| |
|
S21 Vibrações e Análise Modal |
| |
|
Title:
|
AN ASSESSMENT OF GENETIC ALGORITHMS AS APPLIED TO SOME INVERSE PROBLEMS IN ELASTODYNAMICS
|
| |
|
Summary :
|
|
GENETIC ALGORITHMS (GA S) ARE OPTIMIZATION AND SEARCH PROCEDURES BASED ON DARWIN S NATURAL SELECTION PRINCIPLES. THE MAIN PURPOSE OF THIS WORK IS TO PERFORM AN EVALUATION OF THE GA S WHEN APPLIED TO SOME MODEL-BASED PARAMETER IDENTIFICATION PROBLEMS IN ELASTODYNAMICS. SUCH PROBLEMS ARE FORMULATED AS OPTIMIZATION PROBLEMS IN WHICH THE OBJECTIVE FUNCTIONS REPRESENT THE DIFFERENCE BETWEEN THE EXPERIMENTALLY OBSERVED DYNAMIC BEHAVIOR AND ITS ANALYTICAL COUNTERPART. THE FOLLOWING INVERSE PROBLEMS ARE FOCUSED: LOCALIZATION AND EVALUATION OF THE EXTENT OF MODELING ERRORS FROM REAL EIGENVALUES AND EIGENVECTORS AND IDENTIFICATION OF PHYSICAL PARAMETERS OF LINEAR AND NONLINEAR MOUNTS FROM TIME RESPONSES. BASED ON BOTH NUMERICALLY SIMULATED AND EXPERIMENTALLY TESTED SYSTEMS, THE GA S ARE APPRAISED IN TERMS OF THE ACCURACY OF THE SOLUTIONS AND ROBUSTNESS WITH RESPECT TO RANDOM NOISE CONTAMINATING THE DATA.
KEY-WORDS: GENETIC ALGORITHMS, PARAMETER IDENTIFICATION, INVERSE PROBLEMS, MODEL UPDATING
|
| |
|
Author :
|
da Silva, Luciano Afonso
Rade, Domingos ALves
|
| |
|
Paper View :
|
|
