Applications of Artificial Intelligence in Mining and Geotechnical Engineering

By Hoang Nguyen and Jian Zhou

Applications of Artificial Intelligence in Mining, Geotechnical and Geoengineering provides recent advances in mining, geotechnical and geoengineering, as well as applications of artificial intelligence in these areas. It serves as the first book on applications of artificial intelligence in mining, geotechnical and geoengineering, providing an opportunity for researchers, scholars, engineers, practitioners and data scientists from all over the world to understand current developments and applications. Topics covered include slopes, open-pit mines, quarries, shafts, tunnels, caverns, underground mines, metro systems, dams and hydro-electric stations, geothermal energy, petroleum engineering, and radioactive waste disposal.

In the geotechnical and geoengineering aspects, topics of specific interest include, but are not limited to, foundation, dam, tunneling, geohazard, geoenvironmental and petroleum engineering, rock mechanics, geotechnical engineering, soil mechanics and foundation engineering, civil engineering, hydraulic engineering, petroleum engineering, engineering geology, etc.

• Guides readers through the process of gathering, processing, and analyzing datasets specifically tailored for mining, geotechnical, and engineering challenges.
• Examines the evolution and practical implementation of artificial intelligence models in predicting, forecasting, and optimizing solutions for mining, geotechnical, and engineering problems.
• Offers cutting-edge methodologies to address the most demanding and complex issues encountered in the fields of mining, geotechnical studies, and engineering.

• Language: English
• Publisher: Elsevier Science
• Release date: Nov 20, 2023
• ISBN: 9780443187650

Book preview

Preface

It is a privilege to introduce Applications of Artificial Intelligence in Mining and Geotechnical Engineering, a collection of papers that presents cutting-edge artificial intelligence (AI) advancements across both mining and geotechnical engineering domains.

This contemporary research book offers key insights into developmental and practical aspects of AI for both academic and practicing resource engineers, together with data scientists worldwide.

Focal points include the impact of AI techniques on open-pit and underground mining with respect to modeling and optimization in mining backfill design, problems related to blasting, slope stability, truck-haulage systems, mining supply chains, mine planning, mining capital cost, commodity prices, and environment-related applications. The positive effects of AI in geotechnical engineering are also scrutinized, with contemporary analysis of their use in rock mechanics, geotechnical engineering, tunneling, geohazards, soil mechanics, civil engineering, hydraulic engineering, as well as engineering geology research.

As designed, this unique volume introduces the innovative techniques required to address the increasingly multifaceted issues encountered across both mining and geotechnical engineering fields. Accordingly, its contributing authors and editors offer it as a key resource for both new and established professionals seeking to either reach or ideally remain at the forefront of this rapidly evolving software and hardware arena.

Key Features

•Guides readers through the process of gathering, processing, and analyzing datasets specifically tailored for mining, geotechnical, and engineering challenges

•Examines the evolution and practical implementation of artificial intelligence models in predicting, forecasting, and optimizing solutions for mining, geotechnical, and engineering problems

•Offers cutting-edge methodologies to address the most demanding and complex issues encountered in the fields of mining, geotechnical studies, and engineering. The book is edited by internationally reputed scientists: Dr. Hoang Nguyen and Prof. Xuan-Nam Bui (Hanoi University of Mining and Geology, Vietnam), Prof. Erkan Topal (Curtin University, Australia), Assoc. Prof. Jian Zhou (Central South University, China), Prof. Yosoon Choi (Pukyong National University, Korea), and Prof. Wengang Zhang (Chongqing University, China). We hope that the book will be interesting for all readers

Thank you very much.

Hoang Nguyen

Xuan-Nam Bui

Erkan Topal

Jian Zhou

Yosoon Choi

Wengang Zhang

Chapter 1 The role of artificial intelligence in smart mining

Yosoon Choia; Hoang Nguyenb,c    a Department of Energy Resources Engineering, Pukyong National University, Busan, South Korea

b Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology, Hanoi, Vietnam

c Innovations for Sustainable and Responsible Mining (ISRM) Research Group, Hanoi University of Mining and Geology, Hanoi, Vietnam

Abstract

The fourth industrial revolution of manufacturing innovation, which started with Industry 4.0 in Germany in 2011, has developed into the concept of an intelligent society where everything is connected and is spreading to all industrial fields, including artificial intelligence, Internet of Things, cloud computing, big data analysis, smart/wearable devices, and virtual/augmented/mixed reality. As cutting-edge technologies of Industry 4.0, represented by 3D printing, drones, self-driving, and robotics, converge with domain knowledge in each field, innovative changes to improve productivity, safety, and profitability are appearing at various industrial sites. In this chapter, the concept of smart mining technology and its levels of site implementation, along with the role of artificial intelligence technology in top-level site implementation, is explained.

Keywords

Smart mining; Artificial intelligence; Industry 4.0

Acknowledgments

This work was supported by the KETEP grant funded by the Korea Government’s Ministry of Trade, Industry, and Energy (project no. 20227A10100020).

1 Industry 4.0 and smart mining

The paradigm shift occurring because of Industry 4.0 has led to the development of a new technology called smart mining in the mining industry. Smart mining technology improves productivity and safety through the digitization, intelligence, and automation of mining sites and mineral processing factories so that minerals required by the market can be mined, processed economically and safely, and delivered promptly. The scope of application of smart mining technology includes all processes, from ordering products from consuming companies to mining minerals in mines (ore production), processing minerals in factories, and product shipment. Smart mining improves productivity and safety through monitoring, analysis, prediction, diagnosis, optimization, and automation by connecting assets, processes, and people at mining sites based on advanced technologies of Industry 4.0. To realize smart mining sites, cutting-edge information and communication technologies, such as the Internet of Things (IoT), big data, mobile devices, artificial intelligence (AI), virtual/augmented/mixed reality (VR/AR/MR), and robotics, are being introduced to mining sites.

Building a smart mining site has the following requirements, as shown in Fig. 1: (1) technology that links space/state information of a physical mine site, including 3D modeling, smart sensors, and IoT, to a virtual mine model (Physical to Virtual, P2V); (2) intelligence technologies such as AI, big data analysis, and cloud computing to perform analysis, prediction, diagnosis, and optimization in virtual mine models; (3) automation technology, such as drones, autonomous driving, and collaborative robots for control (Virtual to Physical, V2P); (4) technology, such as mobile and wearable devices and VR/AR/MR, that links physical mine sites and virtual mine models centered on people should be combined.

Fig. 1

Fig. 1 Elemental technologies for building smart mining sites.

2 Implementation levels of a smart mining site

The implementation level of a smart mining site can be divided into three stages. Level 1 involves constructing a digital spatial database of a mine site, inputting and visualizing attribute information, and performing preliminary simulations by changing the attribute information. 3D geology/mineral modeling technology and underground space surveying technology are used to construct a spatial database of a mine site (Fig. 2). Currently, unmanned aerial systems, such as drones, are used to build high-resolution spatial databases of mine sites.

Fig. 2

Fig. 2 Smart mining digital twin space database construction using 3D geology/mineral modeling technology and underground space surveying technology.

Level 2 involves the one-to-one matching of the physical mine in the real world with the mine model in the virtual world, in addition to real-time monitoring. As shown in Fig. 3, a mine safety system corresponding to Level 2 of smart mining is developed and supplied to the field. In addition, recently, a technology that can implement smart mining digital twins at a low cost has been developed at a small-scale mining site using a low-power Bluetooth beacon and a smartphone with short-range communication technology (Fig. 4).

Fig. 3

Fig. 3 Example of an ICT-based mine safety management system corresponding to smart mining digital twin [1].

Fig. 4

Fig. 4 Development case of underground mine navigation and production management system using low-power Bluetooth beacon and smartphone [2].

Level 3 of smart mining site implementation uses big data collected from physical mines to perform analysis/prediction/simulation in virtual mines and then optimizes site operation methods based on the results to apply it to real objects in physical mines. To this end, technology is required to reflect the optimization results of virtual mines in actual physical mines. Therefore, as shown in Fig. 5, self-driving robots driven according to the optimization results can be used to perform exploration, transportation, and environmental/safety management of smart mine sites. In addition, as shown in Fig. 6, VR/AR/MR technology and wearable devices can be used to effectively deliver the optimal simulation results of the digital twin to workers in physical mines.

Fig. 5

Fig. 5 Development case of a small self-driving robot for underground mine tunnel mapping [3].

Fig. 6

Fig. 6 VR and AR technology application of smart mining digital twin using a wearable device [4, 5].

3 Role of artificial intelligence in smart mining

To implement Level 3, the top level in the smart mining site, the roles of modeling and simulation (M&S) and AI are essential. M&S optimizes the performance of objects by analyzing/diagnosing the state of physical objects using data collected from the real world and predicting via simulations the conditions under which it is more efficient to operate or the circumstances under which problems may occur. As it is essential to analyze big data collected from physical mines, if M&S is performed for virtual mines, the working conditions of the mines over time can be predicted, along with the key indicators related to productivity and profitability. These prediction results are used as important data for determining the optimal mine operation method.

Recently, attempts have been made to combine the M&S technique, a system science approach, with the AI technique, a data science approach. M&S techniques based on physical theories or operating rules must secure detailed information and knowledge regarding the characteristics of the target object for model development. However, the verification of the reliability of the developed model is essential. AI techniques, such as databased machine learning, require a considerable amount of data on the target object, making it difficult to analyze causal relationships to identify the cause of the problem. Therefore, they cannot be applied to special situations or non-existent systems that have not been learned. Owing to the convergence of AI and M&S, research is actively conducted to address the limitations of each technique.

4 Future perspectives

Recently, smart mining technology has attracted increasing attention worldwide owing to its potential in realizing eco-friendly, highly efficient, low-cost, and accident-free mining sites. Zion Market Research [6] predicted that the smart mining market would expand from $8.6 billion in 2018 to $22.2 billion in 2025, owing to the digital transformation of the mineral industry, with an average annual growth rate of 14.5%. With the emerging necessity to create a new type of business through the convergence of advanced technologies in the mineral industry, the need for developing smart mining technologies is increasing. With the development of smart mining technology, large amounts of data are produced, collected, and shared in real time at mining sites. Accordingly, artificial intelligence technologies, such as machine learning, which can effectively analyze big data at mining sites, are attracting increasing attention. Therefore, the importance of artificial intelligence technologies will increase in the future mining industry.

References

[1] Baek J., Choi Y. Deep neural network for predicting ore production by truck-haulage systems in open-pit mines. Appl. Sci. 2020;10(5):1657.

[2] Baek J., Choi Y., Lee C., Suh J., Lee S. BBUNS: Bluetooth beacon-based underground navigation system to support mine haulage operations. Fortschr. Mineral. 2017;7(11):228.

[3] Kim H., Choi Y. Comparison of three location estimation methods of an autonomous driving robot for underground mines. Appl. Sci. 2020;10(14):4831.

[4] Kim H., Choi Y. Performance comparison of user interface devices for controlling mining software in virtual reality environments. Appl. Sci. 2019;9(13):2584.

[5] Baek J., Choi Y. Smart glasses-based personnel proximity warning system for improving pedestrian safety in construction and mining sites. Int. J. Environ. Res. Public Health. 2020;17(4):1422.

[6] Zion Market Research. Global Smart Mining Market Is Expected To Reach Around USD 22.19 Billion By 2025.https://www.zionmarketresearch.com/news/smart-mining-market. 2019.

Chapter 2 Application of artificial neural networks and UAV-based air quality monitoring sensors for simulating dust emission in quarries

Long Quoc Nguyena,b; Luyen K. Buia,c; Cuong Xuan Caoa; Xuan-Nam Buib,d; Hoang Nguyenb,d; Van-Duc Nguyenb,e; Chang Woo Leee; Dieu Tien Buif    a Faculty of Geomatics and Land Administration, Hanoi University of Mining and Geology, Hanoi, Vietnam

b Innovations for Sustainable and Responsible Mining (ISRM) Research Group, Hanoi University of Mining and Geology, Hanoi, Vietnam

c Geodesy and Environment Research Group, Hanoi University of Mining and Geology, Hanoi, Vietnam

d Department of Surface Mining, Mining Faculty, Hanoi University of Mining and Geology, Hanoi, Vietnam

e Department of Energy and Mineral Resources, College of Engineering, Dong-A University, Busan, Republic of Korea

f GIS Group, Department of Business and IT, University of South-Eastern Norway, Bø i Telemark, Norway

Abstract

This study aims to assess the potential application of low-cost unmanned aerial vehicles (UAVs) for environmental monitoring and assessment at the open-pit mine, with a case study at the Tan My and Thuong Tan quarries in southern Vietnam. Two research objectives were focused on (1) monitoring the dust at the quarries using the UMS-AM and (2) predicting the density of dust components using a multilayer perceptron neural network (MLP neural nets). For this purpose, the DJ Inspire 2 equipped with the Zenmuse X4S camera, and dust sensors were used. A total of three indicators, PM1.0, PM2.5, and PM10 were measured and monitored. In addition, MLP neural nets were employed to predict the density of air components within the quarry when it is expanded in the future with larger production and a more immense ending depth. Finally, 3D models for PM1.0, PM2.5, and PM10 were generated and illustrated. The result demonstrated that dust monitoring at quarries with low-cost UAVs is viable and should be considered for other open-pit mines.

Keywords

Air quality monitoring; MLP neural nets; Open-pit mines; Sensor networks; UAV

Conflicts of Interest

The authors declare no conflicts of interest.

1 Introduction

1.1 Motivations

As one of Vietnam’s most important economic activities, the mining industry has been developing at a high pace. Stone, limestone, and rock are the most common construction materials in high demand due to rapid urbanization. This results in a significant increase in the number of quarries in many regions. To meet the demand for construction, the production of quarries has been supported by the continuous application of advanced technologies. It is essential to continuously enhance the technology of all producing states to achieve all goals of mining production. Mine surveying and environmental management are among the essential activities that have received significant attention from mining managers and scientists.

The mining industry has made a significant contribution to the Vietnam economy. However, this industry also inevitably leads to many environmental problems. In Vietnam, larger-scale open-pit mines are mainly located near populated areas. For example, Tan My and Thuong Tan are among the largest stone quarries in Vietnam, just several kilometers from Tan Uyen town, Binh Duong province. While blasting is an integral part of open-pit mining, it usually causes the emission of particulate materials, for example, dust pollution and gasses potentially hazardous to health [1]. Blasting usually results in airblasts, ground vibration, fly rock, toxic fumes, and particulate matter (PM). All blasting events in mining areas emit the primary residue of PM [1], for example, PM1.0, PM2.5, and PM10. To properly manage air quality in mining areas, it is crucial to establish an effective air quality monitoring system. To meet the high demand for construction materials, mine extension is inevitable. However, the expansion of a quarry is often considered with its potential environmental impacts. An air quality monitoring system can provide important data for the environmental management of quarries not only at present but also in the future when these quarries are expanded.

1.2 Related works

Recently, the rapid development in the unmanned aerial vehicle (UAV) technology has brought many benefits to a wide range of military and civil fields, such as logistics and transportation [2,3], precise agriculture [4,5], forest management and biodiversity conservation [6], hazardous and environmental management [7], and urban management [8–10]. The preliminary successes of UAV applications have proved that this technology could be promising and likely to be employed in the broader field. While the world has witnessed many excellent examples of using UAV technology in the mining industry for topographical surveys, safety investigations, and other works [11], this technology is still relatively new to Vietnam [12]. For instance, UAV technology was used to conduct a topographic survey of slope areas on an open-pit mine [13].

Another UAV-based topographic survey of an ore stockpile was given by Cryderman et al. [14]. These authors used topographic data to estimate ore carrying capacity. Lee and Choi [15] have proved that fixed-wing and rotary-wing UAVs, the most popular ones, can be used effectively in both small-scale and large-scale open-pit mines as a topographic surveying tool [15]. For air quality monitoring in open-pit mine sites, results of the laboratory and field tests conducted by Alvarado et al. [1] demonstrated the feasibility of coupling an optoelectronic dust sensor with UAVs.

In recent years, UAV-based systems have been considered for monitoring dust pollution, including particulate matter PM1.0, PM2.5, and PM10 [16–18]. DJI Matrice 600 Pro with six propellers may be the most popular used due to the ability to carry a payload of up to 6 kg [19,20]. However, the price of more than 5000 USD and relatively large weight (>10 kg) when carrying is a hindrance when applying to open-pit mines. Thus, lighter weight UAV-based systems, that is, Inspire 2, should be investigated. In addition, predicting PM1.0, PM2.5, and PM10 play a vital role in the pollution assessment. The literature review shows that machine learning, that is, support vector machines [21], neural networks [22], and deep learning [22] are state-of-the-art methods used. Nevertheless, applications of machine learning for dust pollution prediction are still rare.

1.3 Contributions

As mentioned above, the application of UAVs has been proven as an alternative tool for air quality monitoring. In this study, a low-cost UAV-based system was employed at Tan My and Thuong Tan quarries in Binh Duong province, one of southern Vietnam’s largest groups of stone quarries. This low-cost UAV system, named as UMS-AM, is designed to collect a variety of data that can be used for optimizing mining operations and control the atmospheric environment. Herein, we focus on monitoring the atmospheric environment at the quarries using the UMS-AM; how the multilayer perceptron neural network (MLP neural nets) could be used to predict three dust pollution, PM1.0, PM2.5, and PM10, and finally, generating 3D model for PM1.0, PM2.5, and PM10.

2 Proposed UMS-AM system

2.1 UAV platform

UAVs are classified based on different but interrelated characteristics such as size and payload, wing types, flight endurance, flight range, altitude, and capabilities [23]. With the wing types, there are two main subtypes, namely, the rotary-wing and fixed-wing UAVs. The latter is suitable for applications with longer flight endurance, but large space is needed for take off and landing. Although the former uses batteries and has shorter flight times [6], it has been increasingly common because of its ability to take off and land vertically in a small space and to maintain position. Therefore, in this study, a rotary-wing UAV was considered a feasible platform for the system. Specifically, its characteristics are given in Table 1.

Table 1

Various UAVs are currently available for the mining industry; however, low price is still considered the main issue. Furthermore, the payload is also a critical issue for this study because several air quality sensors and accessories are mounted on the UAV. For this purpose, DJ Inspire 2 (Fig. 1) was considered to use.

Fig. 1

Fig. 1 DJI Inspire 2. https://www.dji.com

2.2 Sensor networks

Optical sensors (Inspire 2’s camera parameters)

The DJI Inspire 2 is equipped with Zenmuse X4S, a powerful camera featuring a 20-megapixel 1-in. sensor. Dynamic range is increased from the Zenmuse X3 by one stop, with the signal-to-noise ratio and color sensitivity increased by 1.5 stops for next-level image quality. The Zenmuse X4S uses a DJI-designed compact lens with low dispersion and low distortion 24 mm equivalent prime lens. This 84° FOV high-resolution lens make the Zenmuse X4S powerful during aerial imaging as it is on the ground. Combined with CineCore 2.0 and the Inspire 2’s powerful image processing system, the camera can record 4 K/60H.264 and 4 K/30H.265 videos at a 100 Mbps bitrate and an oversample 5.2 K video into 4 K video in real-time, capturing fine image details. Furthermore, in Burst Mode, the Zenmuse X4S supports 14 fps shooting at 20 megapixels in both JPEG and DNG formats, hence the balance between agility and image quality