【摘要】：隨著社會(huì)多元化程度的不斷加快,科技的持續(xù)發(fā)展,我們生活中的各種產(chǎn)品的復(fù)雜程度更是日益提高。而傳統(tǒng)的設(shè)計(jì)優(yōu)化理論在對(duì)復(fù)雜系統(tǒng)進(jìn)行設(shè)計(jì)時(shí)存在設(shè)計(jì)效率低、產(chǎn)品開(kāi)發(fā)周期長(zhǎng)等諸多缺陷。為此,科研設(shè)計(jì)人員開(kāi)發(fā)了新的設(shè)計(jì)優(yōu)化理論,即多學(xué)科設(shè)計(jì)優(yōu)化理論。該理論強(qiáng)調(diào)充分利用各學(xué)科之間的存在的耦合協(xié)同機(jī)制,通過(guò)不斷協(xié)調(diào)各學(xué)科變量的取值,從而獲得系統(tǒng)整體最優(yōu)解。在操作上,該理論要求在設(shè)計(jì)之初就必須充分考慮學(xué)科間存在的耦合關(guān)系,通過(guò)分布式計(jì)算網(wǎng)絡(luò)組織和管理整個(gè)設(shè)計(jì)過(guò)程,從而實(shí)現(xiàn)并行的分析和設(shè)計(jì)。本文深入分析研究了多學(xué)科設(shè)計(jì)優(yōu)化理論及其技術(shù)體系,并對(duì)幾種多學(xué)科設(shè)計(jì)優(yōu)化方法進(jìn)行詳細(xì)的分析與研究,并將這幾種方法應(yīng)用到數(shù)學(xué)算例的求解之中。標(biāo)準(zhǔn)協(xié)同優(yōu)化方在計(jì)算上存在很大的困難性。本文采用二次函數(shù)算例,通過(guò)結(jié)果對(duì)比詳細(xì)探討說(shuō)明了其存在的計(jì)算困難性,并從數(shù)學(xué)理論的角度深入分析探討了導(dǎo)致這種計(jì)算困難的根本原因。針對(duì)其計(jì)算困難的根本原因,本文深入分析研究了三種改進(jìn)措施。將三種改進(jìn)措施應(yīng)用到算例的分析求解之中,并從優(yōu)化結(jié)果的誤差度、迭代次數(shù)、收斂程度等幾個(gè)方面加以評(píng)判。通過(guò)多方面的對(duì)比,表明引入松弛變量的協(xié)同優(yōu)化方法具有很好的計(jì)算性能�？煽慷仁乾F(xiàn)代社會(huì)衡量產(chǎn)品質(zhì)量的一個(gè)重要指標(biāo)。為了提高經(jīng)多學(xué)科設(shè)計(jì)優(yōu)化后產(chǎn)品的可靠度,本文深入研究探討了一階可靠性方法,并將一階可靠性方法引入到多學(xué)科設(shè)計(jì)優(yōu)化的框架之中,從而保證系統(tǒng)產(chǎn)品的可靠度。為此,本文建立了基于可靠性的協(xié)同優(yōu)化方法。通過(guò)數(shù)學(xué)算例表明該方法是行之有效的。而后,將該方法應(yīng)用到齒輪箱減速器的設(shè)計(jì)優(yōu)化之中,通過(guò)結(jié)果對(duì)比表明采用該方法后齒輪箱減速器的可靠度得到了顯著提高。物理規(guī)劃方法因能將工程經(jīng)驗(yàn)等偏好信息融入到產(chǎn)品的設(shè)計(jì)優(yōu)化之中,而深得廣大工程設(shè)計(jì)人員的喜愛(ài)。本文詳細(xì)分析研究了物理規(guī)劃方法的設(shè)計(jì)思想、偏好函數(shù)的建立以及計(jì)算流程等。在此基礎(chǔ)上,本文將物理規(guī)劃方法融合到多學(xué)科設(shè)計(jì)優(yōu)化方法之中,建立了基于物理規(guī)劃的協(xié)同優(yōu)化方法。將該方法應(yīng)用到某型號(hào)戰(zhàn)術(shù)導(dǎo)彈的設(shè)計(jì)優(yōu)化之中,不僅偏好信息等得到了充分的考慮,而且所得到的優(yōu)化結(jié)果也是非常令人滿意的。
[Abstract]:With the rapid development of social diversity and science and technology, the complexity of various products in our life is becoming more and more complex. However, the traditional design optimization theory has many defects, such as low design efficiency and long product development cycle. Therefore, a new design optimization theory, I. e., multidisciplinary design optimization theory, has been developed by researchers and designers. The theory emphasizes to make full use of the coupling and coordination mechanism between disciplines, and to obtain the overall optimal solution of the system by continuously coordinating the values of the variables of each discipline. In operation, the theory requires that the coupling relationship between disciplines must be fully considered at the beginning of design, and the whole design process should be organized and managed by distributed computing network, so that parallel analysis and design can be realized. In this paper, the theory of multidisciplinary design optimization and its technical system are deeply analyzed, and several multidisciplinary design optimization methods are analyzed and studied in detail, and these methods are applied to the solution of mathematical examples. The standard cooperative optimization method is very difficult to calculate. In this paper, an example of quadratic function is used to illustrate the difficulty of calculation in detail by comparing the results, and the fundamental causes of this difficulty are analyzed and discussed from the point of view of mathematical theory. Aiming at the root cause of its difficulty in calculation, three kinds of improvement measures are analyzed and studied in this paper. Three kinds of improvement measures are applied to the analysis and solution of an example, and the error degree, iteration number and convergence degree of the optimization result are evaluated. The comparison shows that the cooperative optimization method with relaxation variables has good computational performance. Reliability is an important index to measure product quality in modern society. In order to improve the reliability of the product after multidisciplinary design optimization, the first order reliability method is studied in this paper, and the first order reliability method is introduced into the framework of multidisciplinary design optimization to ensure the reliability of the system product. Therefore, a reliability-based collaborative optimization method is established in this paper. A mathematical example shows that the method is effective. Then, the method is applied to the design optimization of gearbox reducer. The results show that the reliability of gearbox reducer is improved remarkably. The physical planning method is popular among the engineering designers because it can integrate the engineering experience and other preference information into the design optimization of the product. In this paper, the design idea of physical programming method, the establishment of preference function and the calculation flow are analyzed in detail. On this basis, the physical programming method is integrated into the multidisciplinary design optimization method, and a collaborative optimization method based on physical planning is established. When this method is applied to the design optimization of a certain type of tactical missile, not only the preference information is fully considered, but also the optimization results are very satisfactory.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類(lèi)號(hào)】：TB47;TB114.3

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 劉蔚,崔維成;多學(xué)科設(shè)計(jì)優(yōu)化:載人潛水器設(shè)計(jì)的一種新工具(英文)[J];船舶力學(xué);2004年06期

2 余雄慶,姚衛(wèi)星,薛飛,穆雪峰,劉克龍,黃愛(ài)鳳;關(guān)于多學(xué)科設(shè)計(jì)優(yōu)化計(jì)算框架的探討[J];機(jī)械科學(xué)與技術(shù);2004年03期

3 尤政;李冠華;;多學(xué)科設(shè)計(jì)優(yōu)化方法在微納衛(wèi)星總體設(shè)計(jì)中的應(yīng)用[J];中國(guó)航天;2010年04期

相關(guān)碩士學(xué)位論文 前1條

1 袁亞輝;多學(xué)科設(shè)計(jì)優(yōu)化及其不確定性研究[D];電子科技大學(xué);2009年

，

本文編號(hào)：2291940