【摘要】 人工神经网络是近期发展最快的人工智能领域研究成果之一.通过分析国内外爆破震害预测研究现状和不足,提出一种基于BP神经网络模型的爆破地震效应预测方法,该方法能克服基于最小二乘法的回归公式的局限性,可选取影响爆破振动的多个影响因素作为输入层参数,达到爆破峰值和主频同步预测之目的.利用该方法对实际爆破监测数据进行预测,结果表明人工神经网络方法在爆破地震效应预测中应用是可行的并且是有效的.这为爆破震害预测研究提供了新途径.更多还原

【Abstract】 The artificial neural network is one of the research result of the artificial intelligence field with quickest development in the near future. Through analysis of the present condition and shortage of forecasting the hazards by blasting vibrations, this article proposes a new forecast method. Based on the BP neural network model of blasting vibration, efficiency, the method can overcomes the return formula’s limitation based on the least squares method. It may choose the factors which play important roles in the blasting vibration as the input layer’ s parameters, and achieve simultaneously forecasting the peak value and main frequency. The result of real data indicates that the method is feasible and effective. This way will be a new approach to forecasting the hazards by blasting vibrations.更多还原