Page 28 - Numerical Analysis and Modelling in Geomechanics

SURFACE DISPLACEMENTS OF AN AIRFIELD RUNWAY 9
point 17 could be used to determine the deflection over the undisturbed
subgrade.
For the computer models of the undisturbed camouflet for all seven depths, the
downward vertical deflections of points 17 to 29 inclusive were within the 1%
limit and this deflection was given the value of 100%. The deflection of points 1
to 16 inclusive was related to the 100% value. For the undisturbed model, when
the depth of the camouflet was 8.354 m (column 1 of Table 1.2), the average
deflection of points 17 to 29 was 128.6 mm and recorded as 100%. The average
deflection for points 1 to 16 inclusive was 131.75 mm (102.5%). This shows that
the model of the filled camouflet may give results that for points 1 to 16
inclusive on average overestimate the deflection by 2.5%. To clarify the amount
of change from the 100% value that may be expected, the deflections for points 1,
4, 8, 12 and 16 are included in column 2 of Table 2. Table 2 shows that the
computer model will overestimate the deflection for all detonation depths up to
and including 15.354 m and for all points 1 to 16 inclusive. In this research any
detonation deflection results that fall within the deflection overestimation of
Table 1.2 are accepted as being undetectable, as also is any deflection between
99% and 101% for detonation depths of 15.354 m and 18.354 m.
For the increasing depth of camouflet a similar analysis was performed for
each of the depths of 9.354 m, 10.354 m, 11.354 m, 12.354 m, 15.354 m and 18.
354 m. The deflection results are recorded in columns 2, 3, 4, 5, 6 and 7
respectively of Table 1.2. The results show that as the camouflet depth increases
the average deflection for points 1 to 16 reduces and tends to 100%. Values are
given for points 1, 4, 8, 12 and 16; to illustrate the variation in the individual
results included in the average value over zone 1. Clearly for λ =−1.388 the finite
c
element model shows that the accuracy of the undisturbed subgrade model will
overestimate the deflection for point 1 by up to 4.4%. But as the depth of the
detonation increases this overestimation will reduce to 3.1%, 2.1%, 1.5%, 1%, 0.
3% and 0.0%.
In previous publications a series of nine material sets were used to represent
the variation in the subgrade caused by the detonation [9–14]. These nine
material sets, renumbered, are used in this research with additional material sets
to represent other possible variations in the subgrade. The nine previous material
sets 1 to 9 inclusive have been renumbered as 5, 7, 6, 13, 14, 16, 12, 3 and 1
respectively of Table 1.1.

Finite element program
PAFEC software [40] was used to model the camouflet and perform the
computational analysis. The authors have gained considerable experience in the
use of the software and are confident in its application to the research under
discussion [41].
For computational modelling purposes the camouflet void was assumed to be
spherical and enclosed in a cylinder with a flat surface uppermost. The full