开挖期深基坑变形RBF模型的输入层及中心确定

针对深基坑变形RBF神经网络监测模型,根据在开挖期间深基坑土体的变形特性,利用遗传蠕变来构造输入因子,鉴于目前在中心选择方面没有严格的理论依据,并充分考虑输入因子的特点,提出了一种选择中心的经验方法,经过实例检验,该方法选择中心可以使模型得到较好的预测效果。

On input nodes of RBF model of deep foundation pit deformation in excavation period and its determination of center

Aiming at the RBF neutral network supervision model of the deep foundation pit,the paper,according to the deformation characteristics of the deep foundation pit in the excavation period,adopts the inheritance creeping to compost the input determiner,gives full consideration of the characteristics of the determiner by referring to the no strict theoretic reference for the center selection,and points out the experience for the way of selecting the center,and proves that the method has the expected effect for the model in the center selection.