基于节点曲率和小波分析的梁式桥多尺度损伤识别

摘要

A damage identification method based on nodal curvatures and wavelet analysis (NCWA)was presented with respect to the Xinyihe Bridge.The proposed methodology was based on the moment -curvature relations and the assumption that internal stress resultants keep invariant before and after damage.Making use of the particular advantage of noise eliminating of wavelet analysis,the linear matrix equations of the pre-damage and post-damage nodal curvatures were solved by using singular value decomposition (SVD).The damage index based on nodal curvature was obtained.A simply supported beam model testing was carried out to verify the feasibility of the proposed method.The sensitivity and anti-noise ability of the damage idenfication method based on NCWA were verified by the multi-scale finite element analysis and dynamic loading test on the Xinyihe Bridge.The results show that the locations of different scale damages can be identified by the method of NCWA without considering noise effects,and the damage sensitivity of small scale units is superior to big scale units.If considering noise effects,the locations of small damage scale can be identified approximately by NCWA,and the anti-noise ability of small scale units is superior to big scale units.The results support the idea that the proposed damage identification method based on NCWA has a great potential in the health monitoring of practical engineering structures,and it lays a solid foundation for the damage prognosis (DP)and safety prognosis (DP)of girder structures.%以新沂河大桥为工程背景，提出了一种基于节点曲率和小波分析（NCWA）的梁式桥多尺度损伤识别方法。首先基于结构弯矩－曲率基本关系和结构微损伤对其应力重分布影响很小的假定，结合小波分析的消噪功能，采用奇异值分解（SVD）方法求解节点曲率损伤前后的线形矩阵方程，推导了基于节点曲率的损伤指标，并通过简支梁试验验证了该方法的理论可行性，最后新沂河大桥多尺度数值模型试验和动载试验验证了基于 NCWA 识别方法的损伤敏感性和抗噪性。结果表明：在不考虑噪声干扰作用下，基于节点曲率的损伤识别方法能较好实现结构不同尺度的损伤识别，但小尺度单元区域的识别效果普遍优于大尺度单元区域；在考虑噪声干扰作用下，基于 NCWA 的损伤识别方法基本能够实现结构小尺度下的损伤识别，小尺度单元区域的比大尺度单元区域的损伤识别抗噪性更好。提出的基于 NCWA 的多尺度损伤识别方法具有应用到实际工程健康监测中的潜力，可为梁式结构损伤及安全预后奠定必要的基础。