Application of the dynamic condensation approach to the hybrid FE-SEA model of mid-frequency vibration in complex built-up systems. (February 2020)

• Record Type: Journal Article
• Authors: Gao, Ruxin; Zhang, Yahui; Kennedy, David
• Is Part Of: Computers & structures. Volume 228(2020)
• Publication Date: 2020-02
• Subjects: Mid-frequency -- Complex built-up systems -- Hybrid FE-SEA method -- Dynamic condensation approach -- Dynamic flexibility; Structural engineering -- Data processing -- Periodicals; Electronic data processing -- Structures, Theory of -- Periodicals; 624.171
• DOI: 10.1016/j.compstruc.2019.106156
• Languages: English
• ISSNs: 0045-7949

Abstract

Highlights: An improved hybrid FE-SEA method is established in this paper. Dynamic condensation is introduced to reduce the order of the total dynamic stiffness matrix. A fast inverse algorithm is employed to calculate the dynamic flexibility matrix of the slave degrees of freedom of the deterministic component. The present method avoids the direct inverse computation of a large matrix at each frequency point of interest, resulting in significant time saving. Abstract: The hybrid Finite Element (FE) – Statistical Energy Analysis (SEA) method developed for mid-frequency vibration of complex built-up systems needs to compute the total dynamic flexibility matrix at every frequency, which is very time consuming. This paper presents an improved hybrid FE-SEA method to overcome this problem. In the present method, first, dynamic condensation is introduced to reduce the order of the deterministic FE component, which results in significant reduction of the total dynamic stiffness matrix. Then, noting that the dynamic stiffness matrix of the deterministic component is established by using the FE method, a fast inverse algorithm is employed to calculate the dynamic flexibility matrix of the slave degrees of freedom of the deterministic component generated in the condensation process. These two steps avoid the direct inverse computation of a large matrix at each frequency point of interest, resulting in significant time saving. A numerical example illustrates the efficiency and convergence of the proposed method.