开挖卸荷下层状岩体洞室塌方灾变风险动态定量评估方法研究

Layered rock mass is a typical sensitive rock mass subject to excavation disturbance. It’s easy to be toppled or slipped during underground cavern excavation, which will result in collapse occurrence, cause the safety risk and delay project construction. This project intends to research on a dynamic and quantitative assessment method for catastrophic risk of cavern collapse in layered rock mass due to excavation-induced unloading. With ultrasonic testing, a mount of relaxation depths are captured. Based on intelligent algorithm, the rock mass parameters of the same lithology and stratum are inversed. And then, the probability density function is fitted. On the basis of research on parameters’ spatial variability, a relevant non-Gaussian random field is built by Karhunen-Loève (K-L) series expansion algorithm. The coupling between the parameters stochastic field and the grid model is further realized by unit traversal. Taking a failure approach index of layered rock mass as the assessment index, probability is calculated by a non-invasive stochastic finite difference method based on the probabilistic collocation method and Monte-Carlo method. Finally, according to dynamic fitting for probability density function on different excavation and support stages, a whole set of scientific, reasonable and feasible method for dynamic and quantitative risk assessment is proposed. Using this method can reveal the change rule of quantitative risk on different excavation and support stages, and provide theoretical guidance and technical support for the safety and rapid construction.

层状岩体是一类典型开挖扰动敏感岩体，在洞室开挖卸荷下极易出现弯折或滑移破坏，导致塌方的发生，进而造成施工安全风险并延误工程建设。项目拟针对开挖卸荷下层状岩体洞室塌方灾变过程中的风险开展动态定量评估方法研究：借助声波测试手段，捕捉大量的围岩松弛深度信息；基于智能算法，反演同一岩性、同一岩层的岩体力学参数，并拟合岩体力学参数概率密度分布函数；在岩体力学参数空间变异性的研究基础上，结合Karhunen-Loève (K-L)级数展开算法，构建相关非高斯随机场；采用单元遍历的手段，耦合随机场与数值网格模型；以层状岩体破坏接近度为评价指标，采用概率配点法及蒙特卡洛法，实现非侵入式随机有限差分法的破坏概率计算；最后，根据不同开挖支护阶段的概率密度分布函数的动态拟合，提出一整套科学合理、切实可行的动态定量风险评估方法，揭示不同开挖支护阶段的定量风险变化规律，为洞室安全快速施工提供理论指导和技术支撑。

层状岩体是一种典型开挖扰动的敏感岩体，在开挖卸荷下洞室围岩极易产生弯折或滑移破坏，导致塌方的发生。塌方不仅会迟滞工程建设，也会构成潜在的安全风险。项目以层状岩体开挖卸荷下，洞室塌方灾变风险问题为靶向，通过现场监测、理论研究和数值模拟分析等途径，开展了塌方定量风险评估方法研究。项目取得的成果包括：通过监测洞室开挖过程中的围岩响应信息，采用进化神经网络方法反演了层状岩体力学参数，确定了最优概率密度分布函数，为构建层状岩体力学参数随机场提供了数据支持。从层状岩体力学参数的空间变异性出发，结合力学参数的统计特征，提出了层状岩体力学参数互相关非高斯随机场的模拟方法，该随机场可有效表征岩体力学参数的非高斯性、沿岩层走向与倾向的参数自相关性与多参数间的互相关性。根据岩层产状与洞室的空间位置关系，采用单元遍历的方法，实现了随机场与网格模型的耦合。基于表征层状岩体破坏程度的破坏接近度指标，建立了层状岩体围岩破坏的极限状态方程。结合蒙特卡洛可靠度分析方法，在考虑参数空间变异性的基础上，实现了开挖卸荷下层状岩体洞室塌方风险的定量风险评估。结合概率配点法基本理论，对非侵入式随机有限差分法洞室失稳风险分析方法进行了研究，实现了随机有限差分法与非侵入式方法的一体化。非侵入式随机有限差分法相比蒙特卡洛法，在保证计算精度的前提下能够显著提高计算效率。应用非侵入式随机场有限差分法，开展了不同支护强度下围岩失稳风险分析，揭示了不同支护措施下围岩失稳定量风险的分布规律，以上成果对地下洞室安全快速施工将起到积极作用。