Optimising stope design and quality monitoring with laser scanning

February 15th, 2018, Published in Articles: EE Publishers, Articles: PositionIT

The demand of copper is such that progressively lower grades of copper ore are being mined profitably. One way to do so is by reducing the overall volume of waste material extracted, ensuring the highest concentration of the metal by minimising external dilution. Laser scanning can play a crucial role in achieving this.

From a mineral resource management perspective, strict control on the quality of material that is delivered to the concentrator mill can have a profound effect on the process costs.

Mine surveyors are uniquely positioned to provide the required information needed to maximise extraction and reduce dilution, regardless of the commodity being mined. Intensive and precise geological modelling, combined with high-precision geospatial control and monitoring, enable profitable ore extraction from the thin, low-grade ore bodies that were avoided or abandoned by earlier generations of miners.

The introduction of laser scanners in the mapping of underground cavities that were (and mostly still are) inaccessible brought with it an impressive array of capabilities that was held back only by the processing power of computers that were available. The ability to process high-density point cloud data has improved and subsequently led to applications in mining that were previously untried.

The Lubambe Copper ore body on the Zambian Copperbelt, which is classified as narrow (5,5 m) and low-grade (2,3% Cu), is one mine that utilises this technology.

Control over the stoping operations needs to be of a very high standard in order to generate quality tons, which will in turn lower the production cost that generally equates to improved profitability. Accurate drill hole and stope measurement are central to this control. The aim is to reduce the overall volume of material generated and ensure that the focus is on extracting the contained metal by minimising external dilution.

Ore body delineation and rock mass characterisation

The whole process begins with the ore body delineation and rock mass characterisation. Delineation drilling from the footwall drives provides an ore intersection approximately every 25 m on strike and on dip.

At Lubambe mine, ore development generally does not expose the assay hanging wall. The location is established by drilling 10 m deep holes into the hanging wall at predetermined intervals. Delineation drilling data is combined with the assay footwall model to generate the assay hanging wall wireframe model.

This component is handled by the geologists in conjunction with the survey department. Boreholes are drilled, geo-referenced, logged, and analysed, and the data is then used in the construction of the ore body wireframe model that defines the “ore envelope”. The geo-referencing consists of high-precision collar surveys, which are combined with subsequent “down-the-hole” borehole surveys. This gives a 3D description of the surveyed borehole.

In addition, the geologists mark out the assay footwall contact in the developed ore drives with yellow spray paint. The surveyors perform an accurate detail survey on this assay contact line and generate a string file in a digital format that is used by the geologist to construct the bottom or footwall wireframe of the ore envelope.

This ore envelope is the ore body as defined by the geologist’s wireframe model. There is a need for routine geological mapping and timely interpretations to keep the ore envelope wireframe current, and to ensure that the stope design is based on the dataset that has the highest confidence levels [1].

The data used in the generation of the monthly extraction analysis report depends on these three sources: the ore body model, the designed stope (which is a derivative of the ore body model), and the as-built wireframes whose construction will be described below.

Pre-stope (as-built) and design wireframes

In terms of quality, very close coordination is required between all facets of the mineral resource management departments (planning, survey and geology) and the mining and production department.

The geologists guide the development of the ore drives by ensuring that the drive is mined at an economically optimised location in relation to the ore body. This method of directional control exposes the footwall assay contact, which is in turn sampled with an x-ray fluorescence spectrometer to identify the location of the cut-off grade along strike. The position of the cut-off grade is marked out with a conspicuous yellow paint for subsequent survey. This painted line is what is referred to as the assay contact line.

Surveyors equipped with reflectorless total stations survey the assay contact line to 0,01 m accuracy and provide the geologists and planners with the precise location of this line in the form of a CAD string file. The geologist uses this to update the footwall of the ore body wireframe model, and the planners use it to optimise the design location of drill holes. The surveyors also perform a laser scan of the ore drive and generate a 3D wireframe model which is used in the stope design. This wireframe is called the pre-stope wireframe (Fig. 1).

The pre-stope/development wireframe is used as an input into the stope design. The drill rigs operating parameters and production requirements are assessed against this wireframe. The geological ore body model is superimposed on this and provides an as-built basis for stope design parameters.

Fig. 1: Pre-stope/development wireframe.

Fig. 1: Pre-stope/development wireframe.

Based on the geological model and the pre-stope wireframe provided by the surveyors, the mining engineers in the planning department design a stope which would provide the most economic extraction volume for the ore at that location (Fig. 2). Complete with machine rigging positions and ring orientation, the design is passed on to the survey department for staking out underground. A drilling pattern that corresponds with the rigging positions is issued to the production crew.

Fig. 2: Stope design wireframe.

Fig. 2: Stope design wireframe.

Once the production stope rings are drilled, the survey department performs down-the-hole (which is actually upwards) camera surveys to verify the drilling quality. The collar coordinates, directions of the holes, and hole-depths are surveyed and measured with a borehole survey instrument (PeeWee). Once these are compared with the design, the rings are either recommended for blasting or are re-drilled depending on the quality of drilling. The print generated by the surveyors is then used by the operational officials to make the decision.

In order to ensure compliance, disciplinary action would be taken against a mining official that blasts a production stope without taking this necessary prerequisite. This is a critical aspect that ensures compliance to stope design before the point of no return, i.e. the point at which the blast is taken and nothing can be done to correct any errors in the blast process. It ensures that all the holes in each ring are drilled correctly.

The tool of choice for checking the drill quality is the PeeWee – an electronic, multi-shot borehole survey instrument.

This tool was originally designed to map and track holes drilled using directional core drilling. The system enables geologists to drill through a fairly large volume of host rock from a single drilling site. It provides many benefits, including limited environmental degradation resulting from multiple drilling sites, reduced drilling costs due to the reduced number of setups, and shorter exploration times. It also reduces the total number of required drilling meters, as the designed deviations can be executed at almost any depth in the hole.

This is where the drilling method distinguishes itself from drilling multiple holes at steep inclinations in several directions from the same site.

The PeeWee is a tool that combines a clock, gyroscope, inclinometer and compass to provide accurate position partials for a given point in time. The time is used to “isolate” the readings needed to graphically reproduce the geometric properties of the hole. From the beginning of the survey to the moment the survey is complete, the surveyor synchronises the depth and reading by taking note on the moment the PeeWee was held still in the production drill hole. The data collected can be thought of like a 3D open traverse, commencing vertically at the collar of the drill hole and ending at the toe.

This data is compiled and shared with the blasting engineer, who optimises the timing and charging up sequences of the respective holes. In situations where the drilling quality is sub-optimal, the instruction to re-drill the holes is issued. Sub-optimal holes include short holes, which may be the source of under-break, holes deviating into the hanging or footwalls, creating dilution, or holes that are too long and which might compromise pillar integrity.

Drill rigs automatically drill the production pattern uploaded to it via a memory stick, and saves data from the actual drilled lengths, dips and hole directions. What this equipment does not generally do is determine the amount of down-the-hole deviation resulting from varying rock densities and such.

Once the holes meet the minimum requirements, they are authorised for blasting. And this is where the Lubambe Mine laser scanning regime differs slightly from most operational mines.

Laser scanning and ring comparisons

The Faro laser scanners and Optech Cavity Monitoring System (CMS) are used on a regular basis to conduct a ring-by-ring comparative analysis. The quality control checks at this point shift from being a hole-by-hole inspection. The blast profile is inspected for deviations from the design, dilution, overbreak, underbreak and rib/ crown pillar integrity. Every eight rings, or approximately 17 m, the stope is scanned and the point cloud is checked for creeping or unplanned breaches into upper levels. The regular scans are compiled into a single file that is used to generate the progressive stope volume at the end of the month, which incidentally is the single use that most mines purchase these units for.

The point data is collected using a Faro Focus3D high-speed terrestrial laser scanner and the Optech V500 CMS. The laser scanner offers one the most efficient methods for 3D measurement and 3D image documentation. In only a few minutes, this 3D laser scanner produces dense point clouds containing millions of points that provide incredibly detailed 3D images of large scale geometries. Multiple scans from different positions can then be compiled to create a cohesive point cloud, resembling an exact measureable copy of even the most complex and large structures.

In the case of the stopes at Lubambe, full entrance into and under the stopes is unsafe. Since by design the stopes are not supported, the scanner is mounted on a dolly, which in turn is mounted on a horizontal open lattice truss attached to a monocycle. This contraption is pushed into the stope and the rear end is stabilised by a heavy-duty survey tripod. This enables the approximate levelling of the scanner. The scanner has a tilt correction capability that will compensate for vertical axis errors of up to 5⁰. The scanner is reeled into the stope by rope and pulley, and is activated by remote control (any Android tablet with WiFi and flash player works for this).

The surveyor coordinates a mini-prism by total station and swaps this with a registration sphere. The coordinates are then used to geo-reference the scan, which would otherwise have been in a localised coordinate system, with the scanner location defining the origin. The laser scanner also provides the option of setting the desired scan resolution and measurement quality at this stage.

The scan resolution is defined as the distance between two successive measurements, 10 m away from the scanners position. Due to the radial nature of the measurements, the point density increases with shorter distances and decreases the further away the scan subject is from the scanner. This holds true for both the Faro and the Optech units.

The scan quality is indicated by the level of confidence ascribed to a measurement. For the Faro laser scanner, a single measurement will typically indicate the lowest confidence, while four independent measurements to the same point will have the highest confidence levels. Confidence level therefore directly affect scanning time.

The higher the resolution, the longer the scan takes to complete. A typical scan at 976 000 points per second takes under two minutes to complete. The geo-referencing survey, however, takes longer.

The lowest scan resolution is still sufficient to collect the spatial points necessary to extrapolate the stope geometry, with the drawback being that the minimum distance to the registration spheres needs to be reduced in order to retain a fair number of points on the spheres. The Optech CMS does not have this drawback, as the data collected is independent of the geo-referencing procedure. The scan is oriented by surveying two points mounted on the scanner and the dolly, which in turn allows a real-time transfer of the surveyed coordinate values.

The registration sphere has a known diameter and is placed in close proximity to the scanners position. The surface of the sphere needs to have a sufficient number of scan points on it to ensure the extrapolation of a theoretical circumscribed sphere – whose vertices are the measured scan points. Since the mini-prism and the registration sphere are concentric, they share the same XYZ coordinates that are used in the subsequent geo-referencing. A minimum number of points are required to ensure a large enough sample from which the theoretical sphere is derived. A maximum of about 46% of the sphere can be scanned from a single scanner location. The software recognises this and, using the points measured on the sphere, calculates the arbitrary coordinates of the centre of the circumscribed sphere.

Pre-processing

The Faro laser scanner’s point cloud data is of such high resolution that even at its lowest settings the direct conversion of the scan data to .dxf or .dwg formats makes the resulting file too large to handle in a technically efficient manner.

Once the georeferencing process is completed in the generic software, a XYZ coordinate list of points 0,2 m apart is generated. This equates to discarding well over 80% of the measured data, as scan points are measured as densely as 1 mm apart in areas close to the scanner. This density renders most CAD packages useless in the subsequent processing, hence the need to filter the data. The scan data is run through scan filter software, which is capable of reading the output from the scanner and generating the coordinate list.

The point data is used for two purposes: periodical quality control and progressive volume measurement.

Fig. 3: A typical progressive scan report for a production stope.

Fig. 3: A typical progressive scan report for a production stope.

A progressive scan report for a production stope (Fig. 3) allows the charging patterns and timing to be adjusted, so as to recover ore which would have been lost as well as to prevent dilution.

Wireframe modelling process

Underground data is collected using the Optech V500 CMS, which is then reduced to point cloud data. A sequential inspection at 2 m intervals provides the basis for generating the strings used in the wireframing process (Fig. 4). This also represents the first and only major loss of fidelity in the measurements taken and derived. However, the effect is deemed negligible as the method provides the most accurate volume when time constraints are taken into consideration. The net effect is similar to the output expected when Simpson’s or trapezoidal rules are applied with a 2 m inter-area interval.

Fig. 4: Closed strings wireframe, from which the base wireframe model is generated.

Fig. 4: Closed strings wireframe, from which the base wireframe model is generated.

A view is taken through a 0,2 m thick cross section of the sampled point data. This in turn provides the dataset required to generate cross sections of the stope. Closed strings of the cross sections are created by digitising the points displayed along the section. This process serves to eliminate noise and other undesirable data that is inadvertently collected during the scan which is likely to distort the outcome.

The other operation that takes place at this stage is the interpolation of probable edges. These edges are usually missing datasets that occur as a result of “shadows”. Shadows are formed when an obstacle obstructs the laser beams from the point source and prevents measurements to a desired face or sidewall. Shadows are easily overcome by relocating the scanner. However, in the open stopes this is not always possible due to accessibility issues and other associated hazards.

Fig. 5: Complete base wireframe, representing the mined-out stope.

Fig. 5: Complete base wireframe, representing the mined-out stope.

The outline of the stope as mined is generated every 2 m. This series of closed strings (Fig. 4) is used to generate the base wireframe model (Fig. 5). Using a process called string linking in Datamine software, a wireframe is generated. The base wireframe is the digital representation of the mined-out stope, and a direct-scale model of the excavation under scrutiny. This wireframe then provides a solid model with various properties that can be queried (Fig. 6). By now the volume of the stope as mined is immediately available.

Fig. 6: Stope wireframe (or solid).

Fig. 6: Stope wireframe (or solid).

Two other wireframes are used in the comprehensive stope analysis process: the design wireframe (Fig. 2) and the geological ore body model.

At this point the similarity with most stope reporting procedures ceases.

Comparative analysis and difference models

Built-in Boolean operations in Datamine enable the creation of difference models. These are models created by subtracting one wireframe from another. A comparative analysis between the stope design wireframe and solid stope (actuals) wireframe enables the volumes of the under-break or over-break to be established with remarkable accuracy for time it takes to do so (Fig. 7).

Fig. 7: Stope under-break wireframe model (difference model).

Fig. 7: Stope under-break wireframe model (difference model).

The difference model (Fig. 7) shows the design without the ore which has been mined out, and represents the very first production losses incurred by the mine, since what is left behind cannot be recovered. This happens to be true for Lubambe copper mine, where the mining method does not permit the recovery of this kind of ore.

The comparison further shows the amount of waste introduced into the ore stream from a specific stope –  both hanging wall and footwall dilution mined together with the ore (Fig. 8). Best case scenarios these volumes are zero, as this is the material that pushes up the mining cost due to the volumes to be moved, milled and processed but which pay absolutely nothing.

Fig. 8: Total dilution model.

Fig. 8: Total dilution model.

The point-by-point comparison also enables engineers to identify areas that might either be too tight or may require concrete back fill. The Lubambe survey team took this aspect of the scan analysis and defined the various parameters of the stopes that were non-compliant to the design and derived a series of wireframe models from the stope. Although Datamine software was developed specifically to process geological ore body models, the functionality is perfectly suited to processing stopes.

Conclusion

While the technologies discussed here have been around for years, the application of laser scanning in the calculation of the stoped-out volumes for month-end payments or reconciliation has been limited. Where the laser scanners or CMS units are applied to quality control measures, they pay for themselves within a short period of time by enabling designed tweaks and balances to the operation.

Where an operation has more than one type of drill rig for production drilling, these systems facilitate optimised drill design for an exact match between drill rig and drill site. This ensures that the drill rig is capable of fitting the drill site without having to physically test the setup capabilities in that area. An accurate scan is critical to this design aspect.

The correct application of laser scanning can minimise dilution and so reduce costs, since excessive dilution directly affects the tramming and energy cost of an operation. Downstream processing costs are also held at bay, as every ton of material that runs through mill and floatation tanks will add costs to the concentrator’s recovery process. The scans also assist in pinpointing the exact source of dilution even in areas where this may not be so easy to discern.

An overall reduction in production cost and increase in quality volumes can also reduce the necessary future investments in tramming equipment, especially when capital planning factors in a standard dilution percentage.

The cost of exploration and infill drilling is sometimes difficult to relate to the final blasted product. Fuller utilisation of geological information is accomplished by continuous comparative analyses between production drill holes, the blast outline, the ore body model and the stope design.

Sequential scans and continuous measurements ensure that “shadows” are eliminated and the stope status is updated with a progressively updated geometric shape that leads to more accurate extraction tonnage estimates and mining reconciliation.

Recommendations

The equipment mentioned here is often quite expensive, and since they are considered measurement devices their potential gain is often overlooked. This perspective needs to be altered, and the impact of changing a mine’s cost structure by controlling what comes out of the ground needs to be understood. Surveyors also ought to take advantage of their geometric analysis knowledge to get involved in improving a mine’s bottom line.

References

[1] E. Villaescusa, 2004 “Quantifying Open stope performance”, Western Australia School of mines.
[2] AQM copper Inc, 2013 “Copper fundamentals”, http://www.aqmcopper.com/s/Copper Fundamentals.asp
[3] Mario A. Morin 2001 “Underground Hardrock Mine Design and Planning- A System’s Perspective” Queen’s University Kingston, Ontari, Canada.
[4] Remondino Fabio, “From point clouds to surface”, International Archives of the Photogrametry, Remote sensing and Spatial Information Sciences, Vol XXXIV-5/W10

Contact Kalumba Bwale, Lubambe Copper Mine, kalumbab@lubambe.com