August 2015

Zambia’s Kitumba copper project

  • By M A Habte, A G Smith, G G Grocott and R Bertuzzi, Pells Sullivan Meynink

A geotechnical assessment of sublevel caving

The Kitumba Project, 200 km west northwest of the Zambian capital, Lusaka, is being evaluated by Intrepid Mines. Iron Oxide Copper Gold (IOCG) style mineralisation is interpreted and its geological setting displays similar characteristics to other IOCG deposits.

The mining method selection for the project evolved from a broad assessment of open pit versus underground at scoping level, to a detailed assessment of sub-level open stoping with paste fill and sublevel caving at pre-feasibility level. The sublevel caving (SLC) method was selected as the preferred mining method at the end of the pre-feasibility study.

The footprint of the proposed mining area is about 300 m by 450 m. The depth of mining starts at 200 m below ground level and extends down to 550 m. Access to the SLC areas is being provided via a box cut, linear and spiral declines. Four shafts are proposed to provide airways, ore passes and escape shafts.

Geological setting

The Kitumba Project is situated adjacent to a giant haematite replacement breccia system extending over a 25 km long north-south trending structural corridor, with associated large scale magnetic and gravity anomalies. The hematite breccia outcrops as a prominent north-south trending ridge of the Kitumba Hills, deep leaching of copper minerals has occurred with a supergene enrichment zone forming below 150 m.

The high-grade copper zone is largely associated with the supergene enrichment. Hypogene mineralisation is situated between below 350-400 m vertical depth from surface. Mineralisation is also spatially associated with faults and lineaments that cross-cut the region.

Kitumba is one of the most deeply and intensely weathered IOCG alteration systems in the world with oxidation occurring to 800 m in places. The deepest and most extensive alteration and oxidation occur in haematite breccia due to leaching associated with major faults.

The principal structural feature of the project area is the Kitumba Fault Zone (KFZ) which is a broad (50-100 m wide) north-south trending and steep westerly dipping structure whose western limits approximate the eastern margin of mineralisation.


Eight drilling campaigns have been completed to assess the copper mineralisation. The last two included a significant geotechnical component to assess the caving characteristics of the rock mass and support requirements for development workings. The geotechnical drilling campaigns comprised:

  • Geotechnical logging of 37 fully cored boreholes with a total cored length of 15 000 m (Figure 1).
  • The use of orientated core and acoustic televiewer (ATV) imaging techniques to assess geological structures. The ATV data included 21 boreholes with a total survey length of 6250 m.
  • Recovery of samples for geotechnical field and laboratory testing, which included UCS, triaxial strength, direct shear strength, indirect tensile strength, point load strength, petrographic analysis, drilling rate index, Cerchar abrasivity index and slake durability tests.


High quality coring and down-hole imaging has been critical to understanding the rock mass, particularly that associated with deep weathering.

Geotechnical model

Geotechnical conditions for the project are complex, with country rock comprising calcareous siltstone and argillite intruded by quartz-feldspar porphyry granite which in turn has been extensively intruded by a feldspar porphyry diorite/syenite complex. Significant alteration over-prints all.

The geotechnical model is key to the geotechnical interpretation, numerical analysis, mining method, support design, and is also instrumental in communicating all the above to mine planning engineers and mine management. Elements of the geotechnical model include:

  • descriptions and distribution of rock classes based on the predominant rock mass weathering grade
  • structural domains and their distribution
  • large-scale structures, particularly faults
  • adopted engineering parameters.

The effect of weathering on rock mass quality is a significant geotechnical issue and forms the basis of the rock mass classes (Table 1):

  • Class 1: Fresh and slightly weathered rock
  • Class 2A: Moderately weathered (excluding breccia)
  • Class 2B: Moderately weathered (breccia)
  • Class 3: Extremely to highly weathered country rock
  • Class 4: Fault zones >2 m thick.


These rock mass classes have been divided into structural domains. The structural domains represent zones of rock mass with similar structure, particularly orientation. Figure 2 shows an interpretive geotechnical cross-section showing distribution of the rock mass classes and major structures. This figure shows the rock quality improves towards the west with increasing distance from the KFZ, while the development workings (declines and ventilation shafts) are largely located in Class 1 rock.


The structural model for the Kitumba deposit comprises:

  • Kitumba Fault Zone (KFZ) is a broad zone of general westerly dipping sheared rock mass to the immediate east of the ore zone, and which appears to flatten with depth towards the base of the ore zone. The KFZ is considered to be the principal structural feature dominating the Kitumba deposit. The current interpretation of the KFZ is it’s an anastomosing or anabranching of numerous faults and sheared zones in which fault slices and blocks of relatively intact materials are imbricated and interwoven with sheared zones and broken formations.
  • Two major, flat-lying to gentle westerly dipping faults which pass through the lower elevations of the ore zone.
  • A major easterly dipping fault that forms the contact between sediments and granitic rocks to the west of the ore zone.
  • The dominant joint sets are northeast and southwest dipping.

The rock mass generally contains numerous sheared and crushed zones and seams suggesting that fault-affected rock occurs in many areas outside of the currently interpreted faults.

The overall structural model tends towards supporting a southwest-dominant stress orientation. Currently available evidence suggests the deposit is not affected by high stress conditions, but verification of this aspect will be evaluated during subsequent project stages.


The geotechnical assessment involved the use of both empirical and numerical methods to assess the stability of production openings and the support of development workings.

Preliminary assessment of caving used Laubscher’s empirical mining rock mass rating (MRMR) method, which correlates the rock mass quality and the size of the underground opening required to initiate caving. The results of this empirical assessment indicate caving initiates for 70 m square openings.

Two (2D) and three (3D) dimensional numerical analyses were carried out to assess the overall rock mass response to sublevel caving, taking into account the spatial distribution of rock classes and structure. The analyses provided detailed information on cave initiation, cave propagation, stress redistribution and subsidence zones. Key features of the modelled SLC layout are:

  • sublevel height of 25 m and sublevel drift spacing of 14 m
  • ore drives 6 m wide and 5 m high, resulting in pillar width of 8 m
  • twelve levels of extraction from level 1 at 1205 mRL to level 12 at 930 mRL.

The main objective of the 2D finite element analysis, which used the program Phase2, was to assess caving initiation. An inherent limitation of 2D analyses is that the excavation is assumed to be effectively continuous in the out-of-plane direction. The approach adopted was to correct for this limitation using hydraulic radius to estimate the actual plan area required to initiate caving.

The emphasis of the 3D explicit finite difference analysis using the program FLAC3D was to assess the initiation and propagation of caving and surface subsidence. Criteria based on rock mass yielding, strain and displacement limits were used to predict cave initiation, propagation and the extent of subsidence zones. A large strain continuum modelling approach was adopted to capture large deformations associated with caving. Typical output from the 3D modelling is shown in Figure 3.


Numerical modelling indicates complete caving is likely to occur when the width of the mined area exceeds 60 m to 80 m. The predicted extent of the subsidence zones at the end of mining is:

  • the caved zone is about 300 m wide at the ground surface
  • the zone of large-scale fracturing is about 450 m wide (also incorporating the caved zone)
  • the surface subsidence zone is about 480 m wide (also incorporating the caved zone and fracture zones).

The development workings, including the spiral decline and shafts, are anticipated to be located outside of the subsidence zone. Five support types have been designed for the decline comprising increasing quantities of rock bolts, mesh and shotcrete with decreasing rock mass quality.


Although investigations have provided information on the major structures in the Kitumba deposit, there are limits to our understanding and some uncertainty still exists. This uncertainty can be managed through targeted geotechnical investigations during mine development.

The deep weathering and alteration in the deposit is expected to provide geotechnical challenges irrespective of the mining method adopted. The proposed shafts and declines will be constructed in these units and their economic construction will be important to the success of the project. Improving the understanding of weathering and alteration will greatly benefit their design and construction.

In situ stress is an important determinant of the performance of SLC and of the development drives, shafts and declines. Specific stress measurements early in the mine development will assist in its design.


The permission of Intrepid Mines to publish this paper is acknowledged.

Share This Article