There is renewed talk of mechanisation on South African platinum mines. By tracking attempts to mechanise the labour process in platinum mines, this briefing explains why mechanisation has not yet and is unlikely to rapidly replace conventional, hand-held mechanical rock-drill technologies at the rock face and the rail-track transportation of miners, materials and equipment underground. The engineering literature shows that full mechanisation – via diesel-powered machines with rubber wheels – and hence a ‘trackless’ technology – which would significantly reduce the size of the current mass-based labour force, the rock drill operators (RDOs) in particular – is fraught with difficulties and challenges. This is due primarily to a complex geology which defines mining methods. Mining methods in turn define the most appropriate technologies to employ. The technologies employed in turn powerfully shape the size and composition of the labour force. While the leading corporate miner, Anglo Platinum, intends to divest itself of its older labour-intensive platinum mines and invest in fully mechanised trackless technologies, the evidence presented here suggests that the continuation, as on gold mines, of a mass-base labour force dominated by the RDOs remains the prospect for mining employment for the foreseeable future.
With the platinum belt at its epicentre, and in the wake of the 2012 strike wave, Marikana and the longest-ever strike in South African history in South Africa in 2014, mechanisation and its social consequences are again on the agenda. Allegedly as a direct result of ‘labour militancy’ – and an increased wage bill resulting from collective worker action – talk has again turned to mechanisation (Reuters 2014). One claim goes that: ‘Twenty years from now, the way platinum is mined in South Africa will be radically different’ with companies either ‘mechanising their operations to reduce reliance on a largely unskilled workforce, or mining vast open pits’ (Seccombe 2014). Another report suggests that ‘only a few years from now’ a ‘new era’ of ‘revolutionary change’ will replace the 150-year-old ‘low-wage, high-employment’ scenario with a ‘high-wage, low-employment’ one which will ‘devastate rural areas’ (Sparks 2014). Another journalist reports online that mining companies could mechanise up to 80% of their operations (but which only applies to Anglo Platinum) with disastrous social consequences (Nxumalo 2014). Yet this is not the first time labour-related issues have been cited as reasons for mechanisation. A decade ago the ‘increased urgency’ to introduce mechanised drilling machines was cited as the effect on mining of HIV/AIDS, which had a severe impact on the practically skilled mining workforce (Pickering 2004, 429). The sense of urgency has again recently been expressed (Vogt 2014).
Regarding the current talk, the National Union of Mineworkers (NUM) has noted the 16,000 jobs recently shed and expressed concern about further job cuts ‘as mines mechanise’ (Whittles 2014). The National Council of Trade Unions has warned of ‘more poverty and more inequality’ if mines mechanise simply to replace people (Nxumalo 2014). The Association of Mineworkers and Construction Union ‘understands that mechanisation can undermine its collective power’, suggests researcher Crispen Chinguno (Reuters 2014).
To put these claims into perspective, this briefing tracks the engineering literature to show that, despite serious and sustained attempts, full mechanisation – by way of trackless technologies – has not been readily implemented. Over time, engineers and analysts continually invoke notions such as the ‘human factor’, the ‘social context’ and ‘resistance to change’ as major factors impeding the successful implementation of new mechanised technologies. These construals can only be read as identifying labour's collective resistance to mechanisation. While designed to improve both capital and labour productivity, mechanisation inevitably means cutting back on the size of the labour force. This briefing shows that the engineering attempts to introduce technological change have been considerably more complex than imagined and that platinum mining capital's recent hopes for accelerated mechanisation may have to be deferred.
First to be noted is what aspects of mining can be mechanised in modern capitalist mining enterprises. Mining has excavation at its centre. Excavation underground can be roughly divided into development and stoping. Development is the excavation of haulage tunnels to transport materials and personnel and to access the ore body and has been subject to considerable mechanisation. Stoping is the blasting of the ore-bearing rock face and has not proved readily amenable to mechanisation. The key reason for this is that the geology – the characteristics of the ore body, namely the ore type, reef width, reef grades and its ever-increasing depth – very largely determines the choice of mining methods. This applies to all mining, but to thin-veined ore bodies, such as those of both gold and platinum mining in South Africa, most particularly. Mining methods – whether underground or surface – in turn determine the technologies employed in the socio-technical labour process. This, in turn, determines the composition of the labour force – in terms of both its size and skill requirements (see Capps 2012). It is this causal chain, related factors and the range of practical engineering difficulties which continue to impact directly on mechanising the rock faces of both gold and platinum mines. By way of contrast, the deep seams of coal on South African collieries have been mechanised since the 1920s.
The geology of the Bushveld Igneous Complex
The Bushveld Igneous Complex (BIC) is a vast, 2000 million-year-old, 65,000 square-kilometre geological formation. It takes the shape of a bowed, albeit broken, geographical arc and has been well described in the technical literature (Cawthorn 1999; Cawthorn and Webb 2001; Webb et al. 2004). An outline of the significance, from both a mining and a mineral exploitation perspective, of the ‘limbs’ of the BIC and their ore bodies – the Merensky reef, the Upper Group 2 (UG2) reef and the Platreef – has also previously been briefly described in this journal (Capps 2012). Six elements constitute the platinum metals group found here – platinum, palladium, rhodium, iridium, ruthenium and osmium – with platinum and palladium being by far the most important (Cawthorn 1999, 481). The geological interpretations of the mineral distribution of these composite reefs, critical to mining, however, are not straightforward (see Cawthorn 2011).
The Merensky and UG2 ore-bearing reefs are similar, with the weighted average of the stope width of both two platinum reefs being between 0.6–0.8 metres thick (Bracher, van den Berg, and von der Linden 2003, 35). The stope width is the distance from hanging wall (roof) to footwall (floor). Ideally, only the ore body itself – and no waste rock – is extracted. These are cramped and exacting conditions requiring considerable commitment from the labour which works within these confines. The consequent need for narrow stope widths – achieved to date by the systematic employment of human labour power – has been the chief inhibitor of mechanisation. So similar are these reefs in mineral composition that ‘Anglo Platinum for years ha[d] a single reef processing strategy’ (Luke Zindi, mining consultant, personal communication, 29 October 2014). The Platreef, however, is a ‘massive’ reef often many metres thick and hence lends itself more readily to open-pit ‘truck and shovel’ mechanised mining, but which is also not straightforward (see Bakhtavar, Shahriar, and Mirhassani 2012).
Regarding the proportions of their platinum content, the three main reefs vary (Cawthorn 1999, 482). Estimates as to the grade of these reefs are measured by millgrade (what is actually recovered) or in situ grade (the actual head grade of the ore body); these two measures, with adjustments for losses during mining, are roughly comparable with both measures predicting low grams per ton estimates. The mineral content of these reefs varies on average from between 1.3 and 1.8 grams per ton on the Platreef and 2.1 to 3.2 grams per ton across the eastern and western Bushveld limbs on both the Merensky and UG2 reefs (see Ibid., 483). The value of the UG2 reef seam, however, has been assumed to be 6 grams per ton when specifying mine design criteria (Egerton 2004). The key point is, however, as in gold mining, that hundreds of thousands of tons of rock need to be broken, moved and processed from carefully selected sections of these reefs to ‘win’ profitable amounts of precious metals.
The dominance of conventional mining
The western and eastern ‘limbs’ of the Merensky reef have been mined underground since the 1920s. This was the sole source of platinum in South Africa until the 1970s, since surpassed by the narrowly underlying UG2 reef. While discovered in 1925, the unique Platreef on the northern ‘limb’ was initially mined using open-pit or open-cast mining methods, but was short-lived. In the 1990s open-pit mining was re-established on the Platreef, to which attention is again turning. Underground mining by conventional methods, however, continues to dominate platinum mining.
Conventional mining refers to hand-held drilling underground with rock drills by human operators – the RDOs. Stoping is the heart of underground production. Here, the support, drill and blast cycle of the ore body takes place in an interrupted temporal rhythm known as mining in ‘batches’ – as opposed to the ‘Holy Grail’ of continuous mining such as on the mechanised coal mines. On gold mines, the hand-held pneumatic rock drill technology has remained virtually unchanged for over a century and is the source of the social power of the RDOs who perform this job (Stewart 2013). For, after a century of mining: ‘The match between current technology [i.e. hand-held drilling] and current stoping systems is near perfect’ (Pickering 2004, 424). In the light of the recent strikes initiated by the RDOs, the industry's renewed hopes for mechanisation are not surprising.
Geology and mining methods
From a geological perspective, there is a remarkable continuity of layers of ore-bearing rock within the BIC (Cawthorn 1999), but from a practical mining perspective there are significant irregular geological ‘discontinuities, potholes and continuous undulations’ (Menasce and Thorley 2004, 215) in the Merensky and UG2 reefs. Such geological constraints go a long way in determining mining methods. This provides the short answer why, despite attempts at mechanisation, only roughly 15% of platinum mine rock faces have been fully mechanised using either low-profile or extra (or ultra) low-profile (XLP) trackless technologies. The term ‘low profile’ refers to engineering designs to build the most squat machines possible in order to mine the least amount of waste rock encasing the narrow veins of reef-bearing ore. Currently, for instance, Impala Platinum, a 16-shaft mining complex, has implemented such low-profile, trackless mechanised mining on the Merensky reef on two of its shafts, which account for between 12 and 14% of production (Impala Platinum 2014). Six years ago, Anglo Platinum had 17 sites where trackless mining ‘to varying degrees’ had been implemented (Harrison 2008, 293). While ‘still low’, mechanisation is still more advanced than at Impala Platinum and Lonmin and is certainly higher than ‘on other narrow reef industry segments (e.g. gold mining)’ (Ibid.). The push to mechanise mines nevertheless continues.
The benefits and social consequences of mechanised mining
Mechanisation is ‘safer and more productive than conventional hand held drilling’ (Pickering 2004, 433) as fewer operators are exposed to ‘the sharp end of the production face’ (Harrison 2008, 293) and have lower lost-time injury frequency rates when compared with conventional mining operations (Valicek et al. 2012). While unit labour costs are higher, with a more technically educated, but smaller, workforce, overall labour costs decrease (Willis et al. 2004). Where the integration of mechanised technologies has been achieved, the result is higher face advances (the rate of progress of the daily mining cycle) and better utilisation of both labour and mine infrastructural resources. More rock face is worked faster. This facilitates increasing the rate of introducing new technologies and leads to further face advances, a ‘self-reinforcing cycle of activities’ and reduced costs (Ibid., 117). In short, the issue of safety aside, what drives mechanisation are anticipated increases in labour efficiencies (in terms of square metres of rock mined per employee), the increase in square metres production per month (in terms of ounces and tons) and improved extraction ratios (in terms of grade actually mined in relation to the surveyed head grade of the ore body) (Valicek et al. 2012, 5).
The social consequences of successful mechanisation are twofold. While less labour is required, the negative impact of relocating rural communities asserts itself, when land must be acquired when mining open pits. This has been recognised (Hattingh, Sheer, and du Plessis 2010). With mechanisation, whether underground or on surface, a small, highly skilled and better-paid workforce is employed not only in production, but also in maintaining increasingly sophisticated equipment and machinery. With mechanisation, it has been argued, employment opportunities open up downstream in other industries (Ibid.). What is not recognised, however, is that these technologies are heavily dependent on international original equipment manufacturers. The additional employment resulting from and hence the ‘offsetting’ of having to reduce labour has, it seems, yet to manifest itself in the South African platinum mining industry.
While the geology of deep-level gold mines has stubbornly precluded mechanisation of the stoping rock face for over a century (see Stewart 2013), the experience of platinum mines largely followed this pattern until around 2003. In 2000, 4 million tons were produced by mechanised mines, which had climbed to 45 million tons by 2013 (Majoba 2013). The question still remains, however, of why mechanisation has been relatively slow to take root on the platinum mines and to what extent it is likely to be accelerated as a result of the recent revolt centred on the platinum belt.
Stope width as critical factor in mining
The mechanisation of stoping operations in narrow-reef ore bodies has been underestimated (Harrison 2008, 293). Despite limited successes since the mid 1980s, the mechanisation of narrow, thin-veined reefs has been slow. Mechanisation of existing mining operations is powerfully constrained by not only geological constraints. The infrastructure of mature mines generally precludes the unproblematic introduction of full mechanisation (Ibid., 293–295), which requires changing the entire socio-technical environment embodying a ‘holistic approach’ (Willis et al. 2004, 117). The new ‘selected blast mining’ process – yet to date only on one site – has apparently, however, shown promise that existing mature mines can go over to full mechanisation (Creamer 2014). On the other hand, mechanisation a decade ago under the CEO of Lonmin, Brad Mills, was costly, failed, cost him his job and was reversed, indicating complexity when mechanisation of production stopes is attempted.
The key advantage of conventional, hand-held drilling is the ability of human operators to control stope width. The RDOs are able to blast stope widths of 0.8 metres, which machines have yet to do.
Wider stope widths increase the proportion of waste rock blasted, resulting in ‘dilution’ of the head grade – in situ grams per ton. Mechanisation has historically been associated with wider stope widths and an increase of waste rock (Bracher, van den Berg, and von der Linden 2003; Pickering 2004) and generally ‘compromised the objective to maintain head grade to be as close as possible to the in situ reef grade’ (Bracher, van den Berg, and von der Linden 2003, 35). A key aim of mechanisation is consequently to keep stope widths narrow and hence ‘dilution’ (of the ore-bearing reef in relation to waste rock mined) to a minimum in order to maximise planned head grade (Bracher, van den Berg, and von der Linden 2003). In short, mechanisation is only attempted where aspects of ore bodies are regular and relatively wide (Pickering 2004). What should be clear is that narrow ore-bodies continue very largely to be mined by versatile and relatively cheaply paid crews of RDOs.
Attempts at mechanisation
Mechanisation at the rock face can be divided into explosive and non-explosive techniques (Macfarlane 2001). Non-explosive techniques have only been implemented in pilot projects or remain confined to the laboratory (see Bakhtavar, Shahriar, and Mirhassani 2012). As opposed to the conventional method of blasting, laser rock cutting, for instance, is not about to happen, according to Rod Pickering, a leading developer of mechanisation, of the Centre for Mechanised Mining Systems at the University of the Witwatersrand (Creamer 2014). Explosive techniques remain the norm, including attempts at mechanisation. Mechanisation can further refer to surface and underground operations.
The first trials at mechanising development haulages (access tunnels) underground were initiated on the gold mines in the late 1960s and on platinum mines from 1980 onwards. These attempts met with varying degrees of success (Hattingh, Sheer, and du Plessis 2010; Kendall and Gericke 2000; Willis et al. 2004, 117). This assessment has not significantly changed. Until 2003 it was suggested ‘there has been little mechanisation progress in the extraction of platinum from the Merensky and UG2 reefs’, with even Anglo Platinum, for instance, having then only extracted ‘approximately 4% of underground production by mechanised methods’ (Bracher, van den Berg, and von der Linden 2003, 37).
The first attempts on chrome mines on the Bushveld Complex were limited to cleaning (removing blasted rock) and ore transport operations using the trackless Toro 150 and later Toro 190 and GHH Aardvark loaders (Pickering 2004). The aim, however, is to mechanise the rock face itself and hence the introduction of trackless Stomec and Furukawa drill rigs into the stoping environment on chrome mines during the 1990s. Despite hopes to the contrary which were based on this success, the Tamrock Axera low-profile electro-hydraulic drill rig did not succeed on platinum mines, due not least to ‘different reefs’ on the UG2 on which it was employed (Ibid.). To repeat, while from a geological point of view the ore body is even and consistent, this is simply not the case when actually mining it. Even where hydro-powered drill rigs – which cut back on all labour, especially the number of RDOs required in the stopes – were successfully implemented by 2000, such as at Northam Platinum mines (Macfarlane 2001), in-stope mechanisation has not, even in the past decade and a half, been widely implemented. New technologies were, moreover, introduced ‘because it is fashionable’ and not due to clear analysis as to their potential benefits (Pickering 1999, 2, 2004, 424). Clearly, the introduction of trackless mechanised machinery into South African mining operations has been slow and marred by many challenges and setbacks.
The engineering challenges of mechanising the stope face are legion. For a start, the Merensky and UG2 reefs dip an average of between 18 and 20 degrees. Five years ago, the XLP suite of mechanised trackless (rubber-wheeled) equipment could only be implemented on reefs with a dip of 12 degrees and a stope width of 1.2 metres – an engineering feat nevertheless. The aim to develop XLP technologies to mine reefs with dips above 18 degrees and a stope width of between 0.9 and 1.6 metres on dips between 18 and 30 degrees (Harrison 2008) has yet to be achieved.
A further major challenge is that trackless technologies require a relatively smooth footwall (floor) for vehicles with rubber tyres. Yet, ‘in reality, there are no smooth hanging or footwalls and bumps make the situation worse’ (Menasce and Thorley 2004, 216). In short, where mechanisation has been more generally successful, this has largely been confined to haulages (tunnels) and development ends.
Hybrid forms of conventional and mechanised mining
Hybrid mining combines conventional hand-held technologies in the stopes and rock handling with mechanical scrapers with trackless mechanised technologies in the haulages and development tunnels. While experimental trials on gold mines to mechanise haulages goes back to 1969, these early attempts failed in the production environment. This was not only due to higher costs and the ‘low costs of wages and the inexpensive equipment associated with the hand-held jackhammer method’, but also because the early Promec T 260 SA ‘jumbo’ mechanised hydraulic drill rig could not compete with ‘six highly flexible and low-cost hand-held jack-hammers’ (Wilson and Taylor 1975, 151). While no longer the case, interestingly, a similar reason was cited from 1900 to 1930 (when the mechanical scraper settled the issue) by goldmine managers for continuing with hand-drilling by Jack Hammer Hands (i.e. ‘hammer boys’ – now called RDOs) instead of introducing the hand-held mechanised rock drill (see Moodie 1994).
The mix of conventional and mechanised mining has been successfully implemented on at least one platinum mine on the UG2 since 1984 (Pickering 2004). Pickering is probably referring to Anglo Platinum's Union Section mine, where the UG2 is remarkably consistent in both grade and spatial geometry, where both conventional longwall and trackless technologies have been employed and where favourable geological conditions provided a unique opportunity to compare two very different mining methods. It is not, however, simply geological conditions and engineering challenges which have contributed to the slow pace of mechanisation.
The ‘human factor’
The introduction of any technology traditionally designed to increase productivity and profitability (Bracher, van den Berg, and von der Linden 2003; Kendall and Gericke 2000; Macfarlane 2001) is embedded in a social context and of which the South African mining industry has long been aware (De Bruyn 1981). No fewer than four of six factors cited as ‘barriers to success’ to mechanisation were explicitly noted as ‘soft’ (i.e. human) factors (Willis et al. 2004). Even the ‘best technology’ is likely to fail without ‘fully understanding the human factors affecting the implementation’ of mechanisation (Hattingh, Sheer, and du Plessis 2010, 256). When introducing mechanised trackless technologies, the key managerial issue, Rob Willis, another leading promoter of mechanisation, argues, relates to human agency. An ‘appropriate leader’ or ‘champion’ is ‘vital’ to the success of any mechanisation implementation project (Willis et al. 2004, 120). More broadly than change management and managerial leadership, issues around ‘the availability of skills, training requirements, organisational structure, management, work planning and operation of the mine and relationships with supporting industries’ (Hattingh, Sheer, and du Plessis 2010, 255) are all implicated when mechanisation is introduced. Regarding labour specifically, an impeding factor has been a ‘mechanised mining skills shortage’ and ‘a dire shortage of artisans’ sufficiently skilled to ensure the maintenance of mechanised equipment – identified a decade ago and which continues to be the case (Creamer 2014). Personal experience strongly suggests the Trackless Mechanised Mining Method at JCI's Western Area Gold Mine in the mid 1980s failed not only due to excavating an excessive amount of waste rock, but because there was a lack of skilled diesel mechanics.
Open-pit mining
Given the difficulties of mechanisation underground, it is not surprising that the new hope lies in shifting more decisively to open-pit mining on the Platreef of the northern ‘limb’ of the BIC. The single most significant here is Mogalakwena Mine (previously Potgietersrus Platinum Mine), which covers an area of 137 square kilometres, uses the open-pit ‘truck and shovel’ mining method, excavates the ore-bearing reef down to 240 metres and where the ‘life of mine’ is expected to extend beyond 2060.
The immediate social consequence of open-pit mining, however, is the relocation of communities, which cannot be discussed here. Suffice it to say that, despite the extent of the vast oval of over 200 kilometres under mining rights and that theoretically the outcrop could be mined all way around its rim (where feasible), this is unlikely to happen. The more likely longer-term scenario is mining on surface and then sinking incline shafts to access the extensive ore body below (see Bakhtavar, Shahriar, and Mirhassani 2012).
Conclusion
By way of conclusion, it is not the case that companies will either mechanise underground or mine on surface. Conventional mining will rather continue to dominate until mining methods some time in the future combine open-pit and underground forms of mining – with the latter aiming at full mechanisation. Given this longer-term scenario, the composition of its workforce and the fate of rural communities are likely to be considerably more complex than current reports suggest.
The difficulties mechanisation has faced further suggest mechanised technologies at the rock face are not about to become the norm and replace conventional mining across the Bushveld Complex. Sceptics go further and argue that ‘tough conditions, difficult geology, labour concerns, high up-front costs and previously failed attempts to mechanise’ mean that ‘mechanisation remains a long way off’ (Ferreira-Marques and Lakmidas 2013).
While the sceptics are probably on the right track, the motivation to accelerate the pace of mechanisation, favourable market conditions prevailing, is likely to be enhanced in the decade to follow. The promoters of mechanisation will find greater traction within the industry. The technology to effectively mine narrow hard-rock ore bodies at gradients of between 18 and 30 degrees may even come to fruition. To hazard prediction, however, the pace of mechanisation will pick up, most particularly open-pit mining on the Platreef, but not meaningfully challenge the dominance of conventional mining over the next 20 years. This is despite the opening in June 2014 of Northam Platinum's underground Booysendal ‘bord and pillar’ mechanised ‘block’ or ‘room’ mining, which has cut out the RDOs and has – it is claimed – ‘none of the “exploitation” stigma that accompanies deep-level operations with their necessarily cruder working conditions’ (Hartley 2014, 3). One important social consequence is that ‘[o]ur guys are building houses, not shacks,’ according to Willie Theron, Booysendal's general manager (Ibid.).
Despite the likely emergence of a smaller, better-skilled and decently housed labour force such as at Booysendal, it must nevertheless be concluded that, given the century-long history of stalled mechanisation at the stope face in the gold mines and the experience of a generation of intense effort on the platinum mines, the prospect of fully mechanising the platinum mines is not an immediate one. Mechanisation remains fraught with numerous constraints and challenges and has proven much more difficult to achieve than initially envisaged. The social implication of this, as has been argued in relation to the gold mines (Stewart 2013), is the continuation of a mass-based – though more internally differentiated – mining labour force on the platinum mines in South Africa in the foreseeable future.