北华大学学报：自然科学版2012，Vol.13Issue(3)：369-372,4.
基于神经网络与有限元法相结合的初始地应力场反演分析
Initial Geostress Field Back Analysis Based on Finite Element Method and Neural Network
李建博 1杨保全 1王殿春2
作者信息
- 1. 河海大学地球科学与工程学院,江苏南京210098
- 2. 广东省水电勘测设计研究院,广东广州510635
- 折叠
摘要
Abstract
Currently,the geo-stress measurement is the most straightforward way to provide the initial geo-stress field. Combining with the example of the Fifth Pumped-storage Power Station of Guangdong Province, using method of neural network combined with finite element analysis, we established the rock mass parameters, boundary conditions and the mapping relation of the stress field by BP neural network structure and ADINA three- dimensional finite element model, to invert the initial geostress field. The simulative and experimental results showed that the calculative method of the initial stress field was feasible, the calculation result was reasonable, and this method was also proper for similar engineerings.关键词
地质工程/初始地应力场/神经网络/反演分析Key words
geological engineering/initial geostress field/neural network/inversion analysis分类
天文与地球科学引用本文复制引用
李建博,杨保全,王殿春..基于神经网络与有限元法相结合的初始地应力场反演分析[J].北华大学学报：自然科学版,2012,13(3):369-372,4.
