基于鲁棒性的船舶.doc
约65页DOC格式手机打开展开
基于鲁棒性的船舶,摘要常规优化由于不考虑设计变量的波动，这就使常规优化解处在可行域的边缘上，当设计变量发生波动后，常规优化解将会偏离可行域从而变为不可行解。鲁棒设计由于在设计之初就考虑到了变量的波动性，它对目标函数和约束函数均有鲁棒性的要求，因而鲁棒设计解较常规优化解具有更好的鲁棒性。现代鲁棒设计方法是在田口方法的基础上发展而来的。本文...
内容介绍
此文档由会员 违规屏蔽12 发布
摘要
常规优化由于不考虑设计变量的波动，这就使常规优化解处在可行域的边缘上，当设计变量发生波动后，常规优化解将会偏离可行域从而变为不可行解。鲁棒设计由于在设计之初就考虑到了变量的波动性，它对目标函数和约束函数均有鲁棒性的要求，因而鲁棒设计解较常规优化解具有更好的鲁棒性。
现代鲁棒设计方法是在田口方法的基础上发展而来的。本文主要介绍了基于容差模型法的鲁棒设计，最小灵敏度法和灵敏度分析法。容差模型法考虑到设计变量的变差会引起目标函数和约束条件的变差，要在新的可行域内找到鲁棒解，该法适用于模型线性程度不高以及变差较小的情况；最小灵敏度法和灵敏度分析法是不考虑设计变量容差的方法。
本文将理论和实际相结合，所做主要工作如下：
（1）利用容差模型法对一艘8700DWT散货船中横剖面结构进行了鲁棒优化设计。以板厚为设计变量，整个中横剖面面积最小为目标函数，约束函数采用最坏情况容差进行处理。鲁棒设计结果较规范设计有更好的鲁棒性。
（2）最小灵敏度法不考虑性能参数的余度，因此本文在最小灵敏度法的基础上提出了一种最小波动法。本文所提出的最小波动法将最小灵敏度和性能参数的余度相结合，对性能参数的波动进行量化。通过对4万吨散货船进行中横剖面结构结构的鲁棒优化设计，证明本文所提出的方法较最小灵敏度法有更好的设计结果。同时本文所提出的方法也可以推广到其它的鲁棒优化设计中。
（3）在MSC.Patran环境中建立27.6m双体船全船有限元模型，通过施加边界条件和载荷，经过MSC.Nastran进行计算得出结果。对结果进行分析后，该双体船的结构需要改进。然后借助有限元技术，利用结构分析软件MSC.Patran/Nastran求取27.6m双体船质量、板单元合成应力、梁单元应力、横向刚度以及扭转刚度对板厚的灵敏度。通过综合分析比较，得出了优化方案，并取得了较好的优化结果，说明灵敏度分析用在船舶结构优化设计可以取得较好的结果。本文也可作为其它船型灵敏度分析优化设计时的参考。
关键词：结构鲁棒设计 容差模型 最小灵敏度法 灵敏度分析
Abstract
Conventional optimal design does not take the variations of variables into account, which makes conventional optimal solution in the edge of the feasible region. When the design variables fluctuate, the conventional optimal solution will deviate from the feasible region into unfeasible solution. While robust design takes the variations of variables into account early in the design, and it makes objective function and constraints are robust, therefore robust design optimization solution has better robustness than the conventional solution.
Modern robust design method is based on the Taguchi method and evolved. This paper mainly describes robust method based on the tolerance model, the minimum sensitivity method and sensitivity analysis. Tolerance model method taking into account the variation of design variables which will cause the objective function and constraints have variation, to find a new robust solution in the new feasible region.This method applies to the linear model is not high and the variation is relatively small; minimum sensitivity method and sensitivity analysis are the methods which don’t considere design variable tolerance.
This paper combined theory and practice, and the major work done is as follows:
(1) Use of tolerance model method for 8700DWT bulk carrier’s midship cross-sectional structure in robust optimal design. Plate thickness as design variables, the minimum cross-sectional area as objective function, and constraint functions using the worst-case tolerance for processing. The result is more robust than specification design.
(2)It doesn’t consider the redundancy of the performance parameters, so the method of minimum variation is proposed in this paper. The minimum variation being proposed in this paper cooperates the method of minimum sensitivity and the redundancy of the performance parameters, which will quantify the variations of the performance parameters. By the robust optimization of the midship section structure of 40,000DWT bulk carrier, finding that the method proposed in this paper has better result than the minimum sensitivity. And the method proposed in this paper can also be extended to other robust optimization designs.
(3) In MSC.Patran environment 27.6m catamaran finite element model was build, through the imposition of boundary conditions and load, got the result by the calculation of MSC.Nastran. After analyzing the results, some catamaran's structure needs to be improved. In this paper by using finite element techniques and structural analysis software MSC.Patran/Nastran, got the sensitivity to sheet thickness of mass, the Von.Mises of shell element, the Von.Mises of beam element, lateral stiffness and torsional stiffness. Through comprehensive analysis and comparison, got the optimization program, and achieved good optimization results, indicating the sensitivity analysis used in ship structural optimization can get better result. This paper can also be used as a reference for other ships’ sensitivity analysis and optimization.
Keywords: structure robust design, tolerance model, minimum sensitivity, sensitivity analysis
目录
摘要 Ⅰ
Abstract Ⅰ-br..
常规优化由于不考虑设计变量的波动，这就使常规优化解处在可行域的边缘上，当设计变量发生波动后，常规优化解将会偏离可行域从而变为不可行解。鲁棒设计由于在设计之初就考虑到了变量的波动性，它对目标函数和约束函数均有鲁棒性的要求，因而鲁棒设计解较常规优化解具有更好的鲁棒性。
现代鲁棒设计方法是在田口方法的基础上发展而来的。本文主要介绍了基于容差模型法的鲁棒设计，最小灵敏度法和灵敏度分析法。容差模型法考虑到设计变量的变差会引起目标函数和约束条件的变差，要在新的可行域内找到鲁棒解，该法适用于模型线性程度不高以及变差较小的情况；最小灵敏度法和灵敏度分析法是不考虑设计变量容差的方法。
本文将理论和实际相结合，所做主要工作如下：
（1）利用容差模型法对一艘8700DWT散货船中横剖面结构进行了鲁棒优化设计。以板厚为设计变量，整个中横剖面面积最小为目标函数，约束函数采用最坏情况容差进行处理。鲁棒设计结果较规范设计有更好的鲁棒性。
（2）最小灵敏度法不考虑性能参数的余度，因此本文在最小灵敏度法的基础上提出了一种最小波动法。本文所提出的最小波动法将最小灵敏度和性能参数的余度相结合，对性能参数的波动进行量化。通过对4万吨散货船进行中横剖面结构结构的鲁棒优化设计，证明本文所提出的方法较最小灵敏度法有更好的设计结果。同时本文所提出的方法也可以推广到其它的鲁棒优化设计中。
（3）在MSC.Patran环境中建立27.6m双体船全船有限元模型，通过施加边界条件和载荷，经过MSC.Nastran进行计算得出结果。对结果进行分析后，该双体船的结构需要改进。然后借助有限元技术，利用结构分析软件MSC.Patran/Nastran求取27.6m双体船质量、板单元合成应力、梁单元应力、横向刚度以及扭转刚度对板厚的灵敏度。通过综合分析比较，得出了优化方案，并取得了较好的优化结果，说明灵敏度分析用在船舶结构优化设计可以取得较好的结果。本文也可作为其它船型灵敏度分析优化设计时的参考。
关键词：结构鲁棒设计 容差模型 最小灵敏度法 灵敏度分析
Abstract
Conventional optimal design does not take the variations of variables into account, which makes conventional optimal solution in the edge of the feasible region. When the design variables fluctuate, the conventional optimal solution will deviate from the feasible region into unfeasible solution. While robust design takes the variations of variables into account early in the design, and it makes objective function and constraints are robust, therefore robust design optimization solution has better robustness than the conventional solution.
Modern robust design method is based on the Taguchi method and evolved. This paper mainly describes robust method based on the tolerance model, the minimum sensitivity method and sensitivity analysis. Tolerance model method taking into account the variation of design variables which will cause the objective function and constraints have variation, to find a new robust solution in the new feasible region.This method applies to the linear model is not high and the variation is relatively small; minimum sensitivity method and sensitivity analysis are the methods which don’t considere design variable tolerance.
This paper combined theory and practice, and the major work done is as follows:
(1) Use of tolerance model method for 8700DWT bulk carrier’s midship cross-sectional structure in robust optimal design. Plate thickness as design variables, the minimum cross-sectional area as objective function, and constraint functions using the worst-case tolerance for processing. The result is more robust than specification design.
(2)It doesn’t consider the redundancy of the performance parameters, so the method of minimum variation is proposed in this paper. The minimum variation being proposed in this paper cooperates the method of minimum sensitivity and the redundancy of the performance parameters, which will quantify the variations of the performance parameters. By the robust optimization of the midship section structure of 40,000DWT bulk carrier, finding that the method proposed in this paper has better result than the minimum sensitivity. And the method proposed in this paper can also be extended to other robust optimization designs.
(3) In MSC.Patran environment 27.6m catamaran finite element model was build, through the imposition of boundary conditions and load, got the result by the calculation of MSC.Nastran. After analyzing the results, some catamaran's structure needs to be improved. In this paper by using finite element techniques and structural analysis software MSC.Patran/Nastran, got the sensitivity to sheet thickness of mass, the Von.Mises of shell element, the Von.Mises of beam element, lateral stiffness and torsional stiffness. Through comprehensive analysis and comparison, got the optimization program, and achieved good optimization results, indicating the sensitivity analysis used in ship structural optimization can get better result. This paper can also be used as a reference for other ships’ sensitivity analysis and optimization.
Keywords: structure robust design, tolerance model, minimum sensitivity, sensitivity analysis
目录
摘要 Ⅰ
Abstract Ⅰ-br..
