边坡弹性模量反分析的模拟退火BP网络方法  

Simulated annealing-based BP neural network optimization method for back analysis of modulus of elasticity for slope rock mass

作　　者：石安池[1] 徐卫亚[1] 周家文[1] 张明鸣[1] 张贵科[1] 

机构地区：[1]河海大学岩土工程研究所,江苏南京210098

出　　处：《河海大学学报（自然科学版）》2006年第1期69-73,共5页Journal of Hohai University(Natural Sciences)

基　　金：国家重点基础研究发展规划项目(2002CB412707);国家自然科学基金重点资助项目(50539110)

摘　　要：通过模拟退火算法,对BP网络进行优化并编写了模拟退火BP网络程序,应用该程序对三峡永久船闸高边坡岩体弹性模量进行了位移反分析.结果表明,根据模拟退火BP网络反分析得到的弹性模量计算出的位移值和监测位移值差别较小.因此,模拟退火BP网络方法可以用于边坡岩体力学参数的反分析.The simulated annealing algorithm was adopted for optimization of BP neural network, and BP neural network program was compiled. The application of the program to back analysis of modulus of elasticity for high slope rock mass of the TGP shiplock shows that the calculated result,of the displacement of high slopes according to the back analysis is close to the monitoring data. Therefore, the simulated annealing-based BP neural network optimization method can be used for back analysis of mechanical parameters of slope rock mass.

关 键 词：模拟退火 BP网络 三峡船闸 边坡 弹性模量 反分析 

分 类 号：TU452[建筑科学—岩土工程]