基于神经网络的预裂爆破参数智能设计

预裂爆破具有减振和保护岩体等特点,在爆破工程中得到了广泛运用。通过分析认为预裂爆破成缝理论和成缝计算模型均不完善,同时岩石成缝与岩石性质、地质条件、装药量、孔间距等多种因素有关。目前为止,预裂爆破的设计主要依靠经验公式。运用人工神经网络综合方法,建立了预裂爆破参数设计的BP神经网络模型,并借助Matlab语言,开发了预裂爆破参数优化设计智能系统,且用工程实例进行了验证。结果表明,该智能系统设计参数与实际设计参数相当符合。

INTELLIGENT DESIGN OF PARAMETERS BASED ON NEURAL NETWORK IN PRESPLITTING BLASTING

Presplitting blasting has been applied in engineering widely,with its characteristics of the shock absorption and protection of rock mass.The theory and calculation model of cracks shaped by presplitting blasting have been not perfect to analyze rock mass,and fracture of rock shaped by presplitting blasting is related to property of rock mass,geology,charge weight and explosive distance.Up to now,design of parameters has depend on empirical formula in presplitting blasting.The model of BP neural network was established in order to design presplitting blasting parameters with artificial neural network method.An intelligent system of parameters optimization was built by Matlab and was proved through application of it into engineering practice.The results showed that blasting parameters designed by the intelligent system were consistent with practical design parameters basically.