Shear strength parameters identification of loess interface based on borehole micro static cone penetration system

Background Loess is prone to large deformation and flow slide due to natural and artificial interfaces inside. The strength of these interfaces controls the mechanical properties of loess. Obtaining their mechanical parameters through in-situ testing is essential for evaluating the mechanical stability in loess engineering with interfaces. Methods By developing a borehole micro static cone penetration system and creating various types of loess with interfaces, extensive borehole penetration model tests were conducted to observe changes in cone tip resistance during penetration. The response surface method was used to analyze the impact of various test conditions on the calculated resistance. A three-dimensional surface fitting method was employed to establish the relationship between penetration parameters and shear strength parameters, which was validated through in-situ testing. Results The developed borehole micro static cone penetration system achieves overall miniaturization while providing significant penetration power and ensuring an effective penetration distance. Cone tip resistance development during penetration can be divided into three stages: initial, rapid increase, and slow increase. The transition times between these stages vary for different soils. Calculated resistance is positively correlated with dry density and normal stress and negatively correlated with water content. A quadratic positive correlation was established between calculated resistance and shear strength parameters during penetration. In composite soils, the interaction between water content and normal stress is strong. Compared to intact soil samples, the shear strength parameters of composite soils are more prominently influenced by water content. Conclusion A system for testing interface mechanical parameters was innovatively developed, fulfilling the need to obtain interface shear strength parameters for deep soil. This study can provide support for ensuring the long-term stability of the loess slope or subgrade with interfaces.


Introduction
Loess, a Quaternary deposit, is widely distributed across the Eurasian continent, covering over 600,000 km 2 in China alone (Bao et al. 2023(Bao et al. , 2024a)).In comparison to contemporaneous sediments, loess shows distinct variations in internal material composition, content, and external morphological features (Garcia Giménez et al. 2012;Drewnik et al. 2014).Loess failure along the interface frequently causes highway traffic interruptions, greatly affecting highway construction and later operations in loess areas (Zhang et al. 2024).Currently, scholars primarily investigate loess disasters based on soil mechanic properties combined with internal and external influencing factors.However, they inadequately consider the mechanical effects arising from natural strata and artificial cut-fill interfaces within loess.
Extensive infrastructure construction has occurred in the loess areas of China, resulting in altering terrain and landscape conditions, reshaping geological structures, and creating numerous cut-fill bodies (Li et al. 2014).Cut-fill bodies exhibit unique structures characterized by special hydrological conditions, significant interface effects, and susceptibility to deformation and sliding along interfaces (Zhang et al. 2019).Furthermore, natural interfaces exist within loess of varying periods, scales, properties, and origins, intertwining within the loess body to control its mechanical properties and impact the stability of slope and subgrade engineering (Wang et al. 2019;Wijesinghe et al. 2023).Naturally, soils are predominantly layered, exhibiting notable variations in physical and chemical properties among layers, including density, mineral composition and content, particle size, and distribution characteristics (Bronger and Smolíková 2019;Lan et al. 2023).Due to variations in soil properties, the mechanical performance and deformation mechanisms of different soil interfaces are also complex under external influences, resulting in complex deformation behaviors like dislocation and sliding (Wang et al. 2023).With the expansion of large-scale geotechnical engineering projects, the number of disasters caused by soil interface failure has gradually increased (Lan et al. 2003;Peng et al. 2019).Therefore, considering various factors and investigating the role of shear strength in controlling interfaces has become a crucial task.
The interface, as the separation surface of heterogeneous soil, determines the overall strength of the soil mass (Rashied 2022).Typically, low interface shear strength may result in abrupt shear strain changes in the soil, influencing slope stability and the initiation of landslides (Song and Tan 2020).Research indicates that loess interfaces significantly reduce the shear strength of rock-soil materials, and interface specimens exhibit a notable dilatancy during the shearing process (Stutz and Martinez 2021).Additionally, interface shear strength is affected by several factors, such as temperature, permeability, construction process control, interface soil type, water content, load, dry density, and so on, among which the dry density, water content, and load of the soil are the main controlling factors (Sayeed et al. 2014;Liang et al. 2016;Westgate and DeJong 2023;Bao et al. 2024b).Rainfall, groundwater infiltration, and seasonal humidity variations affect soil water content, leading to changes in pore water pressure.Increased pore water pressure reduces the effective stress of the soil, thereby weakening the shear strength at the interface (Lann et al. 2024;Wu et al. 2018).Load variations primarily change the contact status of the interface, thereby affecting the shear strength of the soil (Wang and Yin 2022).An increase in soil density enhances interface shear strength due to particle interlocking and improved soil cohesion (Ferreira et al. 2015).Therefore, obtaining shear strength parameters of loess interfaces is essential for disaster prevention, control, and management.
Accurately obtaining the physical and mechanical parameters of the soil interface is essential for evaluating its engineering mechanical properties (Castellanos et al. 2016;Moon et al. 2024).The commonly used methods for obtaining mechanical parameters of soil interfaces include numerical simulations and laboratory and in-situ tests.Numerical simulation provides a relatively fast and convenient method for studying interface strength characteristics (Tano et al. 2017).Many scholars use finite element methods to analyze soil interface interaction (Yogarajah and Yeo 1994;El Sawwaf 2007;Bergado and Teerawattanasuk 2008).The discrete element method is also utilized to reveal the micromechanical mechanism of interface shear strength at the particle level (Jensen et al. 2001;Miao et al. 2017).However, numerical simulation methods typically assume ideal conditions for interface study, while actual engineering situations are very complex.Therefore, laboratory and in-situ testing remains an accurate method to obtain interface shear strength parameters.Various laboratory and in-situ testing methods, such as direct shear tests, ring shear tests, improved triaxial compression tests, and borehole shear tests, can be used to obtain interface shear strength parameters (Abu-Farsakh et al. 2007;Vieira et al. 2020;Bao et al. 2020;Shiga et al. 2022).Murphy et al. developed a thermal drilling shear device capable of measuring the influence of temperature on the in-situ shear stressdisplacement curve of interfaces (Murphy and McCartney 2014).Li et al. developed a rotary probing instrument and a wedge probe and discovered that utilizing drilling parameters during rotary probing could better predict soil mechanical parameters (Li and Li 2012).However, no effective tools are currently available for obtaining the strength characteristics of deep soil interfaces.Therefore, innovation in in-situ testing techniques and the development of new equipment are crucial for identifying and evaluating the mechanical parameters of loess interfaces.
This study focuses on loess interfaces and proposes a borehole micro static cone penetration system to test interface shear strength parameters of in-situ deep soil.Mechanical tests were conducted under interface combinations of different soil types, with water content, dry density, and load as variables.The variation trend of cone tip resistance during penetration was studied, and the relationship between calculated resistance and interface shear strength parameters was established and successfully applied in field tests.

Basic structure
Cone Penetration Testing (CPT) is widely used in engineering geology exploration to determine liquefaction resistance, classify soil layers, and ascertain foundation bearing capacity (Yang et al. 2013;Rainer and Szabowicz 2020;Ecemis et al. 2022).However, large penetrometers are cumbersome, imprecise, and expensive.Although Micro Cone Penetration Testing (MCPT) reduces the size of the probe, it remains constrained to vertical soil penetration, thus lacking significant theoretical or technical innovations (Laufer 2013;Lin et al. 2022).Therefore, based on the static penetration mechanism, a borehole micro static cone penetration system driven by a stepper motor was developed to test soil interfaces.As shown in Fig. 1, the borehole micro static cone penetration system mainly comprises a fixed unit, a penetration unit, and a control and data acquisition unit.The system is characterized by its lightweight design and ease of application for soil interface testing.The functions of each unit of the system are as follows.
Figure 2b illustrates that the fixation system of the device consists primarily of upper and lower support cylinders, pistons, claws, and support poles.The cylinder has intake and exhaust ports connected to an external air pump, which controls piston movement by adjusting internal pressure.Piston movement drives the claws and the outer support poles to expand or contract.The fixation system operates as follows: Once the equipment reaches the designated position in the borehole, open the valve of the external air pump to pressurize the support cylinder.Under pressure, the support rods gradually extend, pressing against the borehole wall.When the friction between the support poles and the borehole wall balances with the weight of the device, the device stabilizes at the designated position.If repositioning is necessary after completing a probing operation, start by opening the exhaust valve to release gas from the supporting cylinder.Then, retract the support cylinder poles until reaching the designated position, and repeat the steps mentioned above.
The penetration system of the device primarily comprises a depth stepping deceleration motor, penetration probe rod, and rack-and-pinion transmission mechanism.As depicted in Fig. 2g, the depth stepping deceleration motor and probe rod fixed seat are attached to the motor mounting plate.The shaft of the depth stepping deceleration motor is linked to the gear of the rackand-pinion transmission mechanism, and the end of the rack connects to the tail of the penetration probe.The probe rod fixed seat includes a fixed block positioned in the groove of the rack, enabling linear movement of the Fig. 1 Overall diagram of borehole micro static cone penetration system rack along the groove under the driving mechanism, and it can prevent excessive rack movement to avoid component detachment.The operational procedure of the penetration system is as follows: activate the depth stepping deceleration motor, driving the motion of the penetration probe through the rack-and-pinion transmission mechanism to insert the probe into the soil (Fig. 2c).The penetration probe features a cone tip resistance sensor, as shown in Fig. 2d.
The control system of the device is primarily operated manually through an integrated control box, regulating the forward and backward movement of the penetration probe within the penetration system.It can function at a predetermined penetration speed or be manually adjusted to alter the penetration speed during different stages, while displaying the movement distance of the probe on the monitor screen.The control box is illustrated in Fig. 2f.The data acquisition system of the device consists of sensor cables, a data collector, acquisition software, and a computer.The sensor cables connect to the data collector.The data collector, The YBY-801 dynamicstatic resistance strain eight-channel data collector, collects eight types of data simultaneously (Fig. 2a).The data collector connects to the computer through a connection port and employs self-developed software for data acquisition and logging (Fig. 2e).

Functional parameters
The borehole micro static cone penetration device has dimensions of 150 mm in diameter, and 350 mm in height, weighing approximately 5 kg, with a part machining accuracy of 0.1 mm.The functions achievable by the device include fixation, penetration, control, and data collection.The specific parameters of each function are shown in Table 1 below.

Physical and mechanical properties of soil and test scheme Physical and mechanical properties of soil
This study investigated loess and paleosol samples to obtain their physical and mechanical properties, preparing specimens for direct shear testing.The soil samples used in the test were collected from a loess slope located in Bailu Platform, Xi'an City, China (34° 15′ 46″ N, 109° 4′ 48″ E, elevation 488 m), and belonged to the Quaternary loess.The samples, with a diameter of 61.8 mm, included intact loess samples, intact paleosol samples, and composite samples with loess in contact with paleosol.The composite soil samples were prepared by placing loess in the lower layer and paleosol in the upper layer.Direct shear tests were conducted using a ZLB-1 straincontrolled tri-combined direct shear apparatus, with a shear rate of 0.8 mm/min and a maximum shear displacement of 6 mm according to "JTG3430-2020 Test Methods of Soils for Highway Engineering" (JTG3430 2020).The soil samples were subjected to direct shear tests, and soil type, water content, and dry density were selected as variables.The specific physical and mechanical properties are shown in Table 2. 1.5/1.8indicates that the upper layer is paleosol with a dry density of 1.5 g/cm 3 , and the lower layer is loess with a dry density of 1.8 g/cm 3 .

Penetration test scheme
The equipment utilized for the borehole penetration test primarily comprises the independently developed and processed borehole micro static cone penetration system and the multifunctional ground pressure simulation model box.Since the penetration system has been detailed earlier, only the designed model box will be introduced here.
To conduct indoor penetration tests, a model box designed to accommodate the borehole micro static cone penetration device was created (Fig. 3).The model box consists of two parts, each measuring 400 mm*400 mm*250 mm.By controlling the soil type, dry density, and water content in the upper and lower parts, different soil layer interfaces can be created.The model box is user-friendly and capable of simulating boreholes and soil layer interfaces.By integrating pressure sensors, it accurately simulates the pressure of soil in situ state, effectively addressing the limitations of traditional model boxes in simulating soil layer interfaces and the actual stress of the in-situ soil.The testing conditions for penetration tests are the same as those for direct shear tests.Penetration tests are categorized into two types: intact soil and interface, totaling 63 groups.Dry density selections include 1.5 g/ cm 3 , 1.8 g/cm 3 , and 1.5/1.8g/cm 3 , where 1.5/1.8g/cm 3 indicates that the upper part of the model box is paleosol with a dry density of 1.5 g/cm 3 , while the lower part is loess with a dry density of 1.8 g/cm 3 .Water content selections are 14%, 16%, and 18%.Normal stress selections are 25 kPa, 50 kPa, and 100 kPa.The penetration test scheme is detailed in Table 3, and the test site is shown in Fig. 4.

Cone tip resistance change trend
In accordance with the previously outlined test scheme, borehole micro static cone penetration tests will be conducted to acquire cone tip resistance data under various soil layers and test conditions.Plot the curve of cone tip resistance against penetration distance, as depicted in Fig. 5.The enlarged pentagram points in the figure indicate the turning points of each stage.
Overall, the cone tip resistance of the three soil types exhibits a nonlinear increasing trend with penetration distance under varying test conditions.The penetration process can be divided into initial, rapid increase, and slow increase stages based on the growth rate.Different soil types require varying penetration distances to reach the slow increase stage due to their distinct properties.Paleosol and the composite soil of loesspaleosol tend to stabilize at a penetration depth of approximately 25 mm, whereas loess requires a depth of around 30 mm to show a stable trend.

Influencing factors of cone tip resistance
From the cone tip resistance variation curve in Fig. 5, it can be seen that during the slow increase stage, the cone tip resistance remains fairly constant.Therefore, the cone tip resistance at a penetration distance of 40 mm is used as the calculated resistance.To clarify the variation rule of cone tip resistance under different test conditions, the calculated resistances under various test conditions were summarized and plotted, as shown in Fig. 6.
As shown in Fig. 6a, the calculated resistance of complete soil samples increases with dry density, while for composite soil samples with interface, it follows the same trend with the increase in average dry density of the upper and lower soil layers.Further analysis reveals that an increase in dry density amplifies the effect of normal stress on the calculated resistance.In Fig. 6b, both complete and composite soil samples exhibit a decrease in calculated resistance with increasing water content.An increase in water content

Cone tip resistance identification method of soil mechanical parameters
Fitting shear test data and penetration test data to explore the relationship between shear strength parameters and calculated resistance.Shear strength parameters are independent of normal stress, but calculated resistance is greatly influenced by normal stress.Therefore, when investigating the relationship between calculated resistance and shear strength parameters, the influence of normal stress should be considered.Additionally, because there are some differences in mechanical properties and stress forms between complete soil and interfacial soil, the relationship between calculated resistance and shear strength

Complete soil
① Relationship between paleosol shear strength parameters and calculated resistance Shear strength parameters, calculated resistance, and normal stress for paleosol were fitted into threedimensional surfaces at various water contents and dry densities.This resulted in establishing the relationship between shear strength parameters and calculated resistance under different normal stresses, illustrated in Fig. 7.
Generally, the cohesion and friction angle of the paleosol are positively correlated with calculated resistance.However, as the calculated resistance increases, the cohesion trend changes from rapid to slow, and the friction angle trend changes from slow to rapid.Normal stress does not affect shear strength parameters but only impacts the calculated resistance.The calculated resistance of paleosol is positively correlated with normal stress.The cohesion and calculated resistance fitting model has a root mean square error (RMSE) of 4.56 kPa, while the friction angle and calculated resistance fitting model has an RMSE of 0.31°, indicating good fitting performance for both models.Shear strength parameters, calculated resistance, and normal stress of loess under various test conditions were fitted into three-dimensional surfaces, as depicted in Fig. 8.
The shear strength parameters of loess also maintain a positive correlation with the calculated resistance.As the calculated resistance increases, the trends in cohesion and friction angle change from slow to rapid.Normal stress does not affect shear strength parameters but only influences the calculated resistance.The calculated resistance of loess is positively correlated with normal stress.The cohesion and calculated resistance fitting model has a root mean square error (RMSE) of 4.69 kPa, while the friction angle and calculated resistance fitting model has an RMSE of 1.01°.The results show that the two models fit the data well.

Loess-paleosol composite soil
The study discovered that the friction angle of soil with the interface is strongly influenced by both water content and normal stress, and it does not show a singular monotonous relationship with calculated resistance.Hence, when fitting the shear strength parameters of soil with interfaces to calculated resistance, it's essential to consider the various factors affecting the soil friction angle and separately fit their relationship under different water content and normal stress conditions.As for cohesion, it should align with complete soil to establish the fitting relationship between calculated resistance and shear strength parameters under varying normal stresses.

① Relationship between the cohesion of soil with interface and calculated resistance
The experimental results are used to plot the fitting relationship between cohesion and calculated resistance of soil with interfaces, as shown in Fig. 9.
The fitting model for cohesion and calculated resistance of soil with interface has a root mean square error (RMSE) of 1.32 kPa, indicating satisfactory fitting performance.Additionally, the graph shows that the cohesion of soil with interface exhibits an approximate negative relationship with calculated resistance.With an increase in water content, the effective bonding between the upper and lower soil layers increases, leading to higher soil cohesion.However, as the water content increases, the soil approaches its optimum water content, making it more susceptible to compression, thereby decreasing calculated resistance.Furthermore, with an increase in dry density, the effective bonding between the upper and lower soil layers gradually decreases, leading to decreased soil cohesion.However, as the dry density increases, the cone tip resistance required for unit compression of the  soil also increases.Thus, there is a negative correlation between soil cohesion and calculated resistance.
② Relationship between the friction angle of soil with interface and calculated resistance The friction angle and calculated resistance of soil with interface were linearly fitted under various test conditions, and the corresponding relationship graph is depicted in Fig. 10.
The graph reveals a linear negative correlation (Eq. 1) between the friction angle and calculated resistance of soil with interface at water content levels of 14%, 16%, and 18%.The fitting effect is good, as indicated by the R 2 parameter.
Due to the intricate relationship between calculated resistance and the friction angle of loess-paleosol composite soil, along with various influencing factors, a three-dimensional surface fitting method was utilized to fit ASL and BSL across varying normal stresses and water content levels.The results of the fitting are depicted in Fig. 11.
(1) ϕ = A SL + B SL q c (q c > 0) The fitting results show that parameters A SL and B SL in the correlation between the friction angle and calculated resistance of the loess-paleosol composite soil are significantly affected by the interaction between normal stress and water content, and do not show a simple monotonous trend.Specifically, as the water content increases, A SL initially decreases and then increases, reaching a turning point at approximately 16% water content.In contrast, B SL initially increases and then decreases, reaching a turning point at approximately 14% water content.

In-situ test
The in-situ test site is situated at the Model Test Base in Pucheng County, Weinan City, Shaanxi Province, China, on the Loess Plateau of northern Shaanxi.The terrain is characterized by tableland and has geographical coordinates of 34° 56′ 31″ N, 109° 24′ 34″ E. Micro static cone penetration tests were conducted at the model test site (Fig. 12) in two boreholes (B 1 , B 2 ), each 3 m deep and 160 mm in diameter.The testing depth ranged from 0 to 2.2 m, with a spacing of 0.2 m between penetration tests.The conclusions from the above analysis also apply to average soil, and field tests were conducted on it.
The fitting graph in Fig. 8a shows that the cohesion of loess follows the empirical relationship with calculated resistance as follows: The fitted relationship between the friction angle and calculated resistance of loess, as depicted in Fig. 8b, is expressed as follows: (2) c = 37.44 − 28.02q c + 0.2P + 17.38q 2 c + 0.0008P 2 − 0.23q c P (q c > 0) In practical engineering, the normal stress P acting on a certain depth of soil layer is essentially the stress induced by the weight of the overlying soil, and the actual normal stress in engineering can be calculated using Eq. 4.
In Eq. 4, γ represents the unit weight of the overlying soil, and h denotes the depth of the soil layer or the thickness of the overlying soil layer.
By substituting Eq. 4 into Eq. 2 and Eq. 3, we derive the empirical relationship between the cohesion of loess, calculated resistance, and the thickness of the overlying soil layer, as depicted in Eq. 5. Equation 6 depicts the empirical relationship between the friction angle of loess, calculated resistance, and the thickness of the overlying soil layer.
The relationship curves illustrating penetration distance and cone tip resistance at various depths of the two boreholes are shown in Fig. 13.The in-situ soil at the site has a unit weight of 16 kN/m 3 .The calculated resistance test data are incorporated into the relationship between the calculated resistance and the shear strength parameters of the loess (Eq. 5 and Eq.6), and (3) ϕ = 25.52−1.77qc +0.02P+1.84q 2 c + 0.0001P 2 −0.03q c P(q c > 0) (4)  4 and 5.
The graph in Fig. 13 illustrates that the variation in cone tip resistance conforms to the three stages identified in the probing process outlined earlier: initial stage, rapid increase stage, and slow increase stage.The in-situ tests demonstrate that the borehole micro static cone penetration system is well-coordinated, with each unit able to perform its designated tasks, and the variation curve of cone tip resistance obtained from on-site testing follows fundamental rules.Researchers found that the in-situ soil at the test site consists of Q4 loess and Q3 loess-Malan loess deposits.Triaxial tests on undisturbed samples yielded a friction angle of approximately 25°.Cohesion varied significantly with water content, ranging from 18.24 to 42.0 kPa (Wang et al. 2024).These results are consistent with those obtained by the testing instruments in this study (Tables 4, 5), confirming the practicality of the in-situ micro static penetration system and the feasibility of the method for identifying soil mechanical parameters.

Conclusion
In addressing the challenge of acquiring in-situ mechanical parameters of loess interfaces, an in-situ testing device was developed, and relevant testing methods were established.The conclusions derived from this exploration are as follows: (1) A self-developed borehole micro static cone penetration system was introduced, achieving minia-turization and high precision of the penetration probe, providing significant penetration power, and ensuring effective distance of penetration, meeting the requirement for obtaining in-situ shear strength parameters of deep soil interfaces.
(2) The penetration process can be divided into three stages based on the rate of increase in cone tip resistance: initial stage, rapid increase stage, and slow increase stage.These stages are influenced by water content, dry density, and normal stress.The transition times of each stage vary in different soils, but calculated resistance is positively correlated with dry density and normal stress, and negatively correlated with water content.(3) A relationship between penetration parameters and mechanical parameters was established.The calculated resistance shows a quadratic positive correlation with shear strength parameters.In compos-

Fig. 2
Fig. 2 Each part of borehole micro static cone penetration system: a Whole system; b Fixation system; c Probe penetration; d Probe; e Collection software; f Control box; g Penetration system

Fig. 3
Fig. 3 Model box real object

Fig. 6
Fig. 6 Influence of different factors on calculated resistance: a Calculated resistance curve with dry density; b Calculated resistance curve with water content

Fig. 8
Fig. 8 Relationship between the cohesion and friction angle of loess and the calculated resistance: a Cohesion; b Friction angle

Fig. 9
Fig. 9 Relationship between the cohesion of soil with interface and calculated resistance

Fig. 10
Fig. 10 Relationship between the friction angle of soil with interface and calculated resistance

Fig. 11
Fig. 11 Relationship between fitting parameters, normal stress, and water content: a A SL ; b B SL

Fig. 12
Fig. 12 In-situ borehole micro static cone penetration test

Table 1
Functional parameters of penetration device

Table 2
Mechanical parameters of the soil sample

Table 3
Penetration test scheme

Table 4
B 1 borehole penetration test results

Table 5
B 2 borehole penetration test results ite soils, mechanical parameters are significantly influenced by the interaction of water content and normal stress.Compared to intact soil samples, the shear strength parameters of composite soils are more prominently influenced by water content.