基于遗传算法的桥梁结构传感器优化布置研究
发布时间：2018-10-18 07:33
【摘要】：传感器系统是桥梁健康监测系统中的必不可少的一部分,其对于判定现有桥梁的实时损伤,真实地获取结构整体和局部信息,在特殊时刻提供预警,减少人力资源和财力上的损失和浪费,减少错误信息对处理系统的干扰,实现桥梁的长寿命具有重要的意义。传感器布置作为桥梁健康监测系统中科学研究的重心,人们对它进行的探索一刻也没有停止过,但是由于测量参数的不同,科研人员对传感器优化研究角度的差异,自然得出的结果各异,也没有得出完全唯一的评价准则,存在着许多值得商榷的地方。本论文进行了如下探究:一、建立耿峪河大桥的迈达斯模型,根据挠度识别误差最小的原则,插值拟合建立目标函数,借助遗传算法工具箱在连续梁桥上对传感器布置进行优化,寻找到关心截面的传感器布置最精确的解答。针对遗传算法工具箱中的变异算子和交叉算子对结果的影响进行了深入的研究,发现虽然遗传算法存在随机性,但是在桥梁工程中可以通过多次求平均值予以消除。二、对振型的相似性进行了研究,振型向量分量是一个相对值,在以前的二范数度量振型“相似性”的基础上,引入一范数进行规范化,然后再度量近似度,进而筛选出线性无关的振型以便于分析。三、迈达斯软件给出的是离散节点的分量值,于是介绍了最小二乘法的原理和幂级数、傅里叶级数以及指数级数的拟合方式,最后决定采用精度最高傅里叶级数拟合成振型函数,编制MATLAB程序,利用MAC准则,采用遗传算法工具箱对目标函数进行优化,找出钢桁架梁桥和连续梁桥动力传感器的最优布置位置,并且采用模态运动能法对桥梁传感器进行优化布置,得出在耿峪河大桥采用遗传算法并基于模态置信矩阵的方法布置桥梁传感器优于模态运动能法。
[Abstract]:Sensor system is an indispensable part of the bridge health monitoring system. It can be used to judge the real time damage of existing bridges, obtain the whole and local information of the structure, and provide early warning at special time. It is of great significance to reduce the loss and waste of human resources and financial resources, and to reduce the error information to deal with the interference of the system and to realize the long life of the bridge. Sensor layout is the center of scientific research in the bridge health monitoring system, and the exploration of it has not stopped for a moment. However, due to the difference of measurement parameters, the researchers' research angle of sensor optimization is different. Naturally, the results are different, and there is not a completely unique evaluation criterion, there are many questionable points. The main contents of this paper are as follows: firstly, the Meidas model of Gengyu River Bridge is established. According to the principle of minimum deflection recognition error, the objective function is established by interpolation fitting. By using genetic algorithm toolbox to optimize the sensor arrangement on continuous beam bridge, the most accurate solution to the sensor arrangement of concerned section is found. The influence of mutation operator and crossover operator in genetic algorithm toolbox on the result is deeply studied. It is found that although genetic algorithm has randomness, it can be eliminated in bridge engineering by calculating the average value several times. Secondly, the similarity of modes is studied. The component of modal vector is a relative value. On the basis of the previous two-norm to measure the "similarity" of modes, a norm is introduced to normalize and then measure the degree of approximation. Then the linearly independent modes are screened for easy analysis. Thirdly, the component value of discrete node is given by Midas software, so the principle of least square method and the fitting method of power series, Fourier series and exponential series are introduced. Finally, it is decided to use the highest precision Fourier series quasi-synthetic mode function, to compile the MATLAB program, to optimize the objective function by using the MAC criterion and the genetic algorithm toolbox. The optimal position of dynamic sensors for steel truss girder bridge and continuous beam bridge is found, and the modal motion energy method is used to optimize the layout of the bridge sensor. It is concluded that the genetic algorithm based on the modal confidence matrix is better than the modal motion energy method in the Gengyu River Bridge.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U446
本文编号：2278428
[Abstract]:Sensor system is an indispensable part of the bridge health monitoring system. It can be used to judge the real time damage of existing bridges, obtain the whole and local information of the structure, and provide early warning at special time. It is of great significance to reduce the loss and waste of human resources and financial resources, and to reduce the error information to deal with the interference of the system and to realize the long life of the bridge. Sensor layout is the center of scientific research in the bridge health monitoring system, and the exploration of it has not stopped for a moment. However, due to the difference of measurement parameters, the researchers' research angle of sensor optimization is different. Naturally, the results are different, and there is not a completely unique evaluation criterion, there are many questionable points. The main contents of this paper are as follows: firstly, the Meidas model of Gengyu River Bridge is established. According to the principle of minimum deflection recognition error, the objective function is established by interpolation fitting. By using genetic algorithm toolbox to optimize the sensor arrangement on continuous beam bridge, the most accurate solution to the sensor arrangement of concerned section is found. The influence of mutation operator and crossover operator in genetic algorithm toolbox on the result is deeply studied. It is found that although genetic algorithm has randomness, it can be eliminated in bridge engineering by calculating the average value several times. Secondly, the similarity of modes is studied. The component of modal vector is a relative value. On the basis of the previous two-norm to measure the "similarity" of modes, a norm is introduced to normalize and then measure the degree of approximation. Then the linearly independent modes are screened for easy analysis. Thirdly, the component value of discrete node is given by Midas software, so the principle of least square method and the fitting method of power series, Fourier series and exponential series are introduced. Finally, it is decided to use the highest precision Fourier series quasi-synthetic mode function, to compile the MATLAB program, to optimize the objective function by using the MAC criterion and the genetic algorithm toolbox. The optimal position of dynamic sensors for steel truss girder bridge and continuous beam bridge is found, and the modal motion energy method is used to optimize the layout of the bridge sensor. It is concluded that the genetic algorithm based on the modal confidence matrix is better than the modal motion energy method in the Gengyu River Bridge.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U446
【参考文献】
相关期刊论文 前1条
1 崔飞,袁万城,史家钧;传感器优化布设在桥梁健康监测中的应用[J];同济大学学报(自然科学版);1999年02期
相关硕士学位论文 前2条
1 曹俊;裂纹扩展的实时健康监测技术研究[D];南京航空航天大学;2006年
2 思园园;大跨径连续刚构桥结构健康监测系统研究[D];长安大学;2012年
，本文编号：2278428
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2278428.html
