摘要
矿岩爆破块度是对爆破效果进行定量评价的重要指标,它直接影响到矿山各后续生产工序的效率和采矿生产的总成本。在解决这类动态非线性问题时,需要考虑爆破效果与影响参数之间的复杂关系。论文提出了一种改进粒子群算法的LM(Levenberg-Marquardt,莱文贝格-马夸特)神经网络模型,通过改进粒子群算法可提高LM神经网络对缺陷属性识别的计算精度,从而实现对爆破效果的预测。通过临江冷堡子硅石矿现场爆破实验实例,验证了该神经网络模型及计算方法的可行性及实用性。
Ore rock blasting lumpiness is an important indicator for quantitative evaluation of blasting effectiveness,which directly affects the efficiency of all subsequent production processes in the mine and the total cost of mining production.In solving this kind of dynamic nonlinear problems,the complex relationship between the blasting effect and the influencing parameters needs to be considered.In this paper,a LM neural network model with improved particle swarm algorithm is proposed,which can improve the computational accuracy of LM neural network for defect attribute identification by improving the particle swarm algorithm,so as to realize the prediction of blasting effect.The feasibility and practicability of the neural network model and the calculation method are verified through the on-site blasting experiments of Linjiang Lengbaozi silica mine.
作者
马炳德
张雄天
赵尔丞
MA Bing-de;ZHANG Xiong-tian;ZHAO Er-cheng(Lanzhou Nonferrous Metallurgy Design&Research Institute Co Ltd,Lanzhou 730000,China)
出处
《建材世界》
2024年第1期115-120,共6页
The World of Building Materials
关键词
LM算法
粒子群算法
爆破技术
爆破块度预测
LM algorithm
particle swarm optimization
blasting technology
blasting lumpiness prediction
