Vol 275

In the context of the rapid development of digital technologies and increasing demands on energy efficiency, sustainability and competitiveness of industrial enterprises, the fuel and energy (FEC) and mineral resources (MR) complexes are undergoing major changes. Digital transformation is becoming a key factor in improving the efficiency, reliability, and sustainability of production processes, as well as an important element of the strategy for technological sovereignty and the modernization of production systems. Modern approaches to managing equipment and process chains are based on the use of machine learning methods, big data analysis, digital modeling, and the creation of digital twins, which, in turn, allows not only the optimization of technological and business processes, but also the formation of new control architectures from local systems to industrial metauniverses.

This study investigates the use of Singular value decomposition to decompose time series of electricity consumption from substation feeders. The goal is to identify and classify the electrical load patterns of mining enterprises. The need for continuous improvement in process efficiency is dictated by current trends and tendencies towards increased consumption of fossil fuels and energy resources. The proposed algorithm uses the decomposition results to identify similarities in consumption patterns, enabling the categorization of loads into broader groups. Based on the results of the analysis of electricity consumption data for two independent feeders, the formation of similar recurring characteristic load changes (temporal patterns) with a period of three days was identified. The results facilitate the automated typification and classification of load profiles. This is vital for integrating economic incentives into demand management and for assessing the feasibility and potential of consumer participation in load schedule regulation via demand side management technologies. The proposed algorithms enable the use of these typical consumption profiles to calculate quasi-dynamic electrical modes, supporting tasks related to the long-term development of energy supply systems and energy efficiency improvements for mining enterprises.

Controlling the ventilation in underground mining enterprises (UME), characterized by high inertia and numerous influencing external factors, based on real-time sensor data located in mine workings and on the surface, with a high level of accuracy in regulating air supply by the main ventilation unit (MVU), is feasible only under conditions of a pre-defined sequence of control actions. This task can be classified as an approximate dynamic programming (ADP) problem, which involves synthesizing a suboptimal control function for MVU operation in a predictive modeling mode of air distribution, given a known space of possible states and the selection of the optimal control strategy that meets a specified criterion. A simulation model of a digital twin subsystem for ventilation process control is presented, using the example of two types of UME (potash mines and oil shafts), which can be used to solve ADP tasks. For predictive modeling of air distribution and determining the energy efficiency criterion of the MVU, which consumes up to half of the total electricity of the UME, the digital twin is integrated with external data, based on which energy consumption is evaluated while maintaining the required volume of supplied air. This control approach enables not only safe and energy-efficient management of the ventilation process but also participation in the planning and implementation of measures for price-dependent electricity demand management.

Industrial metaverse is a new direction for development of industrial enterprises. Nowadays the process of full understanding term “industrial metaverse”, its conception, its effectiveness for enterprises is not actualy completed. The methodology, tools, and methods for building an industrial metaverse have not been clearly defined. Therefore, it is advisable to experimentally implement a part of the metaverse on one or several processes with future scaling to other processes. The process of industrial machinery repair and maintenance is proposed as an experimental zone for implementing the industrial metaverse. This process is well suited as an experimental zone. Launching the industrial metaverse concept on it will solve a number of problems, such as the diversity of equipment with unique diagnostic and repair methods, human errors made during repair work, etc. This article presents the concept of building and the architecture of industrial metaverse. A general description of the physical, cyber-physical, and social spaces and the interaction layer between them is provided, without any details of qualitative and quantitative indicators. The avatar of a service engineer is highlighted as one of the elements of the cyber-physical space. The process of creating an avatar of a cyber-physical service engineer is considered: a description of the main functionality is provided, it is shown that a combined system of wearable devices – a glove and a video camera integrated into glasses, a vest, a helmet, or represented by an independent device – is sufficient for its creation. Laboratory experiments were conducted, where the created avatar was tested to determine the task of servicing a centrifugal pump. The results of processing 518 experimental datasets of 10 points, each of which belongs to one of six classes corresponding to a specific technological operation during servicing of a centrifugal pump are shown. Three types of models were obtained (accuracy on training data 0.99; 1.0; 1.0, accuracy on test data 0.625; 0.7; 1.0). It has been shown that achieving 1.0 accuracy on training and test data requires first identifying features representing frequency and temporal characteristics obtained through time series processing. The obtained results allow to make conclusions about the readiness of these technologies for industrial implementation.

Centrifugal pumps, as key components of hydraulic systems, play a fundamental role in ensuring the reliable operation of numerous industrial processes in sectors such as energy, chemical production, and oil refining, where uninterrupted equipment performance is of critical importance. Failures of centrifugal pumps can result in substantial financial losses due to costly repairs and unplanned production downtime. This paper presents an innovative approach to diagnosing and detecting faults in centrifugal pumps. The method is based on the application of Motor Current Signature Analysis (MCSA) in combination with automated machine learning (AutoML) technologies. Such an approach enables efficient and highly accurate identification of early signs of equipment malfunction. The experimental study was conducted using an open dataset collected under conditions close to real-world operation. The proposed method achieved a fault detection accuracy of 89 %, which significantly exceeds the performance of the traditional gradient boosting method. This confirms the advantage of a comprehensive model developed through AutoML. Further improvement in diagnostic accuracy was made possible by applying an enhanced Park’s vector transformation to the raw current and voltage data. This approach makes it possible to detect even subtle anomalies in pump operation, thereby strengthening the capability for early fault prediction. The study not only highlights the potential of MCSA as a non-invasive and scalable tool for equipment condition monitoring but also demonstrates the promise of AutoML for technical diagnostics of industrial pumps.

The prospects for implementing battery-powered dump trucks at open-pit mining enterprises are considered. Main attention is on the problem of charging infrastructure for adopting the unmanned production concept. The suggestion is to use wireless charging stations to join charging with particular technological operations, thereby reducing battery capacity and increasing the utilization rate of electric vehicles. To determine effective charging infrastructure solutions, it is necessary to evaluate the interaction between the dump truck and charging stations. The research aims to develop a model reflecting the power flows between the charging infrastructure and the dump truck battery while an operational process is being executed. The model takes into account the work cycle parameters, dump truck parameters with powertrain options providing energy recovery at braking, and charging infrastructure parameters in three options: one stationary charging station located outside the operating routes designed for the simultaneous charging of several dump trucks (option A); stationary charging stations for one dump truck located at loading points (option B); and a dynamic charging station that charges in motion (option C). The method for determining the power of a single wireless charging station is proposed, along with a related method for determining battery capacity. When establishing capacity, the parameters of the charge-discharge cycle and the charging current ratio are considered. The described model is implemented in MATLAB Simulink using m-files for processing satellite data of route parameters from geographic information systems, as well as elements of the Stateflow and Simscape Electrical libraries. The capabilities of the model were demonstrated on the example of the Lebedinsky GOK, with the BelAZ-7558E selected as a battery-powered dump truck. In the example considered, the total capacity of the wireless charging infrastructure for options A, B, and C was 10.6; 6.3, and 13.5 MW, with option B providing the highest value of the battery average state of charge of 0.65 p.u., with the lowest specific power demand per dump truck of 2.4 MW·h. The simulation results allow us to determine various operating factors of the system, evaluate the power compliance of system elements, compare wireless charging infrastructure options, and make informed design decisions.

Water- and oil-based drilling fluids are polydisperse non-Newtonian systems, the stress state of which is adequately described by the Herschel – Bulkley rheological model. This study hypothesizes that cuttings transport efficiency can be improved by selecting the most effective combination of the three parameters of the rheological model – yield stress, consistency index, and flow behavior index – when designing drilling fluid properties. The effective parameter combination of the Herschel – Bulkley model for achieving a uniform velocity profile was determined using correlation and regression analysis methods as well as machine learning techniques. The computational part of the work was performed in the Wolfram Mathematica symbolic calculation package. Deterministic regions of the dependence of the velocity profile uniformity index on the rheological coefficients were identified. For practical engineering calculations, a linear mathematical model was constructed to represent the relationship between the modified excess coefficient and the parameters of the three-parameter Herschel – Bulkley rheological equation. The proposed methodology can be recommended for designing new drilling fluid systems and testing existing ones under given wellbore cleaning conditions.

We present an innovative approach to simulating flow tests of production wells operating in clastic reservoirs of oil fields in the Perm Region. To solve this issue, modern machine learning solutions (CatBoost, Random Forest, XGBoost, MLP, Gradient Boosting, etc.) were used, which allowed achieving high prediction accuracy. The main parameter for simulating and research is bottomhole pressure at various stages of its recovery during well flow testing. The use of the SHAP model interpretation method for the first time made it possible to assess the impact of geotechnical parameters on bottomhole pressure and identify key ones among them. Analysis of the bottomhole pressure recovery prediction model sensitivity to changes in initial parameters made it possible to evaluate the degree of their influence on the pressure recovery curves (PRC). The uniqueness of the proposed approach lies in studying the significance of parameters at various time stages of bottomhole pressure recovery during flow testing, which allows for a more detailed understanding of the processes occurring under formation conditions. The proposed algorithms made it possible to simulate PRC that are as close as possible to actual data, as well as to study the dynamics of permeability of the remote formation zone in real time. This approach opens up new horizons in simulating flow tests and allows for highly detailed and timely assessment of formation filtration properties across the entire production well stock simultaneously. The process engineering solution is aimed at promptly assessing filtration parameters of remote formation zones and provides the ability to monitor permeability changes, which helps to timely identify areas of reduced oil inflow and develop measures to restore well productivity. This approach significantly reduces economic risks associated with conducting expensive field tests while ensuring reliability and validity of predicted indicators with minimal resource and time costs.

The problem of timely and accurate evaluation of well integrity is becoming increasingly relevant in the context of mature field development, high wellstream water cut, and a growing number of old wells. For production casing diagnostics, geophysical methods are typically used to identify damage and determine its interval. However, high workload of field personnel hinders prompt deployment of wireline crews to survey the integrity of wells. This results in lost oil production, increased water cut, environmental risks, increased non-productive injected volumes, and reduced key economic indices. To address these challenges, a novel approach to evaluation of casing string integrity based on machine learning models has been proposed. The paper presents a procedure for application of interpretable machine learning to detect production casing leakage and provides a comparison of this approach with the ROC-AUC statistical analysis method. The novel approach integrates the LightGBM machine learning algorithm and SHAP analysis to evaluate contribution of key features to well integrity prediction and determine their threshold values. The model training was based on data from 14,318 well surveys conducted between 2000 and 2022. The results indicate that the most important features are sulfate content, solution supersaturation ratio, and water cut. The study confirms the efficiency of interpretable machine learning methods for diagnosing complex technical systems. These results show the potential for application of such models in well integrity monitoring and well workover planning. This approach can also be used in other oil and gas applications, such as prediction of various problems and optimization of well operation conditions.

One promising avenue for reducing CO₂ emissions is through the use of carbon capture, utilization, and storage (CCU|S) technologies, which necessitate capital-intensive capture stage implementation. This study proposes implementing a cluster-based approach to its organization, which enables cost reduction through economies of scale achieved by integrating stationary emission sources into a single network with a shared infrastructure. To evaluate the economic effects of this organizational framework, an optimization model was developed utilizing algorithms (SLSQP, Nelder – Mead method, etc.) that account for: spatial distribution of emission sources, emission volumes, CO₂ partial pressure in flue gas streams. The model was tested using data from 533 Russian industrial enterprises in the energy, cement, and ferrous metallurgy sectors, with aggregate annual emissions exceeding 0.5 billion tons of CO₂. For a preliminary analysis of the spatial and technological data of these enterprises, a methodical approach was developed (based on the DBSCAN algorithm), which made it possible to identify 94 geographical areas of their increased concentration. Information about industrial enterprises forming six largest regions was utilized for modeling 90 configurations of carbon capture and transportation projects with shared infrastructure. The results demonstrated that the cluster-based approach reduced the cost of capture in the considered examples by 6.44-13.51 %, depending on the maximum radius of a cluster. An additional reduction in transportation costs due to the use of joint gas pipelines averaged 37.26 and 57.01 % for a 200 and 500 km distances, respectively. Under the same distances and with a maximum cluster radius of no less than 20 km, the average reduction in aggregate costs across the evaluated configurations amounted to 17.81 %. The results obtained confirm the importance of organizational solutions for scaling up CCU|S projects and establishing novel cross-sectoral technological chains. The proposed methodologies can be effectively employed to identify promising areas for the implementation of CCU|S pilot projects and to design highly efficient local networks for CO₂ capture and transportation with shared infrastructure.

The paper is devoted to the conceptual formalization of the goals, objectives and criteria for managing the digital transformation of mining enterprises. The main idea is to define digital transformation as a process of minimizing or completely excluding human participation from the implementation of production processes in order to build an autonomously functioning cyber-physical industrial system. In order to determine the ways of building and functioning mechanisms of such systems, the necessity of forming an enterprise architecture model has been established, both ensuring the completeness and consistency of the knowledge embedded in it, and providing an instrumental numerical basis for self-organization and self-regulation of the system. The irrelevance of applying existing standards and frameworks for building architecture models is substantiated. Approaches to the construction of an ontological model and ways of its application in the management task of the digital transformation of mining enterprises are proposed. Under the results of the work, a basic ontological model of the mining enterprise architecture was formed based on the OWL descriptive logic language in the Protégé environment, applicable to geospatial natural and technical industrial systems with an open type of production environments. The resulting model was tested using the HermiT logical inference mechanism, confirmed its coherence and consistency, which indicates the potential correctness of the initial hypotheses of the study and the possibility of further research into the formation of a methodology for the digital transformation of mining enterprises.

This article examines the problem of managing ore flow quality at mining enterprises from the perspective of applying big data to improve the efficiency of mineral quality management. It is noted that assessing the feasibility of collecting and processing big data for ore flow quality control requires an optimal quantifiable weight parameter, which determines the data collection discreteness and the effectiveness of their processing. Currently, this parameter is the ore (or concentrate) batch. A scientific-practical approach to determining batch sizes at mining enterprises is proposed, based not on business process conditions, but on the analysis of the distribution of quality parameters within the ore body, considering subsequent methods of mineral raw material transportation. An analysis was conducted on the data from every technological process within the mining technical system, leading to the establishment of principles for calculating the minimum required data samples for each stage of the process. The applicability of the Kotelnikov theorem (Nyquist – Shannon sampling theorem) for determining the optimal quantifiable weight parameter of a mineral raw material batch within quality control frameworks is considered. To obtain a qualitative model, the required scope of quarry operation statistics should range from 16 to 52 months of excavator operation at the face. This range depends on the value of the mineral quality distribution coefficient at the mining enterprise. It was also established that for building a qualitative model, the mentioned coefficient must be considered; the higher its value, the lower the sampling frequency should be when collecting data from technological processing stages.

This article examines the experience of strategic cooperation between a mining software developer and a large mining company in adapting the Mining and Geological Information System (MGIS) to the company’s corporate requirements. The market-out of foreign MGIS from Russia placed large companies in a particularly difficult situation, as they had been building solutions based on imported software products for many years. The task of software import substitution in the mining industry, which deals with complex geotechnical systems, should be considered as a managed interdisciplinary scientific and engineering process requiring a systematic methodological approach. We note the importance of assessing the level of digitalization of existing business processes for engineering support of mining operations in forming a rational plan for software adaptation and modification. Given the requirement for a quick solution to the import substitution issue, we must consider the internal development of MGIS when coordinating with the industrial partner a work plan for functionality modification. This ensures the development of a competitive digital system for engineering support of mining operations not only for the company but for the entire mining industry. We present the main directions for modifying the MGIS functionality in the fields of geology, mine surveying, and geotechnology, along with examples of developed digital tools. We note that experts have mostly resolved the tasks of developing MGIS to meet the requirements of PJSC ALROSA, and the priority has become the development of software tools for medium-term and short-term planning of open-pit and underground mining operations. We provide a functional diagram of the planning unit. For the development of MGIS, we consider building the Mining Geological Digital Platform (MGDP). This platform provides the ability to create working tools (units) through the use of API functions and dynamic attachment of units to the MGDP system core.

The purpose of the work is to review and consider alternative frameworks for object position determination, including for solving dispatching and navigation tasks in technological areas for the operation of highly automated autonomous vehicles without the use of satellite navigation equipment. The main problems associated with the use of satellite navigation equipment for the positioning of vehicles equipped with an automated driving system, as well as loading equipment interacting with them, are considered. The promise and relevance of developing alternative systems and methods for positioning the automated transport component during open-pit mining are shown. The review of technologies is presented, confirming the concept of current research direction related to the digital transformation of the mining industry, ensuring the positioning and position determination of mining equipment at mining enterprises without the use of satellite navigation means. An analysis of existing solutions, their advantages and disadvantages, is carried out. It is proposed to implement the solution to the problem based on machine vision algorithms, the radio direction-finding method, and laser range finding means. Options for the interaction of auxiliary and correcting devices in solving the problems of object orientation in a local coordinate system are provided. The results of field and laboratory studies of radio direction-finding and machine vision methods are presented. A patented, detailed algorithm for determining object coordinates in a designated area, developed by the authors, is described; based on this algorithm, a method for determining the position of loading equipment when interacting with transport vehicles equipped with an automated driving system without the use of global navigation satellite systems is proposed.

This study demonstrates various artificial intelligence methods to predict and optimise blast-induced rock fragmentation at Orapa Diamond Mine in Botswana. These techniques include an artificial neural network (ANN), an adaptive neuro-fuzzy inference system (ANFIS), a genetic algorithm with ANN (GA-ANN), and particle swarm optimization with ANN (PSO-ANN). A collection of data from 120 blasting events with nine input parameters at the mine was utilized for this task. The results indicate that the PSO-ANN model outperforms other models in predicting blast-induced fragmentation. We used the optimal PSO-ANN model to optimise fragmentation, identified using the Monte Carlo method. The optimal model consists of nine inputs, two hidden layers with 65 and 30 neurons, and one output (7-65-30-1). Using gradient descent, we navigated this ten-dimensional solution space to determine the optimised blast design parameters and achieved an optimal fragmentation value of approximately 86 %. Sensitivity analysis results reveal that the most influential input parameters on fragmentation are rock factor (15.3 %), blastability index (14.7 %), and spacing-to-burden ratio (14.7 %). In contrast, the stiffness ratio has the least influence on fragmentation (6.3 %).

The paper considers the features of simulating the blasted rock muckpile formation. We describe various applied approaches and algorithms, as well as discuss the further development of national digital technologies in the mining industry. The study addresses key challenges in simulating explosive impact on rock mass. Due to the significant complexity of mathematical description of rock mass and explosive destruction processes, simulation requires various assumptions that inevitably affect its quality in terms of correspondence to real-world processes. The research compares two approaches to rock fragment dispersion: classical solution based on Newton’s laws and alternative approach assuming that the blasted rock moves as a single indivisible volume at the initial moment of time and fractures only upon contact with the surface. The study demonstrates that, given identical explosive impact and different rock mass representations (2D model with pieces of different sizes and densities), the resulting muckpiles differ significantly. The closest in shape muckpiles for both computation methods are obtained for rock mass simulated with 50 and 100 mm fragments. The obtained results suggest that under certain conditions, it is feasible to use a simplified (alternative) method for simulating the muckpile formation. This approach involves treating the rock movement after explosive impact as a single piece with subsequent fragmentation upon landing.