Novel procedure for measuring the saturation of the pore-pressure system in piezocone tips

Defects in piezocone saturation are routinely listed among likely sources of uncertainty/error in piezocone testing (CPTU) measurements (e.g. Schneider et al., 2008; Löfroth et al., 2010; Ramsey, 2010; Sandven, 2010; Peuchen & Terwindt, 2014, 2015; Kardan et al., 2016). Yet evaluation of porous filters saturation (e.g. Larsson & Mulabdić, 1991; DeJong et al., 2007) and/or cavitation of the saturation fluid (e.g. Ramsey, 2010; De Backer et al., 2022) is limited. Suggested best practices are described in the standards (ASTM, 2020a; ISO, 2021a; ISO, 2021b), but do not necessarily result in well-defined data quality. In engineering practice, therefore, a great variety with respect to filter material, saturation fluid and saturation methodology is found, as reported in DeJong et al. (2007), and pore-water pressure, u₂, response remains a major uncertainty (Tonni et al., 2016).

Good saturation of the piezocone tip is paramount to obtain reliable data for soil characterisation and evaluation of mechanical and/or hydraulic properties. An incorrect measurement of u₂ has cascading effects on the reliability of virtually all data collected, the pore-water pressure being required to obtain the corrected cone resistance from raw data of q_c. Responsiveness in the measurements also relies on good saturation, which is particularly important in layered soil deposits.

Examples where the pore-pressure behaviour is not consistent with the expected response due to a lack of saturation are depicted in Fig. 1. Either desaturation and/or cavitation occurred during penetration or incomplete/incorrect saturation.

This paper describes a device designed to measure a pore-pressure parameter B*, associated with the degree of saturation of an assembled piezocone tip, and discusses preliminary laboratory results obtained on a tip having different initial degrees of saturation. The proposed B* parameter is inspired from the well-established Skempton's pore-pressure coefficient B (Skempton, 1954), which is routinely used in laboratory testing to assess the specimen saturation – for example, prior to triaxial testing. By measuring it prior to and after CPTU testing, the B* parameter could be used to evaluate the reliability of pore-pressure data and the overall test quality. This matches recent suggestions to use reference readings and output stability as non-subjective diagnostic checks to classify test results with respect to quality (Lunne et al., 2017; Soage Santos, 2022).

A piezocone tip manufactured by Delft Geotechnics, with the filter in the u₂ position, was used in this study. The filters fitted onto the cone were 5 μm sintered stainless steel and were purchased from the supplier already saturated with 20 cSt silicon oil, which are conditions typical in current engineering practice, as reported by De Jong et al. (2007).

Fig. 2(a) provides a view of the overall set-up, whereas Fig. 2(c) shows the operating principle of the whole arrangement devised to check the saturation of the piezocone pore-pressure sensing system. The key component is an annular chamber placed around the piezocone filter (Rocchi et al., 2017), with an inner flexible impermeable membrane separating the device from the piezocone (Fig. 2(b)). An indicative sketch of the device is provided in Fig. 2(d). The chamber is pressurised instantaneously with a single stroke of a hand pump, while an external pressure sensor and the pore-water pressure transducer installed in the piezocone tip are both logged through the same acquisition system.

To perform the measurement, the same procedure can be adopted regardless of when the saturation check is performed (i.e. before or after piezocone testing) and the saturation method, filter or fluid employed. After positioning the device on the already assembled piezocone, a pressure impulse is applied and the increment recorded by the pore-pressure sensor placed inside the piezocone is compared to that recorded by the external independent pressure transducer. The pore-pressure parameter B* is thus defined as the ratio of the two increments. This closely mimics what is typically done in the saturation phase of a triaxial test, where an isotropic increment of total stress (i.e. cell pressure) is applied in undrained conditions and the response in terms of pore-water pressure is monitored. Assuming a closed hydraulic system disconnected from the specimen, if the system is not compliant (i.e. infinite stiffness), an isotropic increment of stress must result in an equal increment in the fluid pressure, which corresponds to the pore-water pressure in the soil specimen. Compliance may arise from deformability of the hydraulic system, soil grains or fluid. However, assuming air as the only relevant source of compliance, the two increments are the same for a saturated system – that is, B ∼ 1. Similarly, in the proposed device, air in the porous element or elsewhere in the system is assumed to be the only relevant source of compliance and therefore targeted values of B* ∼ 1 are sought to assess full saturation of the tip.

The tests were conducted after saturation of the piezocone in silicon oil by the methods later described. The device was slid on the piezocone and a 50 kPa was applied to the device chamber. Subsequently, a series of pressure impulses, ranging from 50 to 300 kPa, was applied in steps, each equal to a 50 kPa nominal increment. Increments were applied within a 1 s interval and kept for 30 s before proceeding to the following step. Continuous data logging of both the piezocone and the device was performed during the process.

Saturation was performed before the filters and the seals were mounted onto the tip and the cone was tightened, by using either methods recommended by ISO (2021b); these consist of (1) injecting silicon oil in the piezocone chamber connecting the pore-pressure transducer to the filter, with a syringe (also in ASTM, 2020a), or (2) placing the piezocone under a 85 kPa vacuum in a silicon oil bath together with the seals for 15 min, which is the minimum time recommended by the standards. Furthermore, the results obtained applying a combination of the two methods were also evaluated. Note that for method (2), the tip was assembled while immersed in silicon oil and the device was slid on the piezocone while still in the oil bath. This avoided entrapping air during the process. Furthermore, a small pressure (<5 kPa) was applied to the device chamber to hold the device in place.

To test the different components of the piezocone system, either the filters or the chamber connecting the pore-pressure sensor in the tip to the filters were purposely desaturated. In particular, ahead of saturation 1 ml air was injected with a syringe inside the just mentioned chamber. The tests were then conducted using saturated filters. Additionally, the filters were tested under two reference unsaturated conditions. Desaturation was obtained by placing the filters either under vacuum in air for 24 h or pressurising them with air at 600 kPa for 1 h, leading to 90% and 5% saturated filters, respectively, based on the weight loss recorded. When testing desaturated filters, the saturation procedure was kept the same and corresponded to the combination of the methods described above. Note that compared to Rocchi et al. (2022), the preliminary injection of air in the system guaranteed comparable initial conditions, and therefore improved repeatability.

Preliminary results in Fig. 3 provide clear evidence that filter saturation degree (100%, 90% and 5%) affects the measured pore-water pressure parameter B*. Each data point corresponds to the average value obtained performing the test 3 times, each time using the same procedure but a different filter, while the associated error bars represent the minimum and maximum values measured. The results indicate a relatively good degree of repeatability within the pressure range investigated, regardless of the filter conditions. When testing fully saturated conditions – that is, saturated filters following saturation, the values are high, relatively unaffected by the pressure at which the test is conducted, and are comparable with those required in the laboratory according to ASTM (2020b) (i.e. B > 0·95–0·97 for sands and B > 0·97–0·99 for clays). Inducing slight desaturation to the system, obtained by employing 90% saturated filters, results in a considerable difference in the incremental values of measured pressure, within the range investigated. These differences gradually reduce as the applied pressure is increased during the test, because air bubbles in the tip are compressed and the system becomes stiffer, thus its compliance is reduced. Similarly, the pore-water pressure readings measured by a poorly saturated piezocone in the field generally improve at great depth below the groundwater level. Indeed, only at the high end of the pressure range investigated, the values recorded match those prescribed for laboratory testing of saturated soils. When the filters are almost fully desaturated (5% saturation), it is still possible to obtain a response even though the B* values measured 1 s after the application of the pressure impulse are extremely low. As shown in Fig. 4, the B* value increases with time for 5% saturated filters. However, the increase is not significant and stabilises relatively fast due to the undrained conditions applied by the device. Tests on 90% saturated filters instead show a slight decrease with time, possibly due to relaxation of the system. Even though values after 30 s were used to calculate B* for 5% saturated filters, this still would reduce at increasing pressure applied, as shown in Fig. 3, where measurements after 1 s were used for calculation. Such reduction with increasing pressure could possibly be explained by air progressing further into the pore-pressure measuring system of the piezocone.

Tests on non-saturated piezocone tip/system, though with saturated filters, were also carried out and the relevant data points are also displayed in Fig. 3. In this case, the measured values of B* can be compared with the results plotted in Fig. 5, which were obtained by employing different saturation methods. Again, each data point corresponds to the average value obtained on three tests, each performed with the same procedure but a different saturated filter. The associated error bars are seemingly larger compared to Fig. 3, due to the smaller interval adopted for the vertical axis. When comparing B* with and without saturation of the inner piezocone chamber (both performed with saturated filters), it is clear that the procedures recommended in the standards are highly effective in achieving saturation of the system when employed in laboratory conditions. Finally, the summary plot in Fig. 6, based on the B* values measured at 50 kPa under all conditions tested, highlights that all components must be fully saturated to achieve a satisfactory outcome.

This paper presents a set of preliminary results obtained using an innovative device, that correlate a newly defined B* parameter of the pore-pressure measuring system of an assembled piezocone tip with its overall saturation degree, either before or after the test. The range investigated (50–300 kPa) is typical of routine testing conditions in the field. However, the measured correlation is non-linear and therefore the method proposed loses in applicability when testing at stresses significantly beyond this range, where the relationship enters a tail zone.

The parameter is suggested to be ultimately used as an objective quantitative assessment of data quality in CPTU testing. In applications where the data quality required is particularly high, proceeding to testing could be made conditional to achieving a certain threshold in the B* measured. After the proof of concept herein presented, the results should be confirmed using a range of piezocones, which despite being standardised differ in the fine details of their hydraulic system. Subsequently, the most suitable saturation processes and media can be investigated by comparing their outcome in terms of B* at the laboratory scale. In particular, this is required because the exact B* value measured is a function of the relative stiffness of the fluid and the porous medium. Furthermore, field measurement of u₂ in a same uniform stratigraphy, obtained using the same piezocone having a range of saturation conditions (as assessed by the methodology proposed) is required to verify the correlations obtained in the laboratory. Finally, the quality of CPTU field measurements in different stratigraphic conditions should be investigated to identify threshold values suitable for reliable testing.

B

pore-pressure coefficient (Skempton, 1954)

B*

proposed pore-pressure parameter

q_c

measured tip resistance

u₀

hydrostatic pore-water pressure

u₂

measured pore-water pressure at the shoulder position