The Computational Vibration research team is led by Dr Loic Salles and currently consists of:

  • 1 Research Associate:  Dr Fadi El Haddad
  • 3 PhD Students:  Eve Lian, Jiri Blahos and Mertol Tufekci.
  • Temporary visiting students: Undergraduate, M.Sc.,  and Ph.D levels.

 The main aim of the team’s research is to develop efficient numerical tools for vibration prediction.

Our approaches are based on Finite Element Method, Spectral methods (e.g. the Harmonic Balance Method), Reduced Order Modelling and High Performance Computing for Large Scale Modelling. Current projects are related to:

  • Damping modelling and design of damping devices
  • Geometric Non-linearities
  • Tip-rub events
  • Computational tribology for friction damping prediction
  • Reduced Order Techniques for friction modelling and non-localized nonlinearities
  • High Performance Harmonic Balance Finite Element Method


Fretting-wear and friction modelling of joints


Dr Jason Armand (2013-2017); Dr Luca Pesaresi (2017-2018); Dr Jie Yuan

We have developed a computationally efficient modelling approach which enables considerations of the effect of fretting wear on the nonlinear dynamics. A multiscale strategy is proposed, in which two different time and space scales are used for the contact analysis and dynamic analysis. Thanks to the decoupling of these two analyses, a more realistic representation of the contact interface, which includes surface roughness, is possible. We are also working on a microslip model (e.g. Iwan, Dahl, Valanis and LuGre models; asperity shoulder-to-shoulder contact) for accurate prediction of friction damping in friction joints.


Prof Kai Willner, Friedrich-Alexander-Universität Erlangen-Nürnberg

Oxford UTC 

Harmonic balance method for large scale modelling


Jiri Blahos; Dr Fadi El-Haddad 

We are working on new vibration analysis tool for large-scale finite element models using the multi-harmonic balance method and parallel computing. A  new library in C++ has been developed based on the finite element technique, together with the Fourier Galerkin method using an alternating frequency time procedure. Message passing interface (MPI) programming, direct solvers (MUMPS) and iterative solvers (FETI-like coupled with GMRES) are all employed. Using this approach it is possible to model the frequency response of a full fan blade, considering large deformation of the structure (~ 1.5 million nonlinear equations).


It4i Ostrava (Czech Republic); MSCA-ITN Expertise