BP神经网络在高层建筑沉降预测中的应用  

Application of BP artificial neural network to predit the settlement of tall buildings

作　　者：魏健[1] 胡吉平[1] 谭衢霖[1] 刘万松[1] 

机构地区：[1]北京交通大学土建学院,北京100044

出　　处：《北京测绘》2013年第2期25-27,33,共4页Beijing Surveying and Mapping

基　　金：国家自然科学基金项目(51078020);中央高校基本科研业务费专项项目(2012JBM078)

摘　　要：利用人工神经网络强大的非线性映射和学习能力,提出了基于BP人工神经网络的建筑物沉降预测方法。以某实例工程1期～12期的沉降观测数据为基础,建立网络模型,并对13期～l6期实际观测值与预测值进行了比较,结果比较理想,从而验证了改进的BP人工神经网络对建筑物的沉降预测是可行的,且具有广阔的工程应用价值。Here we proposed a prediction method for the settlement of buildings based on BP artificial neural network withits strong nonlinear mapping and learning ability. We built a network model with the settlement observation data from phase 1 to 12 of an example project as a foundation, and compared the values actually observed with those we predicted from phase 13 to 16. Satisfactory results verified the feasibility of using the improved BP artificial neural network for the settlement prediction and the wide engineering application value.

关 键 词：BP神经网络 建筑物沉降 网络模型 沉降预测 

分 类 号：P258[天文地球—测绘科学与技术]