3188
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
modulus from a falling weight deflectometer or Young’s
modulus from a plate bearing test (Gabr and Cameron 2011).
2 MATERIALS AND RANGE OF TESTS
Two RCA basecourse products, A and B, were tested, along
with a comparable product (A20) of RCA with 20% by mass of
RCM. Further materials were made at UniSA by blending
product B with RCM to 10%, 20% and 30% (B10, B20 and
B30). DPTI permit up to 20% by ma
ss in RCA of “foreign
material” consisting of clay brick tile, crushed rock and
masonry for base course and subbase applications. Finally a
virgin quartzite aggregate (Q) was evaluated, which is
commonly used in Adelaide for construction of Class 1 bases.
The particle size distributions of the materials fell within
DPTI specifications for Class 1 base. All the materials were
well-graded gravel and sand mixtures with silty gravel; GW-
GM according to the Unified Soil Classification System
(USCS). Material A lay close to the coarse specification limit,
while Material B20 crossed between the limits and had a fairly
high proportion of fine sand-sized particles. The fines content of
the two RCA products, A and B were just 5% and 7%
respectively, while the quartzite base material (Q) had 11%.
Tests were conducted in line with the requirements of
current Australian specifications. These included plasticity of
fines, aggregate strength tests, Los Angeles abrasion, CBR tests
on 4 day soaked samples and RLTT to the DTEI 2008 protocol.
In addition, falling head permeability tests and shrinkage tests
were conducted. Some concern has been expressed relating to
the propensity of RCA to exhibit some cementation upon
wetting and compaction. This self-cementation of RCA
materials can produce increase of strength with time, but also
the possibility of reduced permeability (AASHTO 2002) and
shrinkage. Therefore shrinkage was investigated.
3 MATERIAL PREPARATION
All materials were compacted to a target Dry Density Ratio
(DDR) of 98% of maximum dry density under Modified Proctor
compactive effort. Static compaction, which is advocated by
DTEI for unbound granular material, was used for compaction,
largely because of the consistency of preparation. Moulding
moisture variations are indicated for particular tests as follows.
In South Australia, materials are commonly compacted at 80%
of OMC and are allowed to dry back to 60% of OMC.
3.1 Falling Head Permeability
For the falling head permeability tests, the moulding moisture
contents were 100 and 80% of OMC. Blended materials were
tested; A20, B10, B20 and B30.
3.2 Drying Shrinkage
Samples were 200mm high by 100 mm diameter. Triplicate
samples of materials A and B, and duplicate samples of A20
and B20, were prepared OMC. The target moisture content was
reduced to 90% OMC if the material was found to be too fragile
upon extrusion (e.g. samples B & B20). After compaction, the
samples were extruded from the mould and sealed in plastic
bags to cure for 7 days; the samples were then stored in a curing
room (temperature 23±2°C and relative humidity 55±5%).
3.3 Undrained Triaxial and RLTT Testing
Duplicate samples were prepared.The resilient modulus and
permanent deformation behaviour of RCA mixtures were
investigated at different levels of moulding moisture contents,
as was the undrained shear strength. Generally just one day of
curing occurred before de-moulding and testing.
4 RLTT TEST METHODS
In Australia, there are two standard approaches to Repeated
Load Triaxial Testing (RLTT); multi-stage stress testing and
single-stage stress testing, e.g. the DTEI approach. DTEI 2008
specified application of a constant confining stress of 196 kPa
and a vertical deviator stress of 460 kPa, pulsed over 50,000
loading cycles. AUSTROADS established a multi-stage stress
RLTT under drained conditions to determine the permanent
deformation and resilient modulus properties. Both these test
protocols ahave been applied. In the RLTT program,
deformations were measured with two pairs of inductance coils
(“Emu” coils) mounted on the sample (Gabr et al. 2012).
5 INDEX VALUES AND OTHER PROPERTIES
The plasticity of fines of the various materials is indicated in
table 1. The DPTI (2011) specifications call for a maximum
Liquid Limit of 25% for Class 1 and 28% Class 2, and so A20
falls into Class 2, while all other materials would be acceptable
for Class 1 applications.
Los Angeles Abrasion Value (LAA) and Micro Deval help
to evaluate the abrasion resistance/toughness under traffic
loading. The LAA values of the South Australian RCA
examples ranged between 37% and 39%, which failed to meet
the maximum of 30% proposed by DPTI. The Ile de France
specifications (2003) for LAA allowed 35% for roads with the
greatest traffic (GR4), increasing to 40% for GR3 and 45% for
GR2. The values for RCM blends were in the range of 40 to
45%. The French Micro Deval limit of 30% for GR4 was met
by both RCA products as the Micro-Deval values were 30% and
28% for products A and B, respectively. There is however a
further requirement that the sum of LAA and Micro Deval must
not exceed 55 for GR4 and 65 for GR3. Product B was on the
limit for GR3, while product A, exceeded it (69). GR3
corresponds to a road with daily annual traffic of 85.
Average shrinkage curves with time are provided in Figure 1
for the four materials that were tested. Interestingly, shrinkage
strains were similar for the RCA products, as they were for the
blends (20% RCM); however, the addition of crushed masonry
resulted in an appreciable drop in shrinkage. In the case of
product B, a reduction of almost 60% was observed.
The permeability of blended recycled material when
prepared at OMC was approximately 2 x 10-8 m/sec for blends
based on RCA product B, but it was observed that A20 was ten
times more permeable. Compaction to the same density but at
just 80% OMC increased the permeability of all materials
generally by a factor of approximately three.
Table 1. Selected properties of the recycled aggregate blends
Material
A A20 B B10 B20 B30
Liquid Limit (%)
26 27 23 24 23 23
Plastic Index (%)
2
2
1
3
3
2
LAA (%)
39 42 37 42 43 45
Micro Deval (%)
30
-
28
-
-
-
All materials met the specification of a minimum CBR of
80%, despite masonry content reducing the CBR significantly.
Similarly the requirement of a maximum unconfined
compression strength of 1 MPa after curing was met.
A study was undertaken of the matric suction
–
moisture
content relationship, or soil water characteristic curve (SWCC)
of the materials to enable estimation of matric suctions of RLTT
samples from measured moisture contents. Initial matric suction
was determined by the hanging column method for low suctions
and the contact filter paper method (refer Azam and Cameron
