Engineering Rock Mechanics

(Jacob Rumans) #1
168 Anisotropy and inhomogeneity

the location of the minimum value. Moreover, the plot in the question
is a (log k, 0) plot, which accentuates the apparent cusp, as indicated in
the lower-right diagram in which the same information (as given in the
left-hand diagram) is plotted.


Qf 0.7 What does the term 'structural domain' mean in the context
of engineering rock mechanics?


A10.7 Because of the complexity of the inhomogeneity and anisotropy
encountered in rock masses, as conceptually illustrated in A10.2, some
kind of simplification is required for engineering modelling. One ap-
proach is to divide the rock mass into volumetric domains, within each
of which the same rock mechanics properties are essentially constant.
The word 'essentially' means that although the properties will be vary-
ing, they can be assumed to be constant for the rock mass and modelling
in hand. Since all rock masses are different, and the purpose and type of
modelling will vary with the project objective, there can be no absolute
rules for choosing structural domains. Often they are chosen according
to the lithological boundaries because the rock mechanics properties can
be strongly related to the geology; but this is not always the case and
the choice can also be made on the basis of site investigation sampling
schemes supported by geostatistical analyses.


Qf 0.8 During a site investigation for quarry development, geomet-
rical properties of the fractures on a large surface rock exposure
were measured. Eight sampling squares, each 100 m x 100 m, were
established on the rock exposure. One of the set of statistics pro-
duced during the site investigation was the number of fractures with
a mean aperture exceeding 10 mm in each sampling square. These
results were as follows:
Square: 1 2345678

On the basis of these data alone, is there sufficient evidence to
conclude that the sampling squares should be regarded as different
structural domains?

461 397 453 362 389 421 382 423

Af0.8 The purpose of this question is to highlight the issues and
hazards in choosing structural domains for rock engineering design
purposes. The key issues are the following:

The subject of geostatistics is not just the statistical analysis of geological data, but
a form of statistics in which not only the sample value is used but also the location of
the sample. Geostatistical trend analyses are powerful because they can indicate when
different domains should be used. A good introduction to this concept is found in Davis
(1973), see Footnote 6.
The question of whether such variations in data do reflect underlying causes or are a
result of the sampling process pervades much of science and daily lie. In 1999, a crime
prevention officer in the Avon and Somerset police in the UK claimed that the sign of the
zodiac under which a person was born indicates how likely their car is to be stolen. This
Free download pdf