A database was developed in 1990 to store and allow analysis of information on rockfall and rockburst fatal accidents that occur in the tabular orebody gold mines in South Africa. The information populating the database is extracted from the inquiry reports compiled by the inspectorate of the Department of Minerals and Energy. Also included in the database is the annual production, expressed as square metres mined, for each of the reefs mined. The production data is used as a normalizing factor in the analysis of the accident statistics. Various types of analysis of the data are performed with the objectives, inter alia of determining trends, with time, on the occurrence of rock related accidents; and the influence of mining depth on the ratio of rockburst to rockfall accidents, beating in mind the increase in mining induced seismicity with depth. An additional parameter extracted from the database, and important in the design of support, is the distribution of fall-out thicknesses in the different ground control districts. The locations in the mines, where accidents are most likely to occur, are quantified in order to highlight areas of concern and where research effort should be concentrated. Analysis is also carded out to determine the relative risk of mining the different reefs, which have different geootechnical conditions. This also focuses attention on where improved mining practice and research effort are required. The results of these analyses are presented and their significance is discussed.
Gold mining in the Witwatersrand Basin occurs on large tabular ore bodies, many several hundred square kilometres in extent, at depths below surface ranging from a few hundred metres to 3700 m. The average number of people required underground daily to carry out this mining was approximately 400 000 for the period under review. Of serious concern is the high number of accidents that occur. To obtain a better understanding of the causes and factors influencing the accidents, a rock-related, fatal accident database was established in 1993 by CSIR: Division of Mining Technology as part of a SIMRAC research project (Roberts, 1997). The database contains detailed information of all rockfall and rockburst fatal accidents since 1990, and to date encompasses 1950 entries. The database is updated annually from information extracted from the accident inquiry reports compiled by the inspectorate of the Department of Minerals and Energy. The production information is obtained from the mines on a shaft-by-shaft and reef-by-reef basis for each year.
In order to analyze the database in a meaningful way, the annual number of fatalities needs to be normalized with respect to production data. The normalizing parameter used for this purpose is square metres mined on the tabular reefs and is a direct measure of the new hangingwall exposed that has to be supported and has a potential to fall. The fatality data is normalized to a 'rate' expressed as fatalities per million square metres mined. Normalizing the fatality data with square metres mined nullifies the effect of varying production levels during the period under review and for different reefs. The results of various analyses are presented but emphasis is placed on a relative new investigation, which compares the accident rates for different reefs.
