Optimization Method of Sensor Arrangement for Health Monitoring of String Structure
-
摘要: 传感器布置方案设计是结构健康监测中最为基础性和关键性的环节,好的传感器布置方案,既要满足结构健康监测的要求,尽可能地减小噪声等误差因素的影响来得到结构的真实响应,又要满足现场环境的要求,还需要尽可能地降低成本。
针对此问题,提出了一种基于模态置信矩阵的传感器优化布置方法。根据结构的模态振型矩阵,计算得到结构的模态置信度矩阵。基于模态可观测原则,选取模态置信矩阵中非对角元最大值作为评价标准,采用迭代的方法对传感器的布置方案进行优化。具体操作步骤为:1)确定结构健康监测所需要识别的模态数以及所需传感器的数目,根据现场环境确定所有可采取的测点位置;2)选取一组初始传感器测点布置。初始传感器测点布置方案可根据经验确定,初始传感器个数应略少于待识别模态数。计算该组测点所对应的模态置信矩阵并记录其对应的最大非对角元的值;3)在剩余可选测点中选取一个测点增加到当前测点布置方案中,计算新测点方案对应的模态置信矩阵,同时记录模态置信矩阵中的最大非对角元;4)更换所选取的待选测点,重复模态置信矩阵的计算步骤并记录最大非对角元。重复此步骤直至所有待选测点都被计算过。对比所有备选测点所对应的模态置信矩阵最大非大对角元,选择最小的模态置信矩阵最大非对角元对应的测点加入到当前测点布置方案;5)重复3)、4)步骤,直至传感器测点数及模态置信矩阵的最大非对角元满足要求。对一些复杂结构,模态置信矩阵最大非对角元小于0.25。
根据以上方法,以张弦梁结构为例建立数值模型,并模拟实际工程环境,进行了传感器优化布置的试算。经试算,该方法的收敛性较好,计算可行性高;该方法可以有效降低传感器布置的经济成本,降低误差因素的影响,提高模态识别的有效性和准确性。同时根据该算例的试算,提出建议:逐步累加法的迭代过程可分为快速下降段、平稳阶段和迭代末段三个阶段,实际传感器的布置方案宜在平稳阶段过程中选取,尽量避免采用迭代末段的结果。Abstract: Sensor arrangement scheme design is the most basic and key link in structural health monitoring, a good sensor arrangement scheme not only meets the requirements of structural health monitoring, but also needs to reduce the cost as much as possible to get the true response of the structure by minimizing the influence of errors such as noise, etc. as well as the requirements of the field environment.
In order to solve this problem, this paper proposes a method of optimal sensor placement based on modal confidence matrix. According to the modal matrix of the structure, the modal confidence matrix of the structure is calculated. Based on the principle of modal observability, the maximum value of the non diagonal element in the modal confidence matrix is selected as the evaluation standard, and the sensor layout scheme is optimized by iterative method. The specific operation steps are as follows:1)Determine the number of modes needed to be identified and the number of sensors needed for structural health monitoring, and determine the positions of all available measuring points according to the site environment; 2)Select a group of initial sensor measuring point layout. The initial measurement point scheme can be determined according to experience, which is less than the number of modes to be identified. The modal confidence matrix corresponding to the group of measuring points is calculated and the corresponding value of the largest non diagonal element is recorded; 3)Select one of the remaining optional measurement points to add to the current measurement point layout scheme, calculate the modal confidence matrix corresponding to the new measurement point scheme, and record the largest non diagonal element in the modal confidence matrix; 4)Replace the selected measuring points, repeat the calculation steps of the modal confidence matrix and record the maximum non diagonal elements. Repeat this step until all the points to be selected have been calculated. Compared with the maximum value of the non large diagonal elements of the modal confidence matrix corresponding to all the alternative test points, the smallest maximum non diagonal elements of the modal confidence matrix corresponding to the test points are selected to add to the current test point layout scheme; 5)Repeat steps 3) and 4) until the sensor measurement points and the maximum non diagonal element of the modal confidence matrix meet the requirements. For some complex structures, the maximum non diagonal element of the modal confidence matrix is less than 0.25.
According to the above methods, this paper takes the beam string structure as an example to establish the numerical model, and simulates the actual engineering environment, and carries out the trial calculation of the optimal arrangement of sensors. After trial calculation, the convergence of this method is good and the feasibility of calculation is high; it can effectively reduce the economic cost of sensor arrangement, reduce the influence of error factors, and improve the effectiveness and accuracy of modal identification. And according to the trial calculation of the example, this paper proposes:the iterative process of the step-by-step accumulation method can be divided into three stages:fast descent stage, stable stage and iterative end stage. The actual sensor arrangement scheme should be selected in the stable stage process to avoid using the results of iterative end stage as much as possible. -
章国稳. 环境激励下结构模态参数自动识别与算法优化[D].重庆:重庆大学,2012. 马国伟.无线网络传感器在建筑结构健康监测中的应用[J]. 电脑与电信,2009(1):32. 杨雅勋. 基于动力测试的桥梁结构损伤识别与综合评估理论研究[D].西安:长安大学,2008. Breitfeld T. A method for identification of a set of optimal measurement points for experimental modal analysis:modal analysis[J].The International Journal of Analytical and Experimental Modal Analysis,1996(11):254-273. 高玉金,张和豪,颜定宏. 一个新的模态置信判别系数[J]. 振动,测试与诊断,1987(1):1-5. 田向阳. 张弦梁结构动力特性分析[J]. 山西建筑,2008(10):82-83.
点击查看大图
计量
- 文章访问数: 299
- HTML全文浏览量: 32
- PDF下载量: 17
- 被引次数: 0