基于灰色分析和神经网络的爆破振速峰值预测

Prediction of peak velocity of blasting vibration based on pray analysis and neural network

爆破给国家带来了巨大经济效益、给施工提供便利条件的同时避免不了会带来一些负面效应,为此,如何才能既降低爆破振动危害又保证施工进度,己经成为当前隧道工程界亟待解决的一项重要课题。以北京地铁昌平二期05段岩石隧道爆破为例,通过灰色关联分析法确定了对爆破振速峰值有显著影响的指标和输入变量,建立BP神经网络模型,对爆破振动速度峰值进行预测。将结合了灰色关联分析法的BP神经网络模型预报的结果与神经网络模型、传统方法预测的结果相比,其结果为:萨道夫斯基公式的平均误差为18.86%,萨道夫斯基拓展式为16.57%,BP神经网络的误差为14.61%,灰色关联分析神经网络法仅为8.23%。预测结果表明结合灰色关联分析法并运用BP神经网络对爆破振速峰值预测是可行的。

While bringing huge economic benefits to the country and big efficient to the construction,blasting inevitably causes some negative effects.On this condition,how to both reduce the harm of blasting vibration and ensure construction progress,have become an important issue that must be solved by the current tunneling territories.Based on rock tunnel blasting in the second phase of Changping fifth section subway,Beijing,the input variables of the network which have significant effects on blasting vibration velocity peak value were determined by grey relational analysis and the artificial neural network is adopted to predict the peak velocity of blasting vibration,in the analysis.In comparison with the prediction results of traditional formula,traditional artificial neural network and the method combined with neural network and pray relational analysis,the result show that average prediction errors for traditional formula,expanding mode of traditional formula and traditional back-propagation are 18.86%,16.57%and 14.61%,respectively.However,combined method of the grey relational analysis and artificial neural network is only8.23%.According to the prediction results,combined method of the grey relational analysis and artificial neural network has the capabilities of predicting the blasting vibration velocity peak value.