岩爆分类的人工神经网络预测方法  

Artificial neural network for forecasting and classification of rockbursts

作　　者：丁向东[1] 吴继敏[1] 李健[2] 刘成君[1] 

机构地区：[1]河海大学土木工程学院,江苏南京210098 [2]江苏省交通规划设计院,江苏南京210005

出　　处：《河海大学学报（自然科学版）》2003年第4期424-427,共4页Journal of Hohai University(Natural Sciences)

摘　　要：采用人工神经网络原理,选取影响岩爆的一些主要因素,如地应力、岩石抗压强度、抗拉强度等作为输入参数,建立了岩爆分类与预测的神经网络模型.利用国内外一些工程实例作为学习和训练的样本,并用已经训练稳定的样本对某水电站地下厂房岩爆进行预测.研究表明,与其他岩爆预测方法比较,人工神经网络模型更具有客观性和有效性.A neural network model is developed for forecasting and classification of rockbursts by selection of some affecting factors as the inputted parameters, such as the geostress and compressive strength and tensile strength of rocks. With some engineering projects at home and aboard taken as learning and training samples, rockburst forecast is performed for an underground hydropower plant by use of the samples that have been trained stably. The results show that, compared with other forecasting methods, the artificial neural network model is objective and effective.

关 键 词：人工神经网络 岩爆 分类 水电站地下厂房 BP算法 

分 类 号：TV731.6[水利工程—水利水电工程] TU45[建筑科学—岩土工程]