As the time approaches rapidly for mining companies to start reporting on the safety and monitoring of their tailing storage facilities (TSF), several proof-of-concept studies are being undertaken that range from physical site monitoring to a range of remote sensing methods.
Satellite monitoring is one of the more time and cost-effective options that are being considered for continuous monitoring of both active and historic sites.
Satellite based monitoring can be performed for the full life cycle of TSF, from construction, through construction and utilisation to post-closure, says Sonja Goosen MD at Pinkmatter Solutions. The combination of optical and synthetic aperture radar (SAR) methods allows a TSF to be continuously monitored, based on the risk potential that has been assigned to the specific site.
Pinkmatter Solutions has developed the FarEarth Change Monitor platform, that utilises both of these satellite sensor types, making it truly sensor agnostic, and making it a web-based platform that allows users to both view and interrogate satellite data to make informed decisions on time sensitive issues, such as the evaluation of TSF safety or risk detection mitigation.
The advantages of using satellite platforms include: the uninhibited access to anywhere in the world and near fully automated systems. The reduction of travel and of the number of staff required to conduct a survey has cost benefits. It also nearly eliminates the need to physically move around potentially dangerous mine workings and unsafe areas on and around a TSF, adds Goosen.
Using satellite imagery
Satellite imagery or SAR data is used to supplement or validate ground data being generate. Ground investigations range from site investigations that is performed on set intervals such as monthly or quarterly. Instruments are also installed in the TSF, that feed measurement data either in real-time or is collected.
The type of data collected include, pond elevation, piezometric pressures, inclinometer readings and weather conditions. The installation of fibre optic cables (FOC) has allowed real-time observation of tailing facility monitoring up the second, capable of detecting sub-millimetre scale movement. The real-time data is relayed via satellite uplinks or directly to surveillance centres to alert of any critical changes detected. FOC is valuable for catastrophic failure warnings where there are no precursor indicators of potential failure.
High resolution satellite imagery is used at during the planning phase to locate the most suitable site to place a TSF, by generating digital elevation models (DEM) that inform the engineer of conditions of the terrain setting and regional hydrology that would influence the life of the TSF. This is combined with geological data, such as the lithology and local and regional structures, such as faulting and potential seismic activity, notes Dr Nicolaas C. Steenkamp, earth observation manager at Pinkmatter Solutions.
Once construction commences on the TSF, the use of processed imagery volume products, e.g. FarEarth Change Monitor, prove their application value. The survey and engineering departments use the volume products as part of the monthly quality assurance (QA) process and checking the as-built surveys. A Geographical Information System (GIS) integrated platform allows these departments to track the overall construction progress of the dam and water reclaim facilities and evaluate the conformance to design line and grade of the dam.
Monitoring surface water
Surface water accumulation monitoring, both deposited tailing and meteoric water, on a TSF is a critical function from both an operation and safety point and form part of the pool management requirement. The monitoring is performed by regularly scheduled optical satellite image acquisitions. The automated output can be programmed to generate a warning at the monitoring centre if the surface water reaches a dangerous areal extent, continues Steenkamp.
Mine survey departments and the engineer of record also need to be aware of the volume changes of a TSF of time. The best volume estimations are derived from models where a baseline DEM is used to produce high accuracy elevation models and volume change calculations. The FarEarth Change Monitor system receives and automatically orthorectifies the images using precision ground control points (GCPs). Photogrammetry methods applied from stereo high-resolution optical images, produce volume products that are compatible with results from LiDAR surveys, if the previous mentioned criteria are met.
After the TSF has reached its designed capacity, the closure phase will commence with the removal of infrastructure that is no longer required. Stabilisation of the TSF until the start of the next phase, that may entail keeping it in a ready state for later reprocessing or final site rehabilitation. In the case of later reprocessing, satellite imagery is used to monitor the stability of the TSF and risk factors such as the effect of unusual extreme weather conditions or due to illegal removal of material. Once reprocessing commences, a reconciliation of the volumes of material removed and re-deposited can be performed over the duration of the operation. Volume calculations can also be done for historic TSFs, fine residue dumps (RFD) or rock dumps, reducing the risk and cost associated with manual surveys.
During the final rehabilitation phase of the TSF, satellite imagery is used to monitor the establishment of vegetation and stabilisation of the TSF over the short term. Long term monitoring entail vegetation and surface water monitoring and site security. Normalised difference vegetation index (NDVI) gives a good indication of plant health of the vegetation established on an around a TSF.
Failure risk management is done by SAR, says Goosen, as it is able to monitor a site both day and night, is not affected by weather conditions such as cloud cover. SAR data has a horizontal resolution of about 3 meters on commercial platforms and 20 meters on public domain, but has is capable of detecting vertical changes of less than 5 millimetres. The application is however limited by the presence of deep-water bodies in proximity to the TSF, dense vegetation on top of the TSF, for example as part of stabilisation or older sites and very dynamic sites where some active mining is undertaken, notes Steenkamp.
This is mitigated by the installation of corner reflectors that assist in detecting movement on the TSF. Once a sufficient baseline has been established for the TSF, is possible to create more accurate surface movement monitoring (SMM) and modelling, utilising Interferometry SAR (InSAR) or where corner reflectors has been installed the persistent scatterer interferometry (PSI) results.
The output generated on the FarEarth Change Monitor is graphic and highly intuitive to the user, where a heatmap approach is used to delineate areas of movement. Areas with movement in excess of the tolerance or time is indicated in hot colours such as red, with cooler colours indicating acceptable subsidence over time, such as the indicated settlement rate of the tailings material. The results can then be viewed by the user online via the web-based platform. It is also possible to generate a report that can be downloaded from a secure FTP site. Other relevant data such as the inverse velocity graph would then be included, according to Steenkamp.
It is also possible to use SAR data to establish the wetting of the tailing material. As over-wetting of TSF is the leading cause of failures, being able to detect increased wetting over time or rapid increases in wetting in certain areas of the TSF, can be utilised as an early warning proxy.
SMM can also be undertaken for the larger mine lease area, where potential mining induced subsidence due to underground mining or dewatering of underground compartments or placement in dolomitic areas that is prone to the development or karst landscapes. This is especially relevant in areas such as the Witwatersrand in South Africa.
As more satellite constellations are launched in the coming years, it will become possible to move from near real-time monitoring to real-time monitoring of TSF sites along with increased processing power to process imagery, and with the aid of deep learning, or neural networks, identify rapidly developing risks and send alerts to the relevant response teams, concludes Goosen.