随着我国基础设施建设的快速发展，交通、水利水电等重大工程建设迎来了全新的发展机遇，极大的促进了深长隧道(洞)工程的建设。一大批深长隧道工程正在或即将修筑于地形地质复杂的山区与岩溶地区，强富水、高承压或高地应力的地下环境极易诱发重大突涌水灾害。 突涌水灾害难以遏制的重要原因在于对复杂的动力灾变演化机理认识不清、传统的监测技术与预警方法难以实现灾害的有效主动防控。本文以隧道突涌水灾害微震机理与监测分析方法为主题开展了研究，揭示充填介质渗透失稳与岩体渐进破坏演化过程中微震动响应规律，开展了声发射／微震监测、有效信息辨识、灾害状态判定以及灾害预警等研究，采用案例分析、室内试验、现场试验及软件开发等方法获得了以下成果。 (1)从突涌水灾害的赋存环境入手开展相关研究，分别阐述了富水断裂破碎带、充水充泥岩溶体两类地质构造的识别方法，并进一步的分析了两类典型致灾构造的孕灾模式，分析了释放条件、储水条件和诱发条件三方面因素对突涌水灾害的影响规律。对比分析岩爆与突涌水灾害震源机制的差异的基础上，建立了两类灾害微震监测的区别与联系，进而建立了充填介质渗透失稳与岩体渐进破坏演化过程中微震响应机制。结合隧道工程施工经验讨论了微震监测在实际应用中存在的问题，拟从软件层、硬件层和技术层对传统的微震装备进行改造，为适用于隧道突涌水灾害有效微震监测提供了研发改进思路。 (2)通过饱水与天然状态的岩石单轴压缩室内试验，发现了岩石破裂声发射前兆特征，揭示了水体对总输入能、弹性储能的弱化机理。以类岩石材料高压水力破裂的微震监测信息为样本，揭示了突水通道破裂活动的微震信号频率范围。提出了表征微震活动时间演化与能量评估的微震撞击事件概念，圈定了突水通道破裂微震撞击事件发生频次与能量密度的预警阈值，为岩体渐进破坏突涌水灾害微震监测预警提供了借鉴。 (3)物理模型试验还原了既有突水通道内水流、充填物间相互冲击的活动状态，以试验监测的微震信息为分析样本，总结了有效微震信号和噪声微震信号的波形与频谱特征，提出了基于充填块体与通道岩壁撞击微震信号定位突水通道的思路。引入小波包分解理论实现了低信噪比信号于频率域上的分解，圈定了多种微震信号的频率范围，建立了有效信号甄别提取方法，试验分析结果可为突涌水灾害专用微震装备的研发提供参数支撑。 (4)嵌入的小波包分解程序实现了有效微震信号的特征分析与甄别，为突水通道空间定位提供了理论基础。引入方差因子改进了STL／LTA算法，提高了震源初至拾取精度。首创了适应于隧道工程环境的四元平面微震监测方法，实现了震源解析解的求算，规避了目标方程求解的病态问题。以隧道工程为例，设计了四元平面监测网络布置方案，实现地下洞室全方位的立体监测。基于Matlab语言开发了一套集信号读取、滤波、初至拾取和震源定位功能的微震信息分析软件，实现了微震信息的高效率半自动化处理。 (5)以两类典型防突结构破坏全过程的多场物理信息为分析样本，提出了灾害演化状态判识与预警方法。建立了主成分分析及函数态势的灾变全过程的综合分析方法，提出了突涌水灾害多参量的综合评价模型，将突涌水灾害演化状态划分成平静期、发展期、突变期和灾后期，建立了任一时刻下突涌水灾害发生概率的能量计算方法。 (6)以滨莱高速乐疃隧道施工环境为微震监测背景，揭示了背景噪声、爆破作业、锚喷工作、凿岩工作、淋水及疑似突水通道破裂信号伴随的微震信号波形、频谱特征，由隧道生产活动组成的微震监测信号样本可为数据解译与无效信息剥离提供样本支持。 关键词：隧道工程；突涌水灾害；微震技术；信息监测；灾害预警 本文的研究由国家重点基础研究计划(973计划)项目：深长隧道突水突泥重大灾害致灾机理及预测预警与控制理论(2013CB036000)资助完成。

 With the rapid development of infrastructure construction in China, major construction projects such as transportation, water conservancy and hydropower have ushered in new opportunities for development, which greatly promotes the construction of deep and long tunnels. A large number of deep and long tunnel projects are or will be constructed in mountainous and karst areas with complex topography and geology. The underground environment with strong water-rich, high pressure or ground stress is easy to induce serious water inrush disaster. The reason why the water inrush disaster is difficult to contain is that the complex dynamic catastrophe evolution mechanism is not clear and the traditional monitoring technology and early warning methods are difficult to achieve effective active disaster prevention and control. This paper focuses on the topic of microseismic(MS) monitoring technology and early warning method of tunnel inrush water disaster. The MS response of the two typical inrush water disasters is a key scientific problem. The author has carried out research on acoustic emission/MS monitoring, effective information extraction, disaster status identification and disaster early warning. The following results are obtained through case summary analysis, laboratory testing, software development and other methods. (1)The author first carries out relevant research on the occurrence environment of inrush water disaster. The identification methods of water-filled fracture zone and water-filled mudstone karst are expounded respectively. Furthermore, the pregnancy disaster models of these two typical disaster-causing structures are analyzed. The characteristics of the three factors of release conditions, water storage conditions and induced conditions on water inrush disaster are revealed. On the basis of the difference analysis of focal mechanism between rock burst and water inrush, the relationship between MS monitoring of rock burst and water inrush disaster are established. Based on the comparative analysis of the differences in the focal mechanism of rock burst and inrush water disasters, the connection between two types of disaster MS monitoring is established. The focal mechanism of two typical types of inrush water disasters has been proposed. Combined with the experience of tunnel construction, the water inrush MS monitoring problems on the site are discussed. Traditional MS equipment should be upgraded from three aspects: software, hardware and technology. It provides a new idea for the effective MS monitoring of water inrush disaster in tunnels. (2)The weakening mechanism of water to total input energy and elastic energy storage is revealed, the acoustic emission(AE) precursor characteristic of rock failure is proposed, through the laboratory test of saturated and natural rock under uniaxial compression. Taking the MS monitoring information of high-pressure hydraulic fracture of rock-like materials as samples, the main frequency range of effective signals is revealed. The concept of the MS impact event for characterizing the evolution of time and energy of MS events is proposed. The warning thresholds of the frequency and energy density of MS impact events under the failure of water inrush channel are proposed, which provides a reference for the MS monitoring and early warning of rock mass progressive damage water inrush disaster. (3)The physical model test restores the activity state of water flow and filling in the flow channel. The waveform and spectrum characteristic of the effective MS signal and the noise MS signal in the water flow channel are summarized. The idea of locating the water inrush channel based on the impact MS signal of the filling block and the channel rock wall is put forward. The wavelet packet decomposition theory is introduced to realize the decomposition of low SNR signal in the frequency domain. The frequency range of various MS signals is determined. The efficient signal extraction method has been proposed. (4)The wavelet packet filtering theory is embedded in the system to realize the extraction of effective MS signals. The improved STL/LTA algorithm is embedded in the system, which greatly improves the picking accuracy of the source first arrival. A four-element planar MS monitoring method adapted to the tunnel engineering environment has been proposed. This method realizes the accurate calculation of the source and avoids the ill-conditioned problem of solving objective equations. Based on the Matlab language the MS information analysis software is developed, which sets the functions of signal reading, filtering, first arrival pick-up and source location. The analysis software has been successfully applied in the analysis of test data. (5)Based on the whole process evolution information of the two typical water inrush disasters, the fusion analysis and early warning method of multiple information in water inrush disaster is established. A comprehensive analysis method for the whole process of catastrophe based on principal component analysis and functional situation is established. A comprehensive evaluation model for multi-parameters of inrush water disasters is established, and the inrush water disasters are divided into four stages: calm period, development period, mutation period and late stage. Based on the energy theory the probability of the occurrence of water inrush disaster at any time is determined. (6)Based on the construction activities of the Letuan tunnel on the Binlai expressway, the MS signal waveform and spectrum characteristics of the background noise, blasting operation, anchor spray work, rock drilling work, water spraying and suspected water inrush channel failure are revealed. The MS monitoring signal generated by tunnel production activities provide sample support for data interpretation and stripping of invalid information. Keywords: Tunnel engineering; Water inrush disaster; Microseismic monitoring technique; Monitoring information; Disaster early warning This paper was supported by the State Key Development Program for Basic Research of China Grant (973 plan) “The Long Deep Tunnel Water Inrush and Mud Inrush Disaster Mechanism and Forecast and Control Theory” (201 3CB036000).