Hybridizing five neural-metaheuristic paradigms to predict the pillar stress in bord and pillar method

Pillar stability is an important condition for safe work in room-and-pillar mines. The instability of pillars will lead to large-scale collapse hazards, and the accurate estimation of induced stresses at different positions in the pillar is helpful for pillar design and guaranteeing pillar stability. There are many modeling methods to design pillars and evaluate their stability, including empirical and numerical method. However, empirical methods are difficult to be applied to places other than the original environmental characteristics, and numerical methods often simplify the boundary conditions and material properties, which cannot guarantee the stability of the design. Currently, machine learning (ML) algorithms have been successfully applied to pillar stability assessment with higher accuracy. Thus, the study adopted a back-propagation neural network (BPNN) and five elements including the sparrow search algorithm (SSA), gray wolf optimizer (GWO), butterfly optimization algorithm (BOA), tunicate swarm algorithm (TSA), and multi-verse optimizer (MVO). Combining metaheuristic algorithms, five hybrid models were developed to predict the induced stress within the pillar. The weight and threshold of the BPNN model are optimized by metaheuristic algorithms, in which the mean absolute error (MAE) is utilized as the fitness function. A database containing 149 data samples was established, where the input variables were the angle of goafline (A), depth of the working coal seam (H), specific gravity (G), distance of the point from the center of the pillar (C), and distance of the point from goafline (D), and the output variable was the induced stress. Furthermore, the predictive performance of the proposed model is evaluated by five metrics, namely coefficient of determination (R²), root mean squared error (RMSE), variance accounted for (VAF), mean absolute error (MAE), and mean absolute percentage error (MAPE). The results showed that the five hybrid models developed have good prediction performance, especially the GWO-BPNN model performed the best (Training set: R² = 0.9991, RMSE = 0.1535, VAF = 99.91, MAE = 0.0884, MAPE = 0.6107; Test set: R² = 0.9983, RMSE = 0.1783, VAF = 99.83, MAE = 0.1230, MAPE = 0.9253).

Improved mine waste dump planning through integration of geochemical and mineralogical data and mixed integer programming: Reducing acid rock generation from mine waste

Although the environmental significance of acid rock drainage (ARD) generated from mining wastes is well known, selecting the appropriate ARD management strategy can prove a complicated task. Chemical methods are favored for initial mine waste characterization but using these exclusively can overlook key factors, e.g., mineralogy, which controls the formation and elution of ARD. This paper first presents an ARD waste rock classification developed on Triple Characterization Criteria (TCC) which considers three input parameters: neutralizing potential ratio (NPR), net acid generation (NAG pH), and modal mineralogy weathering index (MMWI) values. Second, a new mixed-integer programming (MIP) model to guide waste dump construction with the dual aim of preventing ARD across the life-of-mine (LOM) and reducing waste rock re-handling, is introduced. Last, the spatial distribution of TCC in a planned waste dump is simulated via geo-statistical techniques to evaluate the MIP model. The proposed waste rock classification and dump planning model has been tested at an iron mine. The results of the MIP modeling and simulation of TCC showed the successful prevention of ARD by achieving large values of TCC (NPR ≥2, NAG pH ≥ 4.5, and MMWI ≥4.7) for dump cells, with the planned mine production maintained. The integrated TCC approach introduced in this study is intended to enable mine operators, at the start of the LOM, to effectively forecast ARD from future waste rock. Further, the MIP model will facilitate development of a mine schedule that optimizes the use of the waste materials based on TCC values. If used correctly, the TCC and MIP model have the potential to enable mine operators to reduce their environmental footprint across the entire LOM.

Development of a MIP model to maximize NPV and minimize adverse environmental impact-a heuristic approach

Block sequencing is of great importance in an open-pit mining operation. Sequencing is usually performed to maximize the net present value (NPV). Also, from the environmental viewpoint, the sequence of dumping mined materials is of significant value in the sulfide mines. The potential acid-forming (PAF) waste rocks in these mines can seriously damage the environment due to the formation of acid mine drainage (AMD). To prevent the exposition of the PAF materials, it is essential to design suitable block sequencing. For this purpose, encapsulation of the PAF rocks by non-acid forming (NAF) rocks should be considered during waste dumping. However, this method can impose unnecessary re-handling costs. This issue is due to the determination of the waste-dump sequence based on improper block sequencing obtained from the previous models with the NPV maximization strategy. In the present study, a mixed-integer programming (MIP) model is proposed for generating proper block sequencing taking into account the composition of waste rocks. The main objective of the proposed MIP model is to maximize NPV and minimize the destructive environmental effects of PAF materials dumping. The CPLEX solver was applied to solve the proposed model in small datasets. Then, an artificial bee colony (ABC) is implemented to find out optimum block sequencing and waste dumping (BSWD) on a large scale. The proposed approach was examined employing several sets of data. The obtained results were compared with those of the CPLEX solver as a benchmark. An approximate gap of 2% demonstrates the efficiency of the proposed approach.

Development of GP and GEP models to estimate an environmental issue induced by blasting operation

Air overpressure (AOp) is one of the most adverse effects induced by blasting in the surface mines and civil projects. So, proper evaluation and estimation of the AOp is important for minimizing the environmental problems resulting from blasting. The main aim of this study is to estimate AOp produced by blasting operation in Miduk copper mine, Iran, developing two artificial intelligence models, i.e., genetic programming (GP) and gene expression programming (GEP). Then, the accuracy of the GP and GEP models has been compared to multiple linear regression (MLR) and three empirical models. For this purpose, 92 blasting events were investigated, and subsequently, the AOp values were carefully measured. Moreover, in each operation, the values of maximum charge per delay and distance from blast points, as two effective parameters on the AOp, were measured. After predicting by the predictive models, their performance prediction was checked in terms of variance account for (VAF), coefficient of determination (CoD), and root mean square error (RMSE). Finally, it was found that the GEP with VAF of 94.12%, CoD of 0.941, and RMSE of 0.06 is a more precise model than other predictive models for the AOp prediction in the Miduk copper mine, and it can be introduced as a new powerful tool for estimating the AOp resulting from blasting.

Prediction and Analysis of Blast Parameters Using Artificial Neural Network

In this study an attempt is made to predict the ratio of muck pile profile before and after the blast, fly rock and total explosive used, based on simple field tests as well blast design parameters. Prediction is done by making three different artificial neural network (ANN) models. Comparative statistical analysis is made among these three networks to ensure their performance suitability. Models of ANN were based on Feed Forward Back Propagation network with training functions – Resilient Backpropagation, One Step Secant and Powell-Beale Restarts. Total numbers of datasets chosen were 92 among which 17 were chosen for testing and validation and the rest were used for the training of networks. Statistical analysis is also made for these datasets. Considering performance for all the outputs, the best results are predicted by Powell-Beale Restarts, with an average percentage error of 5.871% for the ratio of muck pile before and after the blast, 5.335% for fly rocks and 5.775% for total explosive used. These parameters are predicted by number of holes to be blasted, hole diameter, pattern (spacing (m) X burden (m)), total volume of rock in a blast, average depth and total drill depth.