增强型神经网络在基坑综合变形预测中的应用  

Application of Enhanced Neural Network in the Synthesized Deformation Prediction of Foundation Pit

作　　者：郑崇启[1] 王现成[1] 冯保民 

机构地区：[1]平顶山工学院,467001 [2]平煤集团七星公司

出　　处：《金属矿山》2009年第2期60-62,78,共4页Metal Mine

摘　　要：为有效改善传统人工神经网络算法的缺陷,精确预测基坑变形,以实际工程为例,采用增强型人工神经网络(ANN)构建神经网络模型,并利用条件相同或相似的已建基坑工程的实测变形资料作为学习样本,实际变形观测值作为期望输出对神经网络进行训练。结果表明:该神经网络训练法的输出结果较为可靠,更接近于实际,是预测基坑变形的一种十分有效的方法,对其他基坑工程具有一定的借鉴价值。t In order to effectively improve the deficiency of conventional artificial neural network＇s algorithm and precisely predict the deformation of foundation pit, this paper establishes a model exemplified by actual projects for neural network by adopting an enhanced artificial neural network. Furthermore, taking actual deformation data of constructed foundation pit with the same or similar conditions as learning sample, and actual deformation measurements as desired output, the neural network is trained. The results obtained show that the output of this training method for neural network is relatively reliable and close to reality. Therefore, such a training method is very effective for predicting the deformation of foundation pit and can be used as a source of reference for other projects of foundation pit.

关 键 词：基坑 支护结构 人工神经网络 变形预测 

分 类 号：TS210.1[轻工技术与工程—粮食、油脂及植物蛋白工程] TV698.11[轻工技术与工程—食品科学与工程]