During the construction process, of an underground excavation, the rockmass is damaged and beyond the damaged area the stresses are modified. These zones are collectively this is known as the excavation damages zones (EDZs). The lab results, including UCS and T as well as crack thresholds CI and CD can be used to determine numerical model input properties. With a large data set it is possible to test the best, mean, and worst case scenarios and to evaluate the results statistically. Using two rock data sets, a limestone and a granite, and two nominal stress re-gimes the influence of the best, mean and worst cases (with hi and lo outliers filtered) were computed to determine the sensitivity of EDZ dimensions. It was demonstrated that both an inner and outer excavation damage zone (EDZi and EDZo) could be dif-ferentiated using the model results by the reversal point in the volumetric strain (con-traction to extension). This indicates the transition between a confined micro-damaged state (EDZo) and a potentially dilated EDZi. The outer boundary of the highly dam-aged zone, HDZ, is related to volumetric strain and a reduction in minor stress con-finement. Guidelines are suggested for determining the EDZs around circular excava-tion models. Cumulative distributions for the dimensions of the EDZs were determined. The highly damaged zoned showed the least variability, were as the EDZo showed the most. This method can be used to determine the depth of a cutoff structure to limit the flow through the damage zone, for example, at the required confidence level.
The design of an underground excavation requires site specific data about the subsur-face. The level of complexity or the design life span will determine the volume of data necessary to adequately design the excavation.