人工神经网络在徐州岩溶地面塌陷评价中的应用

Application of artificial neural network on evaluation of karst ground subsidence in Xuzhou

岩溶地面塌陷的影响因素很多,发展过程也复杂。在众多的对岩溶地面塌陷的评价方法中,神经网络具有自学习、自适应与高度非线性映射的特点,是一种非常有效的评价手段。在徐州岩溶石地面塌陷的评价中,成功地运用了人工神经网络技术,它具有的强大非线性映射能力,能够建立评价因子和评价对象之间的关系,正确选取评价因子,避免主观判断取值,从而得出可靠的预测模型和岩溶塌陷危险性分区图。

There are many influence factors and complex evolvement course for karst ground subsidence. As a method with self-adaptive, self-study and nonlinear projected characters, the artificial neural network is effective among the evaluation methods. The evaluation of karst ground subsidence in Xuzhou shows that the relationship between factors and object of appraisement can be established and subjective judgement can be avoided by artificial neural network. Through establishment of a stable analysis model of artificial neural network and selection of evaluation factors, the perilous regional map of karst ground subsidence in Xuzhou City is obtained on the basis of engineering geological investigation and analysis of formation mechanism.