为对岩爆现有预测及防治技术进行整理和总结，搜集深埋岩爆隧道的大量资料，统计分析已建、在建隧道发生大面积岩爆现象的原因、过程和特征，对岩爆规律、预测方法以及处置措施进行总结和分析。得出结论如下： 1）在岩爆预测阶段，基于岩体物理力学性质的岩爆等级判别式具有一定通用性，判别标准的数值可应用于岩性、埋深相近的隧道；采用多个判别式可相互检验隧道的预测岩爆的准确性及特征，现场监控量测为制定判别标准提供了重要参考和有效性验证。2）细化岩爆防治阶段不同岩爆等级的应力解除方法： 在轻微岩爆段，采用洒水或者灌水； 在严重岩爆段，利用应力孔进行应力释放。3）总结提出不同等级围岩段落的初期支护参数，针对钻爆方法，通过对比总结，提出不同岩爆段落的开挖方法、进尺长度以及相应的等待时间。4）提出利用机器学习、大数据库、无人机、机器视觉、信号采集、多光谱兼热成像技术，实现岩爆预测信息采集及岩爆过程位移场、应变场的全过程监测；提出利用离散元能量迭代算法模拟岩爆过程的新技术。

In this study, a large number of deepburied rockburst tunnel data are collected. The causes, processes, and characteristics of largearea rockburst that occurred in existing and underconstruction tunnels are statistically analyzed, and the rockburst trends, prediction methods, and disposal measures are summarized and analyzed. The drawn conclusions are presented as follows. (1) In the rockburst prediction stage, the rockburstclassification discriminant based on the physicomechanical properties is universal to a certain extent, and the criterion value can be applied to tunnels with similar lithology and depth. The accuracy and characteristics of the prediction results can be checked using multiple discriminants, and the fieldmonitoring measurement provides an important reference and validation for the establishment of the criterion. (2) The stressrelief methods of different rockburst grades in the prevention and control stage are water sprinkling for the slightrockburst section and stress release hole for the strongrockburst section. (3) The primary support parameters of surrounding rocks with different grades are summarized, and based on the comparison and summary of the drilling and blasting methods, the excavation methods, advancement length, and corresponding waiting time of the differentrock blasting sections are proposed. (4) It is recommended that the acquisition of rockburst prediction information and the entire monitoring of the displacement and strain fields in the rockburst process can be realized using the technologies of machine learning, large database, unmanned aerial vehicle, machine vision, signal acquisition, and multispectral and thermal imaging. A new simulation method for rockburst process using discreteelement energy iteration algorithm is proposed.