节点文献
基于遗传算法的棒材孔型多目标优化
The Multi-Objective Optimal Pass Design Based on GA
【作者】 张娜;
【导师】 张文志;
【作者基本信息】 燕山大学 , 机械设计及理论, 2006, 硕士
【摘要】 遗传算法(Genetic Algorithm,简称GA)是近几年发展起来的一种模拟自然界生物进化过程与机制的自组织、自适应的人工智能技术,其理论基础为达尔文的自然选择和孟德尔的遗传变异理论。遗传算法求解优化问题的基本思想是:把优化的目标函数解释为生物种群对环境的适应性,把优化变量对应为生物种群的个体,由当前种群出发,利用合适的复制、杂交、变异与选择操作生成新一代种群,重复这个过程,直到求得合乎要求的种群停止。本文以圆钢棒材轧制为基本实体模型,根据前人的研究成果及作者的探索,选择了相应的温度、宽展、轧制压力、轧制力矩及能耗模型;基于MATLAB遗传算法工具箱,应用遗传算法与惩罚函数相结合的方法编制程序对孔型系统进行多目标优化设计,并利用MATLAB软件绘制各种曲线;利用AutoCAD2006中的Visual LISP集成开发环境进行孔型图的绘制。以三维立体软件Pro/E为平台,对其进行实体建模,并以大型商用有限元软件ANSYS/LS-DYNA为分析工具,对棒材连轧进行模拟,得出轧制过程中棒材变形区的应力、应变的变化情况和轧制力曲线。并将模拟轧制力与优化过程中得到的结果进行比较,验证其计算精度。
【Abstract】 Genetic Algorithm (GA), which simulates natural evolution procedure and mechanism developed in recent years, is self-organizing and self-adapting artificial intelligence technology. The technology bases on Darwinian natural selection and Mendelian genetics. The basic thought of solving optimization problem with GA is as follows. Firstly, the objective function of optimization is interpreted as the capability to environment for biotic-population, and the optimizing variable is corresponded with the individuals of biotic-population. Secondly, the current population is changed to a new one by suitable operation of duplication, hybridization, dissociation and selection. Thirdly, the procedure is duplicated until the fitting population is obtained.The dissertation has been designed using round steel rolling as entity model. Based on the research production of predecessor and investigation of the author, corresponding experience formulas such as roll temperature, spread, roll force, roll moment and roll consume are selected. The pass system is optimized through multi-objective optimization method with MATLAB GA toolbox. Kinds of curves are drawn by MATLAB. The drawings of passes are developed by Visual LISP integrated development environment (IDE) of AutoCAD2006.The three-dimensional models of rolling are established with Pro/E. Based on the finite element analysis software ANSYS/LS-DYNA, the procedure of bar tandem rolling is simulated. The nephograms of stress and strain of bar deformed area and the curve of roll force during rolling are obtained. Compared the result with the other result which comes from optimization, the precision of calculation will be verified.
【Key words】 Genetic Algorithm (GA); Roll-pass design; Multi-objective; Bar; Tandem rolling; Optimization;
- 【网络出版投稿人】 燕山大学 【网络出版年期】2007年 02期
- 【分类号】TG335.62;TP391.7
- 【被引频次】11
- 【下载频次】352