摘要
针对传统曲率模态差法在结构损伤识别应用中振型节点处损伤识别不敏感与损伤程度反映不准确的问题,通过理论推导,分析曲率模态差法存在局限性的原因。经过改进曲率模态差指标,提出归一化叠加曲率模态方差比指标和单元损伤估算方法,从理论层面解决了曲率模态差指标的不足。此外,通过室内简支梁试验及实际连续梁桥有限元算例,验证了新指标较曲率模态差指标在损伤识别及程度分析方面的优势。进一步提出了完整曲线法,对归一化叠加曲率模态方差比指标的使用进行补充,通过模拟无损桥梁模态曲线,为实际桥梁检测时无损状态下模态参数难以获得的问题提供了解决方案,最后研究了噪音对新指标的影响。研究结果表明:1)归一化叠加曲率模态方差比指标能够准确识别结构损伤位置;2)单元损伤估算方法准确度为94.9%,相较于节点损伤估算法提高23.4%,可以较好地反映结构损伤程度;3)完整曲线法拟合复相关系数均值为0.995,与无损结构模态曲线的拟合负相关系数均值为0.935,属于高度相关,表明完整曲线法可以有效模拟桥梁无损状态下的模态曲线;4)噪音作用下,新指标可以准确识别结构损伤位置,损伤程度估算准确度下降2.6%,完整曲线法对无损模态曲线的拟合负相关系数均值为0.866,减小0.069,仍属于高度相关,说明噪音对归一化叠加曲率模态方差比指标、单元损伤估算方法及完整曲线法的影响较小。研究成果可为实际桥梁损伤识别提供参考。
To address the limitations of the traditional curvature mode difference method in structural damage identification,specifically its insensitivity to damage at mode shape nodes and the inaccuracy in damage severity assessment,this study conducted a theoretical analysis to uncover the underlying causes.An improved curvature mode difference index,termed the normalized cumulative curvature mode variance ratio index,was developed,alongside a method for estimating potential risks,thereby theoretically resolving the shortcomings of the original index.The superiority of the new index over the traditional curvature mode difference method in identifying damage and assessing its severity was validated through both laboratory experiments on simply supported beams and finite element simulations of actual continuous beam bridges.Additionally,a complete curve method was proposed to complement the normalized cumulative curvature mode variance ratio index.This method simulated the modal curves of undamaged bridges,providing a practical solution to the challenge of obtaining modal parameters from undamaged conditions during real-world bridge inspections.The impact of noise on the new index was also investigated.The results indicate that:(1)the normalized peak curvature mode variance ratio index accurately identifies structural damage locations;(2)the element damage estimation method achieves an accuracy of 94.9%,representing a 23.4%improvement over adjacent node damage curvature methods,and effectively reflects the extent of structural damage;(3)the complete curve method achieves an average complex correlation coefficient of 0.995 with the undamaged structure’s modal curve and an average fitting negative correlation coefficient of 0.935,indicating a high degree of correlation and confirming the method’s effectiveness in simulating undamaged modal curves;(4)under noise influence,the new index continues to accurately identify damage locations,with damage estimation accuracy declining by only 2.6%.The noise correlation coefficient for the complete curve method on undamaged modal curves is reduced by 0.069 to 0.866,yet still reflects a high degree of correlation.These findings offer valuable insights for practical structural damage identification.
作者
李一鸣
刘屺阳
张戎令
刘德安
张军
LI Yiming;LIU Qiyang;ZHANG Rongling;LIU De’an;ZHANG Jun(Gansu Provincial Institute of Transportation Science Group Co.,Ltd.,Lanzhou 730070,China;Innovative Service Platform for Highway Test and Testing,Gansu Province,Lanzhou 730070,China;Lanzhou Jiaotong University School of Civil Engineering,Lanzhou 730070,China)
出处
《铁道科学与工程学报》
北大核心
2025年第6期2824-2834,共11页
Journal of Railway Science and Engineering
基金
兰州市科技计划项目(2022-5-48)
甘肃省公路交通建设集团有限公司科技项目(2022-ZH-061)。
关键词
桥梁工程
曲率模态
桥梁损伤识别
曲线拟合
损伤程度评定
bridge project
curvature modes
bridge damage identification
curve fitting
damage assessment
