Prediction of uniaxial compressive strength of granitic rocks by various nonlinear tools and comparison of their performances


Yesiloglu-Gultekin N., GÖKÇEOĞLU C., SEZER E.

INTERNATIONAL JOURNAL OF ROCK MECHANICS AND MINING SCIENCES, vol.62, pp.113-122, 2013 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 62
  • Publication Date: 2013
  • Doi Number: 10.1016/j.ijrmms.2013.05.005
  • Journal Name: INTERNATIONAL JOURNAL OF ROCK MECHANICS AND MINING SCIENCES
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.113-122
  • Keywords: Granite, Uniaxial compressive strength, Prediction method, ANFIS, ANN, Multiple regression, ARTIFICIAL NEURAL-NETWORKS, FUZZY INFERENCE SYSTEM, P-WAVE VELOCITY, SLAKE DURABILITY INDEX, BLOCK PUNCH INDEX, POINT-LOAD TEST, TENSILE-STRENGTH, MULTIVARIATE-STATISTICS, IMPACT STRENGTH, SCHMIDT HAMMER
  • Hacettepe University Affiliated: Yes

Abstract

The main goal of this study is to develop some prediction models for the UCS of six different granitic rocks selected from Turkey. During the modeling stage of the study, various approaches such as multiple regression, Artificial Neural Network (ANN), and Adaptive Neuro Fuzzy Inference System (ANFIS) are applied to estimate UCS. Tensile strength (sigma(t)), block punch index (BPI), point load index (Is((50))) and P-wave velocity (V-p) are considered as the input parameters for the models. In the study, total 75 cases including all inputs and output are used. In accordance with the analyses employed in the study, and considering the inputs, three different models are constructed as tensile strength and P-wave velocity (Model 1), BPI and P-wave velocity (Model 2), Is((50)) and P-wave velocity (Model 3) to estimate UCS. Performance assessments show that ANFIS is the better predictive tool than the other methods employed, and Model 1 is the better model for the prediction of UCS. The results show that the models developed can be used as preliminary stages of rock engineering assessments because the models developed herein have high prediction performances. It is evident that such prediction studies provides not only some practical tools but also understanding of the controlling index parameters of UCS of rocks. (C) 2013 Elsevier Ltd. All rights reserved.