作者：凌天龙；武宇；李胜林；梁书锋；柴鹏伟

‍第一作者单位：中国矿业大学(北京)力学与建筑工程学院

摘要：为了得到煤矿巷道光面爆破参数,使用BP神经网络技术对参数进行预测和优选,并用C#语言开发了巷道光面爆破智能设计系统。通过分析影响巷道光面爆破效果的主要因素,将普氏系数、节理裂隙发育情况、炮孔直径、掘进面积等8个因素设置为输入层参数。以光面爆破理论研究和煤矿一线生产调研为基础,建立BP网络神经,进行网络学习,训练样本,并应用该系统对不同矿区的实际巷道进行爆破方案设计。结果表明,预测的参数与实际爆破参数较为一致,在现场试验中也取得较好效果,说明该系统可以提高煤矿巷道光面爆破参数设计的可靠性,对煤矿巷道光面爆破参数设计具有一定的参考价值。

Abstract：BP neural network is used to forecast and optimize the blasting parameters of coal mine roadway, and the intelligent design system for smooth blasting was developed by C# language. By analyzing the effect of main factors on smooth blasting, 8 parameters, such as Protodrakonov scale of hardness, developmental conditions of joint and fracture, diameter of the blast hole and area of the driving section, etc. were chosen as the input layer parameters. Some learning and training samples were established based on expert research results and practical experiences. The system was applied to design blasting scheme of actual roadway in different mining areas. The results show that the predicted parameters are in good agreement with the actual blasting parameters and also achieve good results in the field test, which means that the system has a certain application value to improve the reliability on parameters designing with smooth blasting for coal mine roadway.

关键词：煤矿巷道;光面爆破;人工神经网络;智能设计;

DOI：10.16186/j.cnki.1673-9787.2018.02.005

分类号：TD235.374