基于分布式识别策略的超长线状盾构隧道惟响应损伤定位方法研究

The healthy service of rail transit is vital important for the normal operation of a city. The shield tunnel as the main structural type of rail transit has typical nonlinear characteristics because of its dense joints, complex service environment and traversing through different geological conditions, which lead to the difficulty of its damage identification. This has become the bottleneck restricting the development of the health monitoring of subway. Therefore, the study of the method for long-line shield tunnel structural damage identification is of great value theoretically and practically. This project proposed a method dividing the long-line shield tunnel into linear connection subinterval structures aiming at decentralized identification strategy based on the principle of mutual information; exploiting wavelet packet transform to analyze dynamic response under the action of running trains, extracting local damage-sensitive characteristic parameter and building statistical damage identification index; setting the threshold value by using statistical method to removing the influence of noise and environmental conditions, and discriminating the occurrence and location of damage on subinterval structure by using its space information. The proposed method is an output-only damage identification method based on decentralized identification strategy. It makes full use of the sensitive characteristics of local dynamic response to small damage and has merits of less calculation, high accuracy and significant efficiency, which overcomes the influence of noise and the service environmental change on the accuracy of identification. This project can finally provide accurate and reliable basis for safe service and scientific management and maintenance of shield tunnel structures.

城市轨道交通的健康服役对于城市正常运转至关重要。盾构隧道作为城市轨道交通隧道结构的主要形式，接缝密集，服役环境复杂，穿越不同地质条件，具有典型的非线性特征，造成其损伤识别困难，已成为制约地铁健康监测发展的瓶颈，研究超长线状盾构隧道结构的损伤识别方法具有重要的理论与应用价值。本项目提出基于互信息的面向分布式识别策略的子区间划分方法，将超长线状盾构隧道划分为线性连接的子区间；利用小波包变换分析各子区间在运营列车作用下的加速度响应，提取局部损伤敏感特征参量并构建损伤识别指标；在此基础上，利用统计方法去除噪声和环境变化的影响，实现子区间是损伤状况的判别，并利用子区间的空间信息实现定位。该方法是一种基于分布式识别策略的惟响应损伤识别方法，充分利用了局部动力响应对小损伤敏感的特征，且计算量小，识别效率高，克服了噪声及环境变化对识别精度的影响，可为盾构隧道结构的安全服役和科学管养提供准确与可靠的依据。

城市轨道交通的健康服役对于城市正常运转至关重要。盾构隧道作为城市轨道交通隧道结构的主要形式，接缝密集，服役环境复杂，穿越不同地质条件，具有典型的非线性特征，造成其损伤识别困难，已成为制约地铁健康监测发展的瓶颈，研究超长线状盾构隧道结构的损伤识别方法具有重要的理论与应用价值。本项目提出基于互信息的面向分布式识别策略的子区间划分方法，将超长线状盾构隧道划分为线性连接的子区间；利用小波包变换分析各子区间在运营列车作用下的加速度响应，提取局部损伤敏感特征参量并构建损伤识别指标；在此基础上，利用统计方法去除噪声和环境变化的影响，实现子区间是损伤状况的判别，并利用子区间的空间信息实现定位。通过该项目的研究，形成了一种基于分布式识别策略的惟响应损伤识别方法，充分利用了局部动力响应对小损伤敏感的特征，且计算量小，识别效率高，克服了噪声及环境变化对识别精度的影响，可为盾构隧道结构的安全服役和科学管养提供准确与可靠的依据。该项目共发表论文15篇，其中SCIE检索11篇，中文EI论文1篇，中文核心论文3篇。参加国内外学术会议5次并作报告，发表会议论文4篇。申请发明专利3项，未来有望实现推广应用。有2名博士生和3名硕士生通过该项目的研究，获得了学位，项目组成员有3为职称得到了晋升。