摘要
文章基于MATLAB神经网络工具箱平台,利用BP神经网络建立炸药岩石匹配模型,并根据工程实践中收集的数据和文献中的经典数据作为网络模型的学习训练和检验样本,从岩石爆破特性和爆破效果等因素出发来预测炸药爆炸特性和爆破设计参数,并结合某工程实例数据为输入值对模型进行检验,得到输出值预测误差均在10%以下,仿真效果较好。证明本模型对工程爆破实践有一定的指导意义和参考价值,并为炸药岩石匹配问题的研究提供了较好的思路。
Based on the platform of MATLAB neural network toolbox, this paper presents a matching between Explosive performance and rock strength by using the BP neural network, then taking the data gathered from engineering practice and literatures as the training and check sample, and considering the blast performance and effect of rock, forecasts the blast performance and design parameters of explosives, finally taking the data gathered from an engineering project and the input, obtains that the output value and the forecasted errors are below 10% So the simulation result is fine. It proves that the model makes a certain guidance and reference for blasting projects, as well as provides a better idea for the study on explosives matching.
出处
《辽宁科技学院学报》
2009年第1期1-3,共3页
Journal of Liaoning Institute of Science and Technology
关键词
岩石性质
炸药特性
匹配
神经网络
Rock properties
The characteristics of explosives
Matching
Neural network
作者简介
赵明生（1982-），男，黑龙江双城人，辽宁科技学院资源与建筑工程系助教，硕士。
