Abstract

The mechanical and physical characteristics of rocks hold significant importance across various realms of research and engineering, particularly in the field of Civil Engineering. The utilization of rocks as construction materials hinges on several of their mechanical traits (Los Angeles, Micro Deval under water presence, Young's Modulus, Compression Strength) as well as physical attributes (homogeneity, porosity, etc.). Tests for determining certain of these characteristics are expensive, challenging, and time-consuming. Among these attributes, mention can be made of determining compression strength, which necessitates substantial equipment for proper sample preparation. This renders the process very costly, laborious, and leads to the complete destruction of samples during experimental measurements. Our objective in this study is to assess the uniaxial compression strength of gneiss from southern Togo using non-destructive testing. This would help mitigate the relatively high costs associated with this test. To address this issue, the development of new methods for determining this test using non-destructive testing approaches is necessary. Among the most commonly used reference techniques for characterizing materials and determining their physical and mechanical properties are non-destructive evaluation techniques based on ultrasonic wave propagation and rebound hammer testing. The adopted methodological approach involves the collection of rock samples (amphibole and biotite gneiss) from 33 sites in southern Togo, their proper sampling, and the execution of various tests on the obtained samples. The obtained results have facilitated the examination of several approaches, notably the ANFIS model (Adaptive Neuro-Fuzzy Inference System) and the MLR model (Multiple Linear Regression), for predicting the uniaxial compression strength value based on the sclerometer index and the ultrasonic wave propagation velocity (in parallel or perpendicular orientation to the foliations). It emerges from the ANFIS model, combining ultrasonic waves in the perpendicular orientation to the foliations and the sclerometer index, that the uniaxial compression strength can be predicted with an R² of 0.9884, an RMSE of 2.9271, a MAPE of 1.160, and a VAF of 98.83. In comparison, the MLR model yields an R² of 0.9832, an RMSE of 3.686, a MAPE of 1.402, and a VAF of 98.22. The derived ANFIS models can be utilized to estimate the uniaxial compression strength of gneiss in Togo and beyond. Keywords: Gneiss, Non-destructive Testing, Sclerometer Index, Ultrasonic, Compression Strength.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call