Summary of Masters Thesis
Home ] Up ] Photo Album ] Chronology ] Favourites ]

 

 

Vibration Energy Flow Analysis in Randomly Parametered Coupled Plate Structures using Green’s Function Coupling Technique

Structural dynamic analysis aims to establish relationships between externally applied loads and displacements/strains/stresses fields in the vibrating structures. The most commonly used approach to achieve this consists of discretization of displacement field using finite element method and subsequent solution of equilibrium equations using normal mode expansion. This approach, although, has been successfully applied in a wide variety of contexts, nevertheless, has not been well received in some areas. For instance, the computational demands that the method imposes increase significantly with increases in frequency range of interest. In many engineering applications, such as building acoustics, satellite structural dynamics, automobile and aircraft dynamics, it becomes important to model high frequency vibrational response. It is, in principle, possible to extend analysis to higher frequencies, at the expense of rapidly increasing demands in terms of size of the model, and consequent analysis time and cost. However, as frequencies increase, the results become increasingly dependent upon fine structural detail, including structural connections, which cannot be mathematically represented with sufficient accuracy. In addition, the high frequency vibrational behaviour of individual physical realizations of nominal structures are observed to differ, often greatly, because of the influence of fabrication tolerances. It is clear that the high cost of computational procedures based upon the deterministic models cannot be justified if they do not yield reliable results. In such circumstances other methods may be more appropriate.

One such approach is the Statistical Energy Analysis (SEA) formalism. Here, the primary objective of the response analysis is to predict the space-wise and frequency-wise distribution of total vibrational energy residing in the structure. Detailed spatial distribution, however, is not sought. Instead, vibrational energy fields integrated over certain spatial domains are considered as being of interest. Such a macro-level response description is distinct from the detailed description that finite element method aims to obtain. This method, or more accurately, analysis philosophy, dates from the early 1960's, when engineers sought new analytical tools for dealing with the problems of predicting the response of launcher and payload structures to rocket noise at launch. System parameters were expressed in probabilistic terms, and the objective of an analysis was seen to be the prediction of the ensemble-average behaviour of sets of grossly similar realizations of an archetypal system. System response to vibrational inputs was characterized by time-average vibration energy; energy flows between coupled subsystems were expressed in terms of energy transfer coefficient; and vibration distribution was determined from power balance equations. The basic principle was that the energy transfer between two subsystems is proportional to the difference of their average vibration energies. The proportionality constant was termed as the coupling loss coefficient. This lead to the result that the input vibrational energies are expressible as a linear combination of the averaged subsystem energies and the coupling loss factors, i.e., {P}= [CLF]{E}. Here {P} is the vector of input power averaged over one cycle, {E} is the vector of time-averaged subsystem energy and [CLF] is the matrix of coupling loss factors. This approach to modelling vibrational behaviour can be shown to lead to acceptable solutions only if a set of restrictive assumptions such as linear system behaviour, conservative coupling mechanisms, random uncorrelated forcing, weak coupling between subsystems etc., are satisfied. The implications of these restrictions on the scope of SEA continue to remain subject of current research.

The objective of the present thesis is to examine the validity of some of the mechanistic and probabilistic assumptions underlying SEA formalisms. The study is conducted within the context of vibration energy flow (VEF) analysis of the plate assemblies that are discretely coupled through a set of axially vibrating rods; this class of structure being representative of satellite solar panels tested in laboratory conditions. The thesis is organized into five chapters.

The first chapter provides the details of SEA formalisms and critically examines the underlying assumptions. Literature pertaining to theoretical basis of SEA and methods for determining the SEA parameters are examined. VEF analysis that employs substructure Green's function and a frequency domain coupling technique is also discussed. The merits of this procedure are highlighted. The problem of investigating the effects of far coupling and, also, the problem of confidence band estimation of energy levels in coupled plate systems are observed to have received limited research attention.

A general formulation of VEF models for a set of coupled multi-modal structural systems is outlined in chapter 2. These models make use of the uncoupled subsystem Green's functions. At no stage in the analysis, the global normal modes of the built-up structure are computed. A general-purpose software on MATLAB platform has been developed based on this formulation. This software accepts, as inputs, the subsystem Green function over a given frequency range and produces estimates of spectra of frequency band averaged estimates of the total subsystem energies. The formulation is exemplified by considering VEF models of a structure that consists of a set of three plates that are discretely coupled by a set of axially vibrating rod elements. Results based on this formulation are compared with FEM results (using NISA) that require significantly more computational effort. The basic requirements on vibration energy balance are shown to be satisfied, thereby providing a validation of the software developed.

The procedure developed in chapter 2 is extended in chapter 3 to obtain the descriptors of VEF such as coefficients of coupling loss factor (CLF) matrix. This extension employs the use of power injection method (PIM) that has been developed originally in the context of experimental determination of CLF coefficients. Specific issues on structure of CLF matrix that are associated with far coupling effects in energy flow are investigated. Examples considered again include coupled plate structures.

The problems of characterizing stochastic variability in energy levels and also on the nature of probability distribution function of energy levels are considered in chapter 4. The elastic, geometric and inertial properties of the plate subsystems are modelled as a vector of random variables. Influence of varying damping properties on statistics of energy flow is examined. The study in this part of the chapter employs Monte Carlo simulation (MCS) technique. A new idea to define the satisfactory performance of SEA is considered in the second part. This latter part of the chapter uses the ideas of reliability indices borrowed from literature on structural reliability analysis and compares the results with similar results obtained from MCS technique.

The thesis concludes with chapter 5 wherein a summary of contributions made and a few suggestions for future research are provided.

Appendix A presents the definition reliability index and the algorithm that is used for computation of reliability index.

 

This page was last updated on 04/01/03 08:36:48 PM.