基于卷积神经网络和交叉模型交叉模态方法的钢框架两阶段损伤识别研究

Research on Two-Stage Damage Identification of Steel Frame Based on CNN and CMCM

在结构损伤识别领域中,交叉模型交叉模态方法(CMCM)在求解含噪声的损伤识别问题时,受到系数矩阵不稳定的影响,往往导致损伤识别结果的精度下降。为了解决这一问题,提出了一种两阶段的损伤识别方法,旨在减少CMCM方法系数矩阵的冗余方程,提高CMCM方法的计算性能和噪声鲁棒性。首先,在第一阶段,通过卷积神经网络(CNN)进行结构损伤定位,以消除CMCM方法系数矩阵的冗余方程。其次,在第二阶段,求解已缩减的CMCM系数矩阵方程,得到更准确的损伤识别结果。通过数值和试验研究,验证了所提出的两阶段损伤识别方法的有效性,与传统的CMCM方法相比,所提方法在求解含噪声的损伤识别问题时,显著提高了损伤识别精度。

In the field of structural damage identification,the Cross Model Cross Mode(CMCM)Method is constrained by the rank deficiency of the coefficient matrix when solving noise-inclusive damage identification problems,leading to a decrease in the accuracy of damage identification results.To address this issue,this study proposes a two-stage damage identification method aimed at reducing the redundant equations in the CMCM method’s coefficient matrix,thereby enhancing the computational performance and noise robustness of the CMCM method.In the first stage,a Convolutional Neural Network(CNN)is employed for structural damage localization to eliminate the redundant equations in the coefficient matrix of the CMCM method.Subsequently,in the second stage,the reduced CMCM coefficient matrix equation is solved to obtain more accurate damage identification results.The effectiveness of the proposed two-stage damage identification method is validated through numerical and experimental studies.Compared with the traditional CMCM method,the method proposed in this paper significantly improves the accuracy of damage identification in solving noise-inclusive damage identification problems,demonstrating its superior performance.