【作者】 路伟；

【导师】 孙大鹏；

【作者基本信息】 大连理工大学 ， 港口、海岸及近海工程， 2010， 硕士

【摘要】 进人21世纪后,随着经济全球化涌现的大吨位的LNG、大吞吐量的深水码头、超长距离的跨海大桥等迫切需要海洋工程走向外海深水域,传统的海工建筑物已经不能适应新工程的要求。而开孔直墙式防波堤有着良好的工程应用前景,对海洋事业的发展起着至关重要的作用,对于开孔直墙式防波堤结构及其水动力特性的研究一直备受国内外研究者的关注。针对海床地质良好的工程,完全可以采用明基床形式,使基床顶高程明显高于海床面,减少开孔沉箱的高度,进一步节省造价。但是由于明基床的存在,海底边界条件发生局部变化,势必影响开孔沉箱前的波浪场,有关于考虑基床影响下的反射系数和波浪力的理论研究及可应用于工程实际的计算方法还很少,因此,对明基床上开孔沉箱结构进行系统地分析研究势在必行。本文通过二维物理模型试验,对规则波和不规则波作用下明基床开孔沉箱所受到垂直方向的波浪力、力臂、力矩进行了系统的研究,并结合了相同试验条件下的暗基床开孔沉箱的相关数据进行分析比较。试验是在大连理工大学海岸和近海工程国家重点实验室的浑水水槽中进行的；试验模型设计考虑了不同的消浪室宽度、开孔率及基床高度的影响,为了便于分析比较,分别进行了低基床、中基床上的实体直墙及开孔沉箱结构的试验。通过对试验数据的单因次相关分析,找出相对基床高度、消浪室相对宽度、相对水深、波陡及开孔率等影响因素与浮托力、总垂直力、总垂直力的力臂和总垂直力的力矩之间的关系,采用最小二乘法拟合它们之间的经验关系式。这些经验关系式表达形式简单、直观,计算方便,计算结果与试验数据符合较好,在试验条件变化范围内,可供明基床开孔沉箱进一步研究和工程设计时参考。随着计算机运算能力的迅猛发展,人工智能及模糊理论应用的范围越来越广。人工神经网络是一类基于生理学的智能仿生模型,是由大量处理单元组成非线性自适应动态系统,具有良好的自适应性、自组织性及很强的学习、联想、容错和抗干扰能力。本文也应用神经网络对试验数据进行了分析并且与拟合公式的结果进行了对比。更多还原

【Abstract】 In the 21st century, with the emergence of economic globalization, the large tonnage of LNG, the high-throughput deepwater dock, and projects of ultra-long-distance bridges, traditional marine buildings have not adapt to new project requirements and thus can not meet this urgent need. The perforated caisson has a good application prospect and plays a vital role in the development of the marine industry. There have been many systematic researches, made by both domestic and foreign researchers, on the structure of perforated caissons and their dynamic characteristics. For foundation with high quality, the form of open bedding could be applied, that is the height of bedding being obviously higher than that of the seabed, in order to reduce the height of perforated caissons and to further save costs. As a result, the existence open bedding and the local changes in seabed boundary conditions, will certainly affect the waves in front of perforated caissons. However, the systematic research on reflection coefficients and wave forces under influence of and their computational methods in practical engineering are relatively few, so it becomes imperative to take a systematic research on the perforated caisson with rubble foundation.In this paper, through two-dimensional physical model tests, a systematic research on wave forces, arms and torques on perforated caisson with open bedding under regular and irregular waves has been introduced and comparisons have been made to data of perforated caisson without open bedding under the same experimental conditions. The physical model tests were carried out in the wave tank for suspension material test at the State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology. The test models were designed to consider different influences of the width of wave absorbing room, the porosity and the height of bedding; in order to facilitate analysis and comparison, the solid conventional structure and the perforated caisson each with low and middle foundation were tested respectively.Based on single-dimensional correlation analysis, the relationships between influence factors, such as width of wave absorbing room, the porosity and the height of bedding, etc., and the total horizontal force and the total vertical force were identified. The empirical formulae of these relationships were also presented for theoretical study and practice by use of the Ordinary Least Square Estimation. The resulting expressions were simple, intuitive and easy to calculate; the results were in good agreement with the experimental data; and thus could provide valuable reference to further research and engineering design for perforated caisson.With the rapid development of computing capacity, the Artificial Intelligence and Fuzzy Theory are increasingly widely used. Artificial neural networks are a class of intelligent bionic model based on physiology, and composed of nonlinear adaptive dynamic system with a large number of processing units; it possesses good capabilities of self-organized, powerful learning, fault tolerance and anti-jamming. The present paper also analyzed the obtained experimental results and compared the fitting empirical formulas with the use of neural network analysis.更多还原