A novel hybrid Neumann expansion method for stochastic analysis of mistuned bladed discs. (May 2016)

• Record Type: Journal Article
• Authors: Yuan, Jie; Allegri, Giuliano; Scarpa, Fabrizio; Patsias, Sophoclis; Rajasekaran, Ramesh
• Is Part Of: Mechanical systems and signal processing. Volume 72/73(2016)
• Page Start: 241
• Page End: 253
• Publication Date: 2016-05
• Subjects: Mistuned bladed disc -- Stochastic analysis -- Neumann expansion; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2015.11.011
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: The paper presents a novel hybrid method to enhance the computational efficiency of matrix inversions during the stochastic analysis of mistuned bladed disc systems. The method is based on the use of stochastic Neumann expansion in the frequency domain, coupled with a matrix factorization in the neighbourhood of the resonant frequencies. The number of the expansion terms is used as an indicator to select the matrix inversion technique to be used, without introducing any additional computational cost. The proposed method is validated using two case studies, where the dynamics an aero-engine bladed disc is modelled first using a lumped parameter approach and then with high-fidelity finite element analysis. The frequency responses of the blades are evaluated according to different mistuning patterns via stiffness or mass perturbations under the excitation provided by the engine orders. Results from standard matrix factorization methods are used to benchmark the responses obtained from the proposed hybrid method. Unlike classic Neumann expansion methods, the new technique can effectively update the inversion of an uncertain matrix with no convergence problems during Monte Carlo simulations. The novel hybrid method is more computationally efficient than standard techniques, with no accuracy loss. Highlights: Novel hybrid method for matrix inversions during the stochastic analysis of mistuned bladed disc systems. Technique couples stochastic Neumann expansion in frequency domain and matrix factorization close to resonance. Applied to reduced order and full scale mistunded blade disc finite element models. Results benchmarked against standard LU factorization.