جستجوی مقالات مرتبط با کلیدواژه « Modeling » در نشریات گروه « مهندسی معدن »

Choosing the right drilling and blasting methods is very important to improve mine productivity and increase profits, while increasing the safety of workers and the environment. One of the main goals in Blasting is to estimate the optimal amount of specific cost, which is described as the specific cost required for optimizing rock crushing, air blast, ground shaking, and rock throwing. In this Paper, to further improve the effect of blasting and increase the efficiency of mine production, we optimized the plan by considering the implementation of the blasting plan with respect to the drilling intervals, and a numerical simulation model was created to provide technical guidance for the optimization plan. In this Paper, data collection and analysis were done by the software and it was determined that with R^2=0.969, VAF=9.825, the lowest error is RMSE=0.031 and MAPE=0.323 with unit m^3kg/ . The polynomial function model with incorrect coefficients has a more satisfactory performance and this function model was subjected to sensitivity analysis using the cosine field method. The evaluation and analysis in this way also showed that the selected function model has a more accurate calculation of the specific cost. This proposed function was called in the bat algorithm and the crow search algorithm in MATLAB software, and the optimization operation was performed by the algorithms. In this Paper, the number of 11 data from 31 blasting series that were actually carried out in Sadat iron ore mine of Zarand series was collected. The special cost of this mine was modeled with 6 functions by spss software, and among them, the polynomial function model with incorrect coefficients was chosen to predict the special cost. The special cost optimization operation was performed by two bat algorithms and the crow search algorithm, and the special cost optimized by the algorithms was compared. By sensitivity analysis, the sensitive and effective parameters taken in this mine were identified for special spending. This analysis showed that all the input parameters used in this modeling have a strong sensitivity to the specific cost, and among them, the ratio of the height of the step to the thickness of the bar rock has the highest sensitivity and the total consumption cost has the lowest sensitivity.

In the current research, geological-geomechanical modeling of the Mahallat, Hajiabad Travertine was conducted using 90 boreholes with a total length of 663.35 meters, at a large scale. To achieve the goal, first, 2D geological-geomechanical strip-logs, 3D lithological strip-logs, 3D lithological model of all boreholes, 3D fence diagram of lithology variations for the boundary boreholes and isopach map of the mining deposit were drawn. Afterward, statistical studies of the Core Recovery (RC) and Rock Quality Designation (RQD) data of the boreholes were performed, as well as 3D model of the RC data, 3D statistical map and 3D model of the RQD data were generated.The results of the research show that the average thickness of the deposit (containing travertine and marly travertine) equals to 4.05 meters, and the remaining reserve within the boreholes boundary is estimated 5,868,439 tons, indicating a significant and worthwhile reserve. The average RC of the boreholes is 77.4%, reflecting a relatively high quality of drilling operation and the deposit content. Also, the average RQD in the boreholes is 69.3%, classifying the quality of the Hajiabad Travertine based on the engineering rock classification using the RQD criterion as "fair". Based on the 3D statistical map of RQD data, the highest RQD values belong to the boreholes in the southern half of the area. The 3D model of the RQD also indicates that, in general, the quality of the Hajiabad Travertine stones are better in zones eastern 6 (6E), western 6 (6W), and 14 compared to the other zones.

Prediction of gas emissions is a calculation based on empirical processing in the coal mining. In this study, methane diffusion was investigated coal exploration on parameters such as coal gas content, mining, depth, etc., in two faces in the Tabas Parvadeh central coal mining. Therefore, in the beginning, the initial and final values for the studied faces were modeled based on the neural network. The neural network was calculated methane for P8/TG face, between 0.4476%-0.9921% and P8/MG face are between 0.1636%-0.3379%. The intensity of methane gas emission from the studied mine was done based on the level of coal and the amount of gas volume using MATLAB coding. Also, the type of extraction method in the mine was done on the amount of gas emission caused by mining. According to the amount and aggregate of induced fractures, the permeability shows several degrees of increase. These amounts can evacuation the gas caused by the underground extraction of the mine faster. In this study, the amount of gas emissions due to mining was calculated using coding to the content of the remaining gas matrix In this study, the amount of gas emission due to drilling was 0.04%, blasting was 0.045%, support was 0.040%, muckout was 0.06% and jiggring was 0.25%. Small errors in the measurement of gas content can lead to large errors in the estimated calculations, which were determined using the software. The correlation coefficient of input and output of the studied data was estimated as 0.984%. The modeling done for the 2 face current shows a good result which can be used to improve mine safety.

The cleaner circuit at Miduk copper concentrator consists of 3 parallel flotation columns (4m in diameter and 12m in height). The cleaner concentrate is re-cleaned by 3 parallel flotation columns (3.2m in diameter and 12m in height), when the desired concentrate grade is not reached by the cleaner columns alone. This research work deals with simulation of the cleaner column flotation circuit at Miduk copper concentrator using USIM PAC simulator with the aim of improving the process metallurgical performance. For that purpose, the parameters of the models including flotation rate constants (kf, ks), residence time distribution (RTD), gas hold-up in collection zone (εg), mean bubble size (db), collection (Rc) and froth (Rf) zone recoveries along with some operating and geometrical variables were determined. The flotation rate constants were calculated by fitting the experimental data to the fast and slow floating components model. The residence time distribution of the flotation columns was measured by the tracer injection technique (using saturated NaCl solution as tracer). The gas hold-up and mean bubble size in the collection zone of the cleaner columns were estimated from the pressure difference and the drift flux techniques, respectively. The froth recovery was quantified by measuring the concentrate mass flow rate at different froth depths and extrapolation to the zero froth depth. The cleaner circuit was sampled five times, of which three times were used for calibration and two times for validation of the models. The mass flow rate, copper content and size distribution of the cleaner columns concentrate and tailings were accurately predicted using the simulation models. Increasing the number of operating cleaner columns improved the copper recovery (from 45.67% to 54.64%) at the expense of a reduction in the final concentrate (from 26.17% to 24.22%). The number of recleaner stages in all cases improved the the final concentrate (from 26.17% to 36.99% in the circuit with 2 cleaners and from 24.22% to 36.13% in the circuit with 3 cleaners). Increasing the feed slurry solids concentration reduced the size-by-size fractional copper recovery of the cleaner columns. Increasing the feed slurry solids concentration reduced the cleaner columns copper recovery in all size fractions. The best configuration of the cleaner and recleaner flotation columns was proposed.

Hydrocyclones are the most efficient used classifiers in the grinding circuits. Hydrocyclones are normally modeled and simulated using empirical models. These models can only be used within the range of the experimental data from which the model parameters have been derived. Computational fluid dynamics (CFD) is a powerful tool in simulating fluid flow in hydrocyclones. This research work deals with 3D simulation and modeling of fluid flow in a single phase hydrocyclone using CFD. The main simulation steps include preparing the geometry, meshing it, defining the properties of the materials involved, and setting the boundary layer and conditions. The experimenal data measured in a laboratory hydrocyclone were used for validation of the model. The simulation results indicated that the tangential velocity increased traversing towards the core, before decreasing at the interface with the air core. The liquid axial velocity inside the hydrocyclone varied from -1.59 m/s to 6.52 m/s. The axial velocity is a result of two swirling flows, the inner upward flowing inside the air core and the outer downward flowing near the cyclone wall. The liquid axial velocity inside the hydrocyclone varied from -5.58 m/s to 5.46 m/s. The LES model showed the least error on predicting the velocity profiles, the air core dimensions (7.8%), the pressure drop (7.52%) and the mass split ratio to overflow (0.18%). The effect of various geometric (spigot diameter, vortex diameter and cone angle) and process (feed flow rate) parameters on tangential velocity of the fluid was investigated.

The paper introduces the 3D Cross-Correlation for modeling of total magnetic intensity and its vertical gradient which is the fast way to model data, detect anomalies and estimate their depths and locations. In this approach first, we divide the subsurface space into a 3D regular grid, after computing the correlation value for each node of the grid, these values are plotted. It is noted that the results fall in the range [-1, +1] that represents the mass excess or mass deficit of magnetization (or susceptibility) relative to the magnetization (or susceptibility) of the host volume. This approach is applied to 2 synthetic models. The results show acceptable accuracy of this method in depth estimation and expansion of buried masses. After this method is verified and validated, it will be applied to the Shavvaz mine's total magnetic intensity (TMI) data of Yazd and its vertical gradient, and the results will be discussed.

There are two major approaches for 3D inversion of magnetic data: (i) direct inversion of the density contrast using a linear or nonlinear algorithm, and (ii) modeling of the source distribution in a purely probabilistic sense, in which the results are equivalent physical parameters between +1 and -1. Direct inversion has an inherent problem called Non-uniqueness of solutions and requires a lot of computer memory because of the number of model parameters and data.
In this paper, we introduce and evaluate the 3D Cross-Correlation (CC) method for 3D modeling of magnetic data (or its vertical gradient). This method was applied to 2 different synthetic models and its strengths in modeling of total magnetic field anomaly and vertical magnetic gradient data are discussed and finally, the method was applied to the Shavvaz mine's TMI data and its vertical gradient.
 

Methodology and Approaches:

The 3D cross-correlation approach is a method for modeling the magnetic data (or its vertical gradient) without any external constraints and any linearization.

The results of synthetic examples showed the high accuracy of the CC method in determining the shape and depth of the buried mass. This method is simple and easy to run and there is no need for prior information. In the end, this method was applied to the Shavvaz mine's TMI data and its vertical gradient. These results showed that orebody continues deeper than 40m (estimated from the previous modelling).

In this study, the modeling of vanadium, nickel and copper bioleaching from fuel oil ash ash using artificial neural networks was investigated. In the obtained models, the extraction percentage of metals was investigated as a function of factors such as initial pH (from 1-2.5), initial Fe2+ concentration (from 0 – 9 g/l), initial bacterial inoculation (from 1 – 10%) and process time (from 0-15 day). Three neural network models were presented to estimate the extraction percentage of metals. The propagation error method and Levenberg–Marquardt algorithm were used for training. Furthermore, trial and error method was used to determine the optimal number of neurons. One quarter of the data were used to evaluate the model and were not used for training process. The Mean Relative Errors (MRE) were obtained 5.35%, 3.07% and 2.82% for V, Ni and Cu, respectively. Also the higher 0.99 of R2 indicates the validity of the obtained models.

In the conventional mass balance methods, sampling are carried out to determine the metal content, solid percentage and particle size distribution in the paths (branches) and ultimately the corresponding errors are dispersed. In these methods, there is no information about the amount of feed and the grade (or the metal content) at any moment of time for each mineral processing unit (plant equipment). Thus, in this paper, the quantitative and qualitative values of input feed to each of the operating units in the processing paths were probabilistically investigated using stochastic process theory, and consequently the load distribution in these units was estimated. For this purpose, the processing plant was firstly modeled as a directional and balanced graph. In this graph, each node represents an operating unit and each edge displays a path between two units. The weight of each edge also indicates the passing material weight from the node to another node, which is assigned as a random variable at a specified interval. After modeling the process, the obtained graphical model containing random parameters was analyzed using Markov chain theory. Applying the philosophy outlined in this paper, not only the load distribution, but also the presence probability of each particle or mineral in each of the operational units of the processing plant can be predicted. It is also possible to estimate the tonnage in each unit when the plant is shut down and even the necessary steps are provided to discharge them. In this research, Chadormalu iron ore processing line was modeled using method presented and the results were analyzed and thereafter were reported using probabilistic parameters.

Nowadays, artificial intelligence methods have been broadly developed and applied for variable estimation to facilitate decision making in many fields. Grade estimation is an important issue in evaluating mineral deposits. Geostatistical methods are among the most commonly used approaches for variable estimation. Since these methods are somewhat defective in relation to limited numbers of dispersed nonlinear data, in this study, the support vector regression, a machine learning method, has been used for grade estimation in Esfordi phosphate deposit. The modeling accuracy was 84% according to the test data. Based on the results obtained from the modeling using the support vector regression method, grade estimation has been made within the block model in Esfordi phosphate deposit. The proposed potential areas in the block model can be taken as the the additional borehole sites in the further exploration stage. The tonnage-grade model was also prepared based on the results obtained by using the support vector regression modeling procedure. For example, based on this model, for a 6% cutoff grade, the reserve is about 15.36 million tons with an average grade of 13.59%.

The numerical study of airflow has mainly concentrated on underground mines and road tunnels. Ventilation during the construction of long tunnels, especially the ventilation of TBMs, has received less attention. This paper aims to study airflow pattern in TBMs with regard to safety and as such reduce dead zones area. The results show that the airflow pattern of a TBM ventilation system, which can help to design an effective ventilation system for the TBMs.

The Tunnel Boring Machines (TBMs) has revolutionized the tunneling industry to create underground space safer, healthier and more economical. Ventilation is one of the main components of mechanized tunneling. The airflow quality and the related mass flow rate in the ventilation system should be sufficient to dilute gases and remove the dust inside the tunnel. Since most of TBM crews stay and work in the mainframe area, ventilation of this zone is very important.

In this study, we modeled the TBM ventilation system using CFD method to understand airflow behavior in TBM. Numerical solution of the governing equations and boundary conditions are performed by utilizing the commercial CFD code Ansys CFX 18.1. Tests of mesh-independence were conducted based on four different meshing creations. To define the boundary conditions, airflow velocity sampling was performed using multi-point sampling method in the ductwork outlet. To complete the discretization of the advection term, the high-resolution scheme was computed. Root Means Square (RMS) < was considered as the convergence criterion of mass and momentum equations.

The results show that there is not enough air flow in 89.2% of TBM space in its current state. There are many dead zones from control cabin to the end of mainframe. The main direction of the back airflow moves along the segment feeder and increasing air mass flow has no effect in decreasing dead zones area. The results from the study present show that by increasing the air mass flow rate by 60% the volume of the dead zones in TBM is decreased by 13.42% and has no effective decrease dead zone in the personnel breathing zone. The results from the present study clearly indicate that maximum mass flow capacity of jet fans is not possible to reduce dead zones that will only increase energy costs