【摘要】 选取最大段药量、总药量、水平距离、高差、前排抵抗线、预裂缝穿透率、完整性系数、测点与最小抵抗线方向夹角、炸药爆速9个因素为判别指标,基于非线性分析方法支持向量机理论建立了预测爆破振动幅值、主频率、主频率持续时间等的爆破振动特征参量预测模型。将该方法应用到大冶有色铜绿山矿露天采场爆破振动对砌体结构破坏效应预测问题中,对现场实测的93组数据进行训练和检验,用另外15组现场数据作为预测样本进行测试,预测结果与实测情况吻合较好。研究表明,该方法回判估计的误判率低,判别性能良好,是爆破振动对砌体结构破坏效应预测的一种有效新方法,可以在实际工程中进行推广应用。更多还原

【Abstract】 With some factors,including charge weight of the largest section,total weight of explosive,horizontal distance,elevation difference,front-row burden,the penetration of pre-shear,integrity coefficient of rock mass,the angle between reporting point and the minimum burden,and velocity of denotation,as discriminant indexes,a model has been established,based on nonlinear analysis of support vector machine,for predicting characteristic parameters of blasting vibration,such as blasting vibration amplitude,dominant frequency and duration. This prediction model was applied in the Tonglvshan Mine in Daye,for predicting the damage of vibration to the masonry structure due to blasting in open-pit mine. With 93 groups of in-situ measurement taken as training sets,another 15 groups of in-situ data as samples for prediction,it was found that the prediction results were in consistent with practical measurement. It is concluded that the proposed prediction model,as an effective tool with lower error rate in prediction,can find a wide application in practical engineering project.更多还原