Application analysis of soil nail and cable support in deep excavations of loess regions under complex environmental conditions Authors: Xiaoyi Zhang, Shunfu Zheng, Lulu Liu, Xiaoyan Liu Published: 2026 ABSTRACT AbstractTo evaluate the effectiveness of a prestressed anchor cable combined with soil nailing technology in supporting deep foundation pits in loess regions, comprehensive field monitoring was conducted to measure horizontal and vertical displacements at the pit crest, as well as deep-seated horizontal deformations of the support system. A two-dimensional numerical model was developed using MIDAS GTS NX to facilitate comparative analysis. The results demonstrate that the integrated support system effectively controls deformation during excavation. The maximum displacement at the top of the support structure was recorded as 3.91 mm, while the maximum horizontal displacement below a depth of 0.5 m did not exceed 1.6 mm—both values are substantially below the predefined alarm thresholds. The vertical profile of horizontal displacement exhibits a fluctuating pattern, with relatively smaller displacements observed at the elevations corresponding to soil nails and anchor cables, indicating localized reinforcement effects. Finite element analysis reveals that horizontal displacement within the excavation zone and surface settlement at the pit edge increase progressively with excavation depth, exhibiting an approximately linear trend. In the unsupported scenario, numerical instability occurred at an excavation depth of 3.6 m, beyond which shear failure propagated through the soil mass. Although the simulated displacements prior to failure were higher than those measured in the field, this discrepancy is attributed to the minimal external loading under actual site conditions and the simplified representation of full-length grouting effects in the simulation model, which may underestimate the confinement provided by reinforced soil. In conclusion, the integration of prestressed anchor cables with soil nailing walls demonstrates high performance in deep foundation pit engineering in loess regions, offering effective deformation control and enhanced structural stability.Keywords: Anchor cable, Foundation pits, Loess, Soil nail, SettlementSubject terms: Engineering, Environmental sciences, Solid Earth sciences Introduction With the rapid development of urban high-rise buildings and infrastructure construction, the number of deep foundation pit excavation projects has been steadily increasing. Correspondingly, support technologies have continuously evolved, leading to the emergence of various new types of support structures1–7. Among these, prestressed anchor cable flexible support systems and composite support structures combining internal bracing with prestressed anchor cables are representative examples8–12. The composite support system integrating internal bracing and prestressed anchor cables not only ensures structural safety but also significantly reduces project costs, it can facilitate a shorter construction duration, while effectively controlling foundation pit deformation13–16. As a well-established technique for excavation support, the composite soil nail wall has been extensively employed in challenging engineering contexts, including disaster mitigation, slope stabilization, and foundation pit reinforcement17–22. Within this system, soil nails primarily strengthen the adjacent soil through their adhesive capacity and structural skeleton effect, creating a unified composite mass with the soil via grouting. Meanwhile, prestressed anchor cables are tensioned at the early construction stage to compact the soil between the anchorage zone and the shotcrete facing, thus improving the overall structural performance. In recent years, notable progress has been made in both the theoretical understanding and practical deployment of this technology. Cheuk et al.23 employed an undrained strain-softening soil model to investigate the behavior of soil nail-reinforced loose slopes under rainfall infiltration, demonstrating that soil nails, when combined with appropriate surface protection measures, can effectively reduce deformation induced by rainfall, thus maintaining slope stability. Zhou et al.22 adopted an incremental elastoplastic analysis method incorporating pore water diffusion to study internal deformations of soil nails during and after loading, moisture redistribution within the soil, and the overall performance of the soil nailing system. Johari et al.24, based on actual site conditions and using the Stochastic Finite Element Method (SFEM), derived reliability indices for critical components of soil nail wall systems, including overall stability, lateral displacement stability, tensile strength, and uplift resistance. Guo et al.25 investigated the effects of pile hammer vibration on soil strength and pile-soil interface strength, and analyzed the resulting variations in axial force and displacement of soil nails under vibration. Wang et al.26 conducted a field experiment to monitor the soil nail combined with prestressed anchor rod support system under real conditions, and evaluated the influence of the length of the unbonded sections of prestress and anchor rods. Gui et al.27 simulated a full-scale soil–nail-supported excavation for a 20-story building foundation pit, and examined the influence of different strip foundation configurations on the support performance. In recent years, numerous studies have further advanced the understanding of composite soil nail wall behavior under various geological and construction conditions18,20,28. Han et al.29 investigated the performance of composite soil nail structures in symmetrical silty soil foundation pits, discussing key factors influencing pit stability and associated deformation characteristics. Liu et al.30 conducted a case study on a large-scale deep foundation pit in Xi’an High-Tech Zone, integrating field monitoring data with numerical simulations to systematically analyze variation patterns of lateral displacement in retaining piles, periphery soil, crown beams, ground surface, and adjacent building settlements. Wang et al.21 enhanced the stability analysis method for composite soil nail walls using an energy-based approach and proposed a calculation formula for the safety factor of prestressed anchor bolt-supported composite soil nail walls. Sun and Li31 employed the FLAC 3D three-dimensional finite difference software to analyze the evolution patterns of stratum stress, soil nail internal force, anchor rod axial force, as well as pile stress and displacement during foundation pit excavation and support. Wang et al.32 developed high-precision numerical models for expansive soil slopes supported by various methods, providing in-depth insights into the working mechanisms of different support schemes in treating expansive soil slopes. Imam and Sadeghi33 performed finite element simulations on a typical 20-m-high composite wall under static and seismic loading conditions, validating the results through comparisons with full-scale and physical model test data. Although extensive research has been conducted on soil nail-anchor cable composite support systems in soft soil excavations, their response mechanisms under the unique engineering conditions of loess areas remain poorly understood. In particular, previous studies have predominantly focused on saturated soft soils, overlooking the displacement-control benefits offered by the deep unsaturated conditions typical of loess regions. To address this gap, this paper draws on eight months of field monitoring data from a deep excavation in Luoyang to reveal the deformation evolution pattern of the composite support system in unsaturated loess. In addition, a sensitivity analysis is performed on soil parameters and soil nail stiffness. The findings provide a valuable reference for similar projects. Project overview and geotechnical conditions Project overview The project site encompasses a land area of approximately 76,340.80 m2, with a total constructed floor area of 206,738.61 m2. The development comprises 16 residential buildings, a three-story kindergarten, a neighborhood center ranging from two to five stories, and a two-story commercial block. The primary structural systems are shear-wall and frame structures, founded on independent and pile-raft foundations. The foundation pit measures 445 m in length and 133.5 m in width. Following leveling of the northern section, the ground elevation ranges from 300.3 to 304.0 m, while the garage base elevation is 291.70 m, resulting in an excavation depth of 8.6 to 12.3 m. Due to the proximity of a tomb and an electrical pole at the pit’s northwest corner, a vertical excavation with pile-anchor support was implemented. All other sections employ a composite soil–nail wall support with a 1:0.3 sloped excavation. Overview of engineering geology The project is situated within a designated urban renewal zone in Luoyang City. The site is primarily composed of cultivated land and previously demolished areas. Topographically, it is generally flat with localized undulations. Geomorphologically, the site is classified as a loess hilly slope. The physical and mechanical parameters of the principal soil layers within the excavation influence zone are provided in Table 1. The site’s groundwater level is approximately 60 m below the surface. The perennial water level is found at a depth of about 50 m. Table 1Physical and mechanical properties of soil.Name of soil layerConditionMoisture content ω (%)Porosity ratio eSaturation (%)Natural densityρ(kN/m3)Poisson’s ratioCompression modulusEs(MPa)Shear strengthccq (kPa)Internal friction Angleφcq (o)(2) Slightly moist, moderately denseSlightly moist, moderately dense18.50.836917.10.356.710.021.0(3) Loessal silty clayHard plastic21.30.7258418.70.37.220.510.5(4) Loessal silty clayHard plastic20.60.78418.80.37.621.511.5(5) Silt intercalated with silty clayModerately dense20.20.777517.90.36.610.521.5(6) Silty clayHard plastic22.30.738518.80.37.623.513.0 Physical and mechanical properties of soil. It should be noted that the groundwater table at the site is approximately 60 m deep; consequently, the loess-like silt and silty clay within the excavation depth remain in an unsaturated state. Due to the presence of matric suction, the in-situ stiffness and strength of the soil are higher than the parameters derived from saturated laboratory tests. Therefore, the values presented in Table 1 represent a conservative lower bound for the engineering analysis. The physical and mechanical parameters of the soil were obtained through laboratory tests. Specifically, natural density and water content were determined using the cutting ring method and oven-drying method, respectively; shear strength parameters (cohesion c and internal friction angle φ) were derived from consolidated quick direct shear tests; and the compression modulus was measured via confined compression tests. Surrounding environment of the project As illustrated in Fig. 1, the foundation pit is situated within a notably complex environment containing several existing structures and features. These primarily consist of 500 kV high-voltage transmission lines, a cemetery, orchards, and utility poles. Particular attention must be paid to the interaction between the pit support structure and a high-voltage tower, which maintains a minimum clearance of 25.4 m. A cemetery, containing burials at approximately 4 m depth, is located on the northwest side and falls outside the project’s demolition scope; it will therefore require either relocation or specific protective measures during construction. Additionally, 10 kV high-voltage lines, buried to a depth of roughly 2.5 m, run adjacent to the pit’s western boundary. On-site working condition. (a) Surveying and positioning; (b) Cross-section. Design and verification of foundation pit support schemes The site stratigraphy primarily consists of loess-like silt and silty clay. A competent layer of silty clay, characterized by uniform composition and high structural integrity, is encountered at a depth of approximately 4.0 to 5.0 m, providing a stable stratum for the foundation pit. Hydrogeologically, the site is characterized by a deep groundwater table, with no groundwater encountered within the excavation depth, thereby mitigating seepage risks and eliminating dewatering requirements. The unoccupied surrounding area provides ample space for construction logistics, including heavy machinery operation and material staging. Notably, due to significant original topographic relief, the excavation depth varies considerably, reaching a maximum of 12.3 m on the north side while remaining below 6.0 m on the south side. Collectively, these geological, hydrological, environmental, and topographic conditions establish the critical boundary constraints for the design and construction of the pit, directly influencing the selection and implementation of the support system. The proposed support structure is illustrated in Fig. 2. Schematic diagram of pit support structure. Based on the aforementioned geological and environmental conditions, the foundation pit support design was formulated as follows. Given its proximity to a tomb and a utility pole, the northwest corner employs a vertical excavation system utilizing pile-anchor support (Table 2). For the remaining areas, a composite soil–nail wall with sloped excavation is implemented, using slope ratios of either 1:0.3. This paper focuses on the practical application of the composite soil–nail wall technique. Table 2Parametric case of soil nails and anchor cables.The first row of soil nailingThe second row of soil nailingThe third row of soil nailingThe first row of anchor cableThe second row of anchor cablesThe third row of anchor cablesThe four row of anchor cablesLength (m)9.012.012.015.016.016.015.0Horizontal spacing (m)1.51.51.51.51.51.51.5Vertical spacing (m)1.51.51.51.51.81.81.8Steel bars/steel strands1 φ 221 φ 222 φ 223 15.24 15.24 15.23 15.2Borehole diameter (mm)110110110180180180180Elastic modulus (GPa)200200200195195195195Poisson’s ratio0.30.30.30.30.30.30.3Bulk density (kN/m3)76.9876.9876.9878.578.578.578.5Prestressed (kN)–––301399450.8273 Parametric case of soil nails and anchor cables. Calculations confirm that the proposed anchor-cable composite soil–nail wall meets all requisite safety standards for the foundation pit. The stability analysis was conducted assuming a groundwater table depth of 60 m, sufficiently deep to exert negligible influence, and a uniformly distributed surcharge load of 20 kPa applied at the slope crest. The overall stability was assessed using the Swedish slice method. All other parameters used in this analysis are detailed in Table 3. Table 3Selection of parameters for profile calculation.Parameter nameParameter valuePartial coefficient of soil nail load1.25Empirical coefficient of axial tensile force of anchor supports on the bottom surface of the soil nail wall0.8The width of the soil strips in the strip division method0.5 mThe cut-off depth below the bottom surface of the foundation pit5.0 mThe step size of the bottom surface of the foundation pit is searched by the sliding crack surface1.0 mThe overall stability and safety factor during the construction period is reduced1.0Overall sliding stability safety coefficient1.3 Selection of parameters for profile calculation. The calculated standard values for uplift bearing capacity of the soil nails and anchor cables, from the topmost to the bottommost element, are 168, 230, 229, 509, 560, 568, and 541 kN, respectively. The corresponding standard values for axial tensile force are 44, 83, 132, 215, 285, 322, and 195 kN. The resulting pull-out safety factors, computed as the ratio of bearing capacity to tensile force, are 3.82, 2.77, 1.73, 2.37, 1.96, 1.76, and 2.77, respectively. All calculated safety factors exceed the required minimum of 1.6, confirming the design’s compliance with stability criteria. Numerical simulation analysis of support profile Maida GTS NX is often used for foundation pit support analysis and verification. According to the existing design and construction of foundation pit support, GTS is used for two-dimensional analysis and verification of this section. Establishment of foundation pit support model The model is 30 m wide and 20 m deep. The size of the edge grid and the five layers of silty clay is 0.5 m, and the size of the support surface grid is 0.2 m. The horizontal soil layer and each excavation bottom surface are divided by linear length. The design model is shown in Fig. 3. The soil constitutive adopts the modified Moorn-Coulomb. The required soil model parameters are the same as those in Table 1. Among the nonlinear parameters, the secant stiffness of the triaxial experiment and the secant stiffness of the main compaction loading experiment both adopt the compressive modulus of the soil in the investigation report. Under normal circumstances, the unloading elastic modulus is 2 to 5 times the compressive modulus. In this paper, a threefold relationship is selected for all soil masses. Both the soil nails and anchor cables adopt implantable truss units. According to the foundation pit support construction plan, the excavation depth of each layer from top to bottom is 1.8 m. The construction conditions are shown in Table 4. Computational model diagram. Table 4Modeling conditions.Model operating conditionsConstruction contentInitial stageActivate soil layers, apply boundary constraints, apply gravity load, and apply a 20 kPa surcharge load at the pit edgeExcavation 1- Support 1Excavate to a depth of 1.8 m, and activate the first row of soil nails along with the first layer of shotcrete with wire meshExcavation 2- Support 2Excavate to a depth of 3.6 m, and activate the second row of soil nails along with the second layer of shotcrete with wire meshExcavation 3.1 and 3.2- Support 3Excavate to a depth of 5.4 m, and activate the third row of soil nails along with the third layer of shotcrete with wire meshExcavation 4- Support 4Excavate to a depth of 7.2 m, and activate the fourth row of soil nails along with the fourth layer of shotcrete with wire mesh, applying a prestress of 301 kPa to this support layerExcavation 5- Support 5Excavate to a depth of 9.0 m, and activate the fifth row of soil nails along with the fifth layer of shotcrete with wire mesh, applying a prestress of 399 kPa to this support layerExcavation 6.1 and 6.2- Support 6Excavate to a depth of 10.8 m, and activate the sixth row of soil nails along with the sixth layer of shotcrete with wire mesh, applying a prestress of 450.8 kPa to this support layerExcavation 7- Support 7Excavate to the final depth of 12.3 m, and activate the seventh row of soil nails along with the seventh layer of shotcrete with wire mesh, applying a prestress of 273 kPa to this support layer Modeling conditions. Simulation result analysis According to the described parameters and working conditions, the deformation of the foundation pit was simulated and calculated, and the displacement of the profile model shown in Fig. 4 was obtained. Schematic of pit profile displacements. (a) Support condition I; (b) Support condition II; (c) Support condition III; (d) Support condition IV; (e) Support condition V; (f) Support condition VI; (g) Support condition VII. The simulated horizontal displacement field is shown in Fig. 4. The results indicate that, except for Cases 6 and 7, the maximum displacement of the excavation under all other support conditions occurred at the bottom of the excavation after excavation. This is because the removal of the overlying soil weight during excavation leads to rebound and heave of the underlying soil. As excavation depth increased, the bottom heave values for Cases 1 to 4 were 20.08 mm, 33.23 mm, 42.2 mm, and 49.95 mm, respectively, reaching a near-maximum value of 53.84 mm by Case 5. Further excavation resulted in only minor changes in bottom displacement, exhibiting a “first increase then stabilization” trend. This pattern reflects the combined effect of excavation-induced soil rebound and the restraining action of subsequent anchor cables. For Cases 1 to 7, the horizontal displacements at the top of the supporting structure were 13.39 mm, 22.15 mm, 28.13 mm, 37.47 mm, 49.3 mm, 60.54 mm, and 80.72 mm, respectively. These values show a monotonic increase, indicating that the unloading effect persistently acts on the supporting system throughout the construction process, leading to continuous deformation accumulation without any sign of stabilization or reduction. This behavior aligns with the typical deformation pattern of deep excavations. Furthermore, the data presented in Fig. 5 clarify the relationship between the maximum displacement at the excavation edge and the excavation depth. The analysis reveals that the maximum lateral displacement is approximately proportional to the excavation depth. Deformation of the support structure under various support conditions. As shown in Fig. 5, the horizontal displacement at the top of the slope under the soil nail-anchor cable composite support exhibits a significant nonlinear positive correlation with excavation depth. An exponential function regression analysis was performed on the measured data, yielding a deformation evolution model with a coefficient of determination (R2) as high as 0.997, indicating that the model can accurately describe the growth pattern of top displacement of this support system with increasing excavation depth. The curve shape of the displacement-depth relationship presents a typical concave characteristic, reflecting the nonlinear cumulative deformation effect of the composite support structure. During the shallow excavation stage, the timely installation of soil nails creates a favorable composite reinforcement effect with the soil, effectively restraining deformation and resulting in a relatively gentle increase in horizontal displacement at the slope top. However, as excavation depth continues to increase, the unloading effect in deeper soil layers intensifies significantly. The restraining capacity provided by the anchor cable prestress and soil nails gradually becomes insufficient to counteract the lateral deformation induced by excavation unloading. Consequently, there is a marked acceleration in the growth rate of slope-crest displacement, with the curve gradient progressively steepening as the excavation depth increases. The high goodness-of-fit (R2 = 0.997) of this exponential model confirms that excavation depth is the dominant factor controlling slope-top displacement in soil nail-anchor cable support systems. Moreover, the deformation is not a simple linear accumulation but exhibits an accelerating trend with increasing depth. This finding reveals the deformation response mechanism of soil nail-anchor cable composite support in deep excavation engineering. A shown in Fig. 6, without support, when the excavation reached the first working condition with a depth of 1.8 m, the maximum horizontal displacement at the top of the slope was 46.9 mm, and the slope was in a stable state. When the excavation continued until the first working condition with a depth of 3.6 m, the load on the soil model was too large, and the rock and soil mass entered the plastic yield stage. The excessive deformation led to the failure to meet the convergence conditions. The essence is that the load level of this working condition exceeds the current ultimate bearing capacity of the model, causing the slope to collapse and the calculation to terminate. As can be seen from the Fig. 6, after the excavation reaches the first working condition, the displacement of the soil is mainly concentrated at the excavation surface of the foundation pit. This phenomenon is significantly different from the situation with support. The reason is that the support structure disperses the overturning force of the outer soil to the deep soil on the outer side of the slope through soil nails, anchor cables and grouting soil, causing the deep soil to also undergo displacement. It can be seen from this that the composite support method of soil nails and prestressed anchor cables used in this project has limited the displacement of the slope and is truly effective in ensuring the safety of the project. Schematic diagram of unsupported slope displacement under identical conditions. A comparative numerical analysis was conducted to evaluate the efficacy of the support system. As shown in Fig. 6, for the unsupported slope scenario, excavation to the first stage (1.8 m depth) resulted in a maximum horizontal displacement of 46.9 mm at the slope crest, with the model remaining in a stable state. However, upon excavation to the second stage (3.6 m depth), the induced stresses exceeded the soil’s shear strength, forcing the model into a widespread plastic yield state. This condition produced excessive, non-convergent deformations, representing a numerical failure that corresponds to a physical slope collapse. The calculation terminated as the load demand surpassed the model’s ultimate bearing capacity. Figure 6 also illustrates that upon reaching this failure stage, displacement in the unsupported model is concentrated at the excavation face. This pattern contrasts sharply with the supported case, where displacements are distributed more deeply. The support system—comprising soil nails, prestressed anchor cables, and grouted soil—effectively transfers lateral earth pressures from the near-slope region to more competent soil at depth, mobilizing a larger resisting mass. Consequently, the composite soil–nail and anchor-cable support method employed in this project successfully limits slope displacement and is fundamentally effective in ensuring overall project safety. Monitoring and analysis of foundation pit support displacement Construction of the foundation pit proceeded according to the established support plan. Concurrently, a comprehensive monitoring program was implemented at the section position detailed in this study. Excavation at this location commenced on June 5, 2024, and reached the final pit bottom elevation on July 9, 2024. The pit remained in an open, unsupported state with no backfill as of the latest monitoring date in February 2025. The instrumentation scheme, shown in Fig. 2, consisted of three primary monitoring points at the section: a horizontal displacement monitoring point (W44) at the slope crest, a surface settlement monitoring point (D41) at the slope crest, and an inclinometer (C12) installed within the support structure to measure deep-seated lateral movements. Data analysis of slope top displacement monitoring Figure 7 presents the time-history displacement data recorded at monitoring point W44 at the slope crest. As shown in Fig. 7a, the vertical displacement at the slope crest exhibited a slow, increasing trend following the commencement of excavation. The settlement rate increased after the excavation reached its final depth on July 9, which aligns with the typical deformation behavior of deep excavations. The settlement culminated in a maximum value of 3.91 mm on August 29, approximately 50 days after excavation completion, before gradually stabilizing. This post-construction settlement is attributed to soil creep and consolidation, a process likely accelerated by continuous regional rainfall from July to September. The infiltrating rainwater increased the soil unit weight and potentially reduced its shear strength, leading to additional time-dependent deformation. Schematic of displacement monitoring data at the top of the slope. (a) Vertical displacement; (b) Horizontal displacement. From Fig. 7b, the horizontal displacement at the crest increased incrementally with excavation depth, reaching a maximum of 3.4 mm on October 18. This trend conforms to the expected mechanical response. A subsequent minor rebound to 2.6 mm was recorded on November 7, representing a decrease of 0.8 mm. Given the small magnitude of this change, it may be attributed to minor onsite disturbances, thermal effects on instrumentation, or expected measurement variability. The monitoring results demonstrate that the composite soil–nail and anchor-cable support system effectively controlled deformation, with all recorded displacements remaining within the project’s safety specifications. The system is particularly suitable for this loess-area project due to its minimal dependency on groundwater conditions, which is advantageous given the site’s deep groundwater table. Surface settlement at the pit edge was monitored at a location one meter from the excavation boundary. As shown in Fig. 8, all measured settlements remained within a narrow range of 3.5 mm. The data reveal a distinct two-phase behavior: an initial significant settlement of 2.23 mm occurred on June 20, immediately following the start of excavation, indicating considerable disturbance to the near-surface soil. Subsequently, the settlement rate stabilized dramatically, with variations of no more than 1.0 mm recorded over successive 5 to 10-day intervals. This represents a transition to a steady-state deformation phase as excavation deepened. The cumulative settlement reached a maximum of 3.23 mm on September 28. The minimal magnitude and stabilized trend of the surface settlement confirm the effectiveness of the foundation pit support system. In the loess stratum, where soil stability is inherently favorable in the absence of groundwater influence, the prestressed anchor cable composite soil–nail wall successfully controlled excavation-induced deformations. Schematic of surface settlement monitoring data at the top of the slope. Monitoring and analysis of inclinometry of support structures Inclinometer monitoring data provide a clear representation of the deformation rate and magnitude of the foundation pit support structure. Figure 9 presents a time-series of inclinometer data at key monitoring dates. The displacement-depth profiles exhibit consistent undulating patterns. Crucially, displacement is minimized at the precise depths corresponding to the soil nail and anchor cable heads, confirming their effectiveness in restraining lateral movement. The soil between reinforcement layers shows slightly greater displacement, though the relative displacement across any anchor cable position does not exceed 1.0 mm. This pattern is attributed to the composite action of the support system, where the shotcrete facing provides continuous surface restraint, transferring loads to the discrete reinforcements. The maximum horizontal displacement in each profile consistently occurs at a shallow depth of approximately 0.5 m. This correlates well with the numerical simulation and is explained by the site’s stratigraphy: the uppermost layer is a weak, loess-like silt, and the first row of soil nails is installed at a depth of 1.5 m. Consequently, the restraint of the near-surface soil relies primarily on the stiffness of the shotcrete facing and the zone of soil improved by grouting5,6,34,35. Throughout the entire excavation period, the maximum recorded horizontal displacement of the support structure remained below 1.6 mm, demonstrating the overall efficacy of the designed support system. Schematic diagram of the horizontal displacement monitoring data of the CX12 slope. Comparative analysis of simulation and monitoring data This subsection presents a comparison between the monitored horizontal displacement of the support structure (following pit stabilization) and the corresponding deformation results from the GTS NX numerical model for the final excavation condition. Figure 10 shows the comparison of horizontal displacement monitoring of support structures and finite element analysis. Comparison of horizontal displacement monitoring of support structures and finite element analysis. A comparative analysis was conducted between the in-situ horizontal displacement of the support structure, measured after pit stabilization, and the results from the GTS NX numerical model under final excavation conditions. Field monitoring recorded a maximum horizontal displacement of 1.34 mm, a practically negligible value that remained well below the on-site maximum of 1.6 mm. In stark contrast, the numerical simulation predicted a maximum displacement of 80.72 mm. This substantial discrepancy of over two orders of magnitude necessitates a detailed examination of the factors contributing to the divergence between predicted and observed behavior. The significant discrepancy between simulated and measured displacements can be attributed to limitations inherent in the numerical modeling approach, particularly concerning boundary conditions and the simulation of construction effects. First, the assumed boundary conditions may not fully represent the actual site. The selected pit section borders a village orchard without adjacent roads or permanent structures. While sporadic activity from pedestrians and light agricultural machinery occurs, these transient loads are challenging to quantify. Therefore, the model’s assumption of a uniform, long-term 20 kPa surcharge likely represents an overestimate of the true static load, leading to exaggerated deformation predictions. Second, the numerical model cannot fully capture beneficial ground improvement from the construction process. During construction, three rows of soil nails were installed using full-length grouting with 42.5-grade cement slurry at a pressure ≥ 0.6 MPa. This process actively densified and strengthened the soil matrix around the nails, creating a composite mass with higher stiffness and strength than the native soil assumed in the 2D simulation. The omission of this three-dimensional improvement effect results in the model over-predicting horizontal displacement. Analysis of the monitoring data demonstrates that the composite support system combining soil nailing and anchor cables is particularly suitable for loess regions with low groundwater levels, where minimal wall displacements are observed. On the other hand, finite element analysis using Midas GTS NX reveals a relatively regular pattern of horizontal displacement in the support structure. Specifically, lateral displacement increases progressively with excavation depth and, upon reaching the bottom of the pit, exhibits a decreasing trend from the slope crest downward—a deformation profile consistent with typical slope behavior. A comparative assessment of field data and simulation results highlights notable discrepancies attributable to uncertain site-specific factors, indicating that the current numerical approach tends to yield conservative predictions. To improve the predictive accuracy of finite element models, further refinement of boundary conditions and more precise representation of construction-induced effects—such as grouting—are recommended. Model optimization and analysis Due to deviations between the simulated data and the measured data, the model parameters were adjusted. The adjustments are presented in Table 5. Table 5Adjustment of soil parameters.ParametersAdjustment methodOriginal parametersModify parametersLoessal silty clayLoessal silty claySilt intercalated with silty clayLoessal silty clayLoessal silty claySilt intercalated with silty clayTriaxial experiment tangential stiffnessExpand by four times0.005340.005640.004900.01670.02140.0196Internal friction angle φcq (o)Appropriate adjustment10.511.521.5241824 Adjustment of soil parameters. Three rounds of adjustments were made to the model material parameters. The first adjustment involved the secant stiffness from triaxial tests. The second adjustment, based on the first, appropriately increased the internal friction angles of each soil layer. The third adjustment, based on the second, modified the elastic modulus of the soil nails from the original value of 200–200,000 kN/mm2. The rationale for these adjustments cannot exclude the possibility of limitations in the data from the geological investigation report, raising doubts about whether the model can accurately reflect the actual geological conditions. After parameter adjustment, the model was re-established and simulations were performed, as shown in Fig. 11, which presents the deformation contours of the supporting structure for Case 7 under the three adjustment rounds. Displacement of the supporting structure after parameter adjustments. (a) After adjusting the triaxial secant stiffness; (b) After further adjusting the internal friction angle; (c) After subsequently adjusting the soil nail material stiffness. Through phased optimization of soil stiffness, strength, and soil nail properties, the displacement deformation in the excavation simulation was significantly reduced, from an initial value of 80.72 mm to a final value of 19.36 mm—a reduction of 76%. This indicates that the model parameter adjustments were effective and that the structural stability was greatly improved. Specifically, first, when the soil stiffness was increased to four times its original value, the displacement decreased to 33.59 mm, a reduction of 58.4%, demonstrating that soil stiffness is the primary factor controlling overall deformation. Second, after adjusting the internal friction angles of the soil layers, the displacement further decreased to 23.42 mm, indicating that a reasonable assignment of soil shear strength effectively suppresses the development of plastic zones. Third, following the adjustment of the soil nail elastic modulus, the displacement finally stabilized at 19.36 mm, confirming that correct nail stiffness is crucial for controlling local deformation and ensuring cooperative composite behavior. The final displacement of 19.36 mm is within the acceptable range for engineering practice, and the deformation distribution is reasonable. The soil nails and the soil exhibit good cooperative behavior, with no abnormal concentrated deformation or continuous slip surface. The model thus possesses engineering application value. Sensitivity analysis of each parameter To quantify the sensitivity of support deformation to specific design parameters, a single-factor sensitivity analysis was performed on the triaxial reference stiffness , the internal friction angle φcq, and the elastic modulus of the soil nails within the second layer of loess-like fine soil. During this analysis, each parameter was varied independently while the remaining variables were held constant. As illustrated in Table 6 and Fig. 12, as increased from the baseline value of 0.00534 kN/mm2 to 3.0 times its original value, the maximum horizontal displacement at the crest of the support structure decreased monotonically from 42.48 to 30.35 mm—a total reduction of 28.6%. The relationship between displacement and followed a significant exponential decay (R2 = 0.998). This trend suggests that while initial increases in soil stiffness yield substantial reductions in displacement, the marginal utility of increasing stiffness diminishes as the soil becomes more rigid. Table 6Sensitivity of the line stiffness of the triaxial test to the top displacement of the supporting structure.The initial value of the secant stiffness for the three-axis experiment (kN/mm2)ChangeValue (kN/mm2)Top displacement of the supporting structure (mm)0.00534 × 10.0053442.48 × 1.50.0080136.68 × 20.0106833.69 × 2.50.0133531.75 × 30.0160230.35The data in the table were obtained under the condition of a fixed internal friction angle φcq = 10.5° and an elastic modulus of soil nails of 2.0 × 105 kN/mm2. Only the normal stress stiffness of the second layer of loess-like fine soil was changed. Sensitivity of the line stiffness of the triaxial test to the top displacement of the supporting structure. The data in the table were obtained under the condition of a fixed internal friction angle φcq = 10.5° and an elastic modulus of soil nails of 2.0 × 105 kN/mm2. Only the normal stress stiffness of the second layer of loess-like fine soil was changed. Relationship curve between the secant stiffness of the triaxial test and the maximum horizontal displacement at the top of the support structure (R2 = 0.998). As illustrated in Table 7 and Fig. 13, when the internal friction angle φcq was increased from 10.5° to 21.0°, the displacement exhibited only a marginal decrease from 42.48 to 40.64 mm, representing a total reduction of 4.3%. Although a strong linear negative correlation exists between displacement and φcq (R2= 0.987), the magnitude of this variation is considerably smaller than that induced by stiffness parameters. This suggests that fluctuations in strength parameters within a reasonable range do not exert a significant impact on deformation predictions. Table 7Sensitivity analysis of the internal friction angle φcq on the displacement at the crest of the support structure.Initial value of the internalfriction angle (°)ChangeValue (°)Top displacement of the support structure (mm)10.5 × 110.542.48 × 1.2513.12541.86 × 1.515.7541.42 × 1.7518.37540.94 × 22140.64The secant stiffness of the three-axis test is 0.00534 kN/mm2. and an elastic modulus of soil nails of 2.0 × 105 kN/mm2. Only the internal friction angle φcq of the second layer of loess-like fine soil is changed. Sensitivity analysis of the internal friction angle φcq on the displacement at the crest of the support structure. The secant stiffness of the three-axis test is 0.00534 kN/mm2. and an elastic modulus of soil nails of 2.0 × 105 kN/mm2. Only the internal friction angle φcq of the second layer of loess-like fine soil is changed. Relationship curve between internal friction angle φcq and the maximum horizontal displacement at the top of the support structure (R2 = 0.987). As presented in Table 8, an orders-of-magnitude increase in the elastic modulus of the soil nails—from to 200,000 kN/mm2—resulted in a reduction in displacement from 37.06 to 31.67 mm. Notably, when the modulus was increased beyond 20,000 kN/mm2, the displacement remained nearly constant. This plateau suggests that the contribution of the soil nails’ axial stiffness to the overall support system deformation has reached a saturation point. These findings imply that under the simulated conditions of the embedded truss unit, the load transfer mechanism is primarily governed by the bonding performance at the grout-soil interface rather than the intrinsic elastic modulus of the reinforcement. Table 8Sensitivity of soil nail elastic modulus to the top displacement of the supporting structure.Initial value of the elastic modulus of soil nails (kN/mm2)ChangeValue (kN/mm2)Top displacement of the support structure (mm)200 × 120037.06 × 10200032.33 × 10020,00031.74 × 1000200,00031.67The data in the table were obtained under the condition of a fixed internal friction angle φcq = 10.5° and a triaxial test shear line stiffness = 0.00534 kN/mm2. Only the elastic modulus of the soil nails was changed. Sensitivity of soil nail elastic modulus to the top displacement of the supporting structure. The data in the table were obtained under the condition of a fixed internal friction angle φcq = 10.5° and a triaxial test shear line stiffness = 0.00534 kN/mm2. Only the elastic modulus of the soil nails was changed. Conclusion Based on field monitoring data and numerical simulations of a deep foundation pit project in a loess region, this study evaluates the performance of a prestressed anchor cable composite soil–nail support system. The principal conclusions are as follows: Finite element analysis effectively captures the deformation behavior of the excavation process. Both the settlement at the pit edge and the lateral displacement of the support structure exhibit an approximately linear relationship with increasing excavation depth. The maximum recorded displacements—3.91 mm at the slope crest and less than 1.6 mm at a depth of 0.5 m—are significantly below the predefined monitoring alarm thresholds, thereby confirming the effectiveness of the support system. Finite element analysis effectively captures the deformation behavior of the excavation process. Both the settlement at the pit edge and the lateral displacement of the support structure exhibit an approximately linear relationship with increasing excavation depth. The maximum recorded displacements—3.91 mm at the slope crest and less than 1.6 mm at a depth of 0.5 m—are significantly below the predefined monitoring alarm thresholds, thereby confirming the effectiveness of the support system. Both numerical simulations and field monitoring consistently identify the shallow, weak silt layer with loess-like characteristics as the zone of maximum deformation. This is attributed to the low cohesive strength of the soil and the fact that the first row of soil nails is installed at a depth of 1.5 m, resulting in limited confinement of the near-surface soil mass. Both numerical simulations and field monitoring consistently identify the shallow, weak silt layer with loess-like characteristics as the zone of maximum deformation. This is attributed to the low cohesive strength of the soil and the fact that the first row of soil nails is installed at a depth of 1.5 m, resulting in limited confinement of the near-surface soil mass. Comparative analysis highlights the necessity of the composite support system. In numerical modeling, the unsupported slope fails during the second excavation stage (at 3.6 m depth), when stresses exceed the shear strength of the soil, indicating a potential for physical collapse. In contrast, the supported model remains stable throughout the excavation process. The support system functions by transferring lateral earth pressures through the combined action of soil nails, anchor cables, and the grouted soil mass, thereby mobilizing deeper soil layers and redistributing deformation away from the excavation face. Comparative analysis highlights the necessity of the composite support system. In numerical modeling, the unsupported slope fails during the second excavation stage (at 3.6 m depth), when stresses exceed the shear strength of the soil, indicating a potential for physical collapse. In contrast, the supported model remains stable throughout the excavation process. The support system functions by transferring lateral earth pressures through the combined action of soil nails, anchor cables, and the grouted soil mass, thereby mobilizing deeper soil layers and redistributing deformation away from the excavation face. Field monitoring validates the design assumptions. The horizontal displacement profile with depth exhibits a distinct undulating “sawtooth” pattern, with local minima occurring at the elevations of reinforcement elements. This pattern provides direct evidence that soil nails and anchor cables effectively restrain lateral soil movement at their respective installation levels. Field monitoring validates the design assumptions. The horizontal displacement profile with depth exhibits a distinct undulating “sawtooth” pattern, with local minima occurring at the elevations of reinforcement elements. This pattern provides direct evidence that soil nails and anchor cables effectively restrain lateral soil movement at their respective installation levels. Parameter optimization (soil stiffness, friction angle, nail modulus) reduced the maximum numerical displacement from 80.72 to 19.36 mm (76% reduction), improved internal force distribution and soil nail‑anchor synergy, and suppressed plastic zones. Parameter optimization (soil stiffness, friction angle, nail modulus) reduced the maximum numerical displacement from 80.72 to 19.36 mm (76% reduction), improved internal force distribution and soil nail‑anchor synergy, and suppressed plastic zones. Parameter sensitivity analysis demonstrates that soil stiffness is the primary governing factor for foundation pit deformation, whereas the influence of strength parameters and soil nail axial stiffness is relatively small. Parameter sensitivity analysis demonstrates that soil stiffness is the primary governing factor for foundation pit deformation, whereas the influence of strength parameters and soil nail axial stiffness is relatively small. Acknowledgements National Natural Science Foundation of China (Grant No. 42477209, 42302320, 42577173) and Jiangsu Natural Science Foundation (BK20250207). Author contributions Xiaoyi Zhang and Lulu Liu wrote the main manuscript text and Xiaoyan Liu and Shunfu Zheng prepared all figures. All authors reviewed the manuscript. Funding National Natural Science Foundation of China, Grant No. 42477209, Grant No. 42302320, Grant No. 42577173, Jiangsu Natural Science Foundation, BK20250207. Data availability All data, models, and code generated or used during the study appear in the submitted article. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data, models, and code generated or used during the study appear in the submitted article.