Rock socketed piles are a common foundation solution to transfer heavy loads from structures to the underlying rock mass. Their design methods are traditionally based on conservative empirical rules, some of which incorporate the rock mass quality through a reduction factor, often related to RQD. This paper presents a comprehensive analysis of an international database consisting of 72 in-situ load tests, reviews methods for borehole roughness description and back analysis of borehole roughness height, and presents a validation study for the reviewed procedures for quantitative roughness description. Our analysis reveals that in contrast to intuition, shaft resistance of piles in rock masses with lower RQD values tends to be higher than in rock masses characterized by higher RQD values. We explain this surprising result by the effect of RQD on borehole wall roughness.
1.1. The empirical basis for shaft resistance design
Methods to estimate shaft resistance (fsu) in rocks are largely based on empirical rules that are typically derived from pile design in clays, and are generally expressed as: (mathematical equation available in full paper)
The Uniaxial compressive strength (qu) of saturated clays is equal twice the cohesion (C), therefore Eq. (1) becomes: (mathematical equation available in full paper)
Or in general form: (mathematical equation available in full paper)
where a and ß are empirical fitting parameters the value of which varies among various authors [1-7]. An adhesion factor aq is defined as: where in general qu denotes the uniaxial compressive strength of the weaker member in the pile - rock interface. Many building codes use these simple empirical laws to estimate fsu. For example, an Israel Standard  suggests the values of a = 0.2 and ß = 0.5 which agree well with empirical results published by [1, 2]. The standard also suggests multiplying fsu by a reduction factor which would take into account mechanical and structural characteristics of the rock mass, but it does not specify how its value might be evaluated.
1.2. Previous attempts to integrate RQD into foundation design schemes
From the early 1980's researchers made attempts to upgrade these relationships in order to quantify the influence of rock quality on shaft resistance through the application of a reduction factor, usually referred to as "reduction factor with respect to rock mass effect" or "mass modulus factor", and marked as ß. Here the reduction factor will be labeled MMF (Mass Modulus Factor) to distinguish it from ß used in Eq. (4). The accepted way to evaluate MMF is through its empirical relationship with moduli ratio, namely the ratio between Young's moduli of rock mass (EM) and intact rock (ER). Published trends for these relationships are plotted in Figure 1. As can be inferred from inspection of Figure 1 the published trends between MMF and the moduli ratio are intuitive.
Fig. 1. Mass Modulus Factor (MMF) versus Modulus ratio[after 7, 9, 10, 11]. (available in full paper)
Many researchers suggested to use correlations between the moduli ratio and RQD such as shown in Figure 2 for the purpose of integrating rock mass quality designation into foundation design in rock.