The Council for Scientific and Industrial Research (CSIR), as part of its collaboration with the Mandela Mining Precinct, has either developed or adapted several technologies that could be applied in mining environments to improve safety, productivity and efficiency.
The function of the Mandela Mining Precinct is to coordinate research activities towards the modernisation of South Africa’s mining operations through the development of next-generation mechanised mining systems.
The focus is on achieving zero harm, reducing costs and increasing efficiencies. In parallel to this, the precinct aims to support the development of local supply chains by enhancing local manufacturing capability.
CSIR principal researcher Dr Dave Roberts explains that there has been a particular focus on safety because, “despite significant improvements since the mid-1990s, in 2014, . . . we had reached a bit of a plateau in mining safety . . . more had to be done in terms of zero harm”.
The CSIR showcased some of these tech- nologies – the Monster, Rock Pulse, pedestrian detection system (PDS) and ground- penetrating radar (GPR) technologies – at a media briefing last month.
Prototype technologies have also been adapted for use in a mining environment, including the Locomotive Condition Monitor- ing System (LCMS) and the Dassie.
Speaking at the Mandela Mining Precinct last month, CSIR principal geophysicist Dr Michael van Schoor noted that the cause of most fatalities (40%) underground was fall-of-ground and that GPR could assist in reducing these fatalities.
GPR can be used to improve roof bolting applications and detecting potential faults in a hanging wall.
In terms of general fault detection, Van Schoor explained, GPR worked in a similar way to speed traffic radar detection systems, as GPR transmitted a signal down into the ground and geological features were mapped on a radargram based on the amplification and duration of the return signal.
“Wherever you have a discontinuity or some anomalous property that causes the radar to reflect . . . you can measure the amplification of the signal and the time it takes for the signal to travel from the system to the ‘target’ and back to the system.” He noted that the system could be used from the surface or placed on a hanging wall.
Van Schoor added that the CSIR was developing the technology to produce three-dimensional (3D) maps, which could be integrated in real time to an existing mine plan. There were plans to mount the GPR on a robotic platform to accelerate underground 3D surveys, which was currently done manually.
CSIR principal engineer Shaniel Davrajh notes that GPR is typically a predictive technology used to identify areas in the wall and the stope that are potentially dangerous.
The CSIR has developed a robot platform called Monster, equipped with safety inspection sensors for “early entry examination”, as it is built to detect causes for potential falls of ground.
Monster aims to identify and assess risks for underground mines using thermal imaging and audio sensors. The thermal imaging sensors could be used to detect loose rocks, based on the knowledge that loose rocks cool faster than the hanging wall because of the increased ventilation.
At last month’s briefing, Davrajh noted that the sensor could detect temperature differences as small as 0.1 ºC. The audio sensor works in a similar way to “tapping a watermelon” – a loose rock “has a certain sound,” so, once the thermal sensor has identified that a rock could be loose, the audio sensor, placed on a mechanism that physically taps the rock, will verify that the rock is loose.
The CSIR Monster prototype has been trialled at the precinct’s stope simulation, which has a decline of about 30º.
Davrajh adds that integrating 3D-mapping with range finding into the Monster platform is possible, and that the robot can be used to accommodate GPR and other sensing technologies.
The CSIR, in conjunction with the Department of Viticulture and Oenology and the Institute for Wine Biotechnology at Stellenbosch University, in the Western Cape, developed a robotic platform to inspect and monitor horticultural crops on local farms. Davrajh says the precinct is looking into adapting this technology for tailings recovery.
The automated, intelligent robotic system uses sensors that have been configured to estimate grape yield and monitor plant growth and canopy health. Grapes’ sucrose content causes them to emit a certain frequency. Their health or sucrose content is measured using a hyperspectral camera, which can determine their ripeness by measuring the wavelengths of the light emitted by the grapes.
Davrajh notes that companies can potentially prioritise areas for tailings retreat- ment using this technology. “The hyperspectral camera has 1 000 filters, enabling us to extract any range within the light spectrum . . . If we know what frequency to look at, we can tell whether the grapes are ripe or not.” He says that the concept could be applied in similar fashion when looking for gold in tailings dams.
The CSIR partnered with Transnet engineers to develop the first 13 units of the LCMS to monitor Transnet’s locomotives – using machine learning and computer vision algorithms – following the earlier successful deployment of two prototypes. This has subsequently enabled Transnet to implement predictive maintenance and detect rail infrastructure defects since June.
The LCMS is an automated solution that gathers sensor and fault data through a real-time mechanism, which is then automatically sent to the Transnet servers for evaluation. The system has been designed and tested on all applicable rail standards.
It has a unique communications module that integrates 3G, WiFi, satellite and ultra-high-frequency radio into one module, allowing for communication even in adverse conditions. An integrated battery backup system enables the device to function for at least one hour during a power failure.
The CSIR notes that the platform has been designed to allow for easy expansion of hardware and software capabilities and there are plans to implement on-board, real-time data analysis.
Davrajh notes that the system can be configured to monitor driver and commuter behaviour, as well as carbon emissions. “What we’d like to do is apply the same system to underground mining vehicles to extract information for predictive maintenance.”
The CSIR is also developing a fatigue detection system, which is based on data obtained by monitoring a driver’s eyes, as well as carbon-dioxide levels in the cabin. However, he notes that this is a recent project, with no results to report as yet.
Davrajh says the CSIR has developed an enhanced PDS, which is used to avert vehicle- to-person collisions underground, using algorithms to predict whether a collision is imminent, thereby eliminating unnecessary vehicle stoppages.
He notes that current collision-avoidance systems use radio frequency interface units, which essentially require the workforce to wear a tag which alerts the system (installed on the vehicle) that a person is in range. “This presents the possibility of a false alarm, as the vehicle would still shut down if the person is in range, but not in harm’s way.” This also requires every person to wear a tag at all times when underground.
Davrajh explains that the improved PDS incorporates a range finder and vision detection system, which identifies a person near a vehicle, and determines the distance between the person and the vehicle. Based on the speed at which the person is moving, the distance of the person from the vehicle and the speed of the locomotive, the system detects whether a collision is possible and, therefore, whether vehicle shutdown is necessary.
The RockPulse device is attached to rock in the local environment of mineworkers. It functions similar to a fire alarm by setting off an alarm when it detects the early onset of impending rockfalls through continuous monitoring of fracturing taking place inside the rock.
“The shortest early detection observed in underground coal trials was about 90 seconds before the event, which is typically referred to as ‘goafing’,” notes principal electronic engineer Gideon Ferreira.
He adds that technology has been trialled extensively in various coal mining scenarios where collapse forms part of the regular mining process – with great success.
RockPulse has also been used to collect data in hard rock mines, facilitating site- specific rock engineering investigations. The technology has not been verified with regard to early warning in hard rock mines, owing to the absence of events in mines where it has been installed, thus far. “However, laboratory and coal mining results suggest early warning in hard rock mines could also be successful (the underlying failure process being detected by RockPulse is not dependent on rock type),” says Ferreira.
By continuous real-time monitoring before, during and after the event, RockPulse could also play an important role in safe re-entry into the area.
Roberts notes that the mining industry is “an industry in crisis”. In addition to safety problems, mining struggles with sustainability, as costs are increasing, production is decreasing and issues relating to corporate social responsibility are becoming more prevalent.
The CSIR and the Mandela Mining Precinct therefore focus on addressing challenges pertaining to safety, mechanisation and productivity, thus assisting in improving the South African mining sector, including its longevity.