Multidisciplinary Design Optimization And Its Application On Truss Spar Conceptual

To meet the demand of national economic development and people's fast-growing living condition, China has planned projects of exploiting oil and gas reserves in the South China Sea. However, the manufacturing capability of offshore production platforms in China is still below 500 meters water depth, far from the exploration requirement in South China Sea. Nowadays, the mainstream platforms used in deep water around the world include Floating Production Storage and Offloading System, Semi-submersible, Tension Leg Platform and Spar platform. Among them, Spar platforms have advantages in aspects of functionality, stability, cost-efficiency and safety. Besides, China has dedicated many years'study on Spar platform and has rich technical reserves. In such circumstances, Spar platform will become one of the major equipment for the gas and oil exploration for our country in the deep water, including South China Sea.Presently, Chinese design capability is lagged behind the foreign design enterprises on the design of advanced ocean engineering units, especially in the conceptual design stage. In addition, the technical monopoly brings difficulties for us to develop Spar platforms. To speed up the pace of research on Spar platforms, and to break through the technical bottleneck, we should make innovation in both concept and method for the conceptual design of Spar platform.Firstly, this thesis carries out a Truss Spar conceptual design and structural design task based on the current design method of offshore platforms. The current design method prevailing in China is the series process, based on the purpose to improve the parent unit. This design process has the shortcomings of relying too much on the parent unit and high design cost. Even worse, the results of conceptual design can hardly present the overall performances. The design of Spar platforms is a complex engineering problem involving many disciplines, such as hydrodynamics, structures, reliability, etc. These disciplines are both mutual interaction and contradictions, so the overall conceptual design of Spar platform is a typical Multidisciplinary Design Optimization (MDO) problem. However, the current design method can not consider the mutual influence among these disciplines sufficiently, so it often leads to a suboptimal design instead of optimal design. Besides, the trend for ships and offshore platforms design is moving toward digitalization and virtualization, which will be beneficial for designers to acquire comprehensive information about the system to design in the early stage. Such trend demands the design method be able to highly integrate all the relevant disciplines. Obviously, it can hardly be realized by current method. To speed up the development of deep sea resources exploration, some breakthrough and innovation for the concept and method of design is in urgent need.In such a situation, as a part of the project supported by the Science and Technology Commission of Shanghai Municipality, this thesis introduces Multidisciplinary Design Optimization, emerged from aeronautics and astronautics fields for complex engineering integrated optimization problems, and explores feasibility and applicability of MDO method to the conceptual design of Truss Spars. The thesis mainly consists of four parts.Part I: Review about the related research areaThe significance of applying MDO techniques in the Truss Spar conceptual design is stated. The definition, development and main research contents of MDO are reviewed. Four mainstream MDO algorithms are elaborated and compared in categories including origin and purpose, approximation models, decomposition, convergence and engineering applicability. Two algorithms– collaborative optimization based on approximation models and BLISS 2000 are regarded as suitable methods to apply MDO techniques in the design of ships and offshore platforms.Part II: Study on the current conceptual design method and structural design method of Truss SparThe current Truss Spar conceptual design method is explored. Two aspects are involved during the conceptual design: one deals with the design method and process, and the other concerns the hydrodynamic behavior. The conceptual design process of Spar platform can be described as an interactive flow. A Truss Spar conceptual design process including general arrangement, weight control and selection of main particulars is conducted. The stability and hydrodynamic performances for the Truss Spar are accessed. Based on the results of conceptual design, the structural design is conducted including:(1) The design method for each part of Spar platforms is concluded and the structural scantling process is accomplished based on the main particulars, general arrangement and load distribution; (2) The global strength analysis based on design wave methodology is accomplished. Two critical load criteria are selected and wave-induced load is obtained based on 3-D potential theory.(3) The impact on bottom deck strength for two structural configurations– double bottom and single bottom is studied. The result shows the double bottom can reduces the stress of bottom deck of hard tank in great extent.The Spar designed by current conceptual design method and structural design method is treated as baseline Spar for further comparison with the result of MDO.Part III: Application of MDO on the Truss Spar conceptual designThe design parameters in the design of Spar platform, including main particulars, operation parameters, economics and reliability, are elaborated. Because the thesis focuses on the conceptual design, the main particulars are selected as design variables. Then the varying level of design variables are arranged by the theory of Design of Experiment (DOE) and corresponding design plans are determined by uniform design table.One of the most significant factors for a successful application of MDO is to obtain a reasonable MDO model. A comprehensive MDO model, which includes every discipline related to Truss Spar, will make the optimization process difficult to converge; On the contrary, if only a few disciplines are included, the result of MDO can hardly represent the whole system and be impractical for engineering. To solve this contradiction, an innovative multidisciplinary modeling method is proposed. The disciplines related to Spar platform are divided into three modules: optimization module, constraint module and inspection module. Each module is treated as a different role during the optimization.The optimization module is the main focus for platform owners and designers, and treated as optimization object function. The disciplines of constraint module are the characteristics of Spar platforms. These characteristics are precondition for the realization of platform functions, and treated as constraints in the optimization process. Only a few critical objectives for each discipline are selected and integrated in the optimization model. The inspection module is not involved in the optimization process and is used to provide a comprehensive assessment of the optimized design for checking.In the thesis, the hull weight is treated as optimization object. The short term predictions of heave motion and pitch motion are selected as constraint objects for hydrodynamics discipline; the intact stability coefficient is selected as the constraint object for stability discipline; the von Mises stress of the connection zone of hard tank bottom deck with truss is selected as the constraint object for structure discipline. A hull/mooring coupled analysis in time domain is utilized as the method of inspection module. To set up the MDO model for Truss Spar, numerous hydrodynamic analyses, stability analyses and structural strength analyses are conducted for all design plans defined by DOE.The Collaborative Optimization (CO) based on response surface method is selected as the MDO algorithm and the CO model for Truss Spar conceptual design is established for the first time. Then the CO optimization process is executed and an optimal design is achieved.Part IV: The improvements of the Collaborative Optimization modelAlthough the approximation models can greatly improve the computational efficiency, the accuracy of the models may be not eligible. To enhance the accuracy of the approximation models and consequent optimization quality, two methods are used:(1) Adaptive response surface method (ARSM)– The concept and procedure of ARSM is proposed. The collaborative optimization based on ARSM for Truss Spar conceptual design is accomplished and an optimal design solution, which satisfied all the constraints, is obtained. A discussion on the adaptive history for each design variables is made. The optimized results are compared with the baseline Spar. The comparison shows that the optimized solution by MDO achieves a lighter platform, and that the performances of heave response motion, stability coefficient and strength of the bottom deck are much improved. The only performance degradation for the optimized solution is the pitch motion, while the change is not significant. Finally, the time domain analysis based on 100-year storm of Gulf of Mexico is conducted and the results prove that the optimized design solution by MDO is feasible.(2) Variable-complexity method (VCM)– Although ARSM can effectively improve the accuracy of approximation models, it fails to diminish the difference between approximation models and physical models. VCM is adapted to correct results of CO in part II by using the results of high-fidelity analysis tool.However, VCM generally requires large number of calculations for initial sampling points to ensure precision. To overcome such a deficiency, an improved VCM method based on ARSM is proposed. The application results show that the improved VCM strategy can attain high accuracy with a small number of initial sampling. In summary, four innovative accomplishments have been made in the thesis.(1) A complete conceptual design process for Truss Spar is proposed and structural design is conducted. Design wave methodology is used to check the structural strength of the platform.(2) MDO techniques are applied in Truss Spar conceptual design. The optimized design results by MDO show superiority in performances contrasting to the ones by the current design method. The feasibility of applying MDO techniques in the Spar platform is verified.(3) An innovative multidisciplinary modeling method is proposed. The related disciplines for Spar platform are divided into three modules: optimization module, constraint module and inspection module. Each module is treated as different role during the optimization. Such modeling method decreases the convergence difficulty of optimization and relieves the computational burden. The rationality of the MDO model is also guaranteed.(4) Adaptive response surface method and variable-complexity method are adopted to improve the accuracy of approximation models. An improved VCM method based on ARSM is proposed. The method shows advantages in both improving the accuracy of approximation models and diminishing the difference between approximation models and physical models.The accomplishment of this thesis is an innovation to the current conceptual design method. A successful application of MDO in Truss Spar conceptual design will promote the development of digitalization and virtualization for offshore platforms, make a contribution to the independent research and application of offshore platforms, Spar platform included, in deep sea areas in China.