Please use this identifier to cite or link to this item:
|Title:||Development of real-time computation techniques: A novel reduced-basis method and application to inverse analyses||Authors:||KHIN ZAW||Keywords:||Reduced-basis method, finite element method, asymptotic error estimation, greedy procedure, meshfree methods, smoothed Galerkin projection.||Issue Date:||28-Apr-2009||Citation:||KHIN ZAW (2009-04-28). Development of real-time computation techniques: A novel reduced-basis method and application to inverse analyses. ScholarBank@NUS Repository.||Abstract:||This thesis focuses on the development of real-time computational technique, and it consists of two parts. The first part is to develop a smoothed Galerkin projection reduced-basis method (SGP_RBM) based on the standard reduced-basis method and a smoothed Galerkin projection. A very important property of an upper bound to the exact solution (in energy norm) has been investigated through theoretical study and numerical analyses. Two numerical examples of a cantilever beam problem and a thermal fin problem are conducted. Both theoretical analysis and numerical results have demonstrated that the present method is a very efficient method for real-time solutions providing exact solution bound. The second part of the thesis is to establish rapid and reliable inverse searching procedures using the reduced-basis method as a fast forward solver and two searching procedures of genetic algorithm (GA) and neural network (NN) as inverse searching procedures. In the proposed inverse searching procedures, computational efficiency is increased by the use of the reduced-basis method. The feasible searching procedures of GA and NN are employed for parameter estimation. Numerical examples of (i) crack detection in a cantilever beam in the area of NDE and (ii) identification of elastic constants of interfacial tissues between dental implant surfaces and surrounding bones are conducted. From the numerical results, the present rapid inverse procedures are stable and reliable. Additionally, computational efficiency is significant.||URI:||http://scholarbank.nus.edu.sg/handle/10635/15857|
|Appears in Collections:||Ph.D Theses (Open)|
Show full item record
checked on May 23, 2019
checked on May 23, 2019
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.