Prediction of Rock Compressive Strength Using Machine Learning Algorithms Based on Spectrum Analysis of Geological Hammer

The traditional method to estimate rock compressive strength (RCS) in field operation is dependent on hammering rocks and artificial identification. It is too subjective to get high estimation accuracy. For this reason, the new and non-destructive method uses machine learning algorithms to analyze acoustic characteristics of geological hammer to predict RCS accurately. The hammering sound samples were successively preprocessed by signal enhancement algorithm and double-threshold method to reduce noise and acquire valuable intervals of all. We have also performed the time-frequency domain conversion on sound signal through FFT, which obtained two brand new indexes, amplitude attenuation coefficient and high and low frequency ratio, as the input parameters of models. By contrasting the performance of various models based on k-nearest neighbors, naive Bayes, random forest, artificial neural networks (ANN), and support vector machines (SVM), we uncovered that the prediction accuracy of both SVM and ANN was over 95%, superior to others. Thus, SVM and ANN were better for widespread application in geological surveys and construction acceptance to predict RCS accurately. In addition, characteristic mechanism of acoustic spectrum was explained from microstructure, energy dissipation and filter effect, which indicated why there existed strong correlation between acoustic characteristics and RCS. The current rock mass classification standard was supplemented with the above two characteristic indexes for better identification.

[1]  Michael I. Jordan,et al.  Machine learning: Trends, perspectives, and prospects , 2015, Science.

[2]  Vito Adrian Cantu,et al.  Machine learning for graph-based representations of three-dimensional discrete fracture networks , 2017, Computational Geosciences.

[3]  M. Karakus,et al.  Fuzzy and Multiple Regression Modelling for Evaluation of Intact Rock Strength Based on Point Load, Schmidt Hammer and Sonic Velocity , 2006 .

[4]  John R. Anderson,et al.  MACHINE LEARNING An Artificial Intelligence Approach , 2009 .

[5]  E. Ya¿ar,et al.  Estimation of rock physicomechanical properties using hardness methods , 2004 .

[6]  H. Fattahi,et al.  Prediction of porosity and water saturation using pre-stack seismic attributes: a comparison of Bayesian inversion and computational intelligence methods , 2016, Computational Geosciences.

[7]  Resat Ulusay,et al.  The ISRM suggested methods for rock characterization, testing and monitoring, 2007-2014 , 2015 .

[8]  A. Aydin.  ISRM Suggested method for determination of the Schmidt hammer rebound hardness: Revised version ☆ , 2009 .

[9]  Richard M. Schwartz,et al.  Enhancement of speech corrupted by acoustic noise , 1979, ICASSP.

[10]  Eugenio Oñate,et al.  An empirical comparison of machine learning techniques for dam behaviour modelling , 2015 .

[11]  Tin Kam Ho,et al.  The Random Subspace Method for Constructing Decision Forests , 1998, IEEE Trans. Pattern Anal. Mach. Intell..

[12]  Danial Jahed Armaghani,et al.  Prediction of uniaxial compressive strength of rock samples using hybrid particle swarm optimization-based artificial neural networks , 2015 .

[13]  V. Rodriguez-Galiano,et al.  Machine learning predictive models for mineral prospectivity: an evaluation of neural networks, random forest, regression trees and support vector machines , 2015 .

[14]  S. R. Hencher,et al.  Assessing the Shear Strength of Rock Discontinuities at Laboratory and Field Scales , 2015, Rock Mechanics and Rock Engineering.

[15]  John C. Brigham,et al.  Inverse calculation of in situ stress in rock mass using the Surrogate-Model Accelerated Random Search algorithm , 2014 .

[16]  H.R.G.K. Hack,et al.  ESTIMATING THE INTACT ROCK STRENGTH OF A ROCK MASS BY SIMPLE MEANS , 2002 .

[17]  Qiru Zhou,et al.  Robust endpoint detection and energy normalization for real-time speech and speaker recognition , 2002, IEEE Trans. Speech Audio Process..

[18]  N. Altman.  An Introduction to Kernel and Nearest-Neighbor Nonparametric Regression , 1992 .

[19]  Edy Tonnizam Mohamad,et al.  Prediction of the unconfined compressive strength of soft rocks: a PSO-based ANN approach , 2015, Bulletin of Engineering Geology and the Environment.

[20]  M. Monjezi,et al.  Prediction of Backbreak in Open-Pit Blasting Operations Using the Machine Learning Method , 2013, Rock Mechanics and Rock Engineering.

[21]  A. Kesimal,et al.  A comparative study of Schmidt hammer test methods for estimating the uniaxial compressive strength of rocks , 2015, Bulletin of Engineering Geology and the Environment.

[22]  Antonio Brencich,et al.  Calibration and Reliability of the Rebound (Schmidt) Hammer Test , 2013 .

[23]  D. Thomson,et al.  Spectrum estimation and harmonic analysis , 1982, Proceedings of the IEEE.

[24]  Ethem Alpaydin,et al.  Introduction to machine learning , 2004, Adaptive computation and machine learning.

[25]  K. Kido,et al.  Digital Fourier Analysis: Fundamentals , 2014 .

[26]  Chungyong Lee,et al.  Optimizing feature extraction for speech recognition , 2003, IEEE Trans. Speech Audio Process..

[27]  Kourosh Shahriar,et al.  A support vector regression model for predicting tunnel boring machine penetration rates , 2014 .

[28]  Raúl Rojas,et al.  Neural Networks - A Systematic Introduction , 1996 .

[29]  P. Cao,et al.  Correlation of UCS Rating with Schmidt Hammer Surface Hardness for Rock Mass Classification , 2016, Rock Mechanics and Rock Engineering.

[30]  E. Hoek,et al.  Estimating the geotechnical properties of heterogeneous rock masses such as flysch , 2001 .

[31]  Boualem Boashash,et al.  Time-Frequency Signal Analysis and Processing: A Comprehensive Reference , 2015 .

[32]  Hadi Fattahi,et al.  Applying soft computing methods to predict the uniaxial compressive strength of rocks from schmidt hammer rebound values , 2017, Computational Geosciences.

[33]  Li Zhang,et al.  Hybrid decision tree and naïve Bayes classifiers for multi-class classification tasks , 2014, Expert Syst. Appl..