PREDICTION METHOD OF VERTICAL ULTIMATE BEARING CAPACITY OF PILE BASED ON CHAOS OPTIMAL METHOD AND NEURAL NETWORKS
-
摘要: 分析了BP神经网络和混沌优化的特点,并将混沌优化方法和梯度下降法结合起来构成一种新的组合搜索优化方法。结合珠江三角洲大量试桩资料,建立了基于混沌优化-神经网络的单桩极限承载力预测模型。实例研究表明,预测值与实测值吻合较好,该模型性能良好、预测精度高,能满足工程的要求,具有广泛的应用前景。Abstract: The characteristics of BP neural network and chaos optimal method are analyzed. By integrating chaos optimal method with gradient-decline method, an optimal method of combination search is created. Combining a large amount of piles testing data of the Pearl River Delta, the prediction model of vertical ultimate bearing capacity of single pile is established, based on chaos optimal method and BP algorithm. Real example research shows the predicted data agree well with the measured ones. It shows that the method is feasible, effective and satisfied with engineering demand, which can be used widely in projects.
-
Key words:
- single pile /
- vertical ultimate bearing capacity /
- chaos optimal method /
- BP neural network /
- prediction
-
杨位光.地基及基础.北京:中国建筑工业出版社,2000 [2] 王成华,张薇.人工神经网络在桩基础工程中的应用综述.岩土力学,2002(2):173~178 [3] 靳蕃,范俊波,谭永东.神经网络与神经计算机原理.成都:西南交通大学出版社,1991 [4] 吕金虎,陈君安,陈士华.混沌时间序列分析及其应用.武汉:武汉大学出版社,2002 [5] 曹双华,瑞家枞.基于神经网络的空调负荷混沌优化预测.东华大学学报,2003,29(1):52~56 [6] 刘勇健.基于遗传-BP的单桩竖向极限承载力预测方法研究.岩土力学,2004,25(1):59~63
点击查看大图
计量
- 文章访问数: 73
- HTML全文浏览量: 11
- PDF下载量: 53
- 被引次数: 0