Preliminary assessment of seismic hazard and risk in the Bushveld Complex platinum mines

Transcription

1 Safety in Mines Research Advisory Committee Final Project Report Preliminary assessment of seismic hazard and risk in the Bushveld Complex platinum mines A.v.Z Brink, T.O. Hagan, M.K.C. Roberts, A. Milev Research Agency: CSIR: Division of Mining Technology Project Number: GAP 711 Date: November

2 Executive Summary The objective of this project was to identify conditions that constitute seismic hazards and to assess the risk associated with such hazards in the platinum mines in the Bushveld Complex. A preliminary assessment of the current risk of seismicity as well as an assessment of the future risk was undertaken. A further objective was to review current regional and local support strategies with respect to the risk of damaging seismic events. A rock engineering risk assessment technique was developed. This technique was applied on a trial basis. Mining-induced seismicity results from a complex interaction of ambient and mining-induced stresses acting on a rock mass that is intersected by a variety of geological weaknesses and discontinuities. The challenge is to recognise and quantify the seismic risk associated with all potential sources, such as geological structures, remnants, pillars and face bursting. A further objective is to control (manage) the observed risk by optimum support, planning and sequencing of mining operations. The project assessed the overall seismic risk in the Bushveld Complex by using data from existing mine seismic networks. A limited amount of data was available at the time of compiling this report. No mine was operating a mine wide system, but at least 4 systems were in the process of installation. Interviews were conducted with the rock engineering staff at six mines. The mines with current or potential seismic problems were selected. The interviewees were unanimous in their view that currently there is no significant seismic risk in the Bushveld Complex mines. They were, however, equally unanimous that an increased seismic risk will be experienced in the future when mining at greater depth and with larger mined-out areas. Their opinion was that seismicity was seen to only occur in or around highly stressed pillars/remnants. The interviewees felt that geological structures (faults or dykes) had a small impact on seismicity. Potholes were also not seen as a potential source of dynamic failure but rather areas with poor jointed hanging and therefore vulnerable to rockfall should sufficient seismic energy by released in the vicinity. A fatal accident at Rustenburg Mine A Shaft provided the researchers with the opportunity to install a small Ground Motion Monitor at the accident site. The data recorded at the site provided some useful insights into seismicity at this site, and is different to the opinions of the mine rock engineers as the events tended to cluster along a geological structure. 2

3 Acknowledgements The authors would like to express their gratitude to the Safety in Mines Research Advisory Committee (SIMRAC) for financial support of project GAP711. The excellent co-operation of the rock engineering personnel on the platinum mines, which has contributed towards this project, is gratefully acknowledged. In particular, Anglo Platinum is thanked for permission to publish the results from the investigation into seismicity around the site of a fatal accident 3

4 Table of Contents Executive Summary... 2 Table of Contents... 4 Table of figures Seismic Risk Assessment An Overview Introduction References Earlier seismic work in the Bushveld Introduction Analysis of mining induced seismicity Current status of seismic monitoring in the Bushveld and the perceived seismic risk Introduction Methodology Response to questions Interim conclusions based on interviews Rock Related Risk Assessment Techniques on Platinum Mines Introduction Methodology Conclusions Future Work References Appendix A seismic risk assessment trial on a typical Bushveld mining scenario Introduction An example seismic risk assessment exercise P 1 - Level of ground motion P 2 - Vulnerability of the excavation to ground motion P 3 p 1 Exposure of people P 4 Quality of information

5 5.3 Conclusions References Observations at Rustenburg mine A Shaft Introduction Observations Conclusions Stope Support in Rockburst Prone Platinum Mines Introduction The design of rockburst resistant stope support for the Merensky reef References Conclusions and Recommendations Appendices A site visit report - Rustenburg Mine A - rockburst 29/6/2000 Merensky reef Observations from seismic recordings at Rustenburg Mine A Introduction Objectives Observations from seismic recordings Interpretation Interim report for October Introduction Observations from seismic recordings Observations

6 Table of figures Figure Spatial distribution of seismic events recorded at 10 Shaft Wildebeestfontein North Mine. (from SIMRAC GAP 027)...12 Figure Dip section along raise line W1366 W1666, showing location of seismic events with respect to Merensky Reef elevation (SIMRAC GAP 027)...13 Figure Cumulative number of seismic events during the period 9/93 to 5/ Figure Cumulative seismic moment recorded during the period 9/93 to 5/ Figure A diurnal distribution of seismic activity in the risk assessment area...29 Figure An energy capacity deformation design chart for rockburst conditions...41 Figure SDAII plot of blast-on Ebenhaeser rockburst support system...42 Figure The force deformation curve of the Ebenhaeser yielding timber elongate Figure Schematic presentation of the mine plan for the area visited by Miningtek team on 5 July Figure Gutenberg - Richter display of recorded events and blasts...49 Figure Recorded events for the period 10 July to 30 July Figure A frequency magnitude plot of events...51 Figure A simplified plot of seismicity during October

7 1 Seismic Risk Assessment An Overview 1.1 Introduction The primary output of GAP 711 is the evaluation of current and future seismic hazards, the provision of concepts and/or tools for assessing seismic risk in the Bushveld Complex platinum mines and the effect on support strategies The project was addressed from a risk assessment point of view and in particular the assessment of seismic risk to the safety of the underground worker in platinum mines. As in GAP 608, (Survey and assessment of techniques used to quantify the potential for rock mass instability SIMRAC 2000), it is important to qualify the term seismic risk as opposed to seismic hazard. The generic term, hazard, is defined by the Mine Health and Safety Act No 29 of 1996 and interpreted by the Tripartite Working Group (SIMRAC, 1998) as: Hazard is a physical situation, object or condition, which, under specific circumstances has the potential to cause harm. Risk is seen by the Act (and by the Tripartite Working Group) as a measure of the likelihood that some specific harm arising from an incident will occur. If this general definition is applied to seismic events, the seismic hazard will then be those seismic events that have potential to cause harm. A greater hazard will imply the potential to cause greater harm. A study of what the maximum event magnitude in an area might be is typically a seismic hazard determination. The aspects that constitute seismic risk, their impact on mine safety and input to risk assessment are: Seismic hazard The seismic event Recognition of seismic hazard. Coupling between the seismic source and mine excavation. Characteristics of excavation Local stress distribution Site effect amplification Support behaviour and standards of installation Local ground condition Exposure of people Data acquisition Seismic monitoring Mine layout/design Panel characteristics 7

8 Experience reference Data interpretation / visualisation Seismic early warning Integration of seismic monitoring and modelling. Interpretation skills A possible risk assessment methodology is outlined in the GAP 608 final report. This is shown in Table 1.1.1, in which the proposed procedure for evaluating seismic risk is summarised. The assessment is made in four categories, namely Level of Ground Motion Vulnerability of the Excavation to ground motion Exposure of people Quality of information At present each category has effectively the same relative importance or weighting. An overall seismic risk assessment is achieved by combining individual category ratings. The process of achieving a single rating is discussed in section 5.2. Each category is subdivided into various parameters contributing to that category. The parameters are rated individually and averaged to provide a category rating. The risk rating ranges from 1 to 5, where 1 implies a very low risk and 5 an unacceptably high risk. In this project a preliminary evaluation of the applicability of a similar methodology in the Bushveld Complex platinum mines is carried out. 8

9 } Rating Risk Assessment Category Parameter Parameter rating P 1 Level of Ground Motion M max Distance from source Mean Return Time (Frequency) p 1 p 2 Seismic/Time Distribution p 3 Vulnerability of Excavation ERR Geology Support p 1 p 2 p 3 P 2 to Ground Motion (or Falls of Ground) Ground condition Escape ways (Site Effect Amplification) p 4 p 5 (p 6 ) P 3 Exposure of people People/Time distribution p 1 Mine plans/structure/layout p 1 Quality of Seismic Monitoring Early Warning p 2 p 3 P 4 Information Assessment interval and volume Experience reference p 4 p 5 Communication p 6 Table A proposed risk assessment methodology 1.2 References Brink, A.v.Z., Hagan, T.O., Spottiswoode, S.M., Malan, D.F., (2000) Survey and assessment of the techniques used to quantify the potential for rock mass instability SIMRAC Report GAP 608, Department of Mineral and Energy Affairs, South Africa. 9

10 Haile, A.T., Jager, A.J. (1995) Rock mass condition, behaviour and seismicity in mines of the Bushveld igneous complex, SIMRAC Report GAP 027, Department of Mineral and Energy Affairs, South Africa. 10

11 2 Earlier seismic work in the Bushveld 2.1 Introduction In SIMRAC GAP027 the rock mass condition, behaviour and seismicity in the mines of the Bushveld Igneous Complex was investigated. (Haile and Jager, 1995). In an appendix to this report Van der Merwe reported on some aspects of the seismicity at Wildebeestforntein North Mine and at Bafokeng North Mine. 2.2 Analysis of mining induced seismicity Van der Merwe in GAP027 (Haile and Jager, 1995) described the application of a PSS system at No 10. Shaft Wildebeestfontein (Impala). The system covered an area of 200 m by 200m, at a depth of approximately 900. The system accuracy was excellent. The recorded seismograms were each individually inspected, the P- and S- wave arrivals picked manually, and the foci located. The spatial distribution of all events recorded since commissioning of the PSS seismic network until August 1994, is shown in Figure overleaf. The following observations in terms of spatial distribution, were made by Van der Merwe: The majority (49.5%) of the events located on pillars in mined out areas, indicating that these yield pillars are to some extent failing. Most of the larger events recorded plotted at such locations, and some pillar damage and pillar "punching" into the footwall has been correlated with these events. Some 35.5% of the recorded events located at active stope faces. These were generally quite small events, but some exceptions have occurred, for example an event of magnitude +0.8 was correlated with a substantial fall-of-ground. Some 15% of the events located in the back areas of active and older panels. A number of these could be correlated with falls-of-ground. Most events located relatively close to the reef plane (see Figure 2.2.2). A poor Z- definition of a seismic system would normally tend to increase the vertical scatter of the event locations 11

12 Figure Spatial distribution of seismic events recorded at 10 Shaft Wildebeestfontein North Mine. (from SIMRAC GAP 027) In terms of the temporal distribution of seismic events it was noticed that; The relatively straight slope of the plot of cumulative number of events against time (Figure 2.2.3) shows that the number of events per time unit remained steady over the period from September 1993 to May 1994 (inclusive). However, the graph of the cumulative seismic moment released during the same period shows an increase in the rate of moment release (Figure 2.2.4). This therefore indicates that the average moment released per event increased as mining progressed, which in turn implies a relatively higher risk and perhaps more damage in more mature working areas. 12

13 Figure Dip section along raise line W1366 W1666, showing location of seismic events with respect to Merensky Reef elevation (SIMRAC GAP 027) Figure Cumulative number of seismic events during the period 9/93 to 5/94 (SIMRAC GAP 027) Figure Cumulative seismic moment recorded during the period 9/93 to 5/94 (SIMRAC GAP 027) 13

14 As for the magnitude distribution of seismic events the events ranged from ML= -1,8 to ML= +1,1. A b-value of 0,66 projected a largest event size of ML = 2,2, but Van der Merwe argued that his relative small data set does not allow for a reliable indication of M max. Van der Merwe had some success in using P-S moment and P-S radius and moment tensor inversions to estimate source mechanism. The majority of events fell in a transition zone' between a 'shear' and a 'crush' event. Some of his conclusions are also listed as follows: From September 1993 to August 1994 some 254 mining induced seismic events were recorded at the PSS site. The magnitudes of these events ranged from -1.8 to Frequency / magnitude analysis indicated that the potential for seismic events larger than M max = +1.1 to occur in the study area is limited. Although the largest number of events occurred during the blasting period and directly thereafter, the larger magnitude events tend to occur outside this time period. The majority of the larger events occurred on the pillars and remnants, probably due to the (partial) failure and / or punching into the footwall of these pillars. Some of the larger events occurred in the back areas of stopes, and generally coincided spatially with falls-of-ground. Few larger events occurred at active stope faces, but there was at least one instance where it was observed that an ML = +0.8 event was associated with a significant fall-ofground in the face working area. The up-dip panels between W1566 and W1567 displayed an anomalous seismic pattern, with most events occurring at the side abutments of the panels. Most events located at, or very close to the reef plane. Interpretation of seismic source parameters such as stress drop and energy index, demonstrates the build-up of conditions of higher stress at pillars and at the edges of remnants and regional/stabilising or barrier pillars. Some success was encountered with inversion for the seismic moment tensor, giving information about the seismic source mechanisms: Mainly shear type events located some distance from the reef plane in solid rock, and are probably related to geological structures; "Explosive" type events were all correlated with production blasts; and "Implosive" or "crush" type events could generally be correlated with falls-of-ground or (partial) failure of stope support yield pillars. The relevance of these conclusions in terms of seismic risk can be summarised as follows: The probability of having events larger than M L =1,1 is limited at the mining depth monitored. Extrapolating the observed frequency/magnitude distribution indicates that the probability of having an event in the order of, or larger than, M L =2.0 is very small. Purely from a simplistic event magnitude criterion, an M L =2.0 is significant because this is typically the magnitude of event contributes largely to the accidents and fatalities in the deeper level RSA gold mines. Geological structures seldom pose a significant seismic risk in the Bushveld. 14

15 The majority of events are associated with pillars and remnants therefore only personnel in the vicinity of these pillars/remnants are exposed to the associated risk. An increase is observed in the rate of change in the cumulative seismic moment as in Figure This implies a relative higher risk in more extensively mined areas where induced stresses are higher. 15

16 3 Current status of seismic monitoring in the Bushveld and the perceived seismic risk 3.1 Introduction Seismic monitoring in the Bushveld was initially only provided by the national seismic system provided by the Council for Geosciences in Pretoria. A number of events were recorded during the years, but it is debatable how many of these could be classified as mining induced events. It is well known that events do occur in this region outside the immediate mining districts. Table lists the larger events recorded by Council for Geosciences. ID Date/tim e 6/29/90 3:23:48 10/22/90 2:40:06 1/13/92 5:50:47 2/3/93 1:20:24 3/24/94 5:38:56 11/2/94 2:43:04 4/18/95 23:55:05 8/8/95 7:46:05 7/25/97 23:34:00 3/10/98 0:10:11 10/8/98 22:44:07 Latitud e Longitud e Region_and_Descriptio n Rustenburg_Region Rustenburg_Region Rustenburg_Region THABAZIMBI_REGION Rustenburg_Region Rustenburg_Region THABAZIMBI_REGION Rustenburg_Region Rustenburg_Region Rustenburg_Region Rustenburg_Region 3.2 MI Table A summary of events recorded by Council for Geosciences for the last ten years In the 80's a number of GENTEL systems were installed on various mines. These systems were single site surface installations, but were still effective in recording mining-induced events. Accurate locations (say within 50 m) were not possible. Three GENTEL systems were originally installed at Impala (Bafokeng North #12 Shaft, Wildebeestfontein South #9 Shaft, Wildebeestfontein South #10 Shaft) and at Angloplats (Turffontein, Townlands and Frank shafts). Most of these systems are reported to be in operation. Towards the end of 2000 two PRISM systems (at Amandelbult and at Impala) were in operation. Another PRISM system is being installed (at Union section), while two ISS systems are being installed (at Northam and Rustenburg, Frank Shafts). A number of these mines were visited with the objective of collating some of the experience of the Rock Engineers in terms of seismicity on their respective mines. The discussions only 16

17 related to seismic incidents and their perception of the associated risk as well as future seismic risk. Chapter 4 describes similar interviews with the respective rock engineers but with the emphasis on more general rock engineering risk assessment techniques as practised on their mines. During this project a fatality apparently related to a seismic event occurred at Rustenburg Mine A. The researchers were requested to carry out an accident investigation that included a smallscale seismic monitoring installation around the accident site. With the permission of the mine the results of this investigation are included in this report. 3.2 Methodology The mines visited were: Rustenburg Platinum Mines Rustenburg Section Rustenburg Platinum Mines Amandelbult Section Rustenburg Platinum Mines Union Section Impala Platinum Limited Northam Platinum Mine These mines were selected because they have some history of seismic incidents and are establishing seismic monitoring systems on their mines. Similar questions were asked at all the mines and they can be summarised as follows Do you experience any seismicity? Do you experience any seismic related losses (from a safety and production perspective)? Have you recorded any seismic related accidents or fatalities? Do you record seismic incidents other that those linked with reported accidents? What is/are the seismic event mechanism(s)? What is the influence of potholes on the occurrence of seismicity? Do you believe that seismic risk will change in the future and what would be the driving factors for such change? 3.3 Response to questions The individual rock engineering managers/practitioners that responded to the above questions were Karl Akermann, Martin Pretorius, John Potgieter, Wouter Hartman, Lucas van Aswegen and Gavin Potgieter. General consensus was clear on all the responses and some presenting responses are quoted in this report. Question Answers Do you experience any seismicity? Everyone agreed that there is seismicity, but do not want to call it rockbursting. 17

18 Comment Seismicity, to a larger or smaller extend, is been experienced at every mine included in this interview. There is some 'misunderstanding' in that sometimes only the classic forms of seismicity, i.e. slip on a geological plane of weakness, are considered as seismicity. Face or strain bursting is easily ignored as a seismic problem or even contributing to seismic risk. Question Answers Comment Do you experience any seismic related losses (from a safety and production perspective)? In general the consensus is that no seismic related losses are experienced. Everyone can refer to some seismic incidents but not to associated losses. See above comment. Question Answers Comment Have you recorded any seismic related accidents or fatalities? With the exception of one mine the answers were negative. At least four fatalities were quoted that was referred to as being associated with strain bursting on isolated pillars. The fatality at Rustenburg Mine A shaft is discussed later in this report (Chapter 5). Question Answers Comment Do you record seismic incidents other that those linked with reported accidents? No A similar problem exists in the gold mines. The absence of incident reporting is the main reason for the perception that no seismic losses occur. Question Answers Comment What is/are the seismic event mechanism(s)? All seismicity is associated with pillar/foundation failure The implication of this is that the seismic risk is confined to the back areas and that the underground staff at the working face is not exposed to seismic risk. Question Answers Comment Question What is the influence of potholes in the occurrence of seismicity? All the respondents stated that potholes are not a source of seismicity other than potentially being a highly stressed pillar/remnant. It is also recognised that areas close to potholes will be more vulnerable to dynamic ground motion, due to the relatively bad hangingwall conditions when mining close to potholes. Do you believe that seismic risk will change in the future and what would be the driving factors for such change? 18

19 Answers Comment All respondents agreed that seismicity and therefore seismic risk will increase with increased depth and with extensively mined out areas. This is only based on 'gut feeling' and while the researchers agree with this, they could not find any proof of increased seismic risk with depth. 3.4 Interim conclusions based on interviews The response to the interviews had lead the interviewer to come to the following conclusions: Current seismic risk is negligible in the Bushveld All seismicity is related to the failure of pillars/remnants. Geological features do not cause seismicity Seismic risk will increase in the future The experienced gained from an accident investigation at Rustenburg Mine A Shaft provided a different perspective on the above interview-based conclusions and is discussed later in this report. 19

20 4 Rock Related Risk Assessment Techniques on Platinum Mines 4.1 Introduction In the SIMRAC report GAP 608 the assessment techniques for seismic risk in gold and platinum were described. It was realised that seismic risk comprises more than the probability of a large event occurring, but also the vulnerability of the excavation to a seismic event. Such vulnerability is well defined in the standard rock engineering risk assessment techniques (other that seismic). For completeness the interviews and findings relating to platinum mines are repeated here. 4.2 Methodology The mines covered were: Rustenburg Platinum Mines Rustenburg Section Rustenburg Platinum Mines Amandelbult Section Rustenburg Platinum Mines Union Section Impala Platinum Limited The techniques used are basically similar in that they cover the assessment of both regional and local rock engineering parameters in order to reach a risk assessment level. The complexity, however, varies considerably. Each technique is customised to the particular mine, geotechnical conditions and purpose of the assessment. Weightings are generally used to emphasise the relative risk level of parameters. In most cases the assessment is based largely upon data readily available on surface such as that measurable on a mine plan, seismic data, results of numerical modelling etc. The inclusion of significant, up to date, results of underground inspections is rare. Where this is the case, i.e. where resources exist to collect such data, the results of the assessment and efficiency of the corrective actions appears to be greatly improved. The management of the assessed risk takes a number of forms on the various mines and includes discussions and decisions (at planning meetings) regarding required actions to manage high risk levels; drawing up of summary sheets (normally monthly) to facilitate discussion at planning meetings; input of assessments on to a database mainly for the determination of trends; use of summary sheets as part of the production planning process; reporting of risk ratings to senior management; 20

21 inspections by senior personnel and rock engineers of very high risk panels; declaration of special precautionary areas; presentation of monthly risk assessment results to the Mine Overseer of the section prior to planning meetings; a zero tolerance approach for high risk levels, i.e. panels are stopped; a monitoring programme to check on compliance and efficiency of the recommended actions; use of trained observers underground who are empowered to make on the spot decisions regarding corrective actions or whether it is safe to continue mining; and active involvement of trained production personnel in risk assessment to the extent that they are empowered to recommend corrective action. The first nine points are all important to expedite and facilitate the risk management process. The last three actions are rarely undertaken but are strongly recommended to improve the efficiency of both the assessment and the management of rock related risk. 4.3 Conclusions Listed below are some of the more important shortcomings of the risk assessment and management systems in operation at present. These were highlighted during interviews on the mines: Subjectivity. It is extremely difficult to ensure even reasonable consistency in the assessment of risk levels. This is despite the detailed procedures and careful weighting of parameters etc. that are in place in some cases. The general lack of assessments done by personnel external to the mine. It is strongly recommended that external audits be performed regularly to ensure that risky practices condoned (in some cases unintentionally) by the mine be minimised or eliminated. Lack of sufficient rock engineering resources particularly trained and certificated personnel. In most cases the part of the assessment that is possible, using readily available data on surface, is reasonably well implemented. Other essential data from underground in the form of up to date assessment of rock conditions, support standard compliance and effectiveness etc. is generally not adequately carried out. On two of the mines visited this problem has been largely overcome by the use of trained observers and production personnel. One of the most important factors, namely that of rockburst risk, has proved very difficult to quantify. Lack of follow-up on recommendations and action plans is a problem. Resistance to change has been a problem i.e. after the introduction of new risk assessment and management systems. Uncertainty with regard to the position and potential seismic response of major geological discontinuities may result in excessively high-risk ratings. 21

22 A risk assessment system must not be seen as a black box providing exact solutions. A degree of engineering experience and judgement is still necessary. Results and advantages of rock-related risk assessment and risk management as seen by users in the field at the moment are: Past problems associated with the declaration of special areas have largely been cleared up. Systems are seen as important management tools forming an integral part of the planning process. There has been a marked improvement in communication between production personnel and rock engineers since the introduction of such systems. The systems do not replace underground trips but do allow rock engineers, managers and supervisors to focus on problem areas and be more pro-active. On some mines there has been a marked improvement in rock-related accident rates since the introduction of the systems. On other mines it is still too early to say whether there has been any direct effect. Significant improvement on two mines may have been as a result of the introduction of improved support systems that were brought in at the same time as the risk assessment system. On one mine a risk assessment system has been very useful in determining whether certain remnants are safe to mine or not. The formal approach of assessing risk ensures that problem areas are more effectively highlighted and addressed. 4.4 Future Work The need for future work to improve the systems in place on the various mines was recognised in all cases. This includes: the training of observers, safety and/or production personnel to assist with risk assessment and management; the introduction and implementation of follow-up procedures; initiating regular external audits; improving the techniques by making them less qualitative and more quantitative; inputting all information on to a database to allow trend analyses; introducing direct, up to date, underground data such as rock mass ratings, support compliance and efficiency, into the assessment; using instrumentation to detect unfavourable structures in the hangingwall and ahead of the mine face; finding ways to reduce subjectivity when assessing risk. 22

23 4.5 References Nicolau, N (1999). Stope Panel Rating System. Association of Mine Managers of South Africa. Papers and Discussions. Paper presented in August, To be published. Johannesburg. South Africa. MRAC Task Group. (1999). Guidelines for the Compilation of a Mandatory Code of Practice to Combat Rockfall and Rockburst Accidents in Metalliferous Mines and Mines other than Coal. Published at website Pretoria, 1999 (35 pages). MOSHAB, (1999). Surface Rock Support for Underground Mines Code of Practice. Mines Occupational Safety and Health Advisory Board. Western Australia. February pages. Akermann, K.A. (1999). A support design methodology for shallow scattered mining environments. SAIMM Colloquium on Successful Support Systems for safe and Efficient Mining, Mintek, Randburg, May

24 4.6 Appendix Name of system Contact and/or ref. Aim and background Applicable to:- Risk assessment parameters Amandelbult Section - Rustenburg Platinum Mines Risk Assessment Programme Akermann 1999, Karl Akermann Used to identify high-risk areas requiring special support specifications and recommendations. Shallow, scattered mining environments Main factors considered in the risk assessment are: - Hangingwall pyroxenite beam thickness Rock Mass Rating (RMR) Geological complexity Support spacing Excavation span Mining sequence Pillar size Water Amplats. Rustenburg Platinum mines. Union Section. Stope Panel Risk Assessment Gavin Potgieter A rock engineering resource problem sparked the need for the active involvement of production staff in risk assessment. Two-risk assessment spread sheets i.e. for the UG2 and the Merensky reef horizons were drawn up and are completed monthly by a trained shift supervisor. Aim is to identify high-risk areas requiring special support specifications and recommendations. Shallow, scattered mining environments particularly on the platinum mines. 14 weighted parameters are taken into account, they are: - Mining sequence Lead and lag distances Distances from holing Sidings Regional stability Yield pillar geometry Panel spans between pillars Distance to fault or dyke Mining on side of fault or dyke Orientation of face to geological structure Orientation of major joint set to face General hangingwall conditions Extent of mining spans an area overview Other i.e. anything else affecting risk Risk rating ranks as 1,2,3,4,5 being very good, good, average, poor and very poor respectively. Impala Platinum Limited Rock Engineering risk assessment Les Gardner, Noel Fernandez To identify problem areas and risks in stopes underground and to ensure action is taken to minimise the risk Relatively shallow, scattered mining environment particularly on the platinum mines Risk factors taken into account are: - Panel length Pillar width Leads/lags compared to adjacent panels Rock Mass Ratings (RMR) Support types (actual vs. planned) Compliance to support standards Rock engineering dept visits underground Adjustments for faults, dykes, etc. Impala has, at present, over 30 trained observers who measure support standard compliance and determine RMR s underground. Rustenburg Platinum Mines Rustenburg Section Stope Plans Risk Assessment Johan Lombard, Lando Sloane To identify problem areas and risks in stopes underground and to ensure action is taken to minimise the risk. Relatively shallow, scattered mining environment particularly on the platinum mines. Risk factors taken into account are: - Panel length Pillar width Leads/lags compared to adjacent panels Rock mass ratings Support types (actual vs. planned) Rock engineering dept visits underground Adjustments for faults, dykes, etc. A panel starts with a score of 100. Marks are deducted for noncompliance and the final score is expressed as a percentage. The lower the score the higher the risk. 24

25 Risk management Shortcomings and Future work Results All relevant geological information collected by the geology and rockengineering departments is transferred to a plan and discussed in detail at planning meeting. High-risk panels are declared special precautionary areas and support recommendations are to reduce the risk. Highrisk panels therefore become medium or low risk panels in future risk assessments because special area instructions have been enforced. Compliance is assumed. Wedges/keyblocks bounded by serpentine-filled joints are considered potentially hazardous especially if water is present. Present rock engineering department resources are insufficient to ensure that all potentially hazardous wedges/keyblocks are detected. The drive now is to train the workers and supervisors to identify the hazards and therefore be in a position to take effective precautions. Highly significant safety and productivity improvements since 1996 have been largely attributable to an improved support system incorporating the use of pre-stressed elongates. The introduction of the Risk Assessment Programme together with hazard awareness training has assisted in ensuring that the safety and productivity improvements are maintained. Millionaire Shield achieved twice in two years. The risk assessment sheets are filled in monthly for producing panels. If a very poor rating is determined then the rock-engineering department is consulted immediately. A database is kept of all sheets. They are discussed in detail at a monthly planning meeting where the appropriateness of standard support recommendations to suit particular risk ratings is discussed. The shift overseer is also empowered to make additional recommendations that are checked at the planning meeting. Main problem was the lack of sufficient rock engineering staff to cover the mining area adequately. This prompted the training and use of the shift overseer in this respect. Some staff are more motivated than others to perform the assessment and the quality is affected accordingly. As in the case of the Amandelbult Section significant improvements in safety and productivity were realised when an improved support system was introduced. The current system has rally on been in full use for six months and although it must have contributed of late to the continued improvements, the extent of the contribution is difficult to assess. The form, filled in by the observer at the time of each visit, is checked by the Rock Engineer or Strata Control Officer and goes via the Mine Overseer to the Manager of the section. The detailed support compliance data, RMR s and layout risk factors are combined to determine an overall risk rating for each panel. This is summarised monthly for each section of the mine. The observers are able, at present, to visit 75% of the panels monthly on Impala mine. They are trained to a level where they can stop the operation if deemed necessary and/or make recommendations to minimise any observed risk. The ideal is to visit each panel at least once per month. 75% of the panels are presently visited each month. Maintaining consistency in the assessment of conditions and risks between more than 30 observers has proved somewhat problematical. The spread of observer resources across the mine needs attention. The workload differs significantly between shafts and sections. With the large number of underground measurements and an efficient database it is now possible to detect trends and take more timeous and effective action. The increase in cost of support, due to the fact that it is now complying readily with standard, is far outweighed by the safety advantages that the situation offers. 25

26 Comments Each factor is assigned a maximum index of 10 and the probability of a fall of ground is determined by calculating the average index for all factors. Severity is determined from a weighted average. Geology, beam thickness, span, pillar size and water are weighted. A relationship between probability, severity is then used to categorise stope panels as high medium or low risk. Unstable wedge/keyblock and pillar punching are the main rock related hazards. A seismic system is currently being installed to augment data available for risk assessment. 26

27 5 A seismic risk assessment trial on a typical Bushveld mining scenario 5.1 Introduction For the purpose of applying the risk assessment concepts as suggested in SIMRAC GAP 608, the reader should refer to the default description of individual parameters to chapter 6 of SIMRAC GAP 608 ' Survey and assessment of techniques used to quantify the potential for rock mass instability'. The concept of seismic risk assessment categories and parameters are shown on page 9 by Table The assessment is made in four categories, namely Level of Ground Motion Vulnerability of the Excavation to ground motion Exposure of people Quality of information Each category has effectively the same relative importance or weighting. An overall seismic risk assessment is achieved by combining individual category ratings. The process of achieving a single rating is discussed later. Each category is subdivided in various parameters contributing to that category. The parameters are rated individually and averaged to provide a category rating. Risk ratings range from 1 to 5, where 1 implies a very low risk and 5 an alwaysunacceptable high risk. A typical Bushveld mine was selected to demonstrate the application of seismic risk assessment. The outcome of the assessment only makes sense in relative terms with similar assessment outputs form other areas in the same environment. To view the outcome as an absolute measure may be questionable. 5.2 An example seismic risk assessment exercise The risk assessment category ranges from P 1 to P 4 and each category comprises a number of parameters, p 1 +p 2..+p n. The proposed risk assessment requires an assessment at a specific place for a specified time window. In this trial assessment an area in a production stope, about 50 to 100 m back from the face and at least the same distance from a regional pillar or a large remnant (or a pothole) was selected. 27

28 5.2.1 P 1 - Level of ground motion P 1 p 1 Mmax and distance from the source It is shown in this report that the vast majority of events are associated with pillars (either in the form of foundation failure on strain bursting). GAP 608b suggested the following risk matrix combining M max and the distance to the source. P 1 p 1 values > M max >500 Distance to the source (in metres) Table A possible risk matrix for combining M max and distance to the source With a panel width of 30 meters the total stope area is within 20 m of a pillar that could potentially fail. Seismic monitoring suggested an M max of between Magnitude 0 and Magnitude 1., resulting in a rating for P 1 p 1 of P 1 p 2 Mean return time The mean return time refers to how regularly an event larger than a specific magnitude can be expected. Kijko and Lasocki describes a formal approach towards determining the mean return time based on the cumulative distribution function, seismic event magnitude and the mean activity rate. The output is then used in a simple rating process as follows: Risk rating for P 1 p 2 Description of return period 1 Once within life of mine 2 Longer than 5 years 3 Longer that a year 4 Longer than a month 5 Less than 1 month Table A risk rating based on the mean return period 28

29 In this Bushveld trial case, pillars/remnants were the main source of seismic activity, and it can be assumed that a specific pillar will only fail catastrophically once in its lifetime, therefore implying a rating for P 1 p 2 of 1. However, this approach is not valid due to the fact that a number of pillars (potential seismic sources) with a failure rate of once during the life of the mine exists within, say 20 m) of any position within our stope under consideration. The expected magnitude associated with pillar failure is less than Mag. 1. Having, say six pillars that could fail within 20 m from any position, it is reasonable to say that the mean return time of an event associated with these pillars, are in the order of one year. A rating for P 1 p 2 of 3 is given P 1 p 3 Seismic/Time distribution The assumption of an equal probability for an incident resulting in large ground motion, during a 24-hour period is unnecessarily conservative. Every mine with a history of seismic monitoring has a known hourly distribution function for the occurrence of seismicity at that mine. Number of events Hour of the day Figure A diurnal distribution of seismic activity in the risk assessment area It should be noted that the blasting peak at this mine is significantly less than what is experienced in deep level and hard rock gold mines. It may be explained by the fact that the seismicity is in the pillars/remnants in the back areas. The total system is a stiff system with minimum elastic closure at the time of face advance (blasting). The stress transfer to the pillars may be primarily due to time dependant effects. This has significant implications on the seismic risk distribution in time. 29

30 The method of rating the risk associated with the time distribution of seismicity is based on a magnitude range and source position identified as resulting in the largest ground motion. Risk rating for P 2 p 3 Description of relative seismic event rate Table to 20% of max. seismic event rate 2 20 to 40% of max. seismic event rate 3 40 to 60% of max. seismic event rate 4 60 to 80% of max. seismic event rate 5 80 to 100% of max. seismic event rate Risk rating based on the time distribution of seismic event rate in the area/volume of interest. Seismic events are from the magnitude range and source position identified as resulting in the largest ground motion. Our scenario requires a risk rating during the morning shift. From Figure and Table a rating for P 1 p 3 of 3 should be given P 2 - Vulnerability of the excavation to ground motion The P 2 category includes the following parameters p 1 Energy Release Rate (ERR) p 2 Local geological structure p 3 Support p 4 Ground condition p 5 Escape ways p 6 Local site amplification P 2 p 1 Energy Release Rate (ERR) ERR scales with the stress levels ahead of the mining face and reflects local factors such as stress concentrations due to lead/lags and mining span. ERR does not easily find an appropriate parallel, as for risk assessment, in the platinum mines. The concept is that stress ahead of the mining face will relate to the vulnerability of the excavation to a shock wave from a remote event. Again, with the emphasis on pillar-associated seismicity, the vulnerability of the immediate area around the pillar to the dynamic motion, is difficult to define. The 30

31 extent, orientation and density of fracturing are more likely to be determined by local geotechnical parameters. Day & Godden (2000) described a risk assessment methodology for pillar stability. A safety factor is calculated in pillar design, but the authors pointed out that the peripheral effect of spalling can reduce the safety factor to a level where pillar failure can occur. The assessed risk refers to the probability that a pillar can fail and not the vulnerability of the area around the pillar to such failure. However, this could be included in the future assessment of seismic risk. York, et al, (1998) assessed the likelihood of foundation failure that could also be incorporated in a seismicity probability function. Where ERR relates to the local stress environment at the face, perhaps the closest parameter in the Bushveld mines may be the average pillar stress (APS) in the particular area of interest. APS could therefore also be an indicator of the amount of deformation in the area around the pillar and indirectly suggests the likelihood of hanging wall failure should an amount of seismic energy be released in an pillar. This concept requires further investigation and at this stage the researcher cannot provide a meaningful alternative for ERR as a risk parameter in the platinum mines P 2 p 2 Geology An argument around cause and effect exits in terms of the influence of geology on the risk rating. The emphasis in this case is on the influence of geology as to the vulnerability of the excavation. Input parameters to consider will be the presence, or close vicinity of a dyke/fault contact, the attitude of approach to a significant structure (distance and angle to be considered), and the complexity (number and type of structure). A feature of particular importance in the Bushveld is potholes (or slumps). The proximity to potholes influences the degree of fracturing and fracture orientation, which in turn will negatively, influence the vulnerability of the excavation to large dynamic ground motion. In the case of the mining scenario used for this risk assessment there are two potholes within 50 m of the area of interest. A shear zone runs through the area. A relative poor rating of 4 is given P 2 p 3 Support The appropriateness of the support type and adherence to industry and mine standards are reflected in this risk rating and in most cases will be the result of a support audit on: the support in haulages, access ways, gullies and in the stopes, the adherence to appropriate support types, installation standards for the above areas, and the adherence to temporary support standards A conservative, worst case approach should be adopted in determining the risk associated with support. 31

32 The support in the area of interest is assumed to be adhering in most aspects to the appropriate support standard and installation standard. The area has suffered from a series of small seismic events and a seismic related accident. Although the support is adhering to the standards it can be argued that it is inappropriate support for an area with known seismicity. A rating of 4 is thus given P 2 p 4 Ground condition This parameter describes the vulnerability of the excavation due to the observed ground conditions. Some inputs to consider are: Adequacy of barring Blast damage Sidewall fracturing Bedding planes Rock mass competence This rating is most subjective but it is felt that any rock engineer should have no problem in subjectively rate the general ground condition. Effective methods to rate ground conditions currently used on the mines have been summarised in Appendix 4.6 (page 24) and could be used as part of the rockburst management. In the scenario being rated a rating for P 2 p 4 of 3 is given P 2 p 5 Escape ways The Bushveld mine used for this trial has excellent alternative escape ways and a rating of 1 is given P 3 p 1 Exposure of people For the purpose of this evaluation, the area under consideration is the area within 20m from a pillar that could potentially fail. During day shift this implies to the majority of the people underground and a rating of 4 is given P 4 Quality of information Seismic risk assessment is only possible with structured and easily accessible data/information. The data could be provided by mine design, geological mapping, seismic monitoring, seismic interpretation and numerical modelling. Other relevant factors determining the quality of information is the risk assessment interval and the volume of rock mass considered. Furthermore, effective communication for input when assessing risk as well as the effective communication of the results, is a prerequisite. 32

33 P 4 p 1 Mine plans/structure/layout The researchers were not overly impressed with the quality of information on the abovementioned and a rating of 3 is given P 4 p 2 Seismic monitoring It is impossible to meaningfully access seismic risk without quantified seismic monitoring. Monitoring also implies the ability to process and interpret seismic data. The risk rating of seismic monitoring is primarily based on: the quality of the recorded data, including accuracy of location and the quantification of seismic parameters; the sensitivity of the network; and the quality of processing and interpretation This mine has a seismic monitoring system but the density (therefore sensitivity) is such that a rating of 3 only can be given. (All seismic events manifested as rockbursts recorded with an accuracy well within the dimensions of the risk assessed volume of the rock mass). The average distance between sensors is 500 m P 4 p 3 Seismic early warning No seismic early warning is practised but the system forms the basis for pro-active management decisions in preventing and controlling of rock bursts. A rating of 3 is given P 4 p 4 Assessment interval and assessment volume A long time interval between risk assessment, as well as a larger area of mining being assessed, will lead to an averaging of the risk rating. This averaging will result in being of less value as an input towards managing seismic risk. An average good (say 3) rating for the total mining associated with a shaft, may hide some high-risk anomalies in specific sections or panels. A similar argument can be made for assessing risk at long intervals, such as every six months, it is true that some parameters may change slowly, but others would be more dynamic. 33

34 Risk rating for P 4 p 4 1 Description of assessment interval and assessment volume Risk assessment done on individual working areas, for example a single panel and its gully. Time interval prescribed by the fastest changing parameter, for example an increase in seismicity on a structure close to the working place 3 Risk assessment done monthly per panel/working place 5 Risk assessment done at intervals more than quarterly and/or for all mining associated with a shaft Table Risk assessment rating based on the assessment interval and assessment volume Risk assessment is done monthly per panel per working place and justifies a rating of P 4 p 5 History The importance and relevance of having an experience reference was described in GAP 608 Section (page 21). An effective database should exist and allow for easy access to: earlier risk assessment exercises, risk management decisions and outcomes, seismic and FOG data, seismic damage, seismic and FOG linked accidents, production data, mine face positions and that of major structures. A suggested rating is given in Table 5.2.5, which could be significantly improved by rather defining absolute levels. Risk rating for P 4 p Description of quality of experience reference An excellent experience reference A good experience reference A reasonable experience reference An ineffective experience reference No structured experience reference Table Risk rating based on the quality and availability of historic data. 34

35 The mine has no formal incident recording process and due to staff turnover, therefore, no real 'memory' of incidents. Any accident is investigated independently of any previous incidents. A rating of 5 is given Communication Good communication exists in terms of input to, and output from the risk assessment process and a rating of 2 is given. 35

36 5.3 Conclusions This chapter provides a suggested procedure for evaluating seismic risk. The approach is holistic and attempts to address all the factors contributing to seismic risk and also to provide an appropriate weighting as to their respective importance. It is not supposed to be seen as an 'universal' best practice, but rather a general approach or methodology that is adopted. Different environments may experience the parameters contributing to seismic risk as being of more (or less) relevance. Rating } Risk Assessment Category Parameter Parameter rating P 1 Level of Ground Motion M max Distance from source Mean Return Time (Frequency) 4 3 Seismic/Time Distribution 3 P 2 Vulnerability of Excavation to Ground Motion (or Falls of Ground) ERR Geology Support Ground condition Escape ways N/A P 3 Exposure of people People/Time distribution 4 P 4 Mine plans/structure/layout Seismic Monitoring Quality of Early Warning Information Rating per category P 1 = (p 1 +p 2..+p n )/n Assessment interval and volume Experience reference Communication

37 Overall Risk Rating P combined = P 1 * P 2 * P 3 * P 4 = 3,3 * 3 * 4 * 3.2 = 127 P risk = 3 (if 36# P combined <144) Table A risk assessment output on a trial Bushveld mine The results for the risk assessment for this trial Bushveld mining scenario are given in Table This result should not only be viewed from an absolute perspective. It should be seen as relative to other areas of similar environments. Risk control should concentrate on the negative outliers (or anomalies). 5.4 References Day, A.P., Godden, S.J., (2000) The design of panel pillars on Lonmin Platinum's mines, "Keeping it up in the Bushveld" and "Advances in Support Technology" SANIRE 200 Symposium, South African National Institute of Rock Engineering York, G., Canbulat, I., Kabeya, K.K., Le Bron K., Watson, B.P., Williams, S.B. (1998) Develop guidelines for the design of pillar systems for shallow and intermediate depth, tabular, hard rock mines and provide a methodology for assessing hangingwall stability and support requirements for the panels between pillars, SIMRAC Report GAP 334, Department of Mineral and Energy Affairs, South Africa. 37

38 6 Observations at Rustenburg Mine A Shaft 6.1 Introduction An apparent seismic related fatality occurred at Rustenburg Mine A Shaft. With the permission of the mine this accident and subsequent investigation as well as localised monitoring could be used in this report. The details of the accident and the seismic reports are included as appendices (Chapter 9). The conclusions are of great value in providing a different perspective on pillars being the only source of seismicity in the Bushveld. 6.2 Observations Some of the observations relevant to this report are repeated here The seismic results support the original supposition that the accident was the result of a seismic event. The events clustered around a fault structure. Significant seismicity was recorded in the area of interest. The largest recorded event was a Magnitude 1.4. The damage was caused by a rockfall resulting from the dynamic 'shake down' of the hangingwall. The implication of this observation is that the damage was caused by the failure of the support to counter such dynamic movement. The relative limited extent of mining in the area of interest is resulting in a stiff system with little closure. It is therefore unlikely that a large amount of seismic energy can be released. However, underground personnel are exposed to the risk of being close (less than 20 meters) from a seismic source. Even a small event can be manifested as having large enough ground motion to pose a significant risk in a localised area. 6.3 Conclusions The significance of the accident at Rustenburg Mine A Shaft was that it was caused by failure on a geological structure rather than pillar failure. A further relevant observation was the failure of support to sustain a relative small release of seismic energy. 38

39 7 Stope Support in Rockburst Prone Platinum Mines 7.1 Introduction A recent rockburst, which severely effected a Merensky reef stope, has highlighted the vulnerability of these stopes should a mining induced rockburst occur. The stope support systems in platinum mines are designed to prevent rockfalls and cannot accommodate significant seismic induced closure. In the recent rockburst it was apparent that the stope support system had been destroyed and this could have contributed as much to the damage seen as the dynamic event. The rockburst was caused by slip on a geological structure and appeared not to have been the first of such an occurrence at this site. An inspection of the site revealed the following: The event was widely felt underground and was accompanied by a lot of dust. Damage was widespread along the gully; the hangingwall was severely damaged with exposed and open joints sub parallel to the gully. Excessive closure of + 0,5m was seen in one of the down-dip panels Sticks had failed extensively in the stope and it was reported, although not observed, that there had been footwall heave in the vicinity. It was evident that the sidewall of a ledged area had fragmented and slabbed, and that this had occurred during a dynamic event. RSS grout packs had been severely compressed with broken steel bands that had necked. Some steel bands had also been pushed closer together. 7.2 The design of rockburst resistant stope support for the Merensky reef. Staff from the CSIR division of Mining Technology have had extensive experience evaluating rockburst damage in South African gold mines. The damage seen at the rockburst site described above resembles that seen on the Ventersdorp Contact reef in the hard lava geotechnical area. Large blocks of failed hangingwall can commonly be seen at these sites and these blocks can fail individual elongate units. The type of elongate to be used in any potential support system therefore needs to be selected with care, with the consistency of the force deformation behaviour being important. 39

40 Apart from these direct observations very little else is known about rockburst parameters on the Merensky reef. Because of this the gold mine stope support design parameters will be used. The setting of an energy absorption criterion has been based on back analysis of hangingwall ejection thickness involved in fatal rockburst accidents. There are none in this case and it is suggested that a stope support design be based on an ejection thickness of 1.5 m, the number strongly influenced by observations at the rockburst site described above. As in the gold mines the velocity of ejection is assumed to be 3 m/s and it is also assumed that the hangingwall of the stope displaces downward for an amount of 0,2 m during the rockburst. The amount that the hangingwall is allowed to displace will clearly influence the value of the energy absorption criterion, increasing the criterion if h>0,2 and decreasing the criterion if h<0,2. The Merensky reef can be narrow and the value of 0,2 m used can be justified on practical grounds. Consider a 0,9 m high Merensky reef stope prior to dynamic closure of 0,2 m. Once this closure has occurred the stope is then 0,7 m high which is sufficient to allow miners, with difficulty, to manoeuvre and get out of the workings. Therefore, Energy absorption criterion or E ac = 1/2 mv² + mgh Where m = mass of hangingwall ejection thickness per m 2 (ρ = 3.0) v = 3,0 m/s h = 0,2 m g = 9,81 m/s 2 E ac is therefore kj/m 2 Roberts (1999) proposed a design methodology based on tributary area theory comparing the actual energy absorption of the support system to the respective criteria at any distance from the stope face, taking into account the effects of stope closure. Any support units installed in a stope are immediately acted upon by stope closure. Depending on the force deformation characteristics of the support unit, this closure could either degrade or increase its ability to generate load. In the case of support design for rockburst conditions, energy absorption capacity curves are used. Figure combines the graph of energy capacity with a nomogram in the lower half of the diagram, which relates the amount of stope closure experienced by a support unit installed a specific distance behind the stope face for various rates of closure. Consider some distance behind the stope face, point A. A horizontal line is traced from A until it intersects the line representing the stope closure rate, point B. From B a vertical line is traced to intersect the energy deformation curve of the support system, point C. From C a horizontal line is traced back to the y-axis at D, where the available energy can be read off and compared to the energy criterion. 40

41 Further details of the support design methodology outlined above are given by Roberts (1) where an approach to separating the face and permanent support areas, as well as combining multiple support types, is given. 10 mm/day Figure An energy capacity deformation design chart for rockburst conditions. The stope support design methodology described above uses tributary area theory that defines support unit spacing. An early weakness of the methodology was that the stability of these support unit spans was not tested. It is however essential that the support unit spacing is tested for stability before the support resistancedeformation curve is determined. The methodology to do this is outlined in Daehnke et al (2) and incorporated into the SDAII (3) software This kind of analysis will determine the stability of keyblocks between the support units and the zone of influence around individual support units. The stability of keyblocks delineated by extension and shear fractures is dependent on buckling, shear and/or rotational failure mechanisms. When investigating the stability of keyblocks, the possibility of each of the three failure mechanisms needs to be considered. If the keyblock is unstable in any of the three failure modes, the unsupported span between adjacent support units needs to be decreased until neither buckling, shear nor rotational failure can occur. The zone of support influence is defined as the lateral extent of the vertical stress profile, induced in the hangingwall beam by a loaded support unit. The zone of influence can extend some distance away from the immediate support - hangingwall contact, and hence can contribute towards rock mass stability between adjacent support units. 41

42 Figure SDAII plot of blast-on Ebenhaeser rockburst support system. A SDAII plot of a simple yielding elongate support system for rockburst conditions is shown in Figure Figure shows the force-deformation curve of the yielding timber elongate chosen, the Ebenhaeser prop. Figure shows that the support system does not meet the energy absorption criterion after the fifth row from the stope face. If the sweeping line cannot be confined to the fourth row of support then either the support spacing will be required to be closer or a different support element with a higher yield must be chosen. The immediate area of the stope face also does not meet the energy absorption criterion and therefore the front line of support could be moved closer to the stope face. This iterative process is typical of how stope support systems are designed using the SDAII programme Figure The force deformation curve of the Ebenhaeser yielding timber elongate. 42