Application Of Sequential Collaborative Optimization In The Structural System Design Of Deep Sea Space Station

The deep sea space station is a new concept human occupied vehicle, which provides a broader platform for the exploration and exploitation of ocean. As the most important part of the deep sea space station, the structural system consists of pressure hull, exostructure and faring. Traditionally, the subspaces of pressure hull, exostructure and faring are designed independently. While the designers usually emphasized on the structural analysis and design of pressure hull, those of exostructure and faring were paid little attention. Nevertheless, the latter also have a great influence on the general performance of the deep sea space station. Further more, the design of the three subspaces interacts with each other. Therefore, the structural system should be treated as a whole in the design procedure, which results in a complex design problem involving several disciplines, such as structural mechanics and hydrodynamics. Multidisciplinary design optimization (MDO) emerged from the field of aeronautics and astronautics provides an effective way to settle the problem.As a part of the project supported by the Commission of Science Technology and Industry for National defense, the dissertation explores the application of multidisciplinary design optimization in the structural system design of deep sea space station by using sequential collaborative optimization. Focused on the MDO methods, structural design of pressure hull and collaborative optimization of structural system, the dissertation mainly consists of the following aspects:(1) A literature review referred to the theory of MDO is presented. As well as the history and development, the main contents of MDO are introduced in detail, among which the MDO methods are focused primarily.(2) The basic concept, architecture, mathematical model and sensitivity analysis of Collaborative optimization (CO) are introduced. Both of the advantages and disadvantages of CO are analyzed, and several improvement methods are presented. On the basis of the study of L1 compatibility constraint function, the sequential collaborative optimization (SCO) is brought forward with the concept to remove the absolute label in the L1 compatibility constraint function by introducing additional constraints. SCO eliminates the non-smoothness of the compatibility constraint functions and reduces greatly the probability for subspace optimization problem to convergence to local optimum. Besides, the performance of SCO and CO is compared through two numerical examples.(3) A brief introduction of approximation methods in the field of MDO, including design of experiments and approximation models, are given. The performance of three approximation model—response surface, Kriging and radial basis function network are compared through numerical examples by the method of error analysis.(4) The structural optimization problem of multiple intersecting pressure hull is studied. Based on nonlinear finite element method and Kriging mode, the structural optimization is performed. Based on the analysis of nonlinear FEM results, the simplified equations are introduced and used to the structural optimization of multiple intersecting pressure hull of different materials. In addition, several new concepts of multiple intersecting pressure hull are put forward.(5) The dissertation divides the structural design problem of deep sea space station into four subspaces including pressure hull, exostructure, fairing and performance. The mathematical models of all subspaces are set up, and the inputs and outputs of each model are defined. The relevant calculation programs are compiled.(6) The collaborative optimization architecture of the structural system is established and each subspace is described in detail. Kriging model and the SCO method developed in the dissertation are applied to the MDO problem of structural system design then.The main innovation researches of the dissertation can be concluded that:(1) This is the first time to apply the theory of MDO into the structural system design of deep sea space station. Some mathematical models are established and the applicability of MDO for the design large engineering system is proved.(2) Based on the study of collaborative optimization, the sequential collaborative optimization is developed. By the modifying L1 compatibility constraint function, SCO eliminates the non-smoothness of the compatibility constraint functions and greatly improved the general performance of CO. (3) The collapse modes of multiple intersecting pressure hull are concluded through the analysis of nonlinear FEM results and the simplified equations are introduced to perform structural optimization then.Multidisciplinary design optimization is an emerging discipline and its application in the structural system design for complex deep sea structures is getting start. However, there are many aspects of research should be deepen and they are:(1) Although the sequential collaborative optimization has been tested by numerical examples, it convergence should be inspected and verified in theory. Besides, it is possible that the other kinds of modification for the L1 compatibility constraint function improve the performance better and this is a direction for further research.(2) Because of the lack of detailed information about the overall design, the dissertation takes much assumption and simplification. In the future, the structural system design should be integrated with the overall design to perform multidisciplinary design optimization.(3) The fairing of deep sea space station is assumed to have a form of tear-like revolving body. For the further research, it follows a rational line to use the CFD method to model the 3-D form of faring and take more accurate hydrodynamic analysis.(4) Besides collaborative optimization, there are many other MDO methods with great potential, such as CSSO and BLISS. Which MDO method is more suitable for the structural system design of complex deep sea structures should be further studied.(5) Distributed and parallel computation is the trend of MDO. It is meaningful to set up an distributed and parallel environment by integrating large-scale engineering software through computer network for the multidisciplinary design optimization of large ocean structures.