Application of soft computing in predicting rock fragmentation to reduce environmental blasting side effects

In the blasting operation, risk of facing with undesirable environmental phenomena such as ground vibration, air blast, and flyrock is very high. Blasting pattern should properly be designed to achieve better fragmentation to guarantee the successfulness of the process. A good fragmentation means that the explosive energy has been applied in a right direction. However, many studies indicate that only 20–30 % of the available energy is actually utilized for rock fragmentation. Involvement of various effective parameters has made the problem complicated, advocating application of new approaches such as artificial intelligence-based techniques. In this paper, artificial neural network (ANN) method is used to predict rock fragmentation in the blasting operation of the Sungun copper mine, Iran. The predictive model is developed using eight and three input and output parameters, respectively. Trying various types of the networks, it was found that a trained model with back-propagation algorithm having architecture 8-15-8-3 is the optimum network. Also, performance comparison of the ANN modeling with that of the statistical method was confirmed robustness of the neural networks to predict rock fragmentation in the blasting operation. Finally, sensitivity analysis showed that the most influential parameters on fragmentation are powder factor, burden, and bench height.     

An intelligent neural network model for evaluating performance of immobilized cell biofilter treating hydrogen sulphide vapors

Biofiltration has shown to be a promising technique for handling malodours arising from process industries. The present investigation pertains to the removal of hydrogen sulphide in a lab scale biofilter packed with biomedia, encapsulated by sodium alginate and poly vinyl alcohol. The experimental data obtained under both steady state and shock loaded conditions were modelled using the basic principles of artificial neural networks. Artificial neural networks are powerful data driven modelling tools which has the potential to approximate and interpret complex input/output relationships based on the given sets of data matrix. A predictive computerised approach has been proposed to predict the performance parameters namely, removal efficiency and elimination capacity using inlet concentration, loading rate, flow rate and pressure drop as the input parameters to the artificial neural network model. Earlier, experiments from continuous operation in the biofilter showed removal efficiencies from 50 to 100 % at inlet loading rates varying up to 13 g H₂S/m³h. The internal network parameter of the artificial neural network model during simulation was selected using the 2^k factorial design and the best network topology for the model was thus estimated. The results showed that a multilayer network (4-4-2) with a back propagation algorithm was able to predict biofilter performance effectively with R² values of 0.9157 and 0.9965 for removal efficiency and elimination capacity in the test data. The proposed artificial neural network model for biofilter operation could be used as a potential alternative for knowledge based models through proper training and testing of the state variables.     

Prediction of blasting-induced fragmentation in Meydook copper mine using empirical,statistical, and mutual information models

One of the most important aims of blasting in open pit mines is to reach desirable size of fragmentation. Prediction of fragmentation has great importance in an attempt to prevent economic drawbacks. In this study, blasting data from Meydook mine were used to study the effect of different parameters on fragmentation; 30 blast cycles performed in Meydook mine were selected to predict fragmentation where six more blast cycles are used to validate the results of developed models. In this research, mutual information (MI) method was employed to predict fragmentation. Ten parameters were considered as primary ones in the model. For the sake of comparison, Kuz-Ram empirical model and statistical modeling were also used. Coefficient of determination (R ²), root mean square error (RMSE), and mean absolute error (MAE) were then used to compare the models. Results show that MI model with values of R ², RMSE, and MAE equals 0.81, 10.71, and 9.02, respectively, is found to have more accuracy with better performance comparing to Kuz-Ram and statistical models.     

Prediction of flyrock and backbreak in open pit blasting operation: a neuro-genetic approach

An ideally performed blasting operation enormously influences the mining overall cost. This aim can be achieved by proper prediction and attenuation of flyrock and backbreak. Poor performance of the empirical models has urged the application of new approaches. In this paper, an attempt has been made to develop a new neuro-genetic model for predicting flyrock and backbreak in Sungun copper mine, Iran. Recognition of the optimum model with this method as compared with the classic neural networks is faster and convenient. Genetic algorithm was utilized to optimize neural network parameters. Parameters such as number of neurons in hidden layer, learning rate, and momentum were considered in the model construction. The performance of the model was examined by statistical method in which absolutely higher efficiency of neuro-genetic modeling was proved. Sensitivity analysis showed that the most influential parameters on flyrock are stemming and powder factor, whereas for backbreak, stemming and charge per delay are the most effective parameters.     

Burden prediction in blasting operation using rock geomechanical properties

Burden prediction is a vital task in the production blasting. Both the excessive and insufficient burden can significantly affect the result of blasting operation. The burden which is determined by empirical models is often inaccurate and needs to be adjusted experimentally. In this paper, an attempt was made to develop an artificial neural network (ANN) in order to predict burden in the blasting operation of the Mouteh gold mine, using considering geomechanical properties of rocks as input parameters. As such here, network inputs consist of blastability index (BI), rock quality designation (RQD), unconfined compressive strength (UCS), density, and cohesive strength. To make a database (including 95 datasets), rock samples are used from Iran’s Mouteh goldmine. Trying various types of the networks, a neural network, with architecture 5-15-10-1, was found to be optimum. Superiority of ANN over regression model is proved by calculating. To compare the performance of the ANN modeling with that of multivariable regression analysis (MVRA), mean absolute error (E _a), mean relative error (E _r), and determination coefficient (R ²) between predicted and real values were calculated for both the models. It was observed that the ANN prediction capability is better than that of MVRA. The absolute and relative errors for the ANN model were calculated 0.05 m and 3.85%, respectively, whereas for the regression analysis, these errors were computed 0.11 m and 5.63%, respectively. Moreover, determination coefficient of the ANN model and MVRA were determined 0.987 and 0.924, respectively. Further, a sensitivity analysis shows that while BI and RQD were recognized as the most sensitive and effective parameters, cohesive strength is considered as the least sensitive input parameters on the ANN model output effective on the proposed (burden).     

Application of neural networks to predict net present value in mining projects

Net present value (NPV) is the most popular economic indicator in evaluation of the investment projects. For the mining projects, this criterion is calculated under uncertainty associated with the relevant parameters of say commodity price, discount rate, etc. Accurate prediction of the NPV is a quite difficult process. This paper mainly deals with the development of a new model to predict NPV using artificial neural network (ANN) in the Zarshuran gold mine, Iran. Gold price (as the main product), silver price (as the byproduct), and discount rate were considered as input parameters for the ANN model. To reach an optimum architecture, different types of networks were examined on the basis of a trial and error mechanism. A neural network with architecture 3-15-10-1 and root mean square error of 0.092 is found to be optimum. Prediction capability of the proposed model was examined through computing determination coefficient (R ²?=?0.987) between predicted and real NPVs. Absolute error of US$0.1 million and relative error of 1.4 % also confirmed powerfulness of the developed ANN model. According to sensitivity analysis, it was observed that the gold price is the most effective and discount rate is the least effective parameter on the NPV.     

Comparison of FFNN and ANFIS models for estimating groundwater level

Prediction of water level is an important task for groundwater planning and management when the water balance consistently tends toward negative values. In Maheshwaram watershed situated in the Ranga Reddy District of Andhra Pradesh, groundwater is overexploited, and groundwater resources management requires complete understanding of the dynamic nature of groundwater flow. Yet, the dynamic nature of groundwater flow is continually changing in response to human and climatic stresses, and the groundwater system is too intricate, involving many nonlinear and uncertain factors. Artificial neural network (ANN) models are introduced into groundwater science as a powerful, flexible, statistical modeling technique to address complex pattern recognition problems. This study presents the comparison of two methods, i.e., feed-forward neural network (FFNN) trained with Levenberg–Marquardt (LM) algorithm compared with a fuzzy logic adaptive network-based fuzzy inference system (ANFIS) model for better accuracy of the estimation of the groundwater levels of the Maheshwaram watershed. The statistical indices used in the analysis were the root mean square error (RMSE), regression coefficient (R ²) and error variation (EV).The results show that FFNN-LM and ANFIS models provide better accuracy (RMSE = 4.45 and 4.94, respectively, R ² is 93% for both models) for estimating groundwater levels well in advance for the above location.     

Artificial Neural Network as a Tool for Backbreak Prediction

Backbreak is one of the destructive side effects of the blasting operation. Reducing of this event is very important for economic of a mining project. Involvement of various parameters has made the backbreak analyzing difficult. Currently there is no any specific method to predict or control the phenomenon considering all the effective parameters. In this paper, artificial neural network (ANN) as a powerful tool for solving such complicated problems is used to predict backbreak in blasting operation of the Sangan iron mine, Iran. Network training was fulfilled using a collected database of the practiced operation including blast design details and rock condition. Trying various types of the networks, a network with two hidden layers was found to be optimum. Performance of the ANN model was compared with statistical analysis using datasets which were kept apart from the original database. According to the obtained results, for the ANN model there existed a higher correlation (R² = 0.868) and lesser error (RMSE = 0.495) between the predicted and measured backbreak as compared to the regression model. Also, sensitivity analysis revealed that the inputs rock factor and number of rows are the most and the least sensitive parameters on the output backbreak, respectively.     

Estimation of soil compaction parameters by using statistical analyses and artificial neural networks

This study presents the application of different methods (simple–multiple analysis and artificial neural networks) for the estimation of the compaction parameters (maximum dry unit weight and optimum moisture content) from classification properties of the soils. Compaction parameters can only be defined experimentally by Proctor tests. The data collected from the dams in some areas of Nigde (Turkey) were used for the estimation of soil compaction parameters. Regression analysis and artificial neural network estimation indicated strong correlations (r ² = 0.70–0.95) between the compaction parameters and soil classification properties. It has been shown that the correlation equations obtained as a result of regression analyses are in satisfactory agreement with the test results. It is recommended that the proposed correlations will be useful for a preliminary design of a project where there is a financial limitation and limited time.     

Prediction of longitudinal dispersion coefficient using multivariate adaptive regression splines

In this paper, multivariate adaptive regression splines (MARS) was developed as a novel soft-computing technique for predicting longitudinal dispersion coefficient (D_L) in rivers. As mentioned in the literature, experimental dataset related to D_L was collected and used for preparing MARS model. Results of MARS model were compared with multi-layer neural network model and empirical formulas. To define the most effective parameters on D_L, the Gamma test was used. Performance of MARS model was assessed by calculation of standard error indices. Error indices showed that MARS model has suitable performance and is more accurate compared to multi-layer neural network model and empirical formulas. Results of the Gamma test and MARS model showed that flow depth (H) and ratio of the mean velocity to shear velocity (u/u^?) were the most effective parameters on the D_L.     

基于人工神经网络模型的岩石特性预测

近年来,软计算技术被用作替代的统计工具。如人工神经网络（ANN）被用于开发预测模型来估计所需的参数。在本研究中,通过利用冲击钻进过程中的一些钻进参数（气压、推力、钻头直径、穿透率）和所产生的声级,建立了预测岩石性质的神经网络模型。在实验室中所产生的数据,用于开发预测岩石特性（如单轴抗压强度、耐磨性、抗拉强度和施密特回弹数）的神经网络模型,并使用各种预测性能指标对所建模型进行检验,结果表明人工神经网络模型适用于岩石性质的预测。     

Relationship Between the Geometric Parameters of Rock Fractures,the Size of Percolation Clusters and REV

Modeling fractured rocks with numerical methods requires some derived parameters, among which the fracture network connectivity and the size of the representative elementary volume (REV) are both of crucial importance. Percolation and REV analyses were made by the RepSim code. The program uses input parameters such as fractal dimension of the fracture midpoints (D _c), length exponent (E) and relative dip (α ^r) data. For percolation analysis, the relative sizes of the largest percolation clusters have been calculated by stochastic realizations of the simulated fracture networks with different parameter triplets. Furthermore, fracture networks can be classified into three major types on the basis of their (E,D _c,α ^r) parameters. For the REV calculations, the porosity of the generated fracture network was calculated. The derived REV size of a fracture network depends essentially on input parameters and shows a decreasing tendency with increasing D and E and vice versa. The method mentioned above was tested on both metamorphic samples of the Pannonian Basin and Variscan granitoid rocks of the Mórágy Complex. Percolation values predicted for the Mórágy granite are highly sensitive to alterations in the input parameters. The amphibolite bodies displayed a modeled fracture network with 80 to 90% of all fractures being interconnected, while the largest achievable percolation cluster size of gneiss is less than 10%. The REV size of the amphibolite is about 20 m as a result of connected fractures filling the whole body, while gneiss has lower porosity and higher REV (approximately 70 m).     

An application of artificial intelligence for rainfall-runoff modeling

This study proposes an application of two techniques of artificial intelligence (AI) for rainfall-runoff modeling: the artificial neural networks (ANN) and the evolutionary computation (EC). Two different ANN techniques, the feed forward back propagation (FFBP) and generalized regression neural network (GRNN) methods are compared with one EC method, Gene Expression Programming (GEP) which is a new evolutionary algorithm that evolves computer programs. The daily hydrometeorological data of three rainfall stations and one streamflow station for Juniata River Basin in Pennsylvania state of USA are taken into consideration in the model development. Statistical parameters such as average, standard deviation, coefficient of variation, skewness, minimum and maximum values, as well as criteria such as mean square error (MSE) and determination coefficient (R ²) are used to measure the performance of the models. The results indicate that the proposed genetic programming (GP) formulation performs quite well compared to results obtained by ANNs and is quite practical for use. It is concluded from the results that GEP can be proposed as an alternative to ANN models.     

Application of neural network and ANFIS model for earthquake occurrence in Iran

This study examined the spatial-temporal variations in seismicity parameters for the September 10^th, 2008 Qeshm earthquake in south Iran. To this aim, artificial neural networks and Adaptive Neural Fuzzy Inference System (ANFIS) were applied. The supervised Radial Basis Function (RBF) network and ANFIS model were implemented because they have shown the efficiency in classification and prediction problems. The eight seismicity parameters were calculated to analyze spatial and temporal seismicity pattern. The data preprocessing that included normalization and Principal Component Analysis (PCA) techniques was led before the data was fed into the RBF network and ANFIS model. Although the accuracy of RBF network and ANFIS model could be evaluated rather similar, the RBF exhibited a higher performance than the ANFIS for prediction of the epicenter area and time of occurrence of the 2008 Qeshm main shock. A proper training on the basis of RBF network and ANFIS model might adopt the physical understanding between seismic data and generate more effective results than conventional prediction approaches. The results of the present study indicated that the RBF neural networks and the ANFIS models could be suitable tools for accurate prediction of epicenteral area as well as time of occurrence of forthcoming strong earthquakes in active seismogenic areas.     

New Prediction Models for Mean Particle Size in Rock Blast Fragmentation

The paper refers the reader to a blast data base developed in a previous study. The data base consists of blast design parameters, explosive parameters, modulus of elasticity and in situ block size. A hierarchical cluster analysis was used to separate the blast data into two different groups of similarity based on the intact rock stiffness. The group memberships were confirmed by the discriminant analysis. A part of this blast data was used to train a single-hidden layer back propagation neural network model to predict mean particle size resulting from blast fragmentation for each of the obtained similarity groups. The mean particle size was considered to be a function of seven independent parameters. An extensive analysis was performed to estimate the optimum value for the number of units for the hidden layer for each of the obtained similarity groups. The blast data that were not used for training were used to validate the trained neural network models. For the same two similarity groups, multivariate regression models were also developed to predict mean particle size. Capability of the developed neural network models as well as multivariate regression models was determined by comparing predictions with measured mean particle size values and predictions based on one of the most applied fragmentation prediction models appearing in the blasting literature. Prediction capability of the trained neural network models as well as multivariate regression models was found to be strong and better than the existing most applied fragmentation prediction model. Diversity of the blasts data used is one of the most important aspects of the developed models.     

Ore Grade Prediction Using a Genetic Algorithm and Clustering Based Ensemble Neural Network Model

Accurate prediction of ore grade is essential for many basic mine operations, including mine planning and design, pit optimization, and ore grade control. Preference is given to the neural network over other interpolation techniques for ore grade estimation because of its ability to learn any linear or non-linear relationship between inputs and outputs. In many cases, ensembles of neural networks have been shown, both theoretically and empirically, to outperform a single network. The performance of an ensemble model largely depends on the accuracy and diversity of member networks. In this study, techniques of a genetic algorithm (GA) and k-means clustering are used for the ensemble neural network modeling of a lead–zinc deposit. Two types of ensemble neural network modeling are investigated, a resampling-based neural ensemble and a parameter-based neural ensemble. The k-means clustering is used for selecting diversified ensemble members. The GA is used for improving accuracy by calculating ensemble weights. Results are compared with average ensemble, weighted ensemble, best individual networks, and ordinary kriging models. It is observed that the developed method works fairly well for predicting zinc grades, but shows no significant improvement in predicting lead grades. It is also observed that, while a resampling-based neural ensemble model performs better than the parameter-based neural ensemble model for predicting lead grades, the parameter-based ensemble model performs better for predicting zinc grades.     

Back-propagation neural network and support vector machines for gold mineral prospectivity mapping in the Hatu region,Xinjiang, China

Machine Learning technologies have the potential to deliver new nonlinear mineral prospectivity mapping (MPM) models. In this study, Back Propagation (BP) neural network Support Vector Machine (SVM) methods were applied to MPM in the Hatu region of Xinjiang, northwestern China. First, a conceptual model of mineral prospectivity for Au deposits was constructed by analysis of geological background. Evidential layers were selected and transformed into a binary data format. Then, the processes of selecting samples and parameters were described. For the BP model, the parameters of the network were 9–10???1; for the SVM model, a radial basis function was selected as the kernel function with best C?=?1 and γ = 0.25. MPM models using these parameters were constructed, and threshold values of prediction results were determined by the concentration-area (C-A) method. Finally, prediction results from the BP neural network and SVM model were compared with that of a conventional method that is the weight- of- evidence (W- of- E). The prospectivity efficacy was evaluated by traditional statistical analysis, prediction-area (P-A) plots, and the receiver operating characteristic (ROC) technique. Given the higher intersection position (74% of the known deposits were within 26% of the total area) and the larger AUC values (0.825), the result shows that the model built by the BP neural network algorithm has a relatively better prediction capability for MPM. The BP neural network algorithm applied in MPM can elucidate the next investigative steps in the study area.     

Particle swarm optimization feedforward neural network for modeling runoff

The rainfall-runoff relationship is one of the most complex hydrological phenomena. In recent years, hydrologists have successfully applied backpropagation neural network as a tool to model various nonlinear hydrological processes because of its ability to generalize patterns in imprecise or noisy and ambiguous input and output data sets. However, the backpropagation neural network convergence rate is relatively slow and solutions can be trapped at local minima. Hence, in this study, a new evolutionary algorithm, namely, particle swarm optimization is proposed to train the feedforward neural network. This particle swarm optimization feedforward neural network is applied to model the daily rainfall-runoff relationship in Sungai Bedup Basin, Sarawak, Malaysia. The model performance is measured using the coefficient of correlation and the Nash-Sutcliffe coefficient. The input data to the model are current rainfall, antecedent rainfall and antecedent runoff, while the output is current runoff. Particle swarm optimization feedforward neural network simulated the current runoff accurately with R = 0.872 and E² = 0.775 for the training data set and R = 0.900 and E²= 0.807 for testing data set. Thus, it can be concluded that the particle swarm optimization feedforward neural network method can be successfully used to model the rainfall-runoff relationship in Bedup Basin and it could be to be applied to other basins.     

Flow estimations for the Sohu Stream using artificial neural networks

In this study, daily rainfall–runoff relationships for Sohu Stream were modelled using an artificial neural network (ANN) method by including the feed-forward back-propagation method. The ANN part was divided into two stages. During the first stage, current flows were estimated by using previously measured flow data. The best network architecture was found to utilise two neurons in the input layer (the delayed flows from the first and second days), two hidden layers, and one output layer (the current flow). The coefficient of determination (R ²) in this architecture was 81.4%. During the second stage, the current flows were estimated by using a combination of previously measured values for precipitation, temperature, and flows. The best architecture consisted of an input layer of 2 days of delayed precipitation, 3 days of delayed flows, and temperature of the current. The R ² in this architecture was calculated to be 85.5%. The results of the second stage best reflected the real-world situation because they accounted for more input variables. In all models, the variables with the highest R ² ranked as the previous flow (81.4%), previous precipitation (21.7%), and temperature.     

Prediction of rainfall using artificial neural networks for synoptic station of Mashhad: a case study

In this paper, we have utilized ANN (artificial neural network) modeling for the prediction of monthly rainfall in Mashhad synoptic station which is located in Iran. To achieve this black-box model, we have used monthly rainfall data from 1953 to 2003 for this synoptic station. First, the Hurst rescaled range statistical (R/S) analysis is used to evaluate the predictability of the collected data. Then, to extract the rainfall dynamic of this station using ANN modeling, a three-layer feed-forward perceptron network with back propagation algorithm is utilized. Using this ANN structure as a black-box model, we have realized the complex dynamics of rainfall through the past information of the system. The approach employs the gradient decent algorithm to train the network. Trying different parameters, two structures, M₅₃₁ and M₇₄₁, have been selected which give the best estimation performance. The performance statistical analysis of the obtained models shows with the best tuning of the developed monthly prediction model the correlation coefficient (R), root mean square error (RMSE), and mean absolute error (MAE) are 0.93, 0.99, and 6.02 mm, respectively, which confirms the effectiveness of the developed models.