【作者】 彭公孚；

【导师】 唐小兵；

【作者基本信息】 武汉理工大学 ， 工程力学， 2004， 硕士

【摘要】 在扩大基建、拉动内需的政策指引下，我国桥梁事业蓬勃发展。钢管混凝土拱桥因其造型美观、造价低廉、承载潜力大而得到广泛应用。分析及控制拱肋吊装施工过程中的扣索受力成为设计、施工人员共同关注的问题。扣索控制的优劣直接影响着拱肋线形的好坏，从而对于成桥承载受力有着重要影响。若施工控制不好将对拱肋合龙造成困难，甚至导致事故的发生。为了解决这个问题，设计人员常常采用有限元方法计算扣索受力及拱肋控制标高，并根据计算结果设计扣索吊装位置，在此过程中大量重复编程及计算工作占用了很多工作时间，另外在扣索吊装位置的选定过程中又没有明确的方法，只能根据大量的计算结果人工寻找优化方案，这些都制约着设计工作的开展。近年来大量商用的有限元软件不断推出更新。利用这些现有的软件资源解决设计过程中具有共性的问题成为设计人员关注的方向。本文通过一种高级语言对ANSYS这一运用范围广泛的商用有限元软件进行二次开发，将大量重复的机械计算过程封装成为一个整体。使普通技术人员可以从大量重复计算工作中脱离出来，专注于其他设计工作。 对设计人员来说，有限元计算只是设计工作的一部分。在拱肋的施工过程中，如何合理的分布扣索吊装位置，则涉及到扣索施工的优化处理。若仅以人工的经验处理这样一个非线性问题，势必带来很大的误差。而现今又没有有效的理论处理类似的问题，所以需要对扣索吊装位置预测进行必要的探讨。 本文利用BP网络来建立施工中的关键控制数据(关键受力点的受力大小、扣索控制点标高值、扣索受力最大及最小值等)与扣索及拱肋连接点之间的映射关系，通过网络的自学习功能得到稳定的网络。然后根据设计要求，分析各个关键控制数据，并预测最佳的扣索连接位置，从而减少了完全由人工选择扣索连接点时对设计人员经验和运算能力的要求，尽量减少人为误差造成的计算偏差。 算例结果表明该程序具有计算精度高、节省时间、人机对话、使用方便等特点。论文在ANSYS二次开发、扣索位置预测方法、扣索索力计算方面提出了创新点。更多还原

【Abstract】 In the direct of policy about enlarge capital construction and stimulate civil consume. Development of bridge enterprise rise rapidly. Span Arch Bridge of Concrete-Filled Steel Tubes used wildly because of beautiful sculpt low cost potential hard bearing of weight. Tension analysis and control of cables in the process of lifting construction of span arch is attract the attention of architect and engineer. Cable control effect infect the shape of span arch directly, even infect the bear of weight of bridge. It is difficult or even tragedy of closure of span arch construction if cable control badly. To avoid the probability, engineer calculate the cable tension and the pre-rising elevation with finite element theory, it cost much time in design the program and repeat calculation, the other hard work is to find out the place where the cables located in span arch structure, its must find out the optimized result manually which restrict the design work.Recent years, lots of commercial FEA software develop rapidly. Engineers focus on the solution of commonness problems by the software existed. This paper use a advanced program langue in the application development of a commercial FEA software ANSYS. The program encapsulation lots of calculation work to free the engineer from hard work, and pay more attention in other works.There is only little work for engineers to calculate with FEA. It is referring to optimize the location of cable in span arc during lifting construction. Error could not avoid in this non-linear problem if design base on experience only. So it is necessary to discuss the theory which have no effect method in span arc lifting construction.In this pape, a BP networks is founded to show the relationship of the key control data and the location of cables on span arch construction. Form the stable network through the self-study, and use this networks to forecast the location of cables. This method avoid the error produced by man-made.A given example prove the program has high level precision low cost in time interact of man and computer. It give new method in applicationdevelopment of ANSYS forecast the location of cables calculation of tension of cables.更多还原