In the last two decades the demand for primitive stress measurements has increased in the Southern African mining industry. The main reason for this is recognition that the natural complexity and variability of stress in the Earth's crust renders elementary primitive stress estimates inaccurate for engineering purposes. Consequently, primitive stress measurements in situ are essential for a viable primitive stress model of a mine. This paper summarises and reviews experience gained during the last decade when using the CSIRO HI cell (Australian Commonwealth Scientific and Industrial Research Organisation Hollow Inclusion cell) and the BXCCBO strain cell (BX-sized Compact Conical-ended Borehole Overcore cell) to measure in situ stress in Southern Africa. These two strain cells have proven their reliability, but stress measurements remain complex and the risk of measurement failure can be high. The authors describe techniques to control and minimise the risk of measurement failure. There are no reliable benchmarks for measured primitive stresses; therefore the results obtained must be checked qualitatively against major geological trends at the measurement site, and against a background of consistent primitive stress data from the Southern African Stress Database. The paper concludes that the risk of measurement failure can be reduced to an acceptably low level to assure the acquisition of reliable and useful stress measurement data.
Rock mass structure, rock strength, mining induced stress, and primitive stress are the four most important factors that determine the stability of a mine. Measurements of intact rock strength, together with structures such as joints, faults, folds, intrusions, layering, and any other geological features, help to determine rock mass strength and integrity. The strength of the intact rock material is determined from laboratory testing of intact samples. The frequency and properties of discontinuities, such as joints and faults, are less well known because they are inferred from geotechnical information, and sometimes laboratory measurement data. Modern numerical models can take account of rock mass properties as they are estimated or measured from geotechnical data, and the mining-induced stresses estimated accordingly from an appropriate numerical model of the mine layout. The factor that remains least understood–and least measured–of all is the primitive stress before mining. Without knowledge of the primitive stress from a substantive model backed by measurement data, a complete engineering model of the mine in the rock mass, and hence effective mine design, is not possible.