural Science Foundation of China (11372128, 51175244); The Fundamental Research Funds for the Central Universities (NP2013302, NN2012067); Fundation of Graduate Innovation Center in NUAA (kfjj20130204); Project of Collaborative Innovation Center for Advance Aero-engine

Abstract:

In order to save the great computation effort for high fidelity mistuned bladed disk, a new reduced-order modelling and dynamic response prediction method for mistuned bladed disks is developed in this paper. Taking full advantage of cyclic symmetry of tuned bladed disk, a sector model is used to provide the primary information about dynamic properties in the local coordinate systems. Just doing model analysis on a small number of nodes utilizing the concept of "active nodes", the dimensions of operation matrices are reduced enormously but the model properties of mistuned structure are preserved with very high accuracy. In the process of dynamic response prediction, the response properties on the natural frequencies can be used for the choice of narrow frequency ranges, dangerous blades and nodes which need to calculate response. The main advantages are its efficiency and accuracy, which allow the high precision maximum forced responses. The results show that compared with traditional finite element method, the inverse matrices in this method are reduced from 1.5 million to 384, but the errors of the first 50 natural frequencies are limited in 0.005%. In addition, the computational complexity of maximum response calculation is lowered by more than 99%, which just produces -0.35% deviation.

Key words: mistuning, bladed disk, reduced-order modelling, dynamic, response prediction, active nodes

CLC Number: