Yanbin Li

Prediction of Statistical Energy Analysis Parameters in Thermal Environment

An algorithm is presented to predict the statistical energy analysis (SEA) parameters of structures in a thermal environment. An energy flow model considering thermal effects is established by the ...

Nonstationary Random Vibration Analysis of Wing with Geometric Nonlinearity Under Correlated Excitation

An algorithm that integrates Karhunen-Loeve expansion (KLE), nonlinear finite element method (NFEM), and a sampling technique to quantify the uncertainty is proposed to carry out random vibration a...

Demarcation for the Coupling Strength in the MODENA Approach

MODal ENergy Analysis (MODENA) is an energy-based approach which is proposed recently to provide a pure tone analysis of power flow. The net exchanged power between two coupled oscillators is proportional to the weighted difference of total energies of oscillators. Contrary to Statistical Energy Analysis (SEA) or Statistical modal Energy distribution Analysis (SmEdA), the MODENA approach can deal with strong coupling case where the power can flow from the oscillator with lower energy to the one with higher energy. The level of coupling strength between oscillators may affect the accuracy of the MODENA approach. This research work aims to propose a criterion to determine the level of coupling in the MODENA approach. A non-dimensional parameter named coupling strength factor is defined to clearly demonstrate the level of coupling strength. Two numerical examples: (a) a two-oscillators coupling case and (b) a multi-modal coupling case are conducted to show the effectiveness of the proposed criterion.

Non-Stationary Random Vibration Analysis Using Multi-Correlated Random Processes Excitations

An algorithm that integrates Karhunen-Loeve expansion (KLE) and the finite element method (FEM) is proposed to perform non-stationary random vibration analysis of structures under excitations, represented by multi random processes and that are correlated in both time and spatial domains. In KLE, the auto-covariance functions of random processes are discretized using orthogonal basis functions. The KLE for multi-correlated random processes relies on expansions in terms of correlated sets of random variables reflecting the cross-covariance of the random processes. During the response calculations, the eigenfuntions of KLE used to represent excitations are applied as forcing functions to the structure. The proposed algorithm is applied to a two degree of freedom system and a stiffened panel for both stationary and non-stationary correlated excitations. Two methods are proposed to obtain the structural responses: 1) the modal superposition method, and 2) the direct method. Both the effectiveness and the computational efficiency of the proposed methods are studied through numerical simulations. Results show that both the modal superposition method and the direct method can describe the statistics of the dynamic response with sufficient accuracy. However, the modal superposition method is applicable using finite element programs whereas the direct method is more efficient for complex systems. The structural responses due to same type of correlated random processes are bounded by the response obtained by both perfectly correlated and uncorrelated random process excitations. If the impact of cross-correlation is ignored, the response will be larger for the perfectly correlated case and smaller for the uncorrelated case than that for the actual conditions. The structural response increases with a decrease in the correlation length and with an increase in the correlation magnitude. The proposed methodology can be applied for the analysis of any complex structures and any type of random excitations.