一种基于模型修改的结构多点损伤识别方法

可用于结构损伤识别的方法很多。一般来讲正向方法直接利用结构模态参数的变化，逆向方法则利用模态参数变化反演结构物理参数变化，还有些方法利用了神经网络和模式识别技术。文中利用模型修改的思想，通过逆向方法计算结构单元刚度变化系数来对结构的多点损伤进行识别。以一个七自由度弹簧阻尼质量系统作为研究对象，用数值模拟方法及特征系统实现算法计算系统的模态参数，并用这些模态参数验证所提出方法的可行性，结果表明该方法对多点损伤的识别是简单而可行的。

APPROACH OF STRUCTURAL MULTIPLE DAMAGES IDENTIFICATION BASED ON MODEL UPDATING

Structure damages can be sensed and localized using different approaches, including both the forward and inverse ones. While making use of the change of modal parameters belongs to the forward approach, monitoring the change of structural physical parameters according to modal parameters is the inverse one. Other methods such as neural network and pattern recognition can also be adopted for damage identification. A novel inverse approach that refers to the theory of model updating is proposed. Here the finite element model of structure is established and expressed as a form of status equation. The damage location and quantity can be identified by the change coefficient of the stiffness elements. In order to solve the change coefficient, an error that makes use of orthogonality of modal shapes is defined. With the least square method and modal parameters of damaged structure the coefficient can be obtained conveniently. This method is then applied to detect damage in a system with seven degrees of freedom (DOF) though numerical simulation. Results show that the proposed method is simple and feasible in monitoring multiple damages.