Determination of Elatic Constants of Composite Materials Via Vibration testing with Adaptive Real-parameter Simulated Annealing Genetic Algorithm and Gradient Taguchi method

• Main Authors: Jen-chih Wu, 吳仁志
• Other Authors: Shun-fa Hwang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/85163723066669077400

碩士 === 雲林科技大學 === 機械工程系碩士班 === 96 === After obtaining the natural frequencies of the material by vibration tests, finite element method software (ANSYS) combined with an inverse method is applied to determine the elastic constant of the material. Without destroying the material, this method has the flexibility to apply to the material with complicated boundary conditions and geometric shape. Furthermore, two kinds of different inverse methods, adaptive real-parameter simulated annealing genetic algorithm and gradient taguchi method, are used in this method for the determination of elastic constant of composite material. Adaptive Real-parameter Simulated annealing genetic algorithm is based on real-parameter genetic algorithm and simulated Annealing and includes an adaptive mechanism to decide dynamic genetic probability, in order to improve the performance of genetic algorithm. In order to reduce the execution time of the inverse method, a gradient Taguchi method based on the advantages of both Taguchi method and gradient method is proposed in this work. After verifying the performance such as convergence property, accuracy and dependability of gradient Taguchi method, the result reveals that this inverse method has better performance. To further compare these two inverse methods, reference examples from the literature are used, and aluminium plates, unidirectional laminates, Cross-ply laminates, Woven and printed circuit board are also tested with the free-free boundary condition. The results indicate that both the inverse methods could obtain nice elastic constants of the materials, and the gradient Taguchi method indeed shortens significantly the execution time.