Mechanical parameter inversion in tunnel engineering using support vector regression optimized by multi-strategy artificial fish swarm algorithm

Di Zhuang

doi:10.1016/j.tust.2018.09.027

Return

Mechanical parameter inversion in tunnel engineering using support vector regression optimized by multi-strategy artificial fish swarm algorithm

2019-01-01

PRE

DOI：10.1016/j.tust.2018.09.027

Original

Original request for help

Save

Abstract

En 中文

Fast and efficient determination of the mechanical parameters of surrounding rock masses is vitally important to the calculation and evaluation of the stability of surrounding rock masses in tunnel engineering. In this paper, a displacement back-analysis (DBA) model is proposed to identify the mechanical parameters based on support vector regression (SVR) optimized by multi-strategy artificial fish swarm algorithm (MAFSA). The MAFSA adopts the differential evolution strategy, the particle swarm optimization strategy, the adaptive step size and phased vision strategy on the basis of artificial fish swarm algorithm (AFSA) to enhance the global search capability and improve convergence speed and optimization accuracy. Then, the kernel width and the penalty parameter of SVR are optimized by MAFSA, forming into MAFSA-SVR. Meanwhile, the training and testing samples for MAFSA-SVR are constructed by orthogonal design and forward calculation by FLAC(3D) code. Finally, the DBA model is established based on MAFSA-SVR and applied to the mechanical parameter inversion of surrounding rock masses in the Heshi tunnel with the following conclusion: the relative errors of all the mechanical parameters are less than 8% between the inversed values of the DBA model based on MAFSA-SVR and the actual values. The method proposed in this paper could provide an efficient tool for the mechanical parameter inversion of the tunnel surrounding rock masses.

Keywords:

Displacement back-analysis

Support vector regression

Artificial fish swarm algorithm

Orthogonal test

Mechanical parameter inversion

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.