基于BP神经网络的AM工法桩质量预测研究

发布时间：2018-05-29 10:05

本文选题：AM工法桩 + BP神经网络　；参考：《河北工程大学》2013年硕士论文

【摘要】：AM工法旋挖扩底灌注桩（简称AM工法桩）常用作地下水丰富地区及软土地区桩基工程中，目前在工程管理工作中如何确定AM工法桩质量是急需解决的问题之一。因此，本文基于对AM工法桩质量以及神经网络机理的认识，，提出了质量预测的神经网络模型，采用此方法来解决AM工法桩在质量预测方面的难题。本论文首先通过对影响AM工法桩质量存在的各个影响因素进行全面细致的分析和评价，针对包括持力土层、桩体结构、施工工艺、时间及空间效应等方面考虑，建立影响AM工法桩质量因素的指标体系。然后论述了BP神经网络方法在AM工法桩质量预测问题上存在的可行性和优越性，在此基础上确立以桩体、土层、施工等方面共8个因素作为网络的输入层，单桩竖向承载力、桩身完整性作为网络输出层的三层BP神经网络质量预测模型。最后，在Matlab语言环境下编写了相应的程序，利用收集到的天津地区的试桩数据对所建立的BP神经网络进行训练。通过对检验样本训练结果分析，与现场的实测值做对比，验证了预测值的准确性和可靠性，为AM工法桩质量提供一种精度较高的预测方法。
[Abstract]:AM pile is often used in pile foundation engineering in the area of rich groundwater and soft soil. How to determine the quality of AM pile in engineering management is one of the problems that need to be solved urgently. Therefore, based on the understanding of the quality of AM pile and the mechanism of neural network, this paper puts forward the neural network model of quality prediction, and uses this method to solve the problem of quality prediction of AM pile. In this paper, the factors that affect the quality of AM pile are analyzed and evaluated in detail, including the soil holding layer, pile structure, construction technology, time and space effect and so on. The index system of influencing the quality factors of AM pile is established. Then it discusses the feasibility and superiority of BP neural network method in the quality prediction of AM pile. On the basis of this, 8 factors such as pile body, soil layer and construction are established as the input layer of the network, and the vertical bearing capacity of single pile is established. Pile integrity is a three-layer BP neural network quality prediction model. Finally, the corresponding program is written under the environment of Matlab, and the BP neural network is trained by using the data collected from the test pile in Tianjin area. The accuracy and reliability of the predicted values are verified by analyzing the training results of the test samples and comparing with the field measured values, which provides a high precision prediction method for the quality of AM piles.
【学位授予单位】：河北工程大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;TU753.3

【相似文献】