地质雷达用小波基构造方法及其在隧道超前地质预报中的应用

It is likely that there are some problems in the process of tunnel constructions, such as faults, crush zones, weak intercalated layers, caves and other unfavorable geological conditions. To ensure the safety of tunnel construction，geological prediction is widely-adopted to explore potential geological conditions in front of tunnel face. The analysis software in instrument is based on the Fourier analysis theory; however, it fails to accurately calculate unfavorable geological conditions of space shape, size and geometry dimensions. Therefore the analysis software seriously affects geological prediction effect in practical application. The characteristics of ground penetrating radar signals of geological prediction are complex and non-stationary. According to these characteristics, wavelet bases series which are suitable for characteristics of ground penetrating radar signals are constructed. The unique time-frequency localization ability and zoom spectrum analysis techniques of wavelet theory are used. Ground penetrating radar signals acquired under different geological conditions are analyzed and processed. The characteristics of ground penetrating radar signals are explored under unfavorable geological conditions, such as faults, crush zones, weak intercalated layers and caves. The key scientific issues such as advanced theories and interpretation technology which can rapidly, easily and accurately test, diagnose and presume unfavorable geological conditions are studied. The quantitative models of unfavorable geological conditions and early-warning mechanism of on-line are established. The construction methods, supporting times and supporting structures are reasonable adjusted. The theoretical basis and technical means are provided systematically for timely feedback analysis of dynamic information in the process of tunnel constructions.

隧道施工过程中可能会出现断层、破碎带、软弱夹层与溶洞等不良地质情况，为确保隧道施工安全，普遍运用超前地质预报探明掌子面前方的地质情况。然而，目前采用的仪器自带分析软件普遍基于傅里叶分析理论，无法对不良地质情况的空间形态、大小与几何尺寸进行精确计算，严重影响了超前地质预报的实际应用效果。本项目针对超前地质预报地质雷达信号具有复杂多变、非平稳的特点，构造适合地质雷达信号特征的小波基函数系列，运用小波理论独特的时频局部化能力和细化谱分析技术，对不同赋存条件下获取的地质雷达信号进行分析与处理，探索断层、破碎带、软弱夹层与溶洞等不良地质情况的信号特征，研究能快速、简捷、准确地对不良地质情况进行检测、诊断、推定的先进理论与信号解释技术等关键科学问题，建立不良地质情况的量化模型与在线预警机制，合理地对施工方法、支护时机、支护结构等加以调整，为系统地对隧道施工进行动态信息反馈分析提供了理论依据与技术手段。

随着“一带一路”战略的不断推进，基础设施建设得到了迅速发展。隧道与地下工程开挖面前方地质情况对工程施工影响日益突出，不仅影响工程的掘进速度，而且还会在施工过程中发生塌方、突水和突泥等地质灾害，给工程施工安全与正常运行带来严重的生命财产损失。开展隧道与地下工程超前地质预报与交（竣）工衬砌结构质量检测，对减少施工盲目性、确保工程运行具有重要的意义。.根据地质雷达发射子波的时频特征，分别采用连续小波变换理论、基于高斯函数的曲线拟合以及基于小波变换提升格式理论构造了地质雷达小波基，进而提出了小波变换时能密度法。将其应用于隧道与地下工程开挖面前方断层、溶洞充填物、浅层地表不同管线以及隧道交（竣）工衬砌混凝土不同厚度和病害体（含空洞、脱空与不密实）正演模拟与室内模型实验地质雷达检测信号的定量识别，并与波形分析法、经典小波变换法的识别效果进行分析与比较。.项目取得的主要成果包括：（1）随着介电常数的增大，地质雷达信号衰减逐渐增大，小波包频带能量的衰减不断增加，频带能量的方差与标准差越小；（2）采用小波变换时能密度法识别混凝土不同厚度的准确率高，相对误差小于5%，还给出了钢筋混凝土结构以及保护层厚度的分布情况；（3）不同管线地质雷达信号经F-K偏移后，提高了雷达图像的分辨率，压制了多次波的干扰，使双曲线绕射波能量更好聚集；（4）波形分析法需通过先验知识确定反射界面；经典小波变换法提取的位置与小波基的时频特征有关，相对误差约15%；当充填物的尺寸大于波长时，小波变换时能密度法识别误差小于5%。（5）采集充填物纵、横向多条测线的地质雷达时间剖面，采用F-K偏移技术使绕射波归位，根据不同测线在空间位置上的关联性，对目标体进行识别分析，获得目标体的空间位置与尺寸。.研究成果有助于实际工程中对隧道开挖面前方超前地质预报、浅地层地表管线以及衬砌厚度与质量缺陷地质雷达检测图谱解释提供依据，具有较大的理论意义和工程实用价值。