| 基于分形维数的锚杆无损检测BP神经网络模型 |
| |
| 引用本文： | 李张明,练继建,王生民,曹海斌.基于分形维数的锚杆无损检测BP神经网络模型[J].天津大学学报(自然科学与工程技术版),2008,41(9):1120-1123. |
| |
| 作者姓名： | 李张明 练继建 王生民 曹海斌 |
| |
| 作者单位： | [1]天津大学建筑工程学院,天津300072 [2]长江水利委员会长江勘测规划设计研究院,武汉430010 [3]国网北京电力建设研究院,北京100055 |
| |
| 基金项目： | 国家杰出青年科学基金，国家自然科学基金 |
| |
| 摘 要： | 笔者将分形维数的分析方法引入锚固系统的无损检测中，指出分形维数的大小是反射波在该小波分解频段能量大小的量度．在研究锚固系统无损检测反射波小波分解分形维数特征向量的基础上，确定了以锚杆反射波的原波维数、原波平均波幅以及小波包分解维数的8个波形分形维数共10个数值作为输入参数，以锚杆注浆密实度为输出参数的非线性BP神经网络预测模型．对三峡工程右岸地下电站试验锚杆进行训练，验证了锚杆密实度BP神经网络预测模型的合理性和可靠性．
|
| 关 键 词： | 锚杆 分形维数 小波分析 神经网络 密实度 无损检测 |
BP Neural Network Model for Rock Bolt Nondestructive Detection Based on Fractal Dimension |
| |
| LI Zhang-ming,LIAN Ji-jian,WANG Sheng-min,CAO Hai-bin.BP Neural Network Model for Rock Bolt Nondestructive Detection Based on Fractal Dimension[J].Journal of Tianjin University(Science and Technology),2008,41(9):1120-1123. |
| |
| Authors: | LI Zhang-ming LIAN Ji-jian WANG Sheng-min CAO Hai-bin |
| |
| Institution: | LI Zhang-ming, LIAN Ji-jian, WANG Sheng-min, CAO Hai-bin ( 1. School of Civil Engineering, Tianjin University, Tianjin 300072, China; 2. Changjiang Institute of Survey, Planning, Design and Research, CWRC, Wuhan 430010, China; 3. Beijing Electric Power Construction Research Institute, SG, Beijing 100055, China) |
| |
| Abstract: | The method of fractal dimension analysis was applied to nondestructive detection of anchorage system in this paper, and it is pointed out that the value of fractal dimension was a measurement of the wavelet energy in a frequency range of the reflected wave. Based on the eigenvector of fractal dimension of reflected wave decomposed by wavelet method, a nonlinear BP neural network forecast model was determined, in which the dimension of initial wave, the mean amplitude of initial wave, and the eight fractal dimension of initial wave decomposed by wavelet packet were considered as input parameters, and the denseness of rock bolt as output parameter. The rock bolt underground power station of the Three Gorges right bank was trained, and results showed that the BP neural network forecast model for rock bolt density detection was reasonable and credible. |
| |
| Keywords: | rock bolt fractaldimension wavelet analysis neural network density nondestructive detection |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
