摘要
为了确定隧道围岩的稳定性,选择合适的隧道支护形式及施工方法。利用人工神经网络处理围岩稳定性影响因素非线性能力强的特点,对富溪隧道围岩的稳定性进行了分类,选取了岩石质量指标、岩石单轴饱和抗压强度、岩石完整性系数、结构面强度系数和地下水渗水量5个主要影响因素作为网络的输入节点,输出节点是反映围岩分类结果的定量指标,也选取5个节点。采用规一化法对5个因素进行处理,利用收集到的统计资料,选取样本对围岩分类神经网络进行学习训练,用训练好的网络对富溪隧道各测设段进行了分类和围岩稳定性级别的划分,确定富溪隧道整体围岩稳定性较差,尤其是隧道进口和出口(占隧道全长的27.1%),为极不稳定围岩段。结果证明人工神经网络方法用于围岩稳定性分类结果是可靠的。
In order to determine stability of tunnel surrounding rocks,the appropriate form of tunnel support and construction methods were selected.By use of characteristics of artificial neural networks for dealing with nonlinear problems of surrounding rock stability,Fuxi tunnel surrounding rock stability was classified.Five main factors,such as rock quality,rock uniaxial saturated compressive strength,structural surface intensity coefficient and amount of groundwater seepage were selected as input node of the network;the 5 output nodes were selected,which are quantified indeces describing the surrounding rock classification results.By using the normalized method to deal with the five factors,and using statistical data collected,to select samples for rock classification neural network learning and training.With the trained network,classification of different sections of Fuxi tunnel and measurement of surrounding rock stability level were carried out.The results showed that stability of entire surrounding rocks of Fu-Xi tunnel is poor,especially at the intake and exit of the tunnel(accounting for 27.1% of the tunnel length) extremely unstable section.The results proved that the stability classification results obtained by artificial neural network method are reliable.
出处
《安徽理工大学学报（自然科学版）》
CAS
2012年第3期7-11,共5页
Journal of Anhui University of Science and Technology:Natural Science
基金
国家自然科学基金资助项目(41272278)
关键词
人工神经网络
连拱隧道
围岩分类
围岩稳定性分析
Artificial Neural Network
twin-arch tunnel
surrounding rock classification
surrounding rock stability
