改进的数据驱动的随机子空间算法在桥梁监测中的应用

Application of improved data-driven stochastic subspace algorithms in bridge monitoring

介绍了基于数据驱动的随机子空间算法的改进方法,传统的数据驱动随机子空间算法由于数据量大,处理速度慢而很难用于在线识别和监测,改进的算法在减少数据量的同时仅利用分解得到可观矩阵即可完成整个参数识别,大大提高了运算速度,满足桥梁健康监测的在线模态参数识别和动力模型修正及损伤识别,具有很高的工程应用价值.该改进算法已成功应用于大广高速公路的通用型桥梁健康检测系统.

This paper reviews the improved algorithm based on data-driven stochastic subspace.By dealing with a large number of data,the traditional algorithm based on data-driven stochastic subspace is difficult to apply to on-line system identification and monitoring.However,the improved algorithm can identify all parameters only by decomposed observablity matrix,while large amounts of data were decreased and the operation speed is greatly improved.This has high engineering application value and is applied on bridge health monitoring,dynamic model modification and damage identification.The improved algorithms has been successfully applied to the health monitoring system of general bridges of Daguang-Highway of Heibei Province.