Loading...
Please wait, while we are loading the content...
Similar Documents
Structure Design Optimization Considering Static and Dynamic Responses
| Content Provider | Semantic Scholar |
|---|---|
| Author | Cho, Hee Keun |
| Copyright Year | 2013 |
| Abstract | Structure design has been optimized for physically coupled problems. A non-gradient evolutionary genetic algorithm (GA) is employed to optimize the behaviors of isotropic solid structures. A structure showing thermo-elasticplastic and creep (TEPC) behaviors are considered for optimization under a number of static and dynamic constraints. Complicatedly coupled solid behaviors are numerically calculated by utilizing a FEM. The presented optimization scheme is conducted in conjunction with a FEA and involves newly developed 2D nonlinear TEPC and modal analysis FEA codes. The present evolutionary technique makes it possible to arrive at an optimum design with respect to a solid showing TEPC features with timeand temperature-dependent material properties under dynamic natural frequency constraints. The developed code has been verified through an example problem of a dome structure. The present results, attained for the first time, could be useful to ascertain the synergizing effects of a FEA and a GA in TEPC problems Index Term— FEA, GA(genetic algorithm), optimization, dynamics I. INTR ODU C TIO N Structure design optimization p roblems based on fin ite element analyses in the field of automobiles, aircraft , spacecraft, and other industrial machines have won growing interest in engineering in recent years. Fig. 1 denotes how to optimize structures through the interaction of a number o f detailed technical fields. The development of an optimization algorithm starts from the well known tradit ional gradient-based methods. More recently, a non-gradient probabilistic optimization algorithm-GA and particle swarm optimizat ion algorithm have yielded excellent results in engineering problems and they have been actively applied to structure design. Advanced optimization algorithms that shorten the calculat ion time have recently been developed and have been successfully applied to various techniques. These advanced optimizat ion algorithms have already been utilized for structure design. Structure optimization init ially started as a method to optimize simple design variables such as length or thickness. Recently, however, Hee Keun Cho is with the Faculty of Mechanical Engineering Education, Andong National University, 1375 Kyungdong –ro, Andong, Kyungpook, Republic of Korea (corresponding author, phone +82-54-820-5677 e-mail: hkcho@andong.ac.kr). even complicated geometric structures involving diverse coupled behaviors have been optimized in accordance with the development of FEM techniques. Cho and Rowlands recently combined a GA scheme with a FEA to optimize structures and thereby enhance mechanical performances [1]. Several other representative studies documenting the applicat ion of GAs to structures have also been published [2-7]. Most engineering optimizat ion algorithms use gradient based methods, where the search direction is a function o f the gradient of design variab les [8]. Gradient-based methods, however, are unable to deal efficiently with complex design space or a large number of design variables. These drawbacks can be overcome by using GA optimizat ion. GA procedures also converge well to the global optimum value with respect to both discrete and continuous design variab les. Several researchers have recently recommended GAs for optimizing structures. The present study merges developed nonlinear TEPC analysis FEA codes and a GA with a parallel computing scheme to optimize arbitrary geometric structures. The FEA code accounts for the various design variables in the formulation of a stiffness matrix. Data from the FEA are exchanged, back and forth, with the parallel p rocessing capability of the GA optimization module. Mathemat ical formulations of the fin ite element analysis equilibrium equations are presented in the next section. Fig. 1. Structure design optimization scheme II. RESEARCH METHODS A. Finite Element Analytical Background For this study, several FEA numerical analysis codes for elastic, plastic, creep, heat transfer, and dynamic models are developed and joined together according to their International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:13 No:01 17 137601-4949-IJMME-IJENS © February 2013 IJENS I J E N S interdependent functions. This software accounts for the TEPC characteristics in calculat ing material deformations and stress. Informat ion is parallel-processed freely between the FEA and the optimizat ion module GA. Relevant mathematical details for the nonlinear coupled FEA are outlined below. Stage I: Heat Transfer and Thermal Stress [9-11] The principal idea of the TEPC analysis is first discussed. The code is for 2D 8-node plane stress and/or strain isoparametric elements derived from the virtual work theory. A general non-steady state 2-dimensional heat transfer governing equation is given in Eq. (1). 0 T T T c k k q t x x y y (1) Eq. (1) can be transformed into an FEA equilibrium equation, Eq. (2), by means of a variational method. In Eq. (2), T represents the temperature change rate against time. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.ijens.org/Vol_13_I_01/137601-4949-IJMME-IJENS.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
