基于Kaplan—Meier法的在役梁拱组合体系性能动态评估与趋势预测算法研究  

Research on the Dynamic Performance Assessment and Trend Prediction Algorithm for In-service Arch-Girder Composite Structures based on the Kaplan-Meier Method

作　　者：王亮 乔维 马瑞东 李东 桂成中 段志 WANG Liang;QIAO Wei;MA Rui-dong;LI Dong;GUI Cheng-zhong;DUAN Zhi(Power China Road Bridge Group Co.Ltd.,Beijing 100038,China;China Electric Power Construction(Guangdong)Zhongshan-Kaiping Expressway Co.Ltd.,Jiangmen 529142,China;China Highway Engineering Consulting Corporation,Beijing 100089,China;CHECC Highway Maintenance and Test Technology Co.Ltd.,Beijing 100097,China;Department of Civil Engineering,Institute of Disaster Prevention,Langfang 065201,China;Key Laboratory of Building Failure Mechanism and Disaster Prevention,CEA,Langfang 065201,China)

机构地区：[1]中电建路桥集团有限公司,北京市100038 [2]中电建(广东)中开高速公路有限公司,江门市529142 [3]中国公路工程咨询集团有限公司,北京市100089 [4]中咨公路养护检测技术有限公司,北京市100097 [5]防灾科技学院土木工程学院,廊坊市065201 [6]北京防灾科技有限公司,廊坊市065201

出　　处：《公路》2025年第3期413-421,共9页Highway

基　　金：中国地震局地震科技星火计划项目,项目编号XH24063A;中国电力建设股份有限公司科技项目,项目编号DJ-ZDXM-2022-35;中央高校基本科研业务费专项,项目编号ZY20230202;廊坊市科技计划项目,项目编号2022011066。

摘　　要：为了对在役梁拱组合体系进行性能动态评估与时机预测,开发了一种基于Kaplan—Meier法的评估与预测模型。首先通过数值模拟确定结构构件在3种受力状态下的目标阈值,并据此构建性能指标体系。利用三次样条插值技术对运营期多源异构监测数据进行预处理,并通过关键因子分析筛选关键参数。针对桥梁监测数据的非正态分布特性,实施正态分布检验,建立Kaplan—Meier生存概率模型。进一步地,基于生存概率算法,构建多参数性能拟合回归模型,以预测结构体系的不利性能状态和时机。案例研究验证了该方法的有效性,该模型能为桥梁养护决策提供准确的科学依据,为桥梁的智慧运维提供理论支持和技术路径。A Kaplan-Meier method based evaluation and prediction model is developed for dynamic performance evaluation and timing prediction of beam and arch combination system in service.The target thresholds for structural components under three types of loading conditions are first determined through numerical simulation and a performance indicator system is constructed accordingly.The multi-source heterogeneous monitoring data are preprocessed by cubic spline interpolation,and the key parameters are screened by key factor analysis.According to the non-normal distribution characteristics of bridge monitoring data,the normal distribution test is carried out and Kaplan-Meier survival probability model is established.Furthermore,based on the survival probability algorithm,a multi-parameter performance fitting regression model is established to predict the adverse performance state and timing of the structural system.The case study verifies the effectiveness of the method,and the model can provide accurate scientific basis for bridge maintenance decision.The study provides theoretical support and technical path for intelligent operation and maintenance of Bridges.

关 键 词：桥梁工程 桥梁健康监测 梁拱组合体系 目标阈值 Kaplan—Meier法 性能动态评估 

分 类 号：U446.2[建筑科学—桥梁与隧道工程]