Comparative evaluation of artificial intelligence models for prediction of uniaxial compressive strength of travertine rocks, Case study: Azarshahr area, NW Iran

Machine Learning-Based Intelligent Prediction of Elastic Modulus of Rocks at Thar Coalfield

Elastic modulus (E) is a key parameter in predicting the ability of a material to withstand pressure and plays a critical role in the design of rock engineering projects. E has broad applications in the stability of structures in mining, petroleum, geotechnical engineering, etc. E can be determined directly by conducting laboratory tests, which are time consuming, and require high-quality core samples and costly modern instruments. Thus, devising an indirect estimation method of E has promising prospects. In this study, six novel machine learning (ML)-based intelligent regression models, namely, light gradient boosting machine (LightGBM), support vector machine (SVM), Catboost, gradient boosted tree regressor (GBRT), random forest (RF), and extreme gradient boosting (XGBoost), were developed to predict the impacts of four input parameters, namely, wet density (ρwet) in gm/cm3, moisture (%), dry density (ρd) in gm/cm3, and Brazilian tensile strength (BTS) in MPa on output E (GPa). The associated strengths of every input and output were systematically measured employing a series of fundamental statistical investigation tools to categorize the most dominant and important input parameters. The actual dataset of E was split as 70% for the training and 30% for the testing for each model. In order to enhance the performance of each developed model, an iterative 5-fold cross-validation method was used. Therefore, based on the results of the study, the XGBoost model outperformed the other developed models with a higher accuracy, coefficient of determination (R2 = 0.999), mean absolute error (MAE = 0.0015), mean square error (MSE = 0.0008), root mean square error (RMSE = 0.0089), and a20-index = 0.996 of the test data. In addition, GBRT and RF have also shown high accuracy in predicting E with R2 values of 0.988 and 0.989, respectively, but they can be used conditionally. Based on sensitivity analysis, all parameters were positively correlated, while BTS was the most influential parameter in predicting E. Using an ML-based intelligent approach, this study was able to provide alternative elucidations for predicting E with appropriate accuracy and run time at Thar coalfield, Pakistan.

Rock mass classification prediction model using heuristic algorithms and support vector machines: a case study of Chambishi copper mine

The rock mass is one of the key parameters in engineering design. Accurate rock mass classification is also essential to ensure operational safety. Over the past decades, various models have been proposed to evaluate and predict rock mass. Among these models, artificial intelligence (AI) based models are becoming more popular due to their outstanding prediction results and generalization ability for multiinfluential factors. In order to develop an easy-to-use rock mass classification model, support vector machine (SVM) techniques are adopted as the basic prediction tools, and three types of optimization algorithms, i.e., particle swarm optimization (PSO), genetic algorithm (GA) and grey wolf optimization (GWO), are implemented to improve the prediction classification and optimize the hyper-parameters. A database was assembled, consisting of 80 sets of real engineering data, involving four influencing factors. The three combined models are compared in accuracy, precision, recall, F₁ value and computational time. The results reveal that among three models, the GWO-SVC-based model shows the best classification performance by training. The accuracy of training and testing sets of GWO-SVC are 90.6250% (58/64) and 93.7500% (15/16), respectively. For Grades I, II, III, IV and V, the precision value is 1, 0.93, 0.90, 0.92, 0.83, the recall value is 1, 1, 0.93, 0.73, 0.83, and the F₁ value is 1, 0.96, 0.92, 0.81, 0.83, respectively. Sensitivity analysis is performed to understand the influence of input parameters on rock mass classification. It shows that the sensitive factor in rock mass quality is the RQD. Finally, the GWO-SVC is employed to assess the quality of rocks from the southeastern ore body of the Chambishi copper mine. Overall, the current study demonstrates the potential of using artificial intelligence methods in rock mass assessment, rendering far better results than the previous reports.

Applying Statistical Analysis and Machine Learning for Modeling the UCS from P-Wave Velocity, Density and Porosity on Dry Travertine

In the rock mechanics and rock engineering field, the strength parameter considered to characterize the rock is the uniaxial compressive strength (UCS). It is usually determined in the laboratory through a few statistically representative numbers of specimens, with a recommended minimum of five. The UCS can also be estimated from rock index properties, such as the effective porosity, density, and P-wave velocity. In the case of a porous rock such as travertine, the random distribution of voids inside the test specimen (not detectable in the density-porosity test, but in the compressive strength test) causes large variations on the UCS value, which were found in the range of 62 MPa for this rock. This fact complicates a sufficiently accurate determination of experimental results, also affecting the estimations based on regression analyses. Aiming to solve this problem, statistical analysis, and machine learning models (artificial neural network) was developed to generate a reliable predictive model, through which the best results for a multiple regression model between uniaxial compressive strength (UCS), P-wave velocity and porosity were obtained.

Toward multi-day-ahead forecasting of suspended sediment concentration using ensemble models

This study explores two ideas to made an improvement on the artificial neural network (ANN)-based models for suspended sediment forecasting in several time steps ahead. In this regard, both observed and forecasted time series are incorporated as input variables of the models when applied for more than one lead time. Secondly, least-square ensemble models employing multiple wavelet-ANN models are developed to increase the performance of the single model. For this purpose, different wavelet families are linked with the ANN model and performance of each model is evaluated using error measures. The Skagit River near Mount Vernon in Washington county is selected as the case study. The daily flow discharge and suspended sediment concentration (SSC) in the current day are considered as input variables to predict suspended sediment concentration in the next day. For more lead times, the input structure is updated by adding the forecast of SSC in the previous time step. Results of this study demonstrate that incorporating both observed and predicted variables in the input structure improves performance of conventional models in which those only employ observed time series as input variables. Moreover, ensemble model developed for each lead time outperforms the best single wavelet-ANN model which indicates superiority of the ensemble model over the other one. Findings of this study reveal that acceptable forecasts of daily suspended sediment concentration up to 3 days in advance can be achieved using the proposed methodology.