Partially saturated or unsaturated are named these soils in which the voids are occupied partially by water and partially by air (Figure 1). As such, an unsaturated soil consists of three phases, a solid phase (soil grains), a fluid phase (pore water), and a gas phase (pore air). While the soil grains and the water are considered practically incompressible, the air is very compressible (i.e. its volume reduces when the pressure increases). Due to this peculiarity, unsaturated soils exhibit different behaviour compared to dry and saturated soils and this is the reason why the laboratory testing of these soils is unique.
Partially saturated soils (hereafter mentioned as unsaturated soils) can be found in any place of the world; in fact, almost any soil can exist in an unsaturated state when it is subjected to evaporation and some of the pore water is replaced by air. However, in areas with tropical climates, these soils form most of the ground soil. They are usually located on the upper layers of a soil sequence, near the surface, and always above the groundwater table. Within the unsaturated soil, water is held under tension which indicates that pore water pressure (PWP) is less than pore air pressure (PAP). Since the air pressure is close to the atmospheric pressure in shallow depths, PWP is becoming negative. The difference between the PAP and PWP in a soil specimen is called Suction:
S = Ua – Uw (1)
where S is the suction, Ua is the PAP and Uw is the PWP. Since pore water pressure is negative, suction in unsaturated soils is always positive. However, when the soil becomes fully saturated suction becomes zero as there is no more air in the voids.
Triaxial testing in unsaturated soils
Triaxial testing of unsaturated soils is performed to evaluate their shear strength under various environmental conditions. For example, excessive evaporation will cause soil to become more and more unsaturated and as a result, pore water pressure will turn into more and more negative. That will cause a significant increment in the shear strength of the soil. However, after a rainfall event, the soil becomes fully saturated and suction is lost, so the shear strength reduces significantly. This is the main reason for the occurrence of landslides in tropical and sub-tropical areas.
One of the most difficult procedures in the triaxial test of an unsaturated soil is the determination of its volumetric strain, i.e. ΔV/V, where ΔV is the change in the specimen’s volume and V is the initial volume. The change in the volume of soil happens when its voids reduce or increase (shrinkage or expansion, respectively). When the sample is fully saturated and all the voids are filled with water, any reduction in volume will cause a change in the water volume. In unsaturated soils, though, the reduction in the sample volume will cause a change in the volume of both, water and air, according to the following equation:
ΔV = ΔVa + ΔVw (2)
where ΔVa is the air volume change and ΔVw is the water volume change. While the measurement of the water volume can be easily done with conventional laboratory equipment, the air volume change is not easily measured. This is because air is highly compressible and also dissolves easily into the water. For that reason, the triaxial cells for unsaturated soil testing have been modified appropriately to allow measuring the volume of the sample directly.
When testing unsaturated soils, it is usually required to apply a known value of suction to the specimen. This has practical problems since pressure controllers will not be able to control negative values of pressure. Therefore, a technique called Axis Translation has been applied. According to Equation 1, suction is the difference between the PAP and PWP. Under atmospheric conditions PAP ≈ 0 kPa, therefore suction will be equal to –Uw. As an example, if we would like to apply a suction of 80 kPa to the specimen we would need to apply PWP equal to -80 kPa, which is practically impossible. The solution is to elevate both PAP and PWP by equal amounts so that their difference remains at 80 kPa (i.e. translate the y-axis). As shown in Figure 2 by increasing PAP to 100 kPa we would just need to bring PWP to 20 kPa to get a value of suction equal to 80 kPa. In most cases, PWP remains constant at a small value (10-30 kPa) and so suction is controlled by adjusting the PAP. PWP is controlled in unsaturated soil testing through the HAE disc located on the base pedestal.
This guide intends to provide an introduction to unsaturated triaxial testing for laboratory staff. The main procedures, along with the most common equipment, will be presented. It is not the intention to provide either the theoretical background on the mechanics of unsaturated soils or the theory for the interpretation of the laboratory results; therefore, users must accompany this guide with knowledge about unsaturated soil mechanics which can be retrieved from the books mentioned at the end of this document. For any questions about this guide please send an email to Service@vjtech.co.uk.
Differences in triaxial testing between saturated and unsaturated soils
The main differences between triaxial testing in saturated and unsaturated soils are:
• Saturation: In the triaxial testing of saturated soils, the samples which are due to be subjected into a consolidation stage need to become fully saturated. In unsaturated soils this is optional and usually it is avoided. Furthermore, different procedures needed for the saturation of these two types of specimens.
• High Air Entry discs: While in saturated soil testing we use a coarse porous stone below the sample to allow the water draining out of the sample and measuring pore water pressure, in unsaturated soil testing we use a High Air Entry (HAE) disc instead. The specifications and principles of these discs are discussed below.
• Air pressure supply: While in saturated triaxial specimens we supply only de-aired water to the samples through the top, using the Back Pressure line, in unsaturated soil specimens we supply pressurised air on top. This is done in order to control the pore air pressure and therefore to control suction through the Axis-Translation technique.
• Sample volume measurement: The determination of the volume change of the specimen is critical in the analysis of the unsaturated triaxial test. The only practical way to measure the volume changes of an unsaturated sample is to observe the changes in its dimensions (i.e. height and diameter). This can be done in three ways: a) by immersing the sample in a tube full of mercury and then monitoring the mercury level optically, b) by using on-sample transducers that will accurately measure the dimensions of the sample and c) by monitoring the volume change of the confining fluid which fills the cell that accommodates the sample. Method (a) is very straightforward; however, it has several limitations and as such is not used very often. Method (b) requires the use of small-sized transducers to be mounted onto the sample which can determine accurately the changes in the specimen’s dimensions. This option has, again, some limitations which are discussed in section 3.6. Most commonly, volume change is determined by measuring the volume of the confining fluid that is displaced from the triaxial cell during the test. A device called Volume-Change device is used for that, details can be found below.
• Pore Water Pressure: The pore water pressure that is generated within a saturated sample is usually measured by a PwP transducer. On the contrary, in unsaturated triaxial tests PwP is normally controlled through a hydraulic pressure controller. In this way, a known pore water pressure is induced to the sample and with the Axis-Translation technique suction is applied.
Summary of an unsaturated triaxial test
Unsaturated triaxial tests are not very common tests in geotechnical laboratories and they are usually performed either by advanced geotechnical laboratories or research-type laboratories. Since this is not a standardised test, the procedures that need to be followed can be varied depending on specific requirements. In general terms, an unsaturated soil specimen is placed onto a modified triaxial cell in the same way as in a saturated triaxial cell. Water is supplied to the sample from the base pedestal, through a HAE disc on which the specimen sits, to control the Pore Water Pressure. Air is supplied to the top of the specimen to control the Pore Air Pressure. Initially, the specimen is left to equalise by applying a combination of Net stress and Suction and leaving it to drain freely under these conditions. When all the excess water has been removed from the sample, the specimen can be subjected to consolidation by applying a specific Vertical and Net stress as well as Suction. Finally, the specimen is subjected to a very slow shearing stage (usually drained) until failure. In most cases, more than 1 test must be conducted on the same specimen type with the same initial conditions under varied values of suction or net stress. Multistage tests are also possible with one specimen subjected to several consolidation and shear stages. The unsaturated triaxial test, however, can be modified to meet the requirements of the user.
Apparatus & accessories
A typical unsaturated triaxial system consists of a cell, a load frame (only in Double-Wall systems), hydraulic and pneumatic pressure controllers, a volume-change device and various sensors.
In general, there are two types of triaxial cells used in laboratory testing of unsaturated soils, Double-Wall and Twin-Cell type. They, both, intend to provide a reliable way of measuring the actual soil volume change during the test. Both types consist of two compartments, the inner cell in which the soil specimen is located and the outer cell which surrounds the inner one. During the test, both compartments are filled with de-aired water, and the confining pressure that is applied to the specimen is applied, also, to the outer cell. In this way, the inner cell wall remains under zero differential pressure and is not deforming at all, allowing for accurate measurements of the confining fluid (i.e. de-aired water) volume The two compartments are independent and isolated one from the other; however, they are communicating only through a pipe which passes through a volume change device. When the volume of the soil specimen is increasing (i.e. the specimen swells), water will be pushed out of the inner cell to maintain the pressure. At the same time, this amount of water that has been expelled from the inner cell will move through the volume change device which will record it. The recorded volume corresponds to the volume change of the specimen. When the specimen shrinks, water moves towards the inner cell and it is, again, recorded by the volume change device. Both types of triaxial cells have their inner cell wall made of glass, while the outer wall is made of acrylic material. The use of glass as an inner cell wall has the advantage that preventing water from the inner cell compartment from infiltrating into it through very small cracks that acrylic might develop after some time. That would cause a change in the volume of the de-aired water occupying the inner cell which will be counted as a sample volume change. Moreover, the glass wall deforms much less than the Perspex wall when it is pressurised, which again helps in the accurate measurement of the sample volume change. However, glass is very fragile; therefore the water in the outer cell must be in the same pressure as in the inner cell to achieve zero differential pressure at the two sides of the glass wall.
To enable testing unsaturated soil specimens, both cell types bring a specially designed base pedestal fitted with a High Air Entry value disc. These are ceramic discs with specific pore sizes which are designed to allow only water to pass through up to a specific air pressure (Air-Entry value). In other words, if the HAE disc is fully saturated, then the air cannot pass through it until pressure becomes equal to the nominal Air-Entry value of the disc. The use of HAE discs is very common in unsaturated soil testing as they can separate the two phases, water and air. This is very important because the volume of water moving in and out of the specimen must be very accurately measured and free of air bubbles that can affect the measurement of the volume. HAE discs come in different AE values, usually from 0.5 to 15 Bar, with the highest AE values corresponding to the smallest voids. The selection of the appropriate HAE disc in an unsaturated test must be done based on the soil type that will be tested. In general, the AE value of the disc specifies the maximum suction that can be applied to the specimen. Therefore, coarse-grained soils need HAE discs with a low AE value and fine-grained soils need higher AE value discs.
An illustration of a HAE disc is shown in Figure 3. The soil sample is placed onto the base pedestal which brings the HAE disc. The disc is saturated before the sample installation. Under the disc, a small compartment (usually in the form of a spiral groove) is saturated with de-aired water using the two ports that exist in it. One of the ports is connected to the pressure controller that is used to provide the pore water pressure to the sample and the other port is used to flush water and remove air bubbles from the compartment. An elastic membrane and O-rings are used to seal the sample and prevent any contact with the confining fluid. When the pore air pressure is increased, water and air try to be removed from the sample through the HAE disc, however, due to the properties of the disc only water is allowed to pass. In this way, the sample is desaturating while the pore water passes to the saturated compartment and its volume is measured from the pressure and volume controller.
Furthermore, triaxial cells for unsaturated soil testing include an air supply line that is connected to the specimen through the top cap. An air controller regulates the air pressure in the soil pores (Pore Air Pressure) so that the correct suction is applied to the specimen (see Axis Translation technique in section 1.2). The air supply line is similar to the back pressure line that exists in triaxial cells for saturated soil testing. In some configurations, the top cap has two ports to supply the sample with both, water and air, which is useful for the determination of the unsaturated hydraulic conductivity (i.e. unsaturated permeability).
Double-wall cells (Figure 4) bring a cell wall that consists of glass on the inside and Perspex on the outside. Between the glass and the acrylic sides, there is a small space which is filled with de-aired water. When the sample is set up, the inner and outer cells are filled with de-aired water, and the same water pressure is applied in both. In this way, the glass wall is not deforming and is protected against cracking that the differential pressure could cause. D-W cells are quite convenient as they don’t require excessive work to be assembled before each test. Moreover, they have a bigger space in the inner cell which allows the use of on-sample transducers. Finally, these cells can be flexible regarding the size of the specimen, as the base pedestal can be easily swapped to accommodate samples from 38 mm and up to 100 mm in diameter.
The soil specimen sits on an HAE disc which must be very well saturated before it is used. Under the HAE disc, a very small compartment is filled with de-aired water and is connected to the pressure controller used to control the pore water pressure within the specimen. A coarse porous stone is attached to the top of the specimen and the top cap which provides the specimen with air pressure sits on top. The load cell is located as close to the sample as possible to avoid errors coming from the up thrust due to the confining pressure and the friction between the loading ram and the cell. The specimen is then wrapped with a rubber membrane which will separate the soil from the confining fluid. Load is applied to the sample through a loading frame on which the cell sits. The inner cell communicates with the outer cell only from a pressure line that passes through a volume change device which will measure the volume of the water that is been displaced. An automatic pressure controller will pressurise both compartments of the D-W cell at the same pressure.
Twin cell (Bishop & Wesley cell)
Twin-Cells are modified versions of the Bishop & Wesley cell and can be seen in Figure 5. Originally, Bishop & Wesley cells were used to perform stress path tests in saturated soil samples. Sivakumar et al (2005) presented a modified version of this cell which allows it to be used in the testing of unsaturated soil samples. The most important modification is the replacement of the acrylic inner cell with a high-precision glass wall. Glass offers the advantage that it does not absorb water into microcracks which are usually formed in acrylic cell walls after some time and affect the volume of the confining fluid. Moreover, glass walls are not greatly affected by ambient temperature changes as they do not deform significantly.
Since Twin-Cells are a modified version of the Bishop & Wesley cells, have also the advantage of incorporating a hydraulic loading system underneath which allows to apply the vertical stress to the specimen. Axial Force is exerted on the sample by a piston in the Base driven by a Hydraulic Automatic Pressure Controller. Given this, Twin-Cells do not require a separate loading frame to perform triaxial testing. While the loading ram touches the sample, it is fixed on the top of the cell. A load cell is located between the top cap and the loading ram to measure the axial force applied to the sample. The base pedestal, which brings the HAE disc, is fixed onto a moveable base which is controlled by the hydraulic system at the base of the cell. When the water that occupies the hydraulic loading system is compressed, the base is moving up and the sample is loaded. By calibrating the water pressure against the load and the displacement that it produces, we can apply a very accurate load to the sample or load it at a constant rate. Pore air pressure is controlled from a pneumatic pressure controller, through the top cap that sits on the specimen.
Even though twin cells have the same principle as the D-W cells, they are usually made to accommodate a specific sample size (e.g. 50 mm in diameter). Moreover, the inner cell has generally less volume than the inner cell of the Double-Wall cells which offers better accuracy in the measurement of the volume change. However, the restricted space does not offer the option of using on-sample transducers for local measurements of the sample deformation.
Load frame (only for D-W cells)
A loading frame is used in D-W cell configurations to apply the vertical load to the sample. The load frame must be capable of applying the loading at a very precise rate of axial strain. Usually, the load frame acts also as a data logger which records the raw data during the test (e.g. load, displacement, and pore water pressure, if necessary). Deviator stress is usually applied to the unsaturated specimen at a very low speed to allow the equalisation of suction during shear, in the same manner, pore water pressure needs to be evenly distributed in saturated triaxial tests. Sophisticated load frames can be used to provide the ability to shear at a constant stress rate instead of shearing at a constant speed.
In a DW cell system, two (2) hydraulic pressure controllers should be used for controlling the confining pressure, in both the inner and outer cells and the pore water pressure within the sample. In a Twin-Cell system, another hydraulic pressure controller is needed to be used in the hydraulic loading system of the Bishop & Wesley cell. The water used in these controllers must be strictly de-aired; therefore a very effective de-airing system must accompany them. Apart from the hydraulic pressure controllers, a pneumatic (Air) pressure controller must be used to provide the correct air pressure to the specimen and, therefore, to apply the suction.
Volume change device
As already mentioned, one of the greatest difficulties in triaxial testing of unsaturated soils is the determination of the sample volume. The difficulty comes from the fact that the voids of the specimen are partially filled with water and partially filled with air. When the volume of the specimen changes, the voids are either reducing or increasing in size. If the voids were filled with water, this change in the volume of voids would result in the same amount of water moving in/out of the sample. In a partially saturated specimen, though, that is not the case since there’s also air in the voids which can be compressed. It is, therefore, inevitable to measure the actual volume change of the sample.
A volume change device is used to determine the actual volume change of a triaxial sample. This is installed between the Inner and Outer compartments of an unsaturated triaxial cell to read the volume of the water that is moving in and out of the inner cell. The volume change device incorporates a hydraulic mechanism that moves a piston either upwards or downwards (depending on the direction of flow) when water is passing through it. The movement of the piston is proportional to the volume of the moving water; therefore after correlating the movement of the piston to the water volume, through calibration, the device can provide a very accurate volume of the water that moves in or out of the inner cell. After applying some necessary corrections, this reading will correspond to the change in volume ΔV of the sample during the test. A volume change device is shown in Figure 6.
A data logger device is needed to record all the raw test data and store them for the upcoming analysis. The data collected during a triaxial test in an unsaturated specimen are mainly the displacement input, for the determination of the axial strain, load, for the determination of the deviator stress, and volume change, for the determination of the volumetric strain. In the case of an undrained triaxial test, an additional pore water pressure sensor (for measuring only positive values of pressure) should be used to monitor the excess pore water pressure occurring during the axial loading stage. This sensor is located as close to the base pedestal as possible.
Additional sensors might need to be logged into the data logger, such as on-sample transducers and mid-height PWP transducers. All collected data must be synchronised to produce meaningful results, therefore they must be either logged from an individual logger or they must be logged through a controlling software.
On-sample transducers for volume measurements
As mentioned in section 1.3, the volumetric strain of the sample can be determined also with the use of on-sample transducers mounted on the specimen. These transducers provide very accurate measurements of the actual specimen’s height and diameter during the test and, therefore, the volume of the sample. The axial transducers consist of a set of brackets that are glued on the rubber membrane. A Linear Variable Differential Transformer (LVDT) is then mounted on the brackets and follows any sample height changes which are then recorded. Usually, two axial transducers are used on every sample, mounted at opposite sides (Figure 7b) to record the height changes in these two points. The final height change is taken as the mean value of the two readings. The radial transducer consists of a radial caliper which is glued onto the sample using a set of brackets (Figure 7b). The LVDT is attached to the caliper and measures the changes in its opening while the diameter is changing. Usually, one redial transducer is used on a specimen and is installed in the middle height point.
The use of on-sample transducers provides very accurate measurements of the sample dimensions, however, some limitations make the use of Volume-Change devices more popular for the measurement of the sample volume. The on-sample transducers have usually a very short range (from 2.5 mm to 10 mm) to provide high accuracy. Therefore, the deformation that the sample can accept is reduced and the permitted axial strain is limited. Moreover, the measurement of the diameter change in the middle of the specimen needs the assumption that the sample is bulging uniformly at all its height, which is generally not true. Finally, the use of the on-sample transducers can damage the specimen during their installation which will have an impact on the test results.
On-sample transducers for pore water pressure measurements
Small-sized pressure sensors can be installed at the mid-height of the specimen to provide a local measurement for the pore water pressure. They have a miniature body, on a cylindrical shape, with a diameter of about 10 mm which makes them ideal to be installed on a triaxial specimen without significant disturbance. These sensors have been modified to allow the measurement of suction inside the specimen. For this reason, they bring a small HAE disc on top (Figure 8) which must be very well saturated before they are installed in the soil sample. The advantage of the saturated HAE disc is to separate the gas (air) from the liquid (water) phase and to be able to measure only the water pressure, both positive and negative. Lack of sufficient saturation of the disc will cause them to respond slowly to pressure changes and provide unreliable readings, as small air bubbles that might have remained inside will expand significantly when the pressure becomes negative.
Even though suction is controlled in a triaxial specimen, on-sample transducers provide valuable information about the actual distribution of suction inside the specimen. Also, in undrained tests, the on-sample pore pressure measurements will provide the actual suction conditions in the sample during failure.
Till the day this document was prepared, there was no known standard describing the procedures need to be taken during a triaxial test in an unsaturated soil.
Since there are no standard procedures for the unsaturated triaxial tests, the user has the flexibility to tailor the test according to their needs. However, in most of the cases, the laboratory test should reasonably represent the in situ and loading conditions that are likely to occur in the field. Next, some examples of specific testing procedures are presented.
Unsaturated triaxial tests can follow a similar procedure as the saturated triaxial tests. The test can be characterised as Consolidated if a consolidation procedure takes place before the shearing stage or Unconsolidated if this stage is skipped. If the Unconsolidated approach is to be followed, then the sample needs to be subjected to an equalisation stage, before the shear stage, to allow for equalisation of the two stress components, i.e. suction and net stress. A saturation stage is optional and usually is been avoided, however, it might be useful in the case that suction needs to be reduced to the sample. The shear stage can be characterised as Drained or Undrained, in the same way as in the saturated triaxial tests. Finally, Unconfined and Multistage tests can be performed in an unsaturated soil sample.
Sample preparation and initial procedures
The preparation of the unsaturated triaxial sample is similar to the preparation of a sample for saturated triaxial testing. In the case of an undisturbed sample, this is prepared to the initial dimensions by following the procedures described in any of the known standards (e.g. BS, ASTM, AS, etc). For remoulded samples, the user must be aware of the initial conditions that are needed. For example, if the test is to be performed at a low suction level, the specimen needs to be prepared at a high degree of saturation. If this is not possible, additional testing time will be required to equalise the specimen to the desired suction level.
After the formation of the triaxial specimen, this has to be installed in the triaxial cell. Before the installation of the sample, the HAE disc needs to be saturated by following the procedure suggested by the manufacturer. This process might take several hours, for 0.5 Bar HAE discs, and can reach several days for HAE discs of 15 Bar. The saturation of the HAE disc is a very important step for the unsaturated test and it must not be skipped, otherwise the results of the test will be meaningless. After its saturation, the disc needs to be submerged under water until the moment of the sample installation; otherwise, it could be cavitated, and air to intrude inside the pores. If a series of tests are to be performed, the disc can remain submerged in de-aired water during the installation of the new sample to save time for its saturation. Finally, de-aired water must be used for the saturation to eliminate any air bubbles that can be expanded under high suction.
The sample is seated on the base pedestal, directly onto the HAE disc as the soil needs to be in contact with the saturated disc. A dry, coarse porous stone is placed on top of the sample. The top cap, with the air supply line, is then placed above. A rubber membrane is placed around the sample and two O-rings are used to secure it in place. Then, the inner and outer cells are filled with de-aired water and all the air is removed using the bleeding valves on the top of the cell. Finally, all the hydraulic connections are established and the user must ensure that all the pipes are filled with water, free from any trapped air bubbles. It is, also, very important to make sure that there are no trapped air bubbles in the pore water pressure line, i.e. the path between the pressure controller for pore water pressure and the HAE disc. For this reason, some cells provide the ability to flush water through this line to remove all the air bubbles. Moreover, the base pedestals need to be designed in such a way that will offer the best possible de-airing procedure for the compartment underneath the HAE disc.
Before the start of the test, the following checks need to be performed:
- The inner and the outer cell are communicating through the line that passes from the volume change device;
- Pore air and pore pressure valves are open to the pressure controllers;
- The sample is not pre-stressed and the loading ram is not touching the sample (in some cases, a small seating load might be applied, but this cannot be higher than 0.1% of the maximum deviator stress that the sample can accept);
A full pre-test checking list can be provided by VJTech. Also, it is important to remember that the pore air pressure cannot be higher than the confining pressure at any stage of the test, otherwise, the sample will float and the sealing might be damaged.
It is preferable to keep the pore water pressure constant throughout the test procedure and alter the air pressure when a change in the applied suction is required. This is done to avoid implications in the measurement of the water volume that moves in or out of the sample. A value of 30 kPa or higher is considered adequate.
Application of suction – Equalisation
As mentioned already, suction is controlled within the triaxial sample using the Axis-translation technique. For every soil, the value of suction corresponds to a unique value of its saturation degree. This is a property that is described by the Soil Water Retention Curve (SWRC) or Soil Water Characteristic Curve (SWCC). This curve provides the relationship between suction and the degree of saturation within the sample. In general terms, the higher the suction in the soil the lower becomes its moisture content or degree of saturation. When the specimen remains undisturbed, suction and moisture within it come into equilibrium after some time. Therefore, the application of suction to a soil specimen requires anticipation of the equilibrium; otherwise, there will be flow phenomena inside the soil mass which can affect its behaviour. Figure 9 presents a typical example of a drying and wetting SWC curve for unsaturated soil. For low values of suction, the soil remains fully saturated with the moisture content being at its highest value. As soon as suction increases, water starts being expelled out of the sample and its moisture content reduces. The value of suction that causes the water to start moving out of the sample is called the Air Entry Value and it’s higher for fine-grained soils compared to coarse-grained soils. At high values of suction, moisture content reaches a residual value. If suction reduces, the soil starts gathering water until it comes to a near saturation condition. Further desaturation and saturation cycles will produce scanning curves which will be located between the two original ones, as shown in Figure 9. The SWCC can be determined in a triaxial testing system for unsaturated soils, the procedure is described in section 6.3. At the beginning of a triaxial test in unsaturated soil, the user must specify the level of suction that is required for the specific specimen. Then, it is applied using the hydraulic and pneumatic pressure controllers of the system. If the initial moisture is higher than the one corresponding to the current level of suction, the specimen should be left to equalise by allowing sufficient time to drain out any excess water from the voids. In unsaturated soils, permeability is reduced by several orders of magnitude. So every effort to control and measure the pore water pressure is time-consuming. The procedure will terminate as soon as drainage stops and moisture remains constant.
In some occasions, the specimen needs to be saturated at the start of the test. This happens especially when it is required to determine the SWCC of the sample. Saturation cannot be performed by pressurising water in the specimen through the back pressure line. The most common way to saturate an unsaturated soil sample is by applying zero suction and waiting until the sample gains all the water needed to fill its voids. Air is pushed out of the sample through the pore air pressure line, at the top cap. In this way, water will move through the HAE disc of the base pedestal towards the sample until this becomes fully saturated. However, there is no way to verify the degree of saturation of the sample by using this method, but it can be determined by measuring the volume of the water that is displaced. The procedure needs to be applied very carefully so that the volume of the sample is not changing.
The consolidation procedure in unsaturated triaxial samples is performed to bring the sample to the desired state before the loading. In other words, the sample is subjected to specific values of suction and net confining stress and it is then allowed equalising the effective stress. This is achieved by giving sufficient time to the sample to drain water out and therefore dissipate the excess pore water pressure. Because of the very low permeability of the unsaturated soils, the process is generally very slow and requires constant monitoring of the water volume that is been displaced. Consolidation is considered complete when the pore water stops draining from the sample.
Apart from an isotropic consolidation, anisotropic consolidation can also be performed. This is achieved by applying pre-defined values of suction, net confining stress and deviator stress to the sample under drained conditions. Deviator stress needs to be applied in a slow rate to the sample to allow sufficient drainage and dissipate the excess pore water pressures.
The shear strength of an unsaturated soil is higher than the strength of the same soil in a saturated condition. This is due to the effect of suction which adds an extra component to the soil strength. When the soil becomes saturated (for example, during rainfall) the extra strength component disappears and so the soil becomes more vulnerable to failure.
The three stress components that need to be specified to describe the shear strength of the unsaturated soils are Suction, Net Stress, and Deviator Stress. Usually, a shear test in an unsaturated sample is performed by keeping one stress component (usually the net confining stress) constant while increasing the deviator stress. Suction can be either kept constant in a drained test or not in an undrained test. The sample is then compressed until deviator stress reaches a maximum value or until the axial strain reaches a pre-defined value, usually 20%. The sample is usually compressed under a constant rate of strain, but a constant rate of deviator stress can be chosen as well. Axial load is measured using a load cell and the strain is determined by using an analogue or digital displacement sensor. The volume of the sample is determined either by the volume change device or the measurement from the on-sample transducers.
Most triaxial tests in unsaturated specimens are normally Drained tests in which drainage is allowed through the HAE disc. As the sample is compressed, pore water pressure tends to increase and needs to dissipate fast enough so that suction remains constant all the time. For this reason, the shear rate needs to be very slow during the shear stage. Using a mid-height pore pressure transducer, the user can determine whether or not excess pore pressure builds up and, thus, if suction is changing.
However, triaxial tests in unsaturated soils can also be Undrained tests. During the shear stage in undrained tests, the drainage path is blocked and water is not allowed to drain out of the specimen through the HAE disc. In this case, the speed under which the soil is compressed must be again slow enough to allow the uniform distribution of the excess pore water pressure and thus suction. The recommended axial strain rate for an undrained test is between 1.02 and 1.8 % of strain per minute (Fredlund et al, 2012), however, this can be lowered for fine-grained soils. The measurement of the pore water pressure within the specimen can be done either by using a mid-height pore pressure transducer or with a normal pressure transducer placed very near the HAE disc. The undrained tests are also called Constant-Water-Content tests
For Unconsolidated Undrained tests in unsaturated soils, a normal triaxial cell can be used. The HAE disc is then replaced by an impermeable disc which does not allow any change in the moisture content of the specimen. The sample is then tested at the pre-existed level of suction and is sheared until failure, without prior consolidation. Even though the use of a normal triaxial cell is convenient, this is not going to allow for accurate application of suction into the specimen using the axis-translation technique. Moreover, the variations of the pore water pressure and thus suction during the shear stage will not be determined, therefore the exact stress state of the specimen will be unknown. Finally, Unconfined Compression tests can be performed, again without the need for an unsaturated triaxial cell. In this case, the sample is sheared at a constant rate of strain without any confining pressure applied. Even though this test is performed without a confining liquid, the sample must be wrapped again with a rubber membrane to prevent the evaporation of the pore water which will alter the level of suction.
The most common failure criterion for an unsaturated triaxial specimen is the maximum deviator stress that is achieved during the compression, just as in the case of saturated specimens. Deviator stress can be plotted against the axial strain of the sample to define the strain at which failure occurred. If, however, the deviator stress is not dropping with an increased strain, the maximum deviator stress is defined as the one achieved at an axial strain of 20%.
There are several ways to define the shear strength of the unsaturated soil sample and interpret the laboratory data. These can be found in the literature and their description is outside of the scope for this document.
Advanced Testing Procedures
Next, some more advanced testing procedures are described.
In a multistage test, the specimen is consolidated and then sheared until the deviator stress is imminent. At this point, the axial load is brought back to zero and the sample is re-consolidated on a different suction or net confining stress before it is sheared again. The number of stages can be altered based on the test requirements, but usually will not be more than three.
Unsaturated stress paths
It is already mentioned that the shear strength of the unsaturated soils can be described by three components at the moment of failure, i.e. Suction, Net Stress, and Deviator Stress. During a shear stage, the net confining stress (or suction) is kept constant while the deviator stress increases. However, the sample can also be subjected, to a different stress condition where any of the two stress states remains constant while the third is increasing or decreasing. Another test type can incorporate the change of two or even all three stress components in the sample at the same time. These unsaturated stress paths can be performed in a sample to reproduce a specific stress condition in the field while observing its behaviour. An example of a stress path re-created on an unsaturated soil sample is shown in Figure 10. A special case of an unsaturated stress path is to keep the deviator and the net stress constant while suction is increased in steps, allowing sufficient time for the water to drain. This is called a SWCC test and is described below.
Soil Water Characteristic Curve
Soil Water Characteristic Curve is defined as the relationship between the suction and the moisture content or degree of saturation for the soil sample. This curve can be determined in an unsaturated triaxial cell by performing a series of drained isotropic consolidation stages. At each stage, a different (higher) value of suction is applied to the sample and it is then allowed to drain freely. When drainage stops, the degree of saturation for the sample is determined and one point on the SWCC is obtained. By following the same procedure, a series of points are added to the plot until a well-defined curve is reached. The SWCC can be either a Drying curve if it is determined by applying incremental suction levels or Wetting when the suction is gradually decreased in steps. A special variation of this test is the Stress-Dependent SWCC test in which each series of suction variations is performed under a different level of deviator stress. In this way, each SWCC will correspond to a different level of axial stress.
Bender Element Testing in Unsaturated Soils
Another advanced test that can be performed in an unsaturated triaxial specimen is a Bender Element test to determine the small strain modulus. The Bender Element test is described in more detail in the Introduction to Bender Element Testing document by VJTech. In the particular case of an unsaturated specimen, bender element testing would provide the small strain modulus of the specimen under different stress conditions. Moreover, the effect of suction to the elastic properties of the sample would be determined as well.
To perform a bender element test in a triaxial unsaturated sample, a specially designed bender element set needs to be used which will provide the necessary conditions for the unsaturated testing. Apart from the piezometric elements, the base pedestal needs to be equipped with a HAE disc to allow the water drainage. If such a configuration is used, then the unsaturated procedure will require additional time because the surface from which the soil is drained will be reduced significantly.
For more information on Unsaturated Testing or to speak with one of our experts regarding our testing equipment, please contact us.
Books for further reading
- D. G. Fredlund, H. Rahardjo and M. D. Fredlund, 2012. Unsaturated Soil Mechanics in Engineering Practice. John Wiley & Sons, Inc., Hoboken, New Jersey.
- J-L Briaud, 2013. Geotechnical Engineering: Saturated and Unsaturated Soils. John Wiley & Sons, Inc., Hoboken, New Jersey.
- W. A. Take and M. D. Bolton, 2003. Tensiometer saturation and the reliable measurement of soil suction. Géotechnique, 53:159–172(13)
- D. G. Fredlund, H. Rahardjo and M. D. Fredlund, 2012. Unsaturated Soil Mechanics in Engineering Practice. John Wiley & Sons, Inc., Hoboken, New Jersey
- R. Sivakumar, V. Sivakumar, J. Blatz, and J. Vimalan, 2005. Twin-Cell Stress Path Apparatus for Testing Unsaturated Soils. Geotechnical Testing Journal, Vol. 29, No. 2
- D.G. Toll, 2012. The behaviour of unsaturated soil. In Handb. Trop. Residual Soils Eng., p.117–145. CRC Press