The stability of slopes in open-pit mining operations and quarries is extremely important from both economic and safety points of view. The stability of rock slopes entails the design of safe, economical, and functional excavated slopes to attain equilibrium conditions of natural slopes [1
]. It is generally accepted that during the design of a stable slope, a proper understanding of the geological processes, such as stratigraphy, weathering, geomorphology, petrography, and earthquakes, is necessary. The most significant structures that influence the stability of slopes are joints, bedding planes, and the intersection of joints, faults, and shear zones [2
Instability in rock slopes can be harmful and could result in the loss of human life and damage to basic properties. Failure in slope occurs as a result of the downward movement of materials due to the effects of gravity. However, it is assumed that the sliding of a rock slope will take place where there is an intersection of joint sets. Failure of rock mass is inevitable when the shear stress is greater than the shear strength of the rock [3
]. However, failure of slope walls depends on some activities such as cracking of rock mass, weathering, increase in pore pressure, presence of decomposed clay rock filling materials, leaching, increase in water permeability, strain softening, and change in groundwater dynamics, which causes an increment in shear stress [4
]. Therefore, the nature and behavior of the rock mass must be well-understood to ensure that the design of a pit wall remains stable for the life of the mine while extracting as much ore as safely and economically as possible [5
The most common mode of failure in rock slopes is a plane failure. This type of failure happens when the angle of a structural discontinuity plane such as a bedding plane is smaller than the slope angle and greater than the angle of friction of the discontinuity surface [6
]. Moreover, the water forces acting along the potential failure plane can also destabilize the slope. According to Wang and Niu [7
], the dynamic loading and surcharge forces are other factors that contribute to the driving force that causes failure in rock slopes. The plane failure is influenced by factors such as the geometry, groundwater conditions, dynamic loading, potential failure plane characteristics and surcharge conditions [8
]. However, the presence of major structural features such as faults, major joint planes, and unfavorably oriented bedding planes may also have a significant influence on the stability of the slope.
Over the last few decades, there have been significant advances in slope stability research works that investigated the causes of slope failure and factors that can trigger failure in slope. These factors are categorized by Sha [4
] as internal and external factors. The internal factors that can affect the stability of a sloping wall include the mineral composition of the rock, rock types, and geotechnical and structural strengths. In addition, environmental factors such as earthquakes, rainfall, and weathering that can reduce the strength of the rock mass are also categorized as internal factors, while the external factors are mainly caused by human activity [9
The increase in mineral demand in the 21st century has without a doubt compelled the expansion of mining operations globally, which has resulted in the extraction of minerals in larger capacities and deeper levels. However, it is important that these operations are conducted in a safe and economical manner. Therefore, this means that competent designs and techniques should be adopted to support the ever-growing mining capacities. Despite the analysis of slope stability using conventional techniques, mines are still experiencing slope failures, which have proven to be catastrophic and expensive. Therefore, this requires a better understanding of slope failure and the development of accurate prediction models to forecast this hazard before it occurs. Therefore, this study reports the application of slope stability evaluation techniques using linear equilibrium and numerical modeling, as well as the mechanism of slope failure. Additionally, the common factors known to affect slope stability are discussed, together with the various slope failure cases recorded globally. Last but not least, the application of artificial intelligence in predicting mine slope failure is reviewed, together with recommendations to improve prediction modelling.
Instability and Rock Mass Failure in Slopes
Unstable strata in rock slopes can lead to rock mass movement and cause harmful incidents that can affect mining operations and loss of ore reserves. Consequently, it can result in the premature closure of mines. Nicholas and Sims [9
] reported the risk associated with slope failures in mining operations. These ranged from the loss of equipment to loss of reserves or mine closure and, in the worst cases, loss of life. Hustrulid et al. [10
] argued that instability in rock mass slopes is largely caused by mining activities such as rock drilling, blasting, and the use of heavy machines. Similarly, Read and Stacey [11
] indicated that the presence of groundwater, slope design, complex geology, discontinuities on rock mass, and mining operations are factors that affect the stability of rock slopes in mines. According to Eberhardt [12
], the most common factors that influence the stability of rock slopes are redistribution of in situ stresses, complexity in geology, anisotropy and inhomogeneity of the rock materials, pressure pores and seismic loading. Similarly, Stacey and Swart [13
] reported that the effects of blasting and groundwater are the only two major significant factors that control the stability of slopes. In addition, they indicated that blasting can cause ground vibration that in turn may have a significant influence on the stability of the highwall.
Practical experience in the design of rock slope projects has demonstrated the importance of having a robust design. The basic parameters to be considered in the design are height, overall slope angle, and area of failure surface, as shown in Figure 1
]. There is a correlation between the height and stability of the slope wall; the slope stability decreases with the increment in the height of the wall. Similarly, the slope needs to be steeped to minimize the amount of waste rock mined, thereby reducing mining costs. However, the economic effect of steeping a slope is that some portion of the waste material will be mined but the slope will be stabled [5
The design of rock slopes is a major challenge in open-pit mines as it requires knowledge of the geological structure, lithology, and geotechnical properties of the rock mass [11
]. It is necessary to determine the optimum design parameters to be able to evaluate all geotechnical materials before creating a slope. Read and Stacey [11
] diagrammatically illustrated the design process of a rock slope in Figure 2
, which is grouped into five stages, namely: models, domains, design, analysis, and implementation stages. Prior to the implementation of the design, the site investigation and data play important roles in evaluating the economic viability and stability of a rock slope. Furthermore, if predictions have been made that slope failure will occur, it is very important that the structural characteristics of the rock mass be considered. However, at the early stage of the design, data from the site can be obtained during geological exploration, which provides information on the strength of the rock mass, deformability properties, and geological structure and highlights the presence of major planes of weakness. These parameters could also be used to predict slope stability. Typical samples, as reported by Simataa [5
] from a site investigation as shown in Figure 3
, are selected for laboratory testing to determine the mechanical properties of a rock mass.
5. Slope Failure in Mining Operations
Slope stability is a crucial consideration in the management of mining operations as slope failure compromises the economical and safety aspect of production. Due to the increase in demand for mineral resources and the invention of more sophisticated mining methods as well as machinery, most mines are designed to reap more resources from deeper or steeper mines. Such mines have a higher angle of inclination, which make them more susceptible to slope failure. Slope failure can lead to injury of personnel, damage of mine machinery, and disruption of operations, which all negatively impact the mining performance. In order to prevent such a hazard from occurring, it is imperative that extensive geotechnical studies are conducted to ensure that FOS is within a tolerable range (>1) far from failure. In cases where the FOS is very low, more attention should be given to monitoring such slopes to prevent slope failure.
Slope failures are common in open-pit mines and become unavoidable as excavation is becoming deeper and, as such, more difficult to manage. A review of existing failed slopes around the globe provides a better understanding of the failure modes and factors to be considered in controlling the mechanisms that trigger mass movement of materials in slope engineering [55
]. Mitigation of slope failure risk requires an in-depth understanding of structural geology, rock mass properties, influence of groundwater pressure and other external forces in the area [100
]. Failure in rock slopes occurs due to many reasons, which have been discussed earlier in this paper. During excavation, stripping excavation walls are required to be as steep as possible for operational efficiency. Failure may occur if the slope design is over steeped [101
], although most failures in open-pit mines are triggered by fracture and shear on existing defects. Structures inside the rock mass interact in different ways as the rock deforms, and thus affect the general behavior of the rock. In addition, understanding the in situ stress field pattern is critical in understanding the deformation processes in open-pit mines. For instance, mining of ore using open-pit mining methods expose the surface and the rock adjacent to the pit wall; therefore, it becomes unconfined in the direction normal to the slope face over large areas [102
When a rock mass comprises blocks separated by joints, shear zones and bedding planes, sliding on multiple discontinuity sets as well as tensile and shear failure are bound to occur [103
]. For instance, the report of the slope failure that occurred at Yanqianshan iron mine in China confirmed that the eastern part of the slope collapsed as a result of deformed strata in the area [104
]. The collapsed area comprises of eight build-up stages that are represented by actions from point (a) to point (g). These stages are classified as a cave rock zone, cracking zone, toppling zone, and sliding, as shown in Figure 12
. Based on the report of the investigations, the process of initiation and development of failure zones were categorized into three stages: (i) overlaying area above the goaf, (ii) initiation of the collapsed strata sliding into the pit to form a small landslide and (iii) a large landslide occurring as a result of mining activities and creeping of rock that triggered the mass movement in the northeastern phyllite slope. The disturbance produced by underground mining activities initiated the potential sliding body on the northern phyllite slope and the retaining wall structure gradually tended towards instability.
Similarly, another slope failure was reported in Western Macedonia in Northern Greece as a result of sliding along the sub horizontal direction, unfavorably sloping as shown in Figure 13
. From the analyses, the stability was governed by the interface between lignite and an underlying stiff presence of plastic clay or a marl layer that are very close to the bottom of the slope where there is a significant amount of stress. Based on the study reported by Zevgolis et al. [105
], some factors such as groundwater conditions, pit geometry and shear strength of the critical interfaces between clays and lignite influenced the stability of the slope.
In addition, a massive landslide at the Grasberg gold and copper mine in Indonesia, which forced the pit to suspend operations, was reported by DTE [106
]. In October 2003, a fatal accident happened at the southern wall of the open-pit mine, which collapsed and claimed the life of eight people and injured another five people. This incident moved 2.3 million tons of rock and mud down the slope, thereby engulfing mineworkers and heavy equipment. Similarly, a rock mass slippage happened in the year 2000 at the same mine when the overburden was washed by heavy rainfall into Lake Wanagon [107
]. The main cause of material movement was attributed to ingress of groundwater from a nearby water table, coupled with an abnormally high rainfall season.
Recently, another landslide happened at the Gamsberg mine located in the northern Cape Province in South Africa. The open-pit mine is operated by Vedanta zinc international (VZI), which currently produces 40,000 tons of ore per month [108
]. According to Petley [109
], the landslide incident that occurred at the south pit of Gamsberg was described as a “geotechnical failure”. According to the report, eight mine workers were rescued after the incident, while two people were still missing at the time when the report was published. The failure happened across the road that provides an access ramp. The incident appears to have occurred above the access ramp, but the debris reached the pit floor, as shown in Figure 14
. After the failed slope, Vedanta Zinc International (VZI) suspended all mining-related activities at Gamsberg to carry out a proper investigation into the slope failure.
Another notable incident of slope failure in mining was the movement of the Chiquicamata slope in Chile, which was triggered by an earthquake [110
]. In this study, the authors stated that the mass movement progressed at a more or less steady rate until it fell in 1968. However, the kinematics of the slope movements were not clear but seemed to involve load transfer from the north block to the south block dominated by discontinuities. However, due to the presence of discontinuities, the cohesion that was supposed to prevent the movement of the slope was quite small, with little intact rock. The failure started from the north, i.e., the upper part of the slope, to the south block, which is the toe of the slope, as shown in Figure 15
Possibly the largest landslide of all time was the event that took place at Kennecott Utah copper’s Bingham Canyon Mine in the United States of America. This open-pit mine is regarded as the largest man-made excavation in the world, which measures 1 km deep by 4 km wide [111
]. The Bingham Canyon landslide occurred on 10 April 2013 in the canyon open-pit copper mines, causing a massive movement of the upper half of the northern pit wall. The slide filled the mine floor with thick debris, as shown in Figure 16
. Before the incident, the geotechnical surveillance teams and mine operators were fully aware of the instability and evacuated the mine workers and equipment from the unstable zones. Thus, no fatalities or injuries were recorded. According to Pankow et al. [112
], the seismograph showed that the landslide was triggered by several small earthquakes. Six days after the landslide event, 16 additional seismic events were detected in the mine area. The study by Hibert et al. [113
] indicated that the slide caused the onset of the mobilization of the second slide at a higher elevation.
The failure mechanisms driving the instability and causes of failure are presented in Table 3
However, events of landslide and rock slope failure in underground mining environments have not yet been fully researched [18
]. In most cases, slope failure in underground mine spaces happens during the transition from open-pit to underground mining methods. The transition from open pit to underground usually happened to exploit minerals from deep ore deposits. When mine deposits start from the shallow subsurface and extend to a great depth, sequential use of open-pit and underground mining is an efficient and economical way to maintain mining productivity.
An example of a failed slope during the transition to underground mining was reported by Brummer et al. [117
] for Palabora mine in South Africa. Parabola mine was changing their method of extraction from open-pit to the underground block caving method as the ore extraction was getting deeper, which was no longer economical for the mine. During the transition, there was a failure at the north wall which, is evident by the daylight caving of the zone as a result of intersections of four main faults crossing the pit and three dominant joint sets present at the mine [117
6. Factors Required in the Design of a Stable Slope
The essence of strengthening a rock slope is to prevent rock mass movement and premature closure of mining operations. When making an attempt to stabilize a rock slope that is prone to failure, many solutions are opened to the geotechnical engineering team to decide what to do. According to Niroumand et al. [135
], the method used in preventing slope failure includes a change in geometry, rock–soil drainage, rock material nailing, turfing, shotcreting, geotextiles and the application of retaining the wall. Similarly, Roux et al. [136
] applied techniques such as a regular collection of geotechnical mapping and logging data, the collection of laboratory strength test data, adherence to good housekeeping and the use of a hazard plan and evacuation procedure to achieve a stable slope. The study also implemented precautionary measures, such as surveying the actual excavation profile, measuring the change in groundwater level with a piezometer instrument, monitoring the seepage flow from the toe drains and face, measuring the damage caused by blasting to the rock mass behind the mine design line, inspecting the blast face and crests to allow removal of loose materials, and periodically installing bolt supports, as recommended by the geotechnical engineering team.
The study reported by Singh et al. [137
] indicated that anchored tensioned rocks were installed, along with scaling and trimming of loose rock blocks, to prevent further slope failure. The study mentioned that shotcrete can be sprayed on the slope face with proper drainage galleries to prevent the development of pore water pressure. These techniques protect the slope from degrading agents such as rainwater and serves as part of an effort to strengthen and improve the ground conditions, which in turn increase the FOS and improve the stability of the slope. Another method to stabilize a rock slope involves the use of a leguminous tree planted on the rock slope to control hydrological factors such as erosion and run-off water that have the potential to trigger rock mass movement in rock slope engineering [138
]. The outcomes of the study showed that the Leucaena Leucocephala leguminous plant has the capability to prevent slope failure in open-pit mining operations.
The use of an early warning system (EWS) is another approach to prevent the risk associated with slope stability. Roux et al. [136
] confirmed that the installation of smart technologies such as extensometers aid in monitoring and checking the source of rock slope failures. In deeply weathered rock, the presence of cracks and faults are inevitable and need close monitoring. Extensometers and other smart systems can be installed in open-pit mines to monitor cracks and generate early warning systems. An extensometer integrated with other EWS was installed at Navachab Gold Mine in Namibia to give an early warning prior to failure. A study reported by Roux et al. [136
] indicated that extensometers set off alarms an hour before failure. A similar study reported by Karam et al. [139
] indicated that warning systems in mines should be installed in affected areas and serve as a guide to passive or active countermeasures. The EWS is designed to reduce threat, as shown in Figure 17
6.1. Methods to Improve Slope Stability
There are several ways of achieving stability or strengthening rock slopes in mining operations. These include the construction of drainage and pump installation, slope monitoring, ground control with improvement in geological structures and the application of reinforcement.
6.1.1. Drainage Construction
Since water is one of the factors that promote failure in rock slopes, mine water management systems should be a priority. The presence of groundwater in open-pit mines often creates problems by reducing the stability of pit slopes. Most of the problems are caused by pre-pressure and hydrodynamic shock from blasting operations [140
]. The presence of pore pressure on rock mass reduces the shear strength and seepage pressure, and water in the tension cracks causes an increment in the unit weight, which in turn increases the shear strength of the ground.
A number of researchers confirmed that drainage construction around the mines can reduce the impact and occurrence of groundwater in a mining environment. The three main objectives of drainage systems in mines are to keep working areas dry, stable, and safe, to ensure pit floor workability, and to lower the hydrostatic pressure and increase the effective stress of soil to improve stability [141
]. Some of the drainage systems may include the construction of channels, water collection sumps and pump stations with pipelines to divert water from the surface. In addition, drains may be constructed within the surface to remove excess seepage. Hence, it is necessary for the geotechnical engineering team to know the hydrogeological conditions of the mine. This will assist in the selection of pumping systems required in the mine.
6.1.2. Slope Monitoring
Monitoring of rock slopes is necessary to determine how the rock structure behaves during excavation and if they are a threat to safety. This has been a key technique used to assess instability in rock slopes. The slope monitoring approach can be of value to provide information that is useful in data collection, recording and qualitative and quantitative analysis. The effectiveness of any monitoring system depends on the ability to give warnings before rock mass movement or failure takes place [142
] Traditionally, there are geotechnical instrumentations (piezometers, crack meters, tiltmeters, borehole extensometers, and stress meters) available for the monitoring of slopes in mining and civil engineering. Despite the contribution of these instruments in the monitoring of rock instabilities, they also have some limitations.
Over the years, there have been improvements in geotechnical monitoring instrumentations, which include the use of remote-sensed technological tools. These instruments are commercially available to monitor rock mass movement in a mining environment and are cost-effective. Examples of these are the integration of synthetic aperture radar (SAR) with optical images and interferometric SAR (InSAR), which are currently being used to characterize instabilities in slopes. Advancement in remote sensing technology systems has brought about differential synthetic aperture radar (DInSAR) with high-resolution image processing, which is also being used to monitor instabilities. Similarly, there is a ground-based radar monitoring instrument known as linear SAR (LISA), which is capable of assessing the deformation of an unstable area in slopes that are characterized by high radar reflectivity [139
]. Other advanced geotechnical monitoring instruments are light detection and ranging (LiDAR) and optical satellite images.
6.1.3. Ground Improvement with Enhancement of Geological Structures
The interpretation of the data acquired during site investigation provides information about the structure of the rock mass, strength of the rocks, planes of weakness, and the range of their strengths. This will provide adequate information on the relationship between slope parameters, such as the spatial relationship, face inclination of the rock, and geometrical dimensions of each of the individual elements of a particular sector of the mine shell, such as ramp and geotechnical berm width, bench spill berm width, individual slope height, and slope angle, slope bench, etc.
Apart from the design parameters, the geotechnical engineering team must be familiar with the geology of the mine, i.e., the discontinuities and engineering properties of the rock mass that can influence the stability of the slope. Karama et al. [139
] argued that gravitational load is the most important load affecting the stability of the mine and, notably, the shear strength of the ground, which confers major resistance to failure. To prevent further damage of the rock mass, the shear strength of the ground must be enhanced to strengthen and protect the ground. This strengthening can be completed with the application of backfill, rapid yielding properties, the construction of a retaining wall, the spraying of shotcrete, and application of reticulated micropiles.
6.1.4. Installation of Reinforcement Units
The application of reinforcement in rock slope is a viable approach to increase the stability of unstable slopes. The installation of various support systems increases the shear strength of rocks and reduces sliding effects along the slip surface [143
]. The essence of reinforcement in weathered rock or soil is to stiffen the base and to reduce shear stress magnitudes. The two most commonly used support systems in rock slope and slope engineering measures are rock bolt fame and anchor cable fame [144
]. These reinforcement tools have proven to be effective in resisting the shear stress in rock mass deformation by transferring the load to the bounded rock so that resistance is provided by the rock to balance the deformation. During the installation of bolts, the effects of the inclination angle of bolts must be defined by the designer of the slope for effective reinforcement. According to Sazzad et al. [145
], the FOS of slopes increases with an increase in the angle of inclination. However, the FOS value depends on the overall angle of the slope. Therefore, the relationship between the slope angle and inclination angle for a maximum FOS should be enhanced, i.e., the positioning of the bolts is a key factor in reinforcement design. In addition, when installing a bolt to resist dip bedding slope, the failure mechanism and process occurring in such slope need to be considered. A great number of researchers have acknowledged the effectiveness of bolts as reinforcement tools in controlling the movement of discontinuous layers that are subjected to shearing. The installation of rock bolts in unstable zones can be regarded as one of the major techniques used to improve and control the unstable ground in open-pit mines.
6.2. Role of Artificial Intelligence in the Management of Slope Failure As a Reflection on the Current State of the Art
The monitoring and evaluation of slope stability in mining operations is a crucial necessity to avoid potential damage to both the personnel safety and financial base of mining companies. Despite the effort to reduce slope failure, the geotechnical phenomenon is known to be complex, comprising numerous factors such as ground conditions, geological activities and human actions, which make the prediction of slope failure challenging [146
]. More so, these factors are dynamic and ever-changing, making it cumbersome to continuously measure them. However, the invention of artificial intelligence (AI) has greatly aided engineers in forecasting possible slope failure using detailed analysis rather than solely relying on phenomenological models [148
The study by Kothari and Momayez [148
] compared the prediction performance of inverse velocity (IV) and the artificial neural network (ANN) model to forecast slope failure of an open-cast mine. Twenty-two datasets were collected using radar equipment, and we developed a double-layered feed-forward ANN in the MATLAB environment. The output from the study indicated that ANN prediction was 86% accurate compared to IV. A total of 82% of the slope failure predictions indicated that the slopes were safe, while 18% indicated that the slopes were unsafe. The unsafe predictions were approximated within 5 min before actual failure, proving that the ANN model gave a safer prediction compared to IV. Another study by Chebrolu et al. [149
] utilized 46 datasets (23 obtained from wet slopes and 23 obtained from dry slopes) to predict the FOS of a slope using multi-gene genetic programming (MGGP) and multi-adaptive regression spline (MARS). Thirty-two datasets were used in training the models, and fourteen datasets were used in testing the models. Despite both models proving to have an accurate prediction, MGGP has a better prediction performance, with R-values of 84.38% in training and 85.71% in testing, compared to MARS with R-values of 81.25% in training and 85.71% in testing.
To establish an accurate prediction model for determining the slope failure of a mine in Vietnam, a study by Bui et al. [146
] integrated the M5-rule and genetic algorithms to develop a novel prediction model that was then compared with conventional AI models, including ANN, support vector regression (SVR), the firefly algorithm (FFA), the imperialist competitive algorithm (ICA), artificial bee colony (ABC), and the genetic algorithm (GA). This study utilized bench height, soil unit weight, cohesion, angle of internal friction, and slope angles as input parameters to forecast the factor of safety against failure. The results of this study indicated that prediction performance from the M5-rules GA model gave the best accuracy due to its enhanced optimization prowess. Another study by Du et al. [150
] utilized a ground-based interferometric radar (GB-SAR) to record 150 datasets in the Anjialing open-pit coalmine of China. Twelve input parameters (slope shape coefficient, deformation rate, reverse deformation rate, deformation amplitude, rainfall, temperature, atmospheric pressure, relative humidity, wind speed and direction, and groundwater temperature and level) and one output variable (deformation) were used to develop five prediction models. These models included backpropagation, NN, support vector machine (SVM), recurrent neural network (RNN), adaptive neural fuzzy inference system (ANFIS), and relevant vector machine (RVM). From this result, it was evident that not all models can be reliable in the prediction of slope failure due to high error values. The error values for each model were 4.122 mm for BPNN, 3.612 mm for SVM, 1.660 for ANFIS, 0.578 mm for RNN, and 0.442 mm for RVM. RVM had the lowest root-mean-square error value (RMSE) of 2.64, while BPNN had the highest RSME value of 4.58.
In addition, a study by Ferentinou and Fakir [151
] developed a BP-NN model using 141 databases obtained from worldwide cases of surface mines and eighteen (18) input variables, including environmental factors, rock quality, rock mass characteristics, rock stresses, hydraulic profile, presence of geological features, dimensions of slopes, blast design and previous instability occurrence. This study established slope stability indices, where all slopes were examined to predict failure. The results obtained indicated a mean square error value of 0.0001 converging at 98%, proposing the utilization of BP-ANN as a reliable tool to predict slope failure in feasibility studies. To evaluate and monitor the slope stability of open-cast mines, Luo et al. [147
] developed a particle swarm optimization-cubist algorithm (PSO-CA) to forecast the factor of safety of a Vietnamese mine. This model used five input variables, including bench height, angle of slope and internal friction, cohesion coefficient, and specific weight of the material. The output from this model was then compared with the output obtained from the prediction of the same variables using SVM, classification and regression tree (CART), and k-nearest neighbor. Performance comparison of the models inferred that PSO-CA had the lowest error values (mean absolute error (MAE) of 0.009 and root mean square error (RMSE) of 0.025 and a high correlation coefficient R-value of 0.981. The SVM, CART, and k-NN produced poor prediction performance with MAE values of 0.014 to 0.038, RMSE 0.030–0.056, and R 0.917–0.974.
The reported studies that have utilized AI in determining factors of safety in mines have proven to be efficient and effective in accurately forecasting slope failure before the hazard occurs. Moreover, the models are endowed with the ability to handle large amounts of data at a given time and execute predictions at a high rate compared to conventional stability analysis techniques. The models are designed using various factors known to affect slope stability in order to map out FOS. Various AI models have been used to predict slope failure in mines, but limited studies have been conducted to compare and validate the most accurate model for such a task. Therefore, in order to establish the most robust, versatile, and reliable model, more comparative studies need to be conducted to improve prediction performance.
7. Concluding Remarks
Stability analysis of slopes in open-pit mines and quarries is extremely important from both economic and safety points of view. The effect of unstable ground cannot be over-emphasized as it may lead to the temporary or permanent closure of the mine depending on the level of damage. Instabilities usually occur where there is a presence of geological discontinuities such as fractures, cracks, faults, unfavorably oriented bedding planes, etc. Rock slope failures are triggered when the shear stress exceeds the shear strength of the rock mass. Several factors, such as geometry of the slope, groundwater condition of the rock, lithology, geological structures, cohesion and angle of internal friction, effect of blasting, mining method and equipment selection, have been declared to influence the failure of rocks. The effects of all these factors must be considered during the planning and designing stage of the mine, not only that these factors have effects on the design, but that they also influence the engineering judgement on the stability of the slope. Several techniques, such as construction of drainages, slope monitoring, application of reinforcement and improvement of geological structures, have been used over the years to improve and maintain the stability of the rock slope.
Advancements in technology have brought about improvement in slope monitoring techniques and provide an understanding of slope dynamics through the assessment of stability over time. The accurate prediction of slope failure is an important task in active open-pit mines in order to avoid slope failure so as to prevent injury to personnel and damage to machinery. When determining the stability of a slope, it is imperative to assess the factors affecting slope stability. Different studies have used different AI models with various input variables to predict slope failure in mining. These studies have reported the prowess of AI models as being fast, reliable in mapping out FOS and sturdy to handle large amounts of data at a given time. Despite the usage of such a model, there is a need for comparing the output of various AI models in order to determine the most accurate model for forecasting slope stability. In addition, various elements known to affect slope stability should be captured and used as input variables to forecast slope stability and compare performance of different AI models.