摘要
为克服当前隧道爆破参数选取受人为因素影响的不足,以围岩普氏系数、隧道断面积、实际进尺和炮孔直径等为网络输入参数,以设计进尺、炸药单耗、周边孔距和掘进孔孔距等为网络输出参数,建立了含输入层、输出层和隐含层的神经网络模型,并给出了模型学习算法,提出了基于爆破先验知识的可加快模型求解收敛速度的网络学习约束条件.隧道爆破参数的实例计算结果表明,给出的网络模型及其算法能在借鉴已有爆破资料的基础上准确、快速计算爆破参数,并且获得理想的爆破效果.
To overcome the shortcoming of the parameter design of tunnel blasting, i. e. , they are selected empirically by designers, a three-layer neural network model with an input layer, an output layer and a hidden layer was constructed. The Protodikonov＇s hardness coefficient, tunnel cross-section area, practical advance per round, blast-hole diameter and others are considered as the input parameters of a BP network, and the designed advance per round, powder factor, contour hole spacing, reliever-hole spacing and excavated-hole spacing as the output parameters. A algorithm for the neural network model was given, and the restrain conditions of network study were proposed on the basis of blasting prior knowledge, so solving of the model can be accelerated. The calculating results for a practical example of tunnel blasting design show that with the help of the neural network model and the algorithm, the parameters of tunnel blasting may be calculated accurately and quickly by using the existed blasting data, so an optimal blasting effect can be obtained.
出处
《西南交通大学学报》
EI
CSCD
北大核心
2007年第5期537-541,共5页
Journal of Southwest Jiaotong University
基金
铁道部科技研究开发计划课题资助项目(2004G038)
关键词
隧道工程
爆破参数计算
先验知识
BP神经网络
tunnel engineering
calculation of blasting parameter
prior knowledge
BP neural network
