【摘要】 为提高桥梁健康监测系统的监测能力,完善状态评估系统,对结构特性参数的门槛值进行分级。Fisher最优分割原理可以将有序序列进行最优分类,使得离差平方和最小。建立ANSYS有限元模型,采用实测轴重数据,进行动力分析,对桥梁中跨跨中的横向位移监测指标从数值上按照Fisher最优分割原理进行分级分为:正常值,异常值,报警值。为检验分级效果,将横向位移在各数值段内出现的频率作为分级指标。结果表明,两种分级指标得到的门槛值较吻合,符合统计规律;2.205 mm和2.455 mm可作为横向位移的门槛值。其他监测指标可以采用同样的方法得到。更多还原

【Abstract】 To enhance the capacity of the bridge health monitoring system and perfect the state assessment system,the classification of the threshold values of the structure characteristic parameters was performed.An ordered sequence can be optimally classified several clusters using the principle of Fisher optimal division,with the total sum of squares minimum.The finite model by ANSYS was built,then the dynamic analysis was made adopting measured axle loads.Based on the values of the lateral displacements at the middle of the mid-span of the bridge,three grades were classified,which were the normal value,the abnormal value and the alarm value.To check the results,the frequencies of the lateral displacements in each region were classified as an another index.The results show that the classifications by the two methods are approximate and are up to the statistical law.Then the values of 2.205 m and 2.455 m can be used as the threshold values of the lateral displacements for 3 grades.Another structure characteristic parameters such as stress,strain,acceleration,etc,can be obtained in the same way.更多还原