Disturbance factor (D) was initially proposed to account for factors often not considered in the Hoek-Brown failure criterion and can reduce the overall strength of the rock mass. The main drivers for proposing such a parameter were the various back-analysis and case histories reviewed and referenced by Hoek et al. (2002), where the application of the Hoek-Brown criterion for undisturbed in situ rock masses resulted in rock mass properties that were too optimistic.
Hoek et al. (2002) associated the requirement for a disturbance factor primarily to blast damage and stress relaxation due to removing the overburden. However, a review of other works, including Jefferies et al. (2008) and Cundall et al. (2008), reveals that other factors, such as scale and rock mass heterogeneity effects, can reduce the overall strength of the rock mass.
Some key points must be noted when downgrading the rock mass properties from their original values obtained from laboratory testing and rock mass characterisation. These are outlined below:
Large populations of the previous back-analysis studies were based on the application of simplified constitutive models (e.g. employing Mohr-Coulomb, perfectly plastic, constant elastic modulus, etc.) and simplified numerical modelling methods (e.g. limit-equilibrium, boundary element or implicit finite element). It is questionable how much of the inaccuracies experienced during those works was associated with the chosen rock mass properties, and how much of it was actually related to modelling effects (due to inaccurate constitutive models or modelling methods).
The stress relaxation effect or disturbance caused by mining-induced stresses comprises a significant weight (approximately 70%) in the original guideline provided by Hoek et al. 2002 for selection of the D factor. It is important to note that this effect is mainly controlled by the rock mass’s strain and elastic modulus softening nature which has already been accounted for in more advanced constitutive models like the IUCM.
According to guidelines provided by Hoek et al. (2002), the blasting effect has a much smaller weight than the stress relaxation effect in selecting the D factor. Saiang (2011) conducted a detailed study on the blast-induced damage zone (BIDZ) around underground tunnels and concluded that the BIDZ in most practical cases, on average, ranges from 0.3 m to 0.5 m away from the excavation boundary. It is also important to note that the BIDZ is more likely to influence the more competent rock masses that have fewer discontinuities. Blasting energy in rock masses with more discontinuities (say GSI < 65) can be released through in situ fractures, minimising the induced damage to intact rock bridges. Most other key recent literature, including Hoek and Marinos (2007) and Brown (2008), also suggest that the blasting effect should only be limited to the immediate boundary of the excavation and cannot be applied to the entire rock mass.
The strength anisotropy can also substantially impact the strength properties of both intact rock and rock mass specimens. In anisotropic rocks, the compressive strength of the rock mass or the intact rock can vary by a factor of 6 or more depending upon the direction of loading. Despite this considerable effect, many practitioners still use isotropic constitutive models and often choose the average strength value from laboratory test results. This will often lead to overestimating stability conditions in certain geometrical conditions and underestimation in other conditions. Therefore an arbitrary parameter like the D factor is necessary for certain conditions to reproduce the observed failure in anisotropic rocks. The strength anisotropy is explicitly included in the IUCM.
The scale effect is another essential factor that impacts both the intact rock and rock mass strength. Hoek and Brown (1997) recommend that the Hoek-Brown criterion be applied only in conditions where the intact block size is relatively small compared to the size of the rock mass being analysed. This implies that the Hoek-Brown criterion only accounts partially for the size effect. When the size effect is not appropriately reflected, a reduction factor, such as the D factor, would be required to downgrade the strength parameters accordingly.
Rock mass variability is another factor that can largely influence the overall rock mass strength, particularly in larger scale problems like open pit mining or caving operations. Jefferies et al. (2008) demonstrated that the weakest portions of the heterogeneous rock mass often control the stability of the rock slopes. Therefore, the conventional methods, which apply the median or average rock mass conditions, can overestimate the overall strength of the rock mass. This is, perhaps, another factor contributing to the need for a D factor in the traditional ways of analysis.
Considering the above discussions and based on previous applications of the IUCM in several numerical back-analysis studies, using a D factor is not recommended when using the IUCM in numerical modelling studies. This is mainly because part of the above factors (e.g. strain-softening, anisotropy, etc.) are explicitly accounted for in the IUCM, and the remaining factors (e.g. scale effect, rock mass variability, etc.) are accounted for during the input parameter selection process (e.g. downgrading of UCS or GSI value).
References:
