支持故障预测的传感器优化布置研究

【作者】 代凤娟；

【导师】 吴子燕；

【作者基本信息】 西北工业大学 ， 结构工程， 2007， 硕士

【摘要】 基于振动的结构损伤识别方法在结构康监测领域具有广阔的应用前景，其有效性是以测试的准确性为基础。在测试过程中，传感器布置的位置、数量对监测的结果起着决定性的作用。因此，从提高结构故障预测结果的准确性和可靠性的角度出发，研究如何确定传感器的最佳数目及如何将它们布置在最优位置，具有重要的理论意义和应用价值。 本文主要研究了支持故障预测的传感器优化布置方法。首先，分析了三种传统的传感器优化布置方法：第一种是以测得感兴趣模态保持最大线性无关为目标的有效独立法；第二种是以测得的运动能最大为目标的运动能量法；第三种是以模态对结构刚度的灵敏度最小为目标的灵敏度分析法。其次，在分析比较各方法优缺点的基础上，针对传统方法的不足提出基于信息优化的改进算法-有效独立驱动点残差法。改进的传感器优化布置算法是以反映运动能权重的残差系数C_DPR来修正反映最大线性无关的有效独立法，使传感器布置方案进一步综合优化。最后，以某桥面板结构为数值算例，运用模态保证准则（MAC）、均方差最小和抗噪声性能最好准则验证了本文所提出方法的有效性。 为了进一步验证本文所提出的传感器优化布置方法的有效性和实用性，以161杆空间桁架为研究对象进行了仿真模拟和实验验证，通过大型有限元软件ANSYS建立了完整的空间有限元模型，对其进行了模态分析。在取得传感器优化布置计算所需节点的各阶振型位移模态数据后，进而分别采用有效独立法、运动能量法和本文提出的有效独立驱动点残差法进行传感器优化布置，得到三种不同的优化布置方案。最后，用三个优化评价准则对各个传感器优化布置方案进行了分析研究。研究结果表明，通过有效独立驱动点残差法获得的传感器优化布置方案能得到更全面的结构参数信息，在此基础上拟合出来的结构模态更符合有限元数值计算结果，它是一种更有效的传感器优化布置方法。对于支持故障预测的传感器优化布置问题，有效独立驱动点残差法具有更可靠的参考价值和更高的实用意义。更多还原

【Abstract】 Structural damage identification based on vibration has broad application prospects in the field of structure for health monitoring. But its effectiveness is based on accuracy of modal testing, while the measurement locations and the number of transducers have a decisive effect on testing results. So, for improving the accuracy and reliability of results of failure prediction, the study about how to ensure the optimal number of transducers and how to place those on optimal locations have important theoretic sense and application value.This dissertation mainly studies three methods of optimal sensor placement of supporting failure prediction. The first is the Effective Independence Method (EFI) with the aim of maximizing the data information, the second is the Kinetic Energy Method(KEM) with the aim of measuring the maximal Kinetic Energy, the third is Sensitivity Analysis based on the component that models is sensitive to the stiffness . Based on merits or shortage of those, An improved algorithm- the Effective Independence Driving-Point Residue (EFI-DPR) method is given for the shortages of the above two methods which are the Effective Independence Method (EFI) and the Kinetic Energy Method(KEM). The improved optimal algorithm using Driving-Point Residue coefficient reflecting the weight of the kinetic energy updates the Effective Independence Method with the aim of maximizing spatial independent, the sensor placement is much more integrating optimal than before. Finally, Using Modal Assurance Criterion, Mean Error Minimum Criterion was based on the error between the FE model and the cubic spline interpolated mode shapes and the best ability to withstand the noise pollution Criterion validates the efficiency of this method by numerical analysis for a three span continuous girder deck slab structure. The results of numerical analysis show the optimization method proposed in this dissertation is proved a better method.In order to explain the efficiency and application of the method which this dissertation is proposed, take 161 spatial truss stayed bridge as a r applying research object, First establish the integral finite element model of the spatial truss and do an更多还原