The Department of Mining Engineering in the University of Pretoria’s Faculty of Engineering, Built Environment and Information Technology concluded a successful research year with the presentation of its annual research colloquium on 18 November 2022. The Department, which strives to be a leading research-intensive department of mining engineering, is particularly noted for its strength in rock engineering research.

The purpose of the Department’s annual research colloquium is to showcase the work of its full-time postgraduate students. Welcoming the delegates, Prof Ronny Webber-Youngman, Head of Department, remarked that the dynamic nature of our changing society requires continuous assessments of the status quo, particularly in the volatile, uncertain, complex and ambiguous world of mining. Research plays a significant role in the development of new knowledge. He congratulated Prof Francois Malan, the Department’s Function Head: Research, for his passion and enthusiasm in driving the Department’s research agenda.

Prof Malan introduced the students’ research by providing an overview of research on pillar strength. The subsequent research presentations covered the Department’s four research focus areas: mechanisation and automation, management and leadership, rock engineering and extended reality technology.

The following postgraduate research topics were presented at the colloquium:

• A numerical modelling study of the effect of pillar shape on pillar strength (Jannie Maritz)

Traditionally, mining layouts are typically designed using square pillars to provide local and regional stability. Owing to the poor quality of pillar cutting during the extraction of the ore, pillars are often under size and of an irregular shape, which may compromise safe working environments. Numerical analysis, using a limit equilibrium model in the TEXAN code, was used to study the effect pillar shape can have on the expected pillar strength. This was compared to the theoretical perimeter rule generally applied to pillars that are not square. By keeping the effective width constant for various shapes of pillars, variations in pillar strengths are observed. The modelling study was used to confirm the limited increase in strength predicted for rectangular pillars when using the perimeter rule. The results prove that an increase in strength does not correspond linearly to the rate of increase of the cross-sectional surface area. It can be concluded that, for the application of long dip or strike stabilising pillars, the perimeter rule may not apply, and other factors may determine the stability of the pillar.

• A limit equilibrium model to simulate time-dependent pillar scaling in hard-rock bord-and-pillar mines (Danél Wessels)

This presentation describes the use of a limit equilibrium model to simulate the time-dependent scaling of hard-rock pillars. In the manganese bord-and-pillar mines, extensive scaling is observed for pillars characterised by a high joint density. It appears that the scaling occurs in a time-dependent fashion. Monitoring of selected pillars was conducted in an attempt to quantify the rate of time-dependent scaling. The scaling distance for pillars of different ages could be measured and it seems as if most of the scaling occurred soon after the pillars were formed. Numerical modelling of the time-dependent scaling was conducted using a displacement discontinuity code and a limit equilibrium constitutive model. The postulated exponential decay of the failed rock mass strength at the edges of the pillars resulted in simulated behaviour that is qualitatively similar to the underground observations. This method can be used to investigate the long-term stability of bord-and-pillar excavations.

• A review of the UG2 pillar strength in the Eastern Bushveld (Thomas Oates)

The use of the Hedley and Grant pillar strength formula with a modified K value is common practice for shallow South African hard-rock mining. The Hedley and Grant formula was empirically derived from a database of 28 rib pillars in the uranium mines in the Lake Elliot region of Canada. The use of this empirically derived pillar strength formula in the design of UG2 pillars in the Bushveld Complex is questionable. A recent pillar extraction project at Booysendal North Mine presented the unique opportunity for the compilation of a pillar database for UG2 pillars in the Bushveld Complex. The database of 66 UG2 pillars (seven failed and 59 stable) is the first of its kind. Mining layouts were simulated using the TEXAN displacement discontinuity code to determine the average pillar stress values of the pillars in the database. Using statistical methods, a back analysis of the pillars was done to determine an empirical strength formula for UG2 pillars.

• A review of the subsidence associated with shallow hard-rock mines (Johann Esterhuyse)

Almost no information on the time-dependent failure of hard-rock pillars and the resulting subsidence at South African metalliferous mines is readily available. There are almost no case studies on the collapse of bord-and-pillar workings in South Africa as the pillars are mostly designed conservatively. In comparison, there is ample information, as well as techniques, on subsidence prediction for coal mines. A study was therefore conducted to collect all the available data on subsidence problems at shallow metalliferous mines in South Africa. The findings are given in this presentation. All the collapses and subsequent subsidence events occurred for workings at depths less than 250 m. In contrast to continuous subsidence at coal mines, discontinuous subsidence occurs at metalliferous mines, and large geological structures are typically involved in the mechanism of these collapses.

• Calibration of a limit equilibrium model for numerical modelling using laboratory experiments (Ruan Els)

The limit equilibrium model in boundary element codes has been used by many workers to simulate the behaviour of pillars in underground workings. Good results have been obtained, but the drawback of the model is that many model parameters need to be calibrated. This is difficult to do using underground observations and for this study, physical modelling in the laboratory was done to verify and calibrate the model. For the experiments, an artificial grout material was used and cubes were poured using the standard civil engineering moulds. The friction angle between the artificial “pillar” and the platens of the testing machine was varied by using soap and sandpaper. Different modes of failure were observed depending on the friction angle.

• An investigation into backfill confinements to reinforce pillars in bord-and-pillar layouts (Divine Ile)

The Bushveld Complex is the largest layered igneous complex. It includes approximately 85% of the world’s known platinum group metal (PGM) resources. In these hard-rock operations, about 25% of the reef needs to be left behind as pillars to provide stable excavations. The placement of tailings underground as backfill presents an opportunity to mine the historical pillars and perhaps even use smaller pillars, thus making it possible for the operations to achieve higher extraction ratios. This study explored the effect of backfill on pillar strength, and how this may affect the layouts of bord-and-pillar mining operations in future. The limit equilibrium model in TEXAN gives the user the ability to apply a confinement on the failed pillars. This feature was used for this study.

• Investigation a suitable modelling philosophy for the coal-mining industry (Mesuli Mahlangu)

Several numerical methods are available for evaluating the rock mass response to mining. The development of the displacement discontinuity boundary element method in the coal industry was motivated by the need for a realistic prediction of surface subsidence. It is therefore preferred in collieries. In this method, the seam is treated mathematically as a discontinuity in the host rock mass. The method has generally been implemented as either of the following two models for the surrounding rock mass: the homogeneous isotropic model or a frictionless laminated model. Currently, the South African coal-mining industry predominantly implements the latter. However, it has been found that this model can only be calibrated for realistic pillar stress simulations or for realistic subsidence simulations. The objective of this study was to simulate hypothetical layouts to investigate the advantages and disadvantages of both approaches. Guidelines are provided under which circumstances a homogeneous isotropic model may be as good as the frictionless laminated model.

• The development of an innovative personnel tracking and locator system for underground mines (Larrance Ngwenyama)

The mining industry is still in need of a reliable missing person locator system. This study proposes the integration of a Through-the-Air (TTA) tagging-and-tracking mesh network system and a Through-the-Earth (TTE) communication system to function as a single system. The aim of this study is to develop a proof of concept of the proposed integrated personnel tracking and locator system. The proof of concept is based on the modulation or shift-keying of the low-frequency signals of the TTE communication system to the high-frequency signals of the TTA tagging-and-tracking system, and vice versa. The high-frequency tracking data from the TTA communication system can then be converted to low-frequency signals using a carrier to enable a TTE propagation of the data, between the surface and underground, wirelessly. Furthermore, this study is aimed at developing a prototype of the proof of concept and conducting field tests and trials.

• Developing a tool to mathematically quantify the trajectory of flyrock (Jennifer van der Walt)

Flyrock remains a significant risk to the health and safety of mine employees and infrastructure, as well as to the safety of the neighbouring communities and their property. This research study was initiated after determining that there is a fundamental lack of knowledge on the relationship between blast design parameters and the risk of flyrock. The objective is to develop a technique that can mathematically quantify the trajectory of flyrock, which can be used for future research in this field. The outcome for this technique is envisioned to do the following: Enable mining operations to create an accurate historical flyrock database from their blasting operations; allow research teams to conduct scientific investigations into the impact of various blast design parameters on flyrock risk; and enable the training and education sector to visualise real blasts and flyrock in virtual reality. Further development of this technique is ongoing. It is envisioned that this technique will provide a method whereby future flyrock studies will be comparable and where assumptions will be limited.

• Identifying key requirements for the implementation of blockchain solutions in the mining industry (Keaton Philo)

The value of digital information is ever increasing as more companies utilise digital technologies such as artificial intelligence (AI) and the Internet of Things (IoT) to gain deeper insight into their business operations and drive productivity gains. It is therefore important to safeguard and ensure the integrity of digital information exchange. Blockchain technology was identified as potentially providing the mining industry with a trusted system for securely exchanging digital value. However, there is little evidence or understanding of how blockchain can be implemented and what benefits the industry could obtain. This research provides a fundamental understanding of what the technology is in order to identify the associated capabilities and potential application benefits. The technology’s capabilities are used to see how blockchain value drivers may be mapped to the mining industry’s core value chain processes. This was done to identify potential focus areas within the mining enterprise for the further research and development of blockchain applications.

• Implementing real-time augmented reality solutions in underground operations (Sihle Buthelezi)

Industry 4.0 and smart industries are transforming the role of the operator. This consists of novel interactions between operators and machines, which will aid the modern industrial operator, Operator 4.0. This represents the cognitive enhancement of the operator through different technologies that aid decision making. Augmented reality (AR) is the leading tool for achieving the paradigm of the smart operator by providing context-specific virtual cues. Tracking and registration are essential to the success of any AR application as the virtual information must align with the viewer’s perspective to provide accurate prompts for making the best decision. This is typically done through positioning systems and vision techniques, but these are difficult to implement underground as GPS and Wi-Fi tracking are inadequate and poor lighting is unconducive for cameras. Hybrid systems may provide a solution to these constraints. This project investigates the optimal combination of those necessary for the successful implementation of AR solutions in underground mining.

• A safety risk intervention model for human factors in mining senior management (Bhekumuzi Mkalipi)

This study is based on the outcomes of a critical analysis of human factors in safety risk evaluated from mining senior managers through psychometric risk profiling. To develop habitual safety compliance and participation for senior managers, an intervention model was developed to shape human behaviour in safety risk, predicated on a three-part mind phenomenon. The three-part mind centres on the balance between cognition, conation and affection as a foundation for habit formation. A confirmatory factor analysis was conducted to abstract assimilating human factors from literature and predefined models across high-risk and safety-advanced industries. Complacency, affection and cognitive disconnect, as well as complexity, were identified as predominant factors that characterise the mining industry. Six safety-critical, state-based behavioural factors were identified to form the new eclectic human behaviour model. A system of matrices for measuring safety performance and competence indicators were adapted from the workplace safety competence model developed by JvR Psychometrics to form the taxonomy of the new model.