| 岩石本构模型智能辨识方法研究 |
| |
| 引用本文： | 翟淑花,张亮,丁桂伶,张梅花.岩石本构模型智能辨识方法研究[J].城市地质,2015(Z1). |
| |
| 作者姓名： | 翟淑花 张亮 丁桂伶 张梅花 |
| |
| 作者单位： | 1. 北京市地质研究所,北京,100120;2. 亚太建设科技信息研究院有限公司,北京,100120 |
| |
| 摘 要： | 岩石本构模型是研究岩石力学特征和变形机制的基础,而本构模型或模型中相关参数的识别是本构模型研究中的热点和难点问题。本文基于红板岩室内力学实验数据,分别利用遗传算法、BP神经网络以及遗传规划对红板岩本构模型进行了模式识别,结果表明,遗传算法进行参数识别需要事先假定流变模型的形式,误差较大,而BP神经网络和遗传规划可以一次性同时确定流变模型的结构形式及参数,有效避免模型假定所带来的误差。而遗传规划与BP神经网络相比,具有精度高、收敛快,可视化程度高等特点,为岩石本构参数及模型的智能识别方法的选取提供参考。
|
| 关 键 词： | 本构模型 智能识别 遗传算法 BP神经网络 遗传规划 |
Research on the intelligent identiifcation method of constitutive model |
| |
| ZHAI Shuhua,ZHANG Liang,DING Guiling,ZHANG Meihua.Research on the intelligent identiifcation method of constitutive model[J].City Geology,2015(Z1). |
| |
| Authors: | ZHAI Shuhua ZHANG Liang DING Guiling ZHANG Meihua |
| |
| Abstract: | Constitutive model of rock is the basis for the study of mechanical characteristics and deformation mechanism, while, the identification of model or its parameter is an important problem needing to be solved. In this paper, based on laboratory experiment date of the red slate, genetic algorithm, BP neural network and genetic programming were used to identify the constitutive model, these results show that the results of genetic algorithm have large error, coming from constitutive structure presumption in advance. While, BP neural network and genetic programming can deifne the structure and parameters of constitutive model at single time, which can avoid the error form constitutive structure presumption. Furthermore, genetic programming offers some reference for constitutive model identiifcation, which has more advantages than BP neural network, such as high precision, fast convergence and good visual display. |
| |
| Keywords: | Constitutive model intelligent identiifcation genetic algorithm BP neural network genetic programming |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
