露天采矿爆破振动特征参量的Logistic-ELM预测

Predicting blasting vibration characteristic parameters in open-pit mining based on Logistic-ELM

针对露天采矿爆破过程中特征参量的预测问题,采用Logistic回归分析和极限学习机(ELM)方法,选取总药量、水平距离、高差、前排抵抗线大小、预裂缝穿透率、岩体完整性、传播介质、测点与爆区相对位置、炸药爆速等9个主要影响因素,利用Logistic回归方法分析各个因素的重要程度,提取最主要的因素作为ELM模型的输入,建立基于LogisticELM的露天采矿爆破振动特征参量预测模型。采用露天矿实际爆破过程中测量的100组数据作为学习样本,用于预测模型的训练,使用所得模型对其余15组检验样本进行预测并与真实结果对比。实验结果表明,经过Logistic回归分析提取影响爆破振动特征参量的主要因素后,所得模型可有效预测露天采矿爆破振动的特征参量,误差率较低。

In view of predicting characteristic parameters in the blasting process of open-pit mining,Logistic regression analysis and extreme learning machine（ELM）method was used,total dose,horizontal distance and elevation difference,size of the front resistance wire,pre-splitting fissure permeability,rock mass integrity,transmission medium,relative position of measuring point and blasting area,detonation velocity of explosive were selected as influence factors.Logistic regression method was used to analyze the factors＇ important degree,the main factors were extracted as input of ELM model,then the blasting vibration characteristic parameters prediction model in open-pit mining based on Logistic-ELM was established.Taking 100 groups of open-pit mining blasting data as learning samples and they were used to train the model.The model was used to forecast another15 sets of data and compare with true values.Results show that after extracting main factors influencing the blasting vibration characteristic parameters with Logistic regression analysis,the model can effectively predict the open-pit mining blasting vibration characteristic parameters and the error rate is low.