Introduction

Expansive (or reactive) soils are associated with ground movements due to seasonal moisture content changes [1] that can cause significant damage to infrastructure such as roads, railways, and individual dwellings. The damage caused by expansive soils in the UK exceeds $3.7 billion, China $15 billion, France $2.71 billion, USA $15 billion, Saudi Arabia $300 million, and India $73 million annually [2,3,4]. Australia’s infrastructure has also suffered from expansive soils [5, 6] where six out of eight of Australia’s largest cities are significantly affected by reactive soils.

Engineers have traditionally used a variety of design and construction options to deal with expansive soils, as well as stabilization techniques, including lime and cement, to mitigate the expansive soil problem over roads, railways, foundation bases, and other small structures [7,8,9,10,11,12,13,14,15,16]. However, experience shows that the use of lime and cement may not be effective for specific soil conditions [17, 18] and it may cause environmental issues by increasing the CO2 emission [19].

Seeking an alternative solution [20], recent studies have shown that geosynthetics can reduce the swell of expansive soils and increase their strength. Zornberg and Roodi [21,22,23] reported that a geosynthetics-stabilized base course material was effective in preventing the development of longitudinal cracks on the road compared to the lime stabilization of the expansive subgrade. Badaradinni et al. [24] found that geotextile helped reduce the total swell of expansive soils by more than 70%. Viswanadham et al. [25], noted that if placed vertically, geosynthetics can significantly reduce soil swell, compared to the same material placed horizontally. However, due to the limited number of studies, many uncertainties regarding the effect of geosynthetics on the swell reduction of expansive soils still exist. In particular, it is not clear how geosynthetics affect the swell potential of soil and what type of geosynthetic material can provide better outcomes.

This study seeks to investigate the effect of geosynthetics on the behavior of reactive soil and identify the factors that can reduce swell. A series of large-scale long-term (30 days) experiments were conducted on reactive clay with two types of geosynthetic material. A very high plasticity soil was collected from the Toowoomba area, QLD, Australia, where it has long been associated with road cracking [26]. The performance of this soil with and without geosynthetics during rainfall was evaluated by reproducing an extreme rainfall event that occurred in 2011 in Toowoomba and caused the flooding of a large area [28]. This paper presents and discusses the obtained results.

Materials and Method

This work consists of a series of California Bearing Ratio (CBR) on expansive soil with and without geosynthetics, and three large-scale long-term soil column tests in which the effect of geosynthetics on the vertical swell of soil was studied using simulated rainfall conditions.

Soil Properties

Properties of the studied soil are given in Table 1 below.

Table 1 Soil properties
Full size table

The results of the X-ray diffraction as shown below in Fig. 1. The test revealed the following dominant minerals: quartz, plagioclase, smectite, kaolinite, and illite.

Fig. 1
figure 1

X-ray diffraction result of the soil

Full size image

Geosynthetics Properties

Geosynthetics used in this experiment were chosen based on three categories: (1) ease of use, (2) cost of installation, and (3) availability in the market. Based on the factors, three types of geosynthetics used are non-woven, woven, and geogrid. Materials for the experiment were provided by Geofabrics Australasia Pty Ltd (Gold Coast). Table 2 below show details of the Geosynthetics used for this study.

  • GS1: non-woven geosynthetic is a highly porous material that allows water to pass through while preventing soil migration. It provides excellent filtration, and it is commonly used in the construction of roads, railways, and embankments where the ground is soft and unstable.

  • GS2: woven geosynthetic that is commonly used as the dewatering system for a range of engineering applications including ground stabilization, railroad support, and aggregate separation (Fig. 2a).

  • GS3: geogrid can be used for aggregates and soil stabilization, and it is primarily used in railways and road works for pavement reinforcement (Fig. 2b). In this study, GS3 was combined with the GS1 to increase the stiffness of the non-woven geosynthetic which may result in improvement in the load carrying capacity of the soil [24].

Table 2 Geosynthetics properties
Full size table
Fig. 2
figure 2

a Woven geosynthetic (GS2) and b non-woven geosynthetic (GS1) combined with geogrid (GS3)

Full size image

California Bearing Ratio (CBR)

A series of CBR tests were conducted according to AS1289.6.1.1 (2014). The soil samples were cured at the maximum dry density (MDD) and at optimum moisture content (OMC) for 4 days. During the CBR experiment a load with a constant rate of penetration of 1.0 ± 0.2 mm/min was applied, and load readings were taken at the following penetrations (in mm): 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0, 7.5, 10.0 and 12.5, respectively. The following four initial setup conditions were used:

  • Soil (without geosynthetics).

  • Soil with GS1 combined with GS3.

  • Soil with GS2.

A single layer of geosynthetics was placed at 1/3rd height of the mould following the previous studies in which the effect of geosynthetics on CBR was investigated [27, 29, 30]. For each initial setup, five tests were conducted, and the average value of CBR was reported.

Soil Column Equipment

The soil column used for the experiment was an aluminium-based custom-made column with dimensions of \(1.3 \times 0.3 \times 0.3 \mathrm{m}\) (Fig. 3). The soil mass used in the column was measured at 1.0 m. According to Bergström [32], the minimum ratio for diameter to length of the soil column should be 1:4 and a minimum surface area of the column should be 0.05 m2. The column used in this experiment meets all the minimum criteria. It was fitted with a range of moisture, suction, and displacement sensors as well as a glass front for visual inspection of the soil profile. The following sensors were used:

  • MP 306 Moisture Probe which has a range of 0–100% and an accuracy of ± 2%.

  • Suction Sensor used were Meter Teros-21, which has a range of – 9 kPa to – 2000 kPa with an accuracy of ± 10%. It also measures the temperature of the soil with a range of − 40 to + 60 °C, with an accuracy of ± 1 °C.

  • Panasonic HG-C1100 Digital Laser sensor were used for measuring the vertical displacement (L1 and L2). The sensor has a range of ± 35 mm from an initial measuring distance of 100 mm with an accuracy of 70 μm.

Fig. 3
figure 3

Soil column experiment setup: a sensors and their location, b image of the soil column filled with the studied soil

Full size image

Sample Preparation

To prepare the soil column test, oven-dried soil was mixed with 20% water (by weight) and allowed to cure for 24 h to achieve an even moisture distribution in the soil mass. The moist soil was then placed in the column in several layers and compacted at a rate of 25 blows per 100 mm layer to achieve a desired value of dry density of 1.3 g/cm3. Two aluminium rods (labelled L1 and L2 in Fig. 3b) were placed in two locations: L1 at a depth of 150 mm, and L2 at a depth of 600 mm. Both rods were placed in line with the laser sensors and were used to measure the vertical displacement within the soil.

Three long-term soil column experiments were performed as follows:

  • Test 1: Non-reinforced soil.

  • Test 2: Soil with a single layer of GS1 supported by GS3.

  • Test 3: Soil with a single layer of GS2

The geosynthetic material was placed at a depth of 250 mm from the surface of the soil, where the largest amount of soil swell was expected (based on the result of a soil column test conducted on non-reinforced soil).

Column Test Procedure

This study simulated a severe rainfall event that occurred in Toowoomba, Queensland in 2011 when a total of 320 mm of rainfall precipitation occurred over 6 days. Table 3 shows the amount of water that was introduced in the column to reproduce the event. This was followed by a minimum of 21 days of drying period to observe the changes in the soil mass caused by this simulated rainfall. Water was introduced to the top of the column at the rate of 1 l/30 min. The rate of water introduced to soil column was calculated by multiplying the surface area of the column (\(0.3\times 0.3\mathrm{m}\)) by the precipitation of individual day, this was then converted into litre, l.

Table 3 Amount of water introduced in the soil column for the first 6 days of each experiment
Full size table

Results and Discussion

CBR Test Results

The average values of CBR for different initial conditions are given in Table 4. Based on the obtained results, the main observations can be made as follows: (a) for all tests, the CBR value was observed to be below 2. (b) The addition of geosynthetic materials did not have any substantial effect on the soil’s CBR. This finding seems to contradict the outcomes of the previous research [27, 30, 31] where the inclusion of geosynthetics increased the CBR values of soil. This discrepancy can be attributed to the difference in soil plasticity of the tested soils. Previous studies tested soils with a range of LL from 29–58%, which is significantly low than the LL (79.2%) of the soil used in this study. Considering this, it may be concluded that geosynthetics alone might not have an impact on improving the strength of very high plasticity expansive soils.

Table 4 Results of CBR tests
Full size table

Soil Column Experiment Results

The development and propagation of the wetting front and corresponding changes in the water content during the test without geosynthetics are shown in Fig. 4 below. After the water was introduced to the top of the soil mass, it gradually permeated through the soil as shown in Fig. 3 for day 1 (Fig. 4a1) and day 3 (Fig. 4a2). As more water was introduced in the first 6 days of this experiment, the wetting front increased in day 5 (Fig. 4a3) and eventually reached the bottom of the column after several days (day 17 in Fig. 4a4).

Fig. 4
figure 4

Water infiltration front profile of column experiment: a non-reinforced soil, b GS1, c GS2

Full size image

The corresponding changes in the water content at four depths are given in Fig. 5a. It is evident from this figure that the water content rapidly increased in the top part of the soil mass, (w1 and w2 in Fig. 5a), followed by a gradual increase at greater depths (w3 and w4, respectively). The change in soil suction is given in Fig. 5b, it decreased as the water content increased due to the wetting front propagation.

Fig. 5
figure 5

Results for column test with soil only: a water content changes, b suction changes, and c vertical displacement changes

Full size image

Figure 5c gives the amount of vertical displacement (swell) observed during the test at different depths: the L1 sensor (150 mm below the surface) recorded a total swell of 17 mm, while the L2 sensor (600 mm below the surface) indicated the settlement (consolidation) of 2 mm. This consolidation may be due to the increase in weight of the water and the soil mass as the depth increases.

Soil Column Tests with Geosynthetics

Figure 4 shows the wetting front of the soil profile for the soil column experiments with the geosynthetic material. The non-woven geosynthetic provides better drainage during the first few days of the experiment, compared to the woven and non-reinforced soil. Visual observation also revealed more even moisture distribution in the soil mass with the non-woven geosynthetic and more rapid wetting front propagation through the soil. This was associated with greater values of water content recorded by the water content sensors w1 and w2 on the 3rd day.

The data recorded by the water content sensors (w1 and w2) suggests (Fig. 4) that for all three tests, the water content of the top part of the soil mass rapidly increased to higher values in the first few days of the experiment. However, only for the experiment with GS1, a significant increase in the water content (about 40%) at a greater depth of 400 mm was recorded by the sensor w3 on day 3, while for the non-reinforced soil and the soil with GS2, the water content at a depth of 400 mm (w3) was still very close to its initial value of about 20%.

This finding suggests better drainage provided by non-woven geosynthetic (GS1). This was due to the higher flow rate (Table 2) of non-woven geosynthetic, where water can rapidly permeate down the soil mass instead of accumulating in the top part as observed for the other two experiments. It is noted that for all cases, a slight decrease in water content in the top part of the soil mass was recorded by day 17 when water permeated down the soil mass. For all experiments, the suction change occurred rapidly after day 1 of water influx (Fig. 6b).

Fig. 6
figure 6

a Water content changes measured by w1, b suction changes measured by S2, and c vertical displacement changes measured by L1

Full size image

Soil Mass Swell

The cumulative vertical displacement (swell) recorded by the L1 sensor (150 mm below the ground) for all three large-scale column tests is given in Fig. 7. It is evident from this figure that the largest amount of swell occurred within the first 160 h of each test; that is within 6 days when the water was introduced to the soil mass as part of rainfall simulation. Figure 7 gives the change in vertical displacement over the period of 30 days. Compared to the non-reinforced soil, the total swell reduction in non-woven geosynthetic (GS1) was estimated to be around 44%, and 29% in woven geosynthetic (GS2).

Fig. 7
figure 7

Cumulative vertical displacements over 30 days recorded by L1

Full size image

Figure 8 shows the relationship between the rate of vertical displacement per day in the top part of the soil mass, measured at a depth of 150 mm (L1) and the change in degree of saturation (S) per day from day 1 to day 10, measured at a depth of 200 mm (w2). As can be seen in this figure, the degree of saturation (S) increased, reaching its peak values (typically in a range from 80–85%) in 5–6 days. As a result, the rate of vertical displacement (swell) gradually decreased, and it remained low when the soil was close to its saturation. It is assumed that soil with a greater degree of saturation already has a significant amount of water in the pore space, which limits the further expansion of the soil.

Fig. 8
figure 8

Rate of vertical displacement and degree of saturation (S) over time

Full size image

Further analysis of the data in Fig. 8 shows that the largest amount of swell occurred within the first 2 days of each test. A large amount of swell occurred in test 1 (soil only) after day 1, compared to the soil mass with either GS1 (test 2) or GS2 (test 3). This also correlates with the higher degree of saturation recorded for the non-reinforced soil experiment (S = 80%), compared to the degree of saturation measured in the soil with geosynthetics. This data suggests that there was more water in the top part of the soil mass without geosynthetic (test 1) in the first few days of the experiment, which seems to have enhanced the swelling process. However, when the geosynthetics materials were used (either GS1 or GS2), the top part of the soil mass was less saturated, as moisture was more evenly distributed across the soil mass, while some water could also permeate deeper in the soil profile. As a result, less water in the top part of the soil mass generated a smaller amount of swell in these two experiments where the geosynthetics were used (tests 2 and 3).

Conclusions

This paper seeks to better understand the effect of geosynthetics on the swell of very high plasticity (LL = 79.2%) reactive soil during a severe rainfall event. A series of CBR tests and three long-term large-scale soil column tests were conducted on soil reinforced with non-woven geosynthetic (GS1) and woven geosynthetic (GS2). Based on the obtained results, the following major conclusions can be drawn:

  • Both types of geosynthetics did not improve the CBR of the studied soil. The CBR values obtained for the reinforced and non-reinforced soil were below 2, which can be attributed to extremely high plasticity of the tested soil.

  • Both types of geosynthetics were effective in reducing the total vertical displacement (swell) of the soil mass. The non-woven geosynthetic combined with geogrid provided a greater reduction in vertical swell due to higher flow rate of the material compared to the woven geosynthetic and non-reinforced soil.

  • The reduction in total vertical swell of the soil was directly related to the degree of saturation (S) of the soil mass. When the soil reached a high value of S (80–85%) after the initial wetting phase, the rate of vertical swell significantly decreased which cause a lower swell rate for the remainder of the experiment.

  • The geosynthetic materials used in this study produced better drainage conditions for the soil mass compared to non-reinforced soil. This limited the time of soil–water interaction in the top part of the soil mass resulting in a lower vertical swell observed from the tests.