期刊论文

Prediction of Blasting Vibration Velocity using GA-BP Neural Network

中文

爆破

1001-487X

2020

37

03

148-152

10.3963/j.issn.1001-487X.2020.03.025

50974076:国家自然科学基金 :广东锡源爆破科技股份有限公司项目

本校为第一机构

为了更好的预测爆破振动速度,运用遗传算法(GA)对BP神经网络的权值与阈值进行优化,构建GA-BP神经网络预测模型。结合华润小径湾实际爆破工程监测数据,确定以最大单段药量、爆心距以及测点至爆心的高程差作为输入参数,对爆破振动速度进行预测,并与BP神经网络模型的预测结果进行对比。结果表明: GA-BP神经网络模型预测结果的平均相对误差为5.80%,明显小于BP神经网络模型预测的平均误差14.19%。相比之下,GA-BP神经网络模型的预测精度更高,稳定性更好,可为多因素影响下类似工程爆破振动速度预测提供借鉴经验。

The genetic algorithm(GA)is used to optimize the weight and threshold of BP neural network,and the GA-BP neural network prediction model is constructed to predict the blasting vibration velocity precisely.In the actual blasting engineering of China Resources Xiaojingwan,monitoring data,including the maximum single-dose amount,the blasting distance and the elevation difference from the measuring point to the blasting heart,is taken as input parameters.The blasting vibration velocity predicted by the GA-BP neural network model is compared with the one soluted by the BP neural network.The results...