【作者】 赵敏；

【导师】 崔维成；

【作者基本信息】 上海交通大学 ， 船舶与海洋结构物设计制造， 2009， 博士

【摘要】 深海空间站是一个高新技术密集、多个学科高度综合的复杂系统工程,涉及到多个学科内容:水动力学、结构力学、材料、能源与推进、操纵与控制等。这些学科之间相互联系,有的甚至相互矛盾,使得深海空间站的总体概念设计呈现典型的“多学科”特点。然而,传统设计方法不能充分地考虑学科之间的相互影响,且依赖于总体设计专家的经验,由此得到的设计结果往往只是满足设计要求的解,并非真正的整体最优解,从而导致深海空间站总体性能的下降。此外,传统设计方法存在设计周期长、效率低的问题,且深海空间站尚无设计经验可以依靠,故将传统设计方法直接应用于深海空间站的总体概念设计,会导致整个研发成本的急剧增加,传统设计方法已难以适应深海空间站这类复杂系统的优化设计。为了应对深海空间站总体概念设计中这样的难题,必须寻找到一种新的设计理论和方法,从系统的观点突破传统设计方法存在的缺陷,进而实现深海空间站的整体优化设计。在这种情况下,本文作为国防科工委“十一五”重大科技专项“深海空间站主站关键技术研究”的有机组成部分,引入航空领域迅速发展起来的解决复杂系统优化设计的多学科设计优化方法(Multidisciplinary Design Optimization, MDO),探索其在深海空间站总体概念设计应用上的可行性和适用性。论文围绕多学科设计优化方法及其在深海空间站总体概念设计中的应用,主要进行了以下几方面的研究工作:1.课题相关研究领域的综述针对深海空间站的国内外研究进展以及多学科设计优化研究与应用现状进行了综述。阐述了多学科设计优化在深海空间站总体概念设计中应用的意义,介绍了多学科设计优化的发展概况、难点以及主要研究内容。2.多学科设计优化方法的比较研究选取三种具有代表性的多学科设计优化方法:MDF、CO以及BLISS方法,详细介绍了它们的设计思想以及数学模型。采用NASA提供的多学科设计优化标准算例,并结合多属性决策法对以上三种多学科设计优化方法进行了不同初始点的定量比较。研究结果表明:(1)随着工程对象的不同,所适合的多学科设计优化方法也将不同,没有绝对意义上最好的多学科设计优化方法;(2)基于BLISS和CO提出一种新的多学科设计优化方法,将对深海空间站的总体概念设计有着重要的现实意义。3.结合BLISS和CO的主要特点,提出了两级集成系统协同优化(BLISCO)多学科设计优化方法在保留协同优化方法协同机制的同时,借鉴BLISS将设计变量分为系统级设计变量和子系统设计变量,并用子系统耦合输出响应的加权和代替一致性约束作为子系统优化的目标函数,从而提出了BLISCO多学科设计优化方法。通过NASA提供的两个多学科设计优化标准算例对BLISCO进行了测试。结果表明:BLISCO在满足收敛性的同时,具备较佳的准确性与鲁棒性,是一种有效的多学科设计优化方法,从而为实现深海空间站的总体概念多学科设计优化提供了有力的工具。4.采用BLISCO多学科设计优化方法对载人潜水器(HOV)进行总体概念设计为了更好地实现深海空间站总体概念的多学科设计优化以及探索BLISCO方法在潜器多学科设计优化中的可行性和适用性,采用BLISCO方法成功实现了HOV总体概念的多学科设计优化。优化结果显示:HOV的总体综合性能大大提升,缩短了设计周期,且BLISCO方法对于HOV总体概念多学科设计优化的综合效果优于PGA-CO。因此,BLISCO方法完全可以应用于潜器的多学科设计优化,从而为基于BLISCO方法的深海空间站总体概念多学科设计优化提供工程实例支持。5.深海空间站系统集成模型的研究根据深海空间站系统设计的特点,建立了包含外形、阻力与推进、能源、结构、重量与静水力平衡、以及操纵性六大学科的参数化系统集成模型,明确了各学科的输入输出关系。通过设计结构矩阵的应用,建立了合理的设计流程,并发现系统集成模型存在以下两个特点:(1)学科之间信息的大量交换;(2)只存在设计信息的向前传递,而无信息反馈。因此,深海空间站的总体概念设计是一个顺序执行的过程,属于弱耦合设计问题。本研究为即将开展的基于BLISCO方法的深海空间站总体概念多学科设计优化奠定了坚实的基础。6.BLISCO方法在深海空间站上的应用研究本研究以深海空间站系统集成模型为基础,将深海空间站总体概念设计优化问题分解为水动力子系统、结构子系统以及总体性能子系统,然后采用BLISCO方法对其进行多学科设计优化。BLISCO方法在深海空间站多学科设计优化中的成功应用表明:(1)结合BLISCO思想提出的深海空间站三子系统分解方法是可行的;(2)BLISCO方法可有效应用于深海空间站的总体概念多学科设计优化。本文创新性的工作总结如下:1.结合BLISS和CO的主要特点,提出了BLISCO多学科设计优化方法。通过标准算例的测试,揭示了BLISCO具有诸多优点,是对多学科设计优化方法的丰富和发展;2.针对深海空间站的特点,建立了参数化的系统集成模型,在此基础上将深海空间站的总体概念设计分解为三个子系统,并将本文所建立的BLISCO多学科设计优化方法成功地加以应用,实现了深海空间站总体概念的多学科设计优化;3.针对PGA-CO在载人潜水器总体概念多学科设计优化中存在计算量大和效率低的问题,将BLISCO多学科设计优化方法应用于载人潜水器的总体概念设计,有力提升了载人潜水器的总体性能和设计效率,再次验证了BLISCO多学科设计优化方法的优越性。更多还原

【Abstract】 Deep Sea Space Station is a complex system engineering involving many high-techs and disciplines, such as hydrodynamics, structure, material, energy and propulsion, maneuverability and control etc. There are many mutual interactions and even contradictions among these disciplines, so the overall conceptual design of Deep Sea Space Station (DSSS) is a typical Multidisciplinary Design Optimization (MDO) problem. However, traditional design method can not consider the mutual influence sufficiently among these disciplines, and depends much on the designer’s experience, so it often leads to suboptimal design instead of optimal design and reduces the general performance of DSSS. Besides, there are some disadvantages of traditional design method: long design cycle, low efficiency and no experience for the DSSS design, so it will result in rapid increase of R&D cost with the direct application of traditional design method to the overall conceptual design of DSSS. Therefore, the traditional design method is not fit to the optimal design of the DSSS. In order to solve the difficulty faced by the overall conceptual design of DSSS, new design theory and method is necessary to break through the limitation of the traditional design method and to realize the system optimal design of the DSSS.Under this circumstance, the present study is a part of the National Defense Basic Scientific Research Project–“key techniques research of Deep Sea Space Station”supported by the Commission of Science, Technology and Industry for National Defense of the People’s Republic of China. It introduces Multidisciplinary Design Optimization (MDO) method emerged from aeronautics and astronautics fields for complex engineering integrated optimization problems, and explores feasibility and applicability of MDO method to the overall conceptual design of DSSS. Focusing on MDO method and its application to the overall conceptual design of DSSS, this thesis mainly consists of the following aspects:1. Review about the related research areaProgress of researching Deep Sea Space Station in the world, as well as MDO’s research and application status are overviewed. The significance of the MDO’s application to the overall conceptual design of DSSS is stated. The development, difficulties and main research contents of MDO are reviewed.2. Comparison of different MDO MethodsThree representative MDO methods: MDF, CO and BLISS are studied and described in detail about the mechanism and mathematical models. Combined with multi-attribution decision method, one benchmark example provided by NASA is adopted to compare these three MDO methods quantitatively with different initial points. The result shows that: (1) For different engineering object, the suitable MDO method will be different, there is no absolutely best one; (2) Based on BLISS and CO to propose a new MDO method will contribute a lot to the overall conceptual design of DSSS in practical sense.3. Development of Bi-Level Integrated System Collaborative Optimization (BLISCO) based on the main characteristics of BLISS and COWhile maintaining the collaborative mechanism of CO, BLISCO is proposed with the decomposition of design variables into system design variables and subsystem design variables, and the replacement of compatibility constraint with the sum of coupled output responses as an integrated objective of subsystem, which idea comes from BLISS. Two benchmark examples provided by NASA are adapted to testify the performance of BLISCO. The results show that BLISCO is an effective MDO method with accurate result, reliable robustness and satisfaction of convergence. Furthermore, BLISCO will be a powerful tool to realize the overall conceptual design of DSSS based on MDO.4. Overall conceptual design of HOV based on BLISCOIn order to better realize the overall conceptual design of DSSS based on MDO and to explore feasibility and applicability of BLISCO to the overall conceptual design of submersible, BLISCO is successfully applied to the overall conceptual design of HOV. The result shows that the general performance of HOV is highly improved, design cycle is much decreased, and the integrated effect of BLISCO is better than PGA-CO. Therefore, BLISCO can be applied to the MDO of submersible completely, and the successful overall conceptual design of HOV will provide engineering support for the overall conceptual design of DSSS based on BLISCO.5. Research on the System Synthesis Model of DSSSBased on the characteristics of DSSS, a System Synthesis Model (SSM) is developed for DSSS. It includes six modules, which are hull, resistance & propulsion, energy, structure, weight & hydrostatic equilibrium and maneuverability. The inputs and outputs of different modules are also defined. According to the application with Design Structure Matrix (DSM), a reasonable design sequence is established, and it is found that there are two characteristics of SSM: (1) many data exchanges between different modules; (2) there is only feed forward and no feed back. Therefore, the overall conceptual design of DSSS is a sequential process, which belongs to the weak coupling system. This research has laid a solid foundation for the overall conceptual design of DSSS based on BLISCO.6. Research on the application of BLISCO method to the overall conceptual design of DSSSAfter the DSSS is decomposed into hydrodynamic subsystem, structure subsystem and general performance subsystem based on the SSM, BLISCO is applied to the overall conceptual design of DSSS. The successful application with BLISCO reveals that: (1) decomposition with three subsystems is feasible; (2) BLISCO can be applied to the overall conceptual multidisciplinary design optimization of DSSS efficiently.The innovative research works in this thesis can be summarized as follows:1. Based on the main characteristics of BLISS and CO, BLISCO is developed in this thesis. According to the benchmark tests, the several advantages of BLISCO are revealed, which enriches and develops the MDO method;2. Focused on the characteristics of DSSS, a parametric System Synthesis Model (SSM) is developed. With the decomposition of DSSS into three subsystems based on the SSM, BLISCO is successfully applied to the overall conceptual design of DSSS, which realizes the overall conceptual multidisciplinary design optimization of DSSS.3. In order to overcome the numerous calculations and low efficiency during the application of PGA-CO to overall conceptual design of HOV, BLISCO is applied to the overall conceptual design of HOV. General performance and design efficiency of HOV are much improved by BLISCO, which validates the superiority of BLISCO again.更多还原