Prediction of crack coalescence stress in rock like samples containing non-persistent joint under direct shear test using both of the GWO and GEP algorithms

Document Type : Research Article

Authors

1 Department of Mining Engineering, Hamadan University of Technology

2 tehran university

3 tarbiat modares university

4 Faculty of Engineering, Tarbiat Modares University

10.22034/anm.2024.21039.1619

Abstract

Rock joints and micro cracks have an important effect on the failure mechanism of rock masses. Rock mass failure affect the stability of rock engineering project. Therfore investigation of failure stress of rock mass help engineering to desing the rock engineering structure with high safty. Failure stress that is one of the mechanical propertis of rock masses controls the loading capasity of rock structures. Several parameters have a dominant role on the failure stress of rock masses such as mechanical properties of rock mass, normal stress, ratio of joint surface area to total shear surface area (JC). In this paper, the crack coalescence stress was predicted using both of the GWO and GEP algorithms. Eight import parameters have been used such as normal stress, JC, compressive strength, tensile strength, Poisson ratio, Young modulus, cohesion and friction angel. Totally 450 data obtained from jointed rock like samples tested under direct shear. Three different coefficients of “R2”, “RMSE” and “MAE” were used to evaluate the applicability of algorithms. The coefficients of GWO and GEP on trained data were 0.962 and 0.938 while their values on tested data were 0.996 and 0.981. the GWO has the best applicability on the coalescence stress prediction. The results show that the coefficients of RMSE and MAE in both of the train and test stages in GWO algorithm was less than GEP algorithm. This shows that GWO algorithm has the best application in comparison to GEP. The normal stress and JC has the maximum and minimum effect on the crack coalescence stress, respectively.

Keywords

Main Subjects



Articles in Press, Accepted Manuscript
Available Online from 02 March 2024
  • Receive Date: 06 January 2024
  • Revise Date: 21 February 2024
  • Accept Date: 02 March 2024