基于神经网络的静压PHC管桩单桩竖向极限承载力预测研究-建筑与土木工程专业论文.docxVIP

摘要 摘要 2 2 Abstract PHC pile is a new type of precasting concrete pile. This type of pile has many advantages, such as simple technology, better quality, low cost and convenient detection, so it has been widely used in geotechnical engineering during the recent years. So the study on PHC piles is of great theoretical and practical significance. This article summarizes the impact of single pile ultimate load- capacity factor, the ultimate load- capacity as a gray system, and to study the gray system theory, combining gray GM (1,1) to the extrapolated value established with nonlinear computing power of the BP neural network model used to predict the ultimate load-capacity of single PHC piles. Through the influence of pile load-capacity factors, combined with the characteristics of BP neural networks, established a pile length, pile diameter, pile pressure force, pile embedded depth, the weighted average of the side friction, the pile load- capacity and other parameters the input layer neurons, the middle layer is one, the neurons number of middle layer is 2N +1, the output layer is the one of the BP neural network model(load-capacity). Using the model in the region 62 PHC piles to the sample training, and predicting each simulation of the 9PHC piles, the predicted and measured values were compared. The results showed that calculated and measured results match well, and the error is less than 8%, theoretical models can support the practical engineering. Therefore, the prediction model of PHC piles ultimate load-capacity has some practical significance. Key words: PHC Piles; Ultimate bearing capacity; grey theory; Back-Propagation neural network; 原创性声明 本人声明兹呈交的学位论文是本人在导师指导下完成的研究成 果。论文写作中不包含其他人已经发表或撰写过的研究内容，如参考 他人或集体的科研成果，均在论文中以明确的方式说明。本人依法享 有和承担由此论文所产生的权利和责任。 学位论文作者签名： 日期： 学位论文版权使用授权声明 本人同意授权华侨大学有权保留并向国家机关或机构送交学 位论文和磁盘，允许学位论文被查阅和借阅。 论文作者签名： 指导教师签名： 签 名 日 期： 签 名 日 期： 第一章 第一章 绪论 1 1 第一章 绪论 1.1 引言 预应力砼 PHC 管桩从七十年代初期引入我国后，得到了很大的 发展，特别是近十年来的发展更是迅猛。关键就在于预应力砼 PHC 管桩具有质量