基于粒子群算法和广义回归神经网络的岩爆预测

 岩爆是岩石深部开挖中一种常见的工程地质灾害。为评价岩爆发生的可能性，提出一种基于粒子群算法和广义回归神经网络模型(PSO-GRNN模型)的岩爆预测方法。该方法利用已有岩爆数据，通过神经网络技术建立回归模型，采用粒子群算法对模型参数进行优化，减少人为因素对神经网络设计的影响。据此方法，在能量理论的基础上，选取洞壁围岩最大切向应力、岩石单轴抗压强度、抗拉强度和弹性能量指数作为主要影响因素，利用国内外26组已有工程数据建立岩爆预测的PSO-GRNN模型。通过对苍岭隧道和冬瓜山铜矿岩爆预测的工程实例分析验证该方法的可行性和适用性。所提方法可为类似工程的岩爆预测提供参考。

ROCKBURST PREDICTION USING PARTICLE SWARM OPTIMIZATION ALGORITHM AND GENERAL REGRESSION NEURAL NETWORK

Rockburst is a common engineering geological disaster in deep rock excavations. To evaluate the possibility of rockburst，a rockburst prediction method using the particle swarm optimization(PSO) algorithm and the general regression neural network(GRNN) model is proposed. This approach employs the technology of neural network to build up a regression model based on existing rockburst database，and takes advantage of PSO algorithm to optimize the parameters of the network which is believed to reduce the adverse influence of man-induced factors in model construction. Then，four major influence factors，including the maximum induced tangential stress on the boundaries of tunnels or caverns，the uniaxial compressive strength and the uniaxial tensile strength of the rock，and also the elastic energy index of the rock，are selected as the inputs for establishing the PSO-GRNN model based on the energy theory and the data obtained from 26 practical cases. The generated PSO-GRNN model is finally applied to predict the rockburst for the Cangling tunnel and Dongguashan copper mine，in which the feasibility and applicability of the proposed approach are illustrated. The methodology presented in the paper provides a reference for some similar engineering involving rockburst.