露天煤矿爆破振动的BP神经网络预测

BP neural network prediction for blasting vibration in open-pit coal mine

由于爆破振动效果的影响因素繁多,能够综合考虑各因素并对爆破振动进行合理预测亟待研究。通过分析爆破各影响因子与爆破监测数据间的相互关系,利用神经网络之间的权值训练,找出变量之间的非线性关系,然后用训练好的神经网络对爆破振动进行预测。模型预测结果与实际监测得到的爆破振动数据基本吻合,与常规经验公式的预测效果对比表明,基于BP神经网络的爆破振动预测模型的预测方法更为简单适用,精度高而且误差更小,说明将BP神经网络模型应用到爆破效果预测中是适用的和正确的。利用该模型获得了满足布沼坝露天煤矿西帮生产爆破要求的一组最优参数。

Because the influence factors of blasting vibration effect is various, how to consider all the factorsand predict the blasting vibration effect reasonably is a burning question. Blasting vibration prediction modelwas established by applying the BP neural network,and used in the west slope for the production of blasting.Through the analysis of various influencing factors of blasting and the blasting monitoring data, using weighttraining between the Internet,it is found out the non-linear relation between input and output variables, predictingblasting vibration with the trained neural network. Model prediction results are consistent with the actualmonitoring of the blasting vibration data, thererore, it is feasible that the BP neural network model is applied tothe evaluation process of blasting vibration. Compared with prediction effect of conventional empirical formula,blasting vibration prediction model based on BP neural network is simple and that the prediction deviationsmall and high precision. Finally,a set of optimal parameters is obtained by using the model.