Computational Modal Analysis Of A Half-Scale Business Jet Fuselage Tail Section
Loading...
Authors
Gunawardana, Rochana
Date
Type
thesis
Language
eng
Keyword
Computational Modal Analysis , Finite Element Modelling , Frequency Response Analysis
Alternative Title
Abstract
In aircraft, acoustic noise and structural vibrations are critical aspects of what determines the comfort of passengers throughout the duration of their trip. As such, there is great value in being able to reduce noise and the structural vibrations that cause noise to improve the quality of the on-board experience. To adequately understand how a structure vibrates, the dynamic characteristics such as the natural frequencies and mode shapes of the aircraft structure must be known. This may be accomplished by performing modal analysis to derive those relevant properties. For a business jet, an important location of study for vibration is the tail section, as it supports the engine which is a primary source of structure-borne noise. The objective of this research is to develop and validate an accurate Finite Element (FE) model of a half-scale business jet fuselage tail section that can then be used to predict the behaviour on a real aircraft based on certain excitations or changes to the structure, without having to conduct rigorous experimental testing. This is accomplished by comparing the FE model to responses taken from a corresponding experimental test structure and applying changes to the model if necessary. A preliminary comparison between the Experimental Modal Analysis (EMA) and Computational Modal Analysis (CMA) yielded five correlated modes across the frequency range of 50-400 Hz. This range was chosen based on industry recommendations. The modes showed adequate accuracy given the experimental and computational limitations, with the percent difference averaging to around 10%. The differences in the natural frequencies identified using the two different methods tended increase in the higher frequencies. This discrepancy was improved by applying a model calibration methodology to the materials which improved average natural frequency error by 1-2%.