See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312289757

Shrink-swell index database for Melbourne

Article  in  Australian Geomechanics Journal · September 2016

CITATIONS READS

3 956

4 authors, including:

Jie Li
RMIT University
80 PUBLICATIONS   351 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Unsaturated soil mechanics View project

All content following this page was uploaded by Jie Li on 18 December 2018.

The user has requested enhancement of the downloaded file.

 SHRINK-SWELL INDEX DATABASE FOR MELBOURNE

Jie Li1 , Jian Zou2 , Peter Bayetto3 and Nick Barker4

1 2
Senior Lecturer, Postgraduate Student, School of Civil, Environmental and Chemical Engineering, RMIT University
3
Managing Director, FMG Engineering, 4Senior Engineering Geologist, FMG Engineering

ABSTRACT
A series of laboratory tests have been performed, including shrink-swell, liquid limit, plastic limit, linear shrinkage and
soil suction on soil samples collected from 47 different field sites across 37 suburbs of Melbourne. The sites cover a
wide range of soils and different geological conditions and geographical settings. The results of the laboratory tests
were used to establish a database of tested shrink-swell index (Iss) values. An interactive map showing the distribution
and reactivity of Melbourne soils has been developed that can be used in routine practice by local engineers to assess
and calibrate the shrinkage indices estimated on a visual-tactile basis. The study indicates that there is a significant
variation in Iss results across any particular soil type and that a blanket estimated value based on that soil type may not
be entirely appropriate. Highly reactive soils are mainly present in the western and northern suburbs of Melbourne, and
many of these areas have a shrink-swell index higher than 6% strain /pF. Correlations of shrink-swell index, liquid
limit, plastic limit, plasticity index and linear shrinkage have also been attempted and the results show that simple index
tests are not a reliable substitute for Iss testing.

1 INTRODUCTION
Expansive or reactive soil refers to clay soils that undergo significant volume change in response to changes in soil
moisture or more appropriately, suction. This volume change occurs as swelling upon wetting and shrinkage upon
drying (Li and Cameron, 2002). In Australia, residential footing design and construction practices have been guided by
the national standard AS2870 since 1986. As a routine practice, site classification is based on the predicted design site
surface movement over the life of the house, ys, which is based on the design soil suction change profiles for different
climatic regions of Australia (Li et al., 2014). To calculate ys, it is essential to determine the reactivity or instability
index, Ipt. The value of Ipt is equal to the shrinkage index, Ips, multiplied by a factor α (varying between 1 and 2) to
correct reactivity for the effects of soil cracking and overburden pressure. Australian Standard AS2870 (2011) provides
three testing methods for the estimation of the shrinkage index Ips, namely the shrink-swell test (AS1289.7.1.1, 2003),
the loaded shrinkage test (AS1289.7.1.2, 1998) and the core shrinkage test (AS1289.7.1.3, 1998). The shrink swell test
was used for this study because it has two distinct advantages when compared to the other two methods: (a) both swell
and shrinkage strains are considered so that the sample may be either very wet or very dry; (b) there is no need to
measure soil suction values (Cameron, 2015). The shrink-swell index Iss obtained for each sample was used as an
estimate of the shrinkage index Ips for that sample.
As part of a long term research project on the behaviour of expansive soils in Melbourne, a series of laboratory tests
which include shrink-swell, liquid limit, plastic limit, linear shrinkage, soil-water characteristic curve, soil suction and
X-ray diffraction (XRD) have been performed in the geotechnical laboratory of the School of Civil, Environmental &
Chemical Engineering at RMIT University. The main objective of this research is to establish a database of shrink-swell
indices, which provides information about the nature and distribution of the expansive soils in the greater metropolitan
area of Melbourne and can be used for assessing the values of Ips estimated based on the visual-tactile method by local
geotechnical practitioners. Also, the experimental data was utilised to assess if there is a correlation between shrink-
swell index and traditional soil indices.
This paper describes the sample collection and the procedures for the experimental work, and provides some
preliminary results.

2 STUDY AREA AND SAMPLE COLLECTION

The greater metropolitan area of Melbourne lies on the southern coast of Victoria and is geographically constrained by
Port Phillip Bay to the south with the Dandenong Ranges to the East and North East. This geographical constraint has
led to the recent westwards and northwards expansion of the city onto the volcanic plains. Most of the Melbourne area
is relatively flat and underlain by Silurian Age bedrock of sandstones, siltstones and mudstones. A simplified geologic
map of the Melbourne region is shown in Figure 1. The geology is characterised by its variability and complexity
(Archbold, 1992). Much of the area, especially the western and northern regions are covered by moderate to highly
expansive soils, mostly derived from Quaternary and Tertiary Volcanic deposits (McAndrew, 1965).

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 61

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 61
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 1: Simplified geologic map of the Melbourne region (after Archbold, 1992).
More than 60 soil samples were collected from 47 sites across 37 suburbs in Melbourne, covering a range of geological
and geographical conditions. Figure 2 shows the location of the sites where soil samples were collected. Most of the
samples were collected by RMIT University’s industry partner, FMG Engineering. The ‘undisturbed’ soil samples were
taken by driving thin walled stainless steel tubes (450 mm in length and 50 mm in nominal inner diameter) into the
ground using a hydraulic drilling rig. The steel tubes containing soil samples were sealed and labelled on site,
immediately on extraction from the borehole and then transported to the laboratory. Prior to testing, the sample was
extruded from the tube with a hydraulic jack and then cut and trimmed carefully using a cut ring and a sharp spatula to
suitable size.

Figure 2: The interactive reactive soil map for Melbourne showing the location of the test sites.

62 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

62 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

The Atterberg limits test results on the plasticity chart are shown in Figure 3. It can be seen that most samples are above
the A-line, and are classified as low to high plasticity clay.

Figure 3: Plasticity chart showing results of Atterberg Limits.

3 SHRINK-SWELL TEST
The shrink-swell test consists of a core shrinkage test and a swelling test. It requires two identical soil samples which
have the same initial moisture content and thus the two soil samples (one for core shrinkage test and another for
swelling test) tested should come from the same tube/core sample.
The core shrinkage test requires an undisturbed cylindrical core sample of a diameter of 45-50 mm to be cut and
trimmed to a length within the range of 1.5 to 2 diameters. Drawing pins were placed in the centre of the core at either
end to provide a reference mark for measurements to be taken. The specimen was then placed on a smooth surface,
allowing it to dry out for a period of at least 10 days. In this study, the changes in specimen length, diameter and weight
were monitored twice a day. The distance between the rounded heads of the pins was measured and recorded with a
digit vernier calliper to the nearest 0.01 mm. The specimen was wrapped with plastic membrane overnight to alleviate
moisture gradients within the specimen and reduce surface cracking. After air-drying for about two weeks, the specimen
was placed into an oven and dried to a constant mass at 105oC degrees. Its final mass was then measured, as well as its
final length. After having these data, a plot of axial shrinkage strain vs. moisture content could be generated as shown in
Figure 4 and the specimen was then broken and used to determine the Atterberg limits.

Figure 4: Axial shrinkage strain versus moisture content (sample from Craigieburn).
For the swelling test, a 50 mm diameter sample was cut with a rigid steel ring of 20 mm height and 45 mm in diameter
and trimmed carefully to ensure both ends were flat. The trimmings were collected and used to determine initial water
content and suction. The specimen was then placed in a consolidation cell with two porous stone plates at the top and

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 63

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 63
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

bottom. A seating pressure of 5 kPa was initially applied for about 10 minutes and the displacement transducer was
zeroed under this seating load to allow for a small amount of initial settlement of specimen. The vertical pressure was
then increased to a value equal to the overburden pressure or 25 kPa for a maximum period of 30 minutes
(AS1289.7.1.1, 2003). After recording the initial specimen settlement, which was taken as the datum from which
swelling strain was determined, the specimen was inundated with distilled water and allowed to swell.
Fityus (1996) found that highly reactive soil might swell out of the steel ring during the swelling test, leading to an
inaccurate result in the swell test. To overcome this problem, Fityus (1996) suggested adding an extension annulus on
top of the soil sample to allow for the possible expansion of soil above the consolidation ring. However, for extremely
reactive clay samples, a loss of soil still occurred even if an extension ring was adopted. In a number of swelling tests, it
was observed that the extension ring was lifted from the top edge of the oedometer ring by highly reactive soil samples.
Soil leaking from the bottom of the consolidation ring was also observed in some cases. In order to prevent the soil
from swelling out of the oedometer ring, a few different techniques had been tried in this study. It was decided to add
two extension rings, one at top and one at bottom. The height of the bottom ring is the same as the thickness of the
bottom porous stone so that the height of the specimen was kept at about 20 mm. Before swelling testing, the extension
rings were tied up with a hose clamp to stop them from moving (refer to Figure 5).

Figure 5: Extension rings used in the swelling tests.

In this study, the swelling specimen was allowed to heave for a period of at least of 10 days and the swelling
displacement was recorded with intervals of 5 seconds by displacement transducers connected to a PC computer. The
apparatus used for the swelling tests are shown in Figure 6. At the completion of the swelling test, the specimen was
extracted from the ring and the final water content and suction were determined.

Figure 6: Apparatus used for the swelling tests.

The typical results of swelling tests are presented in Figures 7-8. It is worth mentioning that only 60% of total swelling
strain of the Williams landing specimen was completed in the first 48 hours compared with 74% of total swelling strain

64 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

64 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

for the South Morang specimen. The results clearly indicate that for highly reactive clay it may need more than 10 days
to finish a swelling test. If the swelling test terminated after 8 days, the total swelling strain would be underestimated by
10% for the Williams landing specimen and 6% for the South Morang specimen.

Figure 7: The swelling strain versus time (sample from Williams Landing, depth = 0.65 m).

Figure 8: The swelling strain versus time (sample from South Morang, depth = 0.75 m).
Knowing the shrinkage strain and swelling strain, the shrink-swell index can be calculated by the following equation
(AS1289.7.1.1, 2003)

(1)

Two empirical parameters are introduced in Equation 1, i.e. a correction factor of 2 for axial swelling test and assumed
soil suction change range of 1.8 pF (Fityus et al., 2005). In this study, a series of laboratory tests have also been carried
out to evaluate these two parameters. Results will be published in the near future.

4 INTERACTIVE REACTIVE SOIL MAP

Based on the results of laboratory tests, an interactive reactive soil map which is based on Google Maps API (refer to
Figure 2), has been generated as part of this study. This map provides (1) the site location, (2) depths of soil samples

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 65

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 65
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

collected, (3) site geology information, (4) shrink-swelling indices (Iss), (5) liquid limit, (6) plastic limit and (7) linear
shrinkage. It is intended to be a useful tool for local engineers and the building industry as a whole. Obviously the
interactive database requires large amounts of soil data, covering a wide range of soils, geological and geographical
conditions across metropolitan Melbourne. Subsequently this research project will be continued and futher input is
needed as the database is an ongoing resource that is being built with the help of RMIT researchers in collaboration
with Industry.

5 EXPERIMENTAL FINDINGS
The results of a total of 60 shrink-swell tests are summarised in Table 1. There is a significant variation in values of the
shrink-swell index, Iss, which vary from 0.34 to 11 % strain/pF with a mean of 3.73 % strain/pF. The results indicate
that highly reactive soils are common in the western and northern suburbs of Melbourne. The soil samples from
Williams Landing, Truganina, Tarneit, Wollert and Diggers Rest have a shrink-swell index higher than 6 % strain/pF
while Dimboola, 330 km north-west of Melbourne, has the lowest shrink swell index of 0.34 % strain /pF.
Figure 9 illustrates the frequency (% of population) against ranges of shrink-swell index (% strain /pF). The range has
been set at 1.0 interval and from this point the frequency was calculated. It is apparent that the shrink-swell indices, Iss
of clay soils in greater metropolitan area of Melbourne span a wide range from very low to very high values.

Figure 9: Histogram of all shrink-swell indices.

66 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

66 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Table 1: Summary of shrink-swell test results

Depth Geological Iss Depth Geological Iss
NO Location NO Location
(m) Code* (%/pF) (m) Code* (%/pF)
1 Armstrong Creek 0.75 Neo 5.05 31 Ocean Grove 0.75 Nb 2.46
2 Ascot Vale 0.50 Qc1 1.11 32 Officer 0.65 Qa1 1.91
3 Braybrook 0.75 Neo 4.53 33 Plumpton 0.70 Neo 2.21
4 Brookfield 0.60 Neo 1.47 34 Point Cook 0.60 Neo 6.16
5 Clyde North 0.75 Nbr 1.87 35 Point Cook 0.65 Neo 5.60
6 Clyde North 0.95 Nbr 4.21 36 Point Cook 0.40 Neo 2.83
7 Clyde North 1.50 Nbr 3.50 37 Point Cook 0.60 Neo 1.59
8 Craigieburn 0.90 Neo 3.55 38 Reservoir 2.50 Neo 4.60
9 Cranbourne 0.75 Qm1 2.42 39 South Morang 1.00 Neo 5.73
10 Cranbourne North 0.65 Nbr 2.11 40 South Morang 0.75 Neo 6.16
11 Diggers Rest 0.65 Neo 6.77 41 Tarneit 0.40 Neo 6.96
12 Dimboola 1.20 Nwl 0.34 42 Tarneit 0.65 Neo 6.62
13 Doreen 0.40 Neo 4.02 43 Templestowe 0.65 Sxa 1.27
14 Epping 0.40 Neo 4.30 44 Torquay 0.65 Nt 3.06
15 Glenroy 0.85 Neo 4.40 45 Truganina 0.40 Neo 7.15
16 Glenroy 1.75 Neo 3.04 46 Tullamarine 0.20 Neo 1.13
17 Glenroy 2.00 Neo 2.87 47 Tullamarine 0.40 Neo 3.46
18 Hastings 0.65 Nbr 2.55 48 Tullamarine 0.75 Neo 2.56
19 Horsham-1 0.70 Neo 2.22 49 Tullamarine 1.75 Neo 2.88
20 Horsham-2 1.70 Neo 2.57 50 Vermont 0.75 Sxa 1.75
21 Horsham-3 2.70 Neo 1.69 51 Vermont 1.45 Sxa 1.57
22 Horsham-4 3.75 Neo 2.32 52 Warragul 0.70 Nun 1.08
23 Kilmore 1.00 Dxh 2.25 53 Warragul 0.75 Nun 3.28
24 Kinglake 0.60 Dxh 2.80 54 Williams Landing 0.65 Neo 11.0
25 Lyndhurst 1.00 Neo 3.76 55 Williamstown 0.30 Neo 5.96
26 Lyndhurst 1.60 Neo 3.23 56 Williamstown 1.20 Neo 6.32
27 Macleod 0.80 Sxm 2.51 57 Williamstown 2.00 Neo 7.02
28 Melton West 0.50 Neo 5.55 58 Wollert 0.75 Neo 2.95
29 Melton West 1.20 Neo 5.51 59 Wollert 0.75 Neo 6.88
30 Mernda 0.40 Neo 2.70 60 Wollert 0.50 Neo 6.37
* refer to Table 2
The distribution of shrink-swell indices for different soil types is presented in Table 2 and Figure 10. The results
indicate that there is a significant variation in Iss values (varying from 1.13 to 11 % strain/pF) for Newer Volcanic soils.

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 67

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 67
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

As a consequence, frequent laboratory testing for Iss to calibrate visual tactile logging is required or a significant
database needs to be available and referred to.
Figure 11 displays the average shrink-swell indices and frequency for different basaltic clays. The New Volcanic
basaltic samples (mainly from the Western Suburbs of Melbourne) are consistently recorded as having a shrink-swell
index in the range of 4 to 6.5 % strain /pF, higher than that of the Old Volcanic basalts. Figure 12 shows that the
quaternary soils comprise 73% of the data with an average shrink-swell index of 4.25 % strain/pF, which is obviously
much higher than that of the Neogene and Silurian soils.
Table 2: Iss range for different soil types.

Geology Iss (% strain/pF) Range

Material Age Code Min Max No.
Melbourne Formation Silurian Sxm 2.51 2.51 1
Anderson Creek Formation Silurian Sxa 1.27 1.75 3
Humevale Siltstone Devonian Dxh 2.25 2.80 2
Brighton Group Neogene Nb 2.46 2.46 1
Red Bluff Sandstone Neogene Nbr 1.87 4.21 5
Torquay Group Neogene Nt 3.06 3.06 1
Loxton Sand Neogene Nwl 0.34 0.34 1
Neerim Volcanic Group Neogene Nun 1.08 3.28 2
Alluvium Quaternary Qa1 1.91 1.91 1
Colluvium Quaternary Qc1 1.11 1.11 1
Swamp and Lake Deposits Quaternary Qm1 2.42 2.42 1
Newer Volcanic Group Quaternary Neo 1.13 11.00 41

Figure 10: Distribution of shrink-swell indices for different soil types.

68 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

68 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 11: Distribution of shrink-swell indices for different basaltic soils.

Figure 12: Distribution of shrink-swell indices for sedimentary soils based on the soil age.
Soil suction measurement was taken not only before each shrink-swell test (i.e. initial soil suction), but also at the end
of the swell test and core shrinkage test. Measurements were taken in a constant temperature room using a Wescor HR-
33T Dew Point Microvoltmeter with C52 sample chambers (range: 3.2 – 5 pF or -155 to -9800 kPa). Each reported
suction was taken as the average of three sub-samples.
The state of soil suction at the end of shrink-swell testing is plotted in Figure 13. The values of total suction at the end
of swelling after a minimum of ten days of testing range from 2.82 (65 kPa) to 3.83 pF (660 kPa) with a mean value of
3.36 pF (224 kPa), which is larger than the value of 2.2-2.5 pF (16-31 kPa) reported in Fredlund & Rahardjo (1993).
This indicates that the osmotic suction of Melbourne’s expansive soils may range from 50 kPa to 630 kPa. After air
drying for a period of at least 10 days, the suction of the core shrinkage specimens was measured by using a WP4 Dew
Point PotentiaMeter (range: 0 to -300 MPa with an accuracy of 0.1 MPa). Its working principle is similar to Wescor
HR-33T, i.e. the relative humidity in the sample determines the suction of soils. The total suctions at the shrinkage limit
vary between 4.88 to 6.32 pF with an average of 5.9 pF, which is consistent with the total suction value of 6 pF (air
dry) reported in Lytton (1994).
In Figures 14-15, shrink-swell indices are plotted against the initial and final moisture contents, respectively. While
there is a generally increasing trend in Iss with increasing initial and final moisture contents, the correlations are very
scattered. The patterns of results are similar to those reported by Fityus et al. (1996) and Young and Parmer (1999).

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 69

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 69
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 16 shows the variation in Iss with depth. While most soil samples were taken from within the top 1.5 m depth, a
small number of samples were taken from deeper levels. There is no correlation of shrink-swell index with sampling
depth as shown in Figure 16.

Figure 13: The measured soil suction values at the end of shrink-swell testing .

Figure 14: Shrink-swell index versus initial moisture content.

70 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

70 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 15: Shrink-swell index versus final moisture content.

Figure 16: Shrink-swell index versus sampling depth.

6 CORRELATION BETWEEN THE SHRINK-SWELL INDEX AND TRADITIONAL

SOIL INDICES
For most engineering purposes, the properties of soil can be defined by index tests such as plasticity index, linear
shrinkage or liquid limit. Several studies have been carried out in an attempt to establish possible relationships between
these traditional soil indices and the shrink-swell. Cameron et al. (1999) found that there was a reasonable correlation
between the core shrinkage index and plastic index. Jayasekera and Mohajerani (2003) reported that there were strong
correlations between Iss and plasticity index, liquid limit and Free Swell Index. Mitchell and Avalle (1984) found linear
shrinkage to be a reasonable indicator of soil reactivity. A study undertaken by Young and Parmar (1999) in the Sydney
region shows no obvious correlation existed between traditional soil indices and Iss. A number of correlations were
attempted in this study between the shrink-swell test and other soil tests including liquid limit (LL), plastic limit (PL),
plastic index (PI) and linear shrinkage (LS). The results are presented in Figures 17-20.

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 71

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 71
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

For a correlation to be considered as a reliable mean for estimating the shrink swell index of a soil, the coefficient of
determination, R2 of the trend line needs to exceed 0.8. It is apparent that there is a poor correlation between Iss and LL,
PL, PI and LS since the strength of the correlations is only 19.6% - 53.3% (i.e. R2 = 0.196 - 0.533).

Figure 17: Shrink-swell index versus liquid limit (all samples).

Figure 18: Shrink-swell index versus plastic limit (all samples).

72 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

72 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 19: Shrink-swell index versus plasticity index (all samples).

Figure 20: Shrink-swell index versus linear shrinkage (all samples).

Cameron (1989) suggests that greater conformance would be achieved if a correlation were confined to a single soil
type. Based on the results of the laboratory testing on soils samples collected from twenty field sites in the Newcastle-
Hunter region, Delaney et al. (2005) found that the traditional index tests were of little general value in assessing the
reactivity of expansive soil unless the correlation was confined to a particular soil type.
In the current study, the New Volcanic basaltic samples comprise 68% of the samples tested. A number of correlations
which were confined to the New Volcanic basaltic samples were attempted. The results are presented in Figures 21-24.
Although an improved correlation can be observed, an unambiguous, reliable relationship does not appear to exist.

Figure 21: Shrink-swell index versus liquid limit (New Volcanic samples).

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 73

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 73
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Figure 22: Shrink-swell index versus plastic limit (New Volcanic samples).

Figure 23: Shrink-swell index versus plasticity index (New Volcanic samples).    
 

Figure 24: Shrink-swell index versus linear shrinkage (New Volcanic samples).
Table 3 summarises the range of Iss and average values of traditional soil indices. The results are also plotted in Figure
25. Generally, the higher traditional soil indices, the greater reactivity of soil to moisture changes. However the results
reported here show that no obvious correlations exist between the shrink-swell index and other traditional soil indices.

74 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

74 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016
SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

Table 3 Summary of shrink-swelling test results vs average values of LL, PL, PT and LS.

Other soil indices

LL (%) PL (%) PI (%) LS (%)
Range of Iss

0-2 (%/pF) 50.05 20.69 31.39 13.21

2-4 (%/pF) 50.53 21.42 29.11 14.50
4-6 (%/pF) 67.13 24.74 42.39 18.78
＞6 (%/pF) 80.33 27.54 52.79 21.25

Figure 25: Range of Iss test results versus average values of traditional soil indices test results.

7 CONCLUSIONS
As part of a long term research project on the nature and distribution of expansive soils and the performance of
residential slabs in the Melbourne metropolitan area, a series of laboratory tests have been performed on soil samples
collected from 47 different field sites across 37 suburbs of Melbourne and surrounds, which cover a wide range of soils,
geological conditions and geographic settings. The results of the laboratory tests were used to establish a database of
shrink-swell index values.
The results reported in this study have confirmed the variability of the reactivity of Melbourne soils which may be
difficult to assess by visual-tactile means. Therefore there is a need for more frequent benchmark laboratory testing by
the local consulting industry.
The results also show that for the soil samples tested, there is no obvious correlation between the shrink-swell index
values and the values of traditional soil indices such liquid limit, plastic limit, plasticity index and linear shrinkage.
Hence the traditional soil index alone cannot be relied upon as an accurate indication of the soil reactivity.
The database and the interactive reactive-soil map is intended to be useful to geotechnical practitioners as it can be used
to assess or calibrate the shrinkage index estimated on a visual-tactile basis by local practitioners.
Only limited numbers of shrink-swell tests have been conducted so far. More tests are required for the database and the
interactive soil map also requires further input and analysis.

8 ACKNOWLEDGEMENTS
The authors would like to gratefully acknowledge Raymond Tamstil and Lijing Wang for their contribution to this
project. The assistance given by FMG Engineering in collecting soil samples is greatly appreciated.

AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016 75

AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016 75
 SHRINK-SWELL INDEX DATABASE FOR MELBOURNE LI ET AL.

9 REFERENCES
Archbold, N.W. 1992. Outline of the Stratigraphy of the Melbourne Region, Engineering Geology of Melbourne, Eds.
Peck et al., A A Balkema,
Australian Standard AS 2870. 2011 Residential slabs. and footings. Standards Australia.
Australian Standard AS 1289.7.1.1. 2003 Soil reactivity tests - Determination of the shrinkage index of a soil – Shrink-
swell index. Standards Australia.
Australian Standard AS1289.7.1.2. 1998. Soil reactivity tests - Determination of the shrinkage limit of a soil – Loaded
shrinkage index. Standards Australia.
Australian Standard AS1289.7.1.3. 1998. Determination of the shrinkage limit of a soil – Core shrinkage index.
Standards Australia.
Cameron D.A. 1989. Tests for reactivity and prediction of ground movement. Civil Engineering Transactions,
Institution of Engineers, Australia, , CE 31, no.3, pp. 121-132.
Cameron, D.A. 2015. Management of expansive clay soils in Australia. Invited lecture, International symposium,
Shrink Swell processes in soils, Climate and Constructions, Paris, IFFSTAR, pp. 15-36.
Cameron D.A., Buttignol C., Nguyen T. 1999. A database of shrinkage indices for Adelaide. Proceedings, 8th ANZ
Conference on Geomechanics, Australian Geomechanics Society, Hobart, Tasmania.
Delaney, M. G., Li, J. and Fityus, 2005. S. G. Field Monitoring of Expansive Soil Behaviour in the Newcastle-Hunter
Region. Journal of Australian Geomechanics, Vol. 40, No. 2, pp. 3-14.
Fityus S.G. 1996. The effect of initial moisture content and remoulding on the shrink-swell index, Iss. Proceedings, 7th
ANZ Conference on Geomechanics, IE Australia, Adelaide, pp. 388 - 393.
Fityus, S. G. and Welbourne, J. C. 1996. Trends in Shrink-Swell Test Results in the Newcastle Region, Proceedings of
the 7th Australian-New Zealand Conference on Geomechanics, Institution of Engineers Australia, Adelaide, pp.
394–399.
Fityus S.G., Cameron D.A., Walsh P.F. 2005. The shrink swell test. ASTM Geotechnical Testing Journal, 28 (1), pp. 1-
10.
Fredlund, D .G. and Rahardjo, H. 1993. Soil Mechanics for Unsaturated Soils, John Wylie & Sons, New York.
Jayasekera, S. and Mohajerani, A. 2003. Some relationships between shrink-swell index, liquid limit, plasticity index,
activity and free swell index, Australian Geomechanics, 38(2): 53-58.
Li, J. and Cameron, D.A. 2002. A case study of a courtyard house damaged by expansive soils. ASCE Journal of
Performance of Constructed Facilities, Vol. 16 (6), p 169-175.
Li, J., Cameron, D.A. and Ren, G. 2014. Case study and back analysis of a residential building damaged by expansive
soils. Computers and Geotechnics, Vol. 56, p 89-99.
Lytton, R. L. 1994. Prediction of movement in expansive clays. Geotechnical Special Publication No. 40, Yeung, A.T.,
and Felio, G.Y., eds. ASCE, New York, Vol.2,1827-1845.
Mitchell, P and Avalle, D 1. 1984. A technique to predict expansive soil movements, Proceedings of the Fifth
International Conference on Expansive Soils, Adelaide, South Australia, pp. 21-23.
McAndrew, J. 1965. The Geological Map of Victoria in Geology of Australia Ore Deposits. The Eighth Commonwealth
Mining and Metallurgical Congress Australia and New Zealand, The Australian Institute of Mining and
Metallurgy.
Young G.S, 1999. Parmar M. Shrink-swell testing in the Sydney region. In: Proceedings of the 8th ANZ conference on
geomechanics, Hobert, Tasmania, pp 221–225.

76 AUSTRALIAN GEOMECHANICS VOLUME 51 NO. 3 SEPTEMBER 2016

76 AUSTRALIAN GEOMECHANICS | VOLUME 51: NO.3 SEPTEMBER 2016

View publication stats