cerchar abrasivity index

The abrasivity of materials is of vital importance in selecting an affordable excavation method in the early stages of a tunneling project. The amount of specific energy and the Mode I and Mode II stress intensity factors in the eroded disc is more than the non-eroded disc. Vickers hardness number rock (VHNR), rock abrasivity index (RAI), Cerchar abrasivity index (CAI), LCPC abrasivity coefficient, and Norwegian University of Science and Technology (NTNU) abrasion test are among the methods for estimating rock abrasivity. Correlations were proposed in this study for estimating the CAI of sandstone and tuff. To achieve reliable comprehensive results, various kinds of sandstone and tuff with different geological properties were selected. Sandstone and tuff samples were collected from different regions. Various tests were then carried out on prepared rock specimens to determine the uniaxial compressive strength (UCS), tensile strength, longitudinal waves velocity, Schmidt hammer rebound hardness, and the equivalent quartz content (EQC). Two correlations were presented by analyzing the experimental results with the help of SPSS and Excel. The first correlation estimated the CAI of sandstone based on the UCS and EQC with a coefficient of determination (R2) of 0.81. Moreover, using data analysis in Excel, another correlation was proposed to estimate the CAI of tuff. The second correlation estimated the Schmidt hammer rebound hardness of tuff with an R2 of 0.88.

The CERCHAR abrasivity test is very popular for determination of rock abrasivity. An accurate estimation of the CERCHAR abrasivity index (CAI) is useful for excavation operation costs. This paper presents a model to calculate CAI based on the gene expression programming (GEP) approach. This model is trained and tested based on a database collected from the experimental results available in the literature. The proposed GEP model predicts CAI based on two basic geomechanical properties of rocks, i.e. rock abrasivity index (RAI) and Brazilian tensile strength (BTS). Root mean square error (RMSE), mean absolute error (MAE), Nash-Sutcliffe efficiency (NSE), and coefficient of determination (R2</sup>) are used to measure the model performance. Furthermore, the developed GEP model is compared with linear and non-linear multiple regression and other existing models in the literature. The results obtained show that GEP is a strong technique for the prediction of CAI.