位移反分析在地下厂房开挖中的应用

Application of Displacement Back Analysis in the Excavation of Underground Powerhouse

针对官地水电站地下厂房,采用三维快速拉格朗日法(FLAC3D)对开挖及支护过程进行弹塑性有限元模拟,基于开挖过程中地下厂房围岩实测位移,采用BP网络和遗传算法对围岩变形模量进行反演,并采用反演模量值对地下厂房后续开挖过程中的变形进行了预测。研究表明,采用BP网络和遗传算法进行反演,可以提高反分析计算速度。反演得到的岩体变形模量为19.95GPa,在参数建议值范围内。采用反演参数对地下洞室后续开挖过程中的位移进行有限元计算和预测,预测位移与现场监测位移较为符合,说明反演结果是合理的。研究成果可为类似地下工程开挖变形预测提供参考。

The excavation and support processes of the underground caverns in the Guandi Hydropower Project are simulated bythe elastic-plastic finite element method using FLAC3D. Based on the measured deformations of the surrounding rocks in the underground powerhouse during the excavation, the deformation modulus of the surrounding rocks is back analyzed using the BP networks and genetic algorithms, and the estimated deformation modulus is used to predict the future deformations of the underground caverns during the following excavatiorL The results show that the application of the BP networks and genetic algorithms can accelerate the process of back analysis. The deformation modulus of the rocks is 19. 95 GPa according to the back analysis, and it is in the range of the suggested values. The predicted displacements of the underground caverns during the future excavation using the inversely estimated parameters are similar to the measured displacements. Therefore, the back analysis method is reasonable and can provide references for the deformation prediction of similar excavation projects.