Assessing building damage during station break-out in urban railway construction: real monitoring data

Abstract

The rapid expansion of urban railway networks has necessitated the extensive construction of underground stations within densely populated urban environments, where adjacent buildings are often situated in close proximity to excavation zones. Among the various stages of underground station construction, the station break-out phase—during which the tunnel boring machine (TBM) transitions between the station box and the running tunnel—represents one of the most critical and high-risk operations. This phase is characterized by complex ground–structure interactions, abrupt alterations in stress redistribution, and an elevated likelihood of ground deformation. Such conditions may induce adverse effects on surrounding structures, including cracking, differential settlement, and, in extreme cases, partial or complete structural failure. Despite notable advancements in excavation technologies and ground improvement techniques, unforeseen damage to nearby buildings continues to be reported, underscoring the limitations of existing predictive approaches and the need for enhanced analytical frameworks. Accordingly, this study aims to investigate the underlying mechanisms and contributing factors associated with building damage during station break-out operations, while proposing more robust strategies for risk mitigation and infrastructure protection in urban tunneling projects. The findings reveal a strong interdependence between the stability of excavation chamber pressure and the magnitude of ground response, thereby highlighting the critical importance of rigorous face support control, alongside the implementation of continuous, high-resolution monitoring systems to ensure construction safety and minimize structural impacts.

Introduction

Constructing underground metro stations in densely built urban zones presents significant geotechnical and structural challenges, particularly due to ground movement and soil–structure interaction effects^1,2. In many global cities, including Cairo, tunnelling is performed in proximity to existing buildings and infrastructure, raising concerns about differential settlement, tilting, and structural damage³. The implementation of shield tunnel boring machines (TBMs) introduces complex ground displacement patterns that conventional empirical models, such as Peck’s Gaussian approach, may not fully capture—especially in curved alignments or variable soil conditions⁴. In Cairo’s urban metro expansion—particularly Lines 2 and 3—numerous studies have documented TBM-induced settlement, echoing global observations. For instance,³ performed detailed monitoring of twin stacked tunnels, identifying that face, shield, and tail losses all contribute to surface subsidence, and adjusted volume-loss models accordingly. ⁵ presents results of numerical modeling of tunneling using mechanized tunnel boring machines.⁶ assessed the applicability of current analytical models in predicting the likelihood of splitting in precast concrete tunnel segments during the thrust phase of tunnel boring.⁷ investigated the tunneling in fractured zones significantly that affects surface and building deformation.⁸ made a sensitivity analysis for a metro tunnel excavation.⁹ illustrates a new procedure for analyzing the behavior of the micro-tunnel support system, based on widely used and tested analytical calculation methods. ¹⁰ presented the results obtained from an extensive laboratory test campaign that allowed to fully characterize the two-component filling material during the setting period.

The impact on overlying structures can be substantial. Recently, in the Zamalek district near Safaa Hegazy (formerly Zamalek) Station, settlement events led to the evacuation of the historic Sharbatly building when cracks and tilting were observed during TBM break-out stage. ¹¹ investigated subsurface soil variability in Cairo using integrated geophysical and geotechnical techniques, including MASW, GPR, and electrical resistivity, to assess the causes of building tilt and settlement. Advancements in mitigation and modeling techniques provide promising pathways for addressing settlement challenges. Approaches such as vertical or portal partitioning have been shown to reduce settlements by up to 70%¹². Additionally, deep-learning surrogate models like DeepONet have begun fusing simulation and monitoring data to enable real-time settlement predictions¹³. Recent advances in monitoring systems have significantly enhanced the early detection and management of risks associated with ground instability and infrastructure deformation in underground construction. Modern monitoring frameworks increasingly integrate real-time sensor networks, including inclinometers, extensometers, automated total stations, and satellite-based techniques such as Interferometric Synthetic Aperture Radar (InSAR), enabling continuous tracking of ground and structural responses during excavation processes. For instance,¹⁴ demonstrated the effectiveness of satellite InSAR for monitoring urban subsidence and infrastructure deformation, while¹⁵ highlighted the application of distributed fiber optic sensing technologies in capturing strain evolution along tunnel linings with high spatial resolution. In addition, the integration of Internet of Things (IoT)-based monitoring platforms allows real-time data acquisition and transmission, improving responsiveness to abnormal deformation patterns¹⁶. More recently, machine learning-based early warning systems have been developed to predict ground movement and structural risks by combining monitoring data with numerical simulations. For example,¹⁷ developed an ensemble learning model based on finite element–generated datasets to predict surface settlement induced by NATM excavation. Their model incorporates key geotechnical parameters—such as cohesion, internal friction angle, Poisson’s ratio, and elastic modulus—along with tunnel geometry variables including diameter and cover depth, demonstrating the effectiveness of machine learning techniques in capturing complex soil–structure interaction mechanisms.

According to the previously mentioned issues, it can be said that the development of new infrastructure, including underground facilities and high-rise buildings, often requires deep foundation excavations in close proximity to existing subway tunnels. The disturb which can cause due to the tunnel’s excavation can influence the surrounding buildings leading to a reduction in soil strength. As a result, various issues may arise, such as differential settlement, water leakage, segment cracking, and other forms of structural and geotechnical distress^18–20. Current studies on the disturbance effects of foundation pit excavation can generally be divided into three main approaches: physical model testing, numerical simulation, and theoretical analysis. Among these, model testing provides direct insight into soil–structure interaction under controlled conditions. For example,²¹ conducted centrifuge model tests to investigate excavations adjacent to existing tunnels in soft ground, examining key responses such as diaphragm wall bending, ground settlement, tunnel deformation, and surrounding soil pressure. In parallel, numerical simulation has become a widely adopted tool for evaluating construction-induced effects due to its flexibility and efficiency.²² for instance, performed a finite element parametric study to analyze tunnel behavior influenced by nearby deep excavations, considering factors such as tunnel position, excavation geometry, and protective measures. Their work highlights the capability of numerical models to predict deformation patterns under varying design conditions. Theoretical analysis, on the other hand, offers simplified yet effective solutions for estimating tunnel response with reduced computational effort.²³ developed an analytical model based on Mindlin’s displacement solution to predict tunnel deformation caused by circular excavation, demonstrating good agreement with experimental and analytical benchmarks. However, this study examines the tunnel boring machine (TBM) break-out stage during the construction of Zamalek Station (Safaa Hegazy Station) on Cairo Metro Line 3, with particular emphasis on its impact on the settlement behavior of adjacent buildings. A detailed account of the TBM break-out process is provided, alongside an assessment of its influence on building settlements and the subsequent repair measures undertaken. The analysis is grounded in real-time monitoring data, including records of TBM advance rates and pressure losses. Furthermore, the temporal relationship between TBM pressure variations and observed building settlements is documented, offering critical insights into ground–structure interaction during TBM operations in dense urban environments.

A critical research gap exists in understanding the coupled mechanisms governing building response during the TBM station break-out stage, a highly transient phase characterized by abrupt stress redistribution, unstable boundary conditions, and rapid pressure fluctuations. Unlike steady-state tunnelling, where ground–structure interaction is relatively well understood, the break-out stage remains insufficiently investigated—particularly in terms of establishing direct, time-resolved relationships between TBM operational behavior and observed structural deformation. Existing studies tend to treat settlement as a function of soil properties or generalized excavation processes, while overlooking the dynamic and short-duration operational events that can critically alter confinement conditions. This gap becomes particularly evident when examining real-world cases, such as the settlement-induced damage observed in buildings in the Zamalek district near Safaa Hegazy Station during the construction of Cairo Metro Line 3. Detailed analysis of TBM operational data reveals that loss of confinement pressure is not a gradual process, but rather the result of rapid, interrelated hydraulic and mechanical events occurring within seconds. Specifically, sudden increases in advance rate during diaphragm wall crossing, combined with instability in slurry flow balance, can trigger abrupt leakage of bentonite into the steel bell. This leads to a sharp drop in confinement pressure, accompanied by significant volumetric loss and the onset of hydraulic transients (water hammer effects) within the slurry circuit. Subsequent emergency operational responses, such as valve closures and bypass activation, may further induce secondary pressure surges, exacerbating instability. Despite the critical importance of these mechanisms, the temporal and mechanistic linkage between such TBM operational anomalies (e.g., Δflow spikes, slurry loss, pump pressure fluctuations) and the resulting ground and structural response remains poorly quantified in the literature. Accordingly, this study addresses this gap by integrating high-resolution, real-time TBM operational parameters with field monitoring data to elucidate the sequence, causality, and impact of confinement pressure loss on building settlement. The findings aim to advance the mechanistic understanding of ground–structure interaction during critical tunnelling stages and to improve risk assessment and mitigation strategies in dense urban environments.

Accordingly, the central purpose of this study is to examine how TBM operational conditions during break-out influence the settlement of adjacent buildings. The main research question is: how do variations in TBM face pressure, slurry flow, and advance rate translate into measurable ground deformation and structural response in dense urban environments? To address this, real-time monitoring data from Safaa Hegazy Station (Cairo Metro Line 3) are analyzed, including building settlement, crack observations, TBM advance rate, and pressure records, using a temporal correlation approach to interpret ground–structure interaction mechanisms. The main contribution of this study is to establish a direct cause–effect relationship between TBM pressure loss during break-out and building settlement, moving beyond descriptive incident reporting and providing field-based insights for improved risk assessment in urban tunnelling.

Despite the increasing availability of advanced monitoring technologies, a critical limitation in current tunnelling practice lies in the lack of integration between TBM operational data and real-time ground response. Existing approaches typically interpret settlement as a function of soil properties or excavation geometry, while treating operational parameters—such as face pressure, slurry flow imbalance, and advance rate—as secondary or independent variables. This study addresses this limitation by proposing a data-driven and mechanistic framework that directly links transient TBM operational behavior to observed ground deformation and structural response during the break-out stage. Unlike conventional analyses, the present work focuses on short-duration, high-impact events, where loss of confinement occurs within seconds and governs the entire deformation mechanism. In this context, the study contributes to the development of real-time interpretative models for tunnelling monitoring, enabling a shift from descriptive assessment toward predictive and operationally informed risk management in dense urban environments.

Project background

The Zamalek metro station is situated on Zamalek Island, as confirmed by the accompanying aerial imagery. The principal structures located in the immediate vicinity of the tunnel alignment include the following:

• Building Z-24: A residential high-rise consisting of a ground floor and twelve upper floors (GF + 12F).

• Building Z-27: The premises of the Embassy of the Kingdom of Bahrain, which also houses the embassy’s security entrance office.

• Buildings Z-26 and Z-23: The main administrative and functional buildings of the Embassy of the Kingdom of Bahrain.

These structures represent key components of the built environment surrounding the station area and are of significance in relation to both urban context and engineering considerations. Figure 1 shows the layout of the station and the surrounding building.

Fig. 1.

The layout of the station and the surrounding buildings.

The used TBM properties shown in Table 1.

Table 1.

TBM properties.

Specification	Value
Type of tunnelling	Single shield TBM
Excavation diameter	9.46 m
Overall length (including back-up)	175 m in both modes
Number of gantries in backup	12
Max number of 17″ disc cutters	32
Max number of drag bits	224
Drive unit power	2 436 kW
Nominal thrust	65 000 kN
Unlocking thrust	74 300 kN
Nominal torque	21 000 kN·m
Unlocking torque	27 300 kN·m
Variable speed (both directions)	0 to 3 rpm

The grout mix exhibited performance parameters consistent with the requirements for ground stabilization in urban tunneling applications. The measured flowability, assessed using the Marsh cone, ranged between 30 and 45 s, indicating an adequately fluid mix suitable for efficient pumping and penetration into the surrounding soil matrix. The initial setting time was recorded between 5 and 20 s, reflecting a rapid-setting behavior essential for controlling ground movement during excavation. Bleeding was maintained below 2%, demonstrating the grout’s stability and its ability to retain homogeneity without significant water separation. The density of the mix ranged from 1.5 to 1.7 t/m³, aligning with typical values for cementitious grouts used in tunnel face stabilization. In terms of strength development, the unconfined compressive strength (UCS) ranged from 0.3 to 0.8 MPa after 1 day of curing, increasing to between 1 and 5 MPa at 28 days. These strength values confirm the grout’s capacity to provide both immediate and long-term support to the treated ground, thereby contributing to the overall stability of the excavation environment.

Soil investigation and geological situation

Geotechnical investigation program at Zamalek station

A comprehensive subsurface investigation was conducted in the vicinity of Zamalek Station to establish the geotechnical framework required for the design and safe implementation of tunneling works. This investigation integrated historical records from earlier campaigns undertaken between 2001 and 2011 with newly acquired data from boreholes and in-situ tests completed in 2017. The latter were specifically executed to support the construction of Phase 3 of the Greater Cairo Metro Line 3.

Borehole distribution and testing methodology

The geotechnical investigation program comprised a combination of boreholes, cone penetration tests (CPTs), and pressuremeter tests, strategically distributed along the tunnel alignment beneath Ismail Mohamed Street and in proximity to sensitive structures, including residential buildings, embassies, and schools. As illustrated in the borehole location plan, the testing points were positioned longitudinally along the anticipated tunnel boring machine (TBM) trajectory. This arrangement facilitated the characterization of spatial variability in subsurface conditions and supported the evaluation of potential ground settlement risks. The boreholes were classified according to their primary function:

• Cored boreholes (e.g., BH-Z-21, BH-Z-22, BH-Z-23, BH-12D, BH-13D, BH-13DB, BH-32) yielded continuous stratigraphic profiles and samples for detailed laboratory testing.

• SPT boreholes were employed to determine the relative density and strength characteristics of granular deposits through standard penetration testing.

• Cone penetration tests (CPTs) (CPT-Z-07 and CPT-Z-08) provided high-resolution subsurface profiling of sandy and silty strata to depths of approximately 25 m.

• Pressuremeter tests (as indicated in the investigation legend) were conducted to evaluate in-situ stress–strain behavior and the deformability of soils.

The depth of the subsurface investigation extended from approximately 34.7 m (BH-13D) to 69.1 m (BH-12D and BH-13DB), thereby encompassing the full sequence of strata relevant to both the station box excavation and the adjacent TBM tunnel alignment. The boreholes completed in 2017 (BH-Z-21, BH-Z-22, BH-Z-23) reached depths ranging between 46.5 and 60 m, and were supplemented by shallow CPT soundings to depths of approximately 25 m. This combination facilitated a detailed delineation of soil horizons within the upper sand layers, where the TBM operations were primarily concentrated. In parallel, earlier boreholes (BH-12D, BH-13D, BH-13DB, BH-32), drilled between 2001 and 2011, provided essential baseline information on the deeper stratigraphy. The integration of these datasets ensured that both immediate construction-related ground responses and long-term soil behavior could be effectively characterized and evaluated. Figure 2 shows the borehole distribution. Table 2 shows the boreholes coordinates, depth and date.

Fig. 2.

The soil borehole distribution.

Table 2.

The selected boreholes.

Borehole N°	X-coordinate (m)	Y-Coordinate (m)	Ground Level (m asl)	Depth (m)	Start/End Date
BH-Z-21	636,236.4	817,012.1	19.71	60.1	09/04–15/04/17
BH-Z-22	636,268.4	816,945.5	20.23	46.5	08/04–13/04/17
BH-Z-23	636,367.4	816,908.4	19.92	60.0	16/04–21/04/17
CPT-Z-07	636,372.8	816,929.3	21.06	25.3	4/3/2017
CPT-Z-08	636,227.2	816,951.8	20.12	25.2	4/3/2017
BH-12D	636,377.0	816,932.0	19.3	69.1	3/5/2010
BH-13D	636,237.0	817,018.0	19.2	34.7	3/1/2010
BH-13DB	636,239.0	817,006.0	19.2	69.1	12/1/2011
BH-32	636,392.0	816,930.0	19.3	40.2	4/1/2001

Figure 3 illustrates the temporal variation of groundwater levels recorded at Zamalek Station in relation to the average Nile water level, over the monitoring period from November 2016 to February 2018. The data series include maximum, minimum, and average groundwater levels, in addition to localized measurements from boreholes BH-Z-21 and BH-Z-23. The results indicate that groundwater levels generally fluctuated between 15.5 and 17.5 m (m) above sea level, with discernible seasonal variations corresponding to the Nile water regime. The average groundwater level remained within a narrow band of approximately 16.3–16.8 m above sea level for much of the year, while maximum levels peaked at approximately 17.2–17.4 m above sea level during the summer months (June–August 2017). Conversely, minimum levels were consistently lower, dropping to about 15.5 m above sea level, with notable fluctuations during the early months of monitoring (January–March 2017). The borehole-specific data from BH-Z-21 and BH-Z-23 show close correspondence with the average groundwater level, particularly in the upper sand strata, thereby confirming the representativeness of the monitoring network.

Fig. 3.

The ground water levels variation.

The Surface settlements were continuously monitored using a comprehensive and multi-tiered system. A 24-h, 7-day operational monitoring protocol was adopted, combining both visual inspection and instrumental measurements. Continuous visual surveillance was carried out by a dedicated spotter to ensure immediate detection of any abnormal ground behavior. In parallel, surface settlement points (SSPs) were systematically measured at regular intervals of every 4 h under normal operating conditions. During critical construction activities, the monitoring frequency was significantly increased to ensure enhanced safety and data resolution. Specifically, SSP measurements were conducted at 1-h intervals during key operations, including: (i) emptying of the steel bell to atmospheric pressure, (ii) removal of the final 5 m steel bell segment, (iii) TBM advancement, (iv) side removal of the steel bell, (v) dismantling of Ring 1626, and (vi) any other operations identified as critical based on real-time risk assessment. Furthermore, the Z-24 building was instrumented with an automatic monitoring system, providing high-resolution displacement data at 15-min intervals. This integrated monitoring framework ensured both real-time detection and accurate quantification of ground and structural responses throughout the construction process.

Soil tests

Figure 4 presents the variation of fines content (%) with depth, expressed as elevation in meters above sea level, based on soil samples collected from six boreholes (BH-Z-21, BH-Z-23, BH-12D, BH-13D, BH-32, and BH-13DB). The stratigraphic reference lines L1 to L4 denote significant geological or hydrogeological boundaries. The data demonstrate a pronounced stratification in the subsurface profile. At shallow to intermediate elevations, approximately between + 5 m and + 20 m, the fines content exceeds 60%, with several samples approaching values of 90–100%. This suggests the predominance of cohesive soil deposits—primarily clays and silts—within the upper strata. Such a high proportion of fine-grained materials has significant geotechnical implications, as it tends to reduce the frictional resistance of the soil mass. Consequently, these conditions can adversely affect the stability and bearing capacity of foundations, increasing the potential risk of structural deformation or failure in overlying buildings.

Fig. 4.

The fines content distribution.

In contrast, deeper horizons, particularly below − 5 m, are characterized by fines contents generally less than 15%, suggesting the prevalence of coarse-grained materials such as sands and gravels. Notably, lateral variability is observed among the boreholes. While the overall stratification pattern is consistent, the fines content at corresponding elevations shows local variations, reflecting heterogeneity in depositional processes or localized geological features. These findings are of particular importance for assessing soil permeability and hydrogeological behavior, as the distribution of cohesive versus granular strata directly influences groundwater flow, potential settlement characteristics, and contaminant migration pathways. In summary, it can be concluded that the surficial layers contain a relatively high proportion of fines. This content demonstrates a progressive decrease with increasing depth.

Figure 5 presents the classification of soil samples from six boreholes (BH-Z-21, BH-Z-23, BH-12D, BH-13D, BH-13DB, and BH-32) on a plasticity chart plotting liquid limit (WL) against plasticity index (PI). The chart delineates the Unified Soil Classification System (USCS) boundaries, where the A-line (diagonal) separates clays (above the line) from silts (below the line), and the vertical line at WL = 50% distinguishes high plasticity (H) from low plasticity (L) soils. Data points above the A-line and to the left of the vertical boundary are classified as low plasticity clays (CL), while those above the A-line and to the right are high plasticity clays (CH). Samples below the A-line are identified as silts, with ML and MH denoting low and high plasticity, respectively. Organic silts and clays (OL and OH) are also indicated. The majority of samples plot near or above the A-line, suggesting a predominance of clayey soils in the investigated strata, with several samples in the high plasticity region (CH, MH). This indicates potential challenges in engineering due to the swelling behavior of the soil which may lead to building failure, this should be considered in foundation design and slope stability analysis.

Fig. 5.

The liquid limits and plasticity limits for the boreholes.

Figure 6 illustrates the variation of soil density (kN/m³) with depth (m) for samples collected from boreholes BH-Z-21, BH-Z-23, and BH-32. The horizontal lines (L1 to L4) represent key stratigraphic boundaries identified in the subsurface profile. The data show a general trend of increasing density with depth, ranging from approximately 18.5 kN/m³ near the surface to over 21 kN/m³ at depths greater than 30 m below ground level. Samples from shallow depths (above L2) exhibit lower densities (18.5–19.5 kN/m³), likely associated with recent alluvial deposits or soft clayey materials. Below L3, higher densities (20–21 kN/m³) are observed, suggesting the presence of more consolidated sediments such as dense sands or stiff clays. The observed variation indicates a stratified subsurface with distinct mechanical properties, which is critical for evaluating the geotechnical behavior of the site, particularly for foundation design and groundwater flow considerations.

Fig. 6.

The soil density and level.

Figure 7 shows the variation of undrained shear strength (Cu, in kPa) with depth (m) for soil samples obtained from unconfined compression tests (blue dots, qu) and pocket penetrometer tests (red dots, pp). The horizontal lines (L1 to L4) represent key stratigraphic boundaries identified in the study area. Cu values range between approximately 25 kPa and 150 kPa, with most samples concentrated between 10 and 17 m asl. A general increase in undrained shear strength is observed with depth, reflecting the progression from softer, near-surface materials to more consolidated strata at lower levels. Slight variations between Cu values from qu and pp methods are evident, with pocket penetrometer results tending to show slightly higher strength estimates in some cases. This trend highlights the importance of method selection and potential heterogeneity in soil properties. The data are critical for assessing the bearing capacity and stability of the subsurface, particularly within the upper soil layers where construction activities are concentrated.

Fig. 7.

The undrained shear strength values.

Figure 8 presents the Standard Penetration Test (SPT) results for the Zamalek Station investigation, comparing uncorrected N-values (left) with the corresponding corrected values, N₆₀ (right). The data are shown as a function of depth, expressed in elevation (m), for boreholes BH-Z-21, BH-Z-23, BH-32, and BH-12D. Stratigraphic boundaries (L1–L4) are indicated for reference. In the uncorrected dataset, N values display a wide scatter, with some measurements exceeding 100 blows/30 cm, particularly at deeper strata below − 20 m. However, these values do not account for variations in hammer energy efficiency, overburden pressure, or equipment-specific factors, which may lead to overestimation of soil density and strength. The corrected SPT results (N₆₀) provide a more realistic representation of in-situ soil resistance. After correction, the distribution of values shifts downward, with most measurements falling below 90 blows/30 cm. The corrected data also exhibit a clearer stratigraphic trend:

• In the upper layers (above ~ 0 m), N₆₀ values are generally low to moderate (< 30), reflecting weaker, fine-grained soils.

• Between 0 and − 20 m, values increase gradually (30–60), corresponding to medium-dense sandy strata.

• Below − 20 m, N₆₀ values commonly exceed 60, indicating dense to very dense sands and gravels. This is accordance with the fine content and density analysis.

Fig. 8.

Standard penetration test (SPT) correlations and applied corrections.

This correction significantly improves the reliability of geotechnical interpretation by aligning the SPT results with established empirical correlations for soil strength and deformation properties. The corrected N₆₀ values are therefore more appropriate for use in foundation design, settlement prediction, and tunnel stability analyses.

Main problem

The central problem is associated with the following stages, each of which will be accompanied by a series of explanatory steps.

Steel bell door arrangement

The sequence of figures illustrates the failure process associated with bentonite inflow into a steel bell during tunnel boring machine (TBM) excavation. Initially, following 1.5 months of steel bell installation, bleeding of sand cement produced an air gap at the upper part of the bell, indicating reduced integrity of the fill material. At 11:16:10 on 26/07/2020, the TBM operated with a crown pressure of approximately 1.95 bar, with the tail skin and shield in contact with the grout. Shortly after, between 11:16:13 and 11:16:34, bentonite slurry infiltrated the steel bell, leading to a gradual decrease in the bubble level, reflecting pressure equalization and loss of confinement. Subsequently, within a span of 7–10 s (11:16:48–55), a water hammer effect occurred at the top of the steel bell, displacing approximately 20 m³ of bentonite into the gap. This sudden inflow caused a rapid pressure drop (from 1.8 to 1.2 bar), highlighting the instability of the grout–bell interface and the failure of confinement, which ultimately facilitated uncontrolled slurry migration into the structure. The figure illustrates the progression of instability within the steel bell and cutterhead system during TBM excavation on 26/07/2020 at 11:17:10 (T₀ + 1 min 3 s). At this stage, the crown pressure dropped significantly to 0.9 bar, coinciding with the potential inflow of sand and water into both the cutterhead and the steel bell. The suction from the cutterhead was halted, indicating a loss of control over material flow. A bypass valve was triggered in response to the sudden imbalance, but a critical water hammer event occurred in the slurry line, generating a pressure spike of 15 bars due to leakage and rapid fluid displacement. This event highlights the combined failure mechanisms of sand-cement confinement breakdown, uncontrolled bentonite and water inflow, and hydraulic shock in the slurry circuit, ultimately reflecting the loss of face stability and increased risk of ground failure or blowout. The whole details shown in Fig. 9. Thus, it can be said that the incident occurred during the TBM break-in phase at the station, where a steel belt system had been installed to stabilize the tunnel face and maintain the required face pressure in accordance with the design specifications. However, due to shortcomings in the performance of the steel belt system—as detailed in the manuscript—the system was unable to sustain the prescribed pressure levels. Consequently, instability in the face pressure developed, allowing soil to migrate from the ground ahead of the TBM toward the steel belt zone. This uncontrolled soil movement resulted in ground loss, which subsequently manifested as surface subsidence. Accordingly, the observed settlement mechanism is attributed to the loss of face pressure during the break-in operation, rather than to any deficiency or malfunction in the grouting system.

Fig. 9.

The first phases of breakthrough and the flow of sand and water to the cutter head.

Figure 10a presents the excavation conditions at the completion of ring R1619 on 26 July 2020 at 17:30. The TBM crown pressure was limited to 0.5 bar due to persistent leakage at the steel door, restricting the ability to maintain adequate face support. Evidence of sand and water ingress into the cutterhead and steel bell indicates high soil permeability and sensitivity in this section. To mitigate ground loosening and inflow risks, supplementary stabilization measures were applied, including 3 m³ of over-injection and 10.9 m³ of semi-active mortar injection behind the lining. These interventions were essential to enhance face stability, reduce ground loss, and control settlement under adverse conditions. Figure 10b shows the excavation conditions at ring R1620 on 27 July 2020 at 03:41. Crown pressure remained limited to 0.5 bar due to persistent leakage at the steel door, restricting adequate face support. Excavation was further disrupted by cutterhead blockage, resulting in a ~ 3-h delay and prolonged face exposure, which increased the risk of ground loosening and inflow. Evidence of sand and water ingress into the cutterhead and steel bell confirms the high permeability and sensitivity of the soil. To compensate for the pressure deficit, stabilization measures included 7 m³ of over-injection and 14.4 m³ of active mortar injection behind the lining. These interventions were essential to restore stability, control ground loss, and limit surface settlement under adverse conditions. Figure 10c presents the excavation conditions at ring R1621 on 27 July 2020 at 11:58, showing improved face support compared to previous rings. The installation of an additional beam at the steel door enabled an increase in crown pressure to 1.0 bar (1.15 bar at the face), approaching hydrostatic conditions and enhancing stability. Despite these improvements, the risk of sand and water ingress into the cutterhead and steel bell persisted, indicating continued ground sensitivity. To mitigate soil loosening, 7.3 m³ of active mortar injection was applied behind the lining, contributing to improved ground stability and reduced settlement. Figure 10d presents the excavation conditions at ring R1622 on 28 July 2020 at 15:38. Crown pressure was maintained at 0.95 bar, approaching hydrostatic conditions but remaining insufficient to fully prevent sand and water inflow into the cutterhead and steel bell. Excavation was significantly disrupted by a suction system blockage, causing a 26-h downtime due to the accumulation of fiberglass and fine gravel. To mitigate stability risks associated with prolonged face exposure, 1.5 m³ of over-injection and approximately 8 m³ of active mortar injection were applied, although injection efficiency was affected by the malfunction. These measures were essential to control ground loss and maintain stability under adverse operational conditions.

Fig. 10.

TBM advancement from 1619 to 1622 ring.

The excavation sequence from ring R1619 to R1627 demonstrates the strong interdependence between face pressure control, ground response, and stabilization measures under permeable and sensitive soil conditions. In the initial stages, crown pressure was limited to approximately 0.5 bar due to persistent leakage at the steel door, resulting in insufficient confinement and increasing the risk of sand and water inflow into both the cutterhead and the steel bell. This condition, combined with operational disruptions such as cutterhead blockage and suction system failure, led to prolonged face exposure, ground loosening, and elevated instability risks. To compensate for the pressure deficit, extensive stabilization measures were implemented, including over-injection and active mortar injection behind the lining segments. These interventions played a critical role in controlling ground loss and limiting surface settlement, particularly during periods of extended downtime. Subsequent improvements in sealing and structural reinforcement of the steel door enabled a gradual increase in crown pressure toward hydrostatic conditions (approximately 1.0–1.26 bar), resulting in enhanced face stability and reduced leakage effects. Correspondingly, excavation operations became more stable, although the persistent susceptibility of the ground to inflow required continuous application of targeted mortar injection. Despite these improvements, the system remained highly sensitive to localized failures. A critical incident occurred when a new leak in the steel bell caused a sudden pressure drop to 0.2 bar, leading to washout of injected mortar and a rapid loss of confinement. This event underscores the vulnerability of TBM operations to sealing deficiencies and highlights the necessity of continuous monitoring and rapid remedial actions. Overall, the excavation sequence illustrates that maintaining crown pressure close to hydrostatic equilibrium, in combination with systematic grouting and mortar injection, is essential for controlling soil–structure interaction, mitigating ground loss, and ensuring excavation stability in challenging hydrogeological conditions. Figure 11 shows the TBM advancement from 1622 to 1627.

Fig. 11.

TBM advancement from 1622 to 1627.

Analysis of as built steel bell and interpretation

Figure 12 presents the recorded TBM operational parameters during a critical incident on 26 July 2020, highlighting the dynamic interaction between slurry circulation, advance rate, and chamber pressure. The diagram is divided into three panels showing (i) advance speed and slurry pump pressures, (ii) slurry flow rates, and (iii) excavation chamber pressures and bentonite level variations. In the upper panel, the advance speed increased following the completion of the diaphragm wall (D-wall) crossing. At approximately 11:16:40, suction and discharge pressures in the slurry pump system exhibited abrupt fluctuations, coinciding with the triggering of the bypass system (V31 opened, V030 and V032 closed), which halted material extraction. A significant water hammer effect was recorded in the slurry pipeline, with transient pressures reaching 15 bar. The middle panel illustrates slurry flow rates. Initially, inflow and outflow were balanced, but at around 11:16:40 a sudden increase in slurry flow (ΔQ) was observed, indicating a disruption in equilibrium. This surge was followed by gradual stabilization, reflecting a disturbed but recovering flow regime. In the lower panel, crown pressure in excavation chamber 1 was initially maintained at approximately 2.6 bar. However, following the slurry system disruption, a sudden drop in confinement pressure occurred. Simultaneously, a rapid decrease in bentonite level was observed, with a loss exceeding 20 m³ within 10 s, equivalent to approximately 6500 m³/h. This rapid depletion demonstrates the sudden loss of confinement capacity and the associated risk of ground inflow. Overall, the figure captures the sequence of events leading to a transient instability episode: the increase in advance speed, followed by slurry system disruption, water hammer in the pipeline, loss of confinement pressure, and rapid bentonite level reduction. These coupled processes underline the sensitivity of TBM operations to hydraulic transients and the critical importance of maintaining slurry balance for excavation stability.

Fig. 12.

The pressure, flow, and the advanced speed for the TBM.

The key variables in this paper includes: (i) pressure measurements from sensors located within the excavation chamber, (ii) bentonite slurry level in the bubble chamber, (iii) flow rates in both the slurry and feed lines, and (iv) discharge and suction pressures in pump P2.1 associated with the slurry circuit. The results reveal a sequence of rapid and interrelated changes occurring within a time span of less than two minutes. At 11:16:13, the TBM advance rate increased from approximately 7–8 mm/min to 10–15 mm/min, indicating the completion of diaphragm wall (D-wall) crossing. During this phase, the flow balance (Δflow = feed – slurry) remained stable at approximately zero. Between 11:16:13 and 11:16:34, a gradual decrease in bentonite level was observed (from 0.51 to – 0.30 m), while the confinement pressure at the crown remained relatively stable, with fluctuations below 0.05 bar, likely compensated by the activation of Samson valves.

At 11:16:34, an increase in Δflow was recorded, suggesting that the bentonite slurry was no longer retained within the excavation chamber but instead infiltrated the steel bell. A critical event occurred at 11:16:49, where Δflow abruptly increased from 284 to 480 m³ within one second, indicating a sudden and significant leakage of bentonite. This was immediately followed, at 11:16:53, by a sharp drop in confinement pressure (from 1.74 to 1.34 bar within one second), accompanied by a rapid decrease in bentonite volume (approximately 20 m³ lost within 10 s into the steel bell). This event is identified as the primary rupture, likely inducing a water hammer effect within the steel bell due to the sudden acceleration of slurry flow.

Subsequently, between 11:16:53 and 11:17:04, the bentonite level continued to decline significantly (from 1.13 to 2.76 m), while the confinement pressure progressively decreased despite compensatory valve actions, eventually stabilizing at approximately 1.0 bar. At 11:17:10, the system initiated a bypass operation, closing valves V32 and V31 (feed and slurry lines within the excavation chamber) and opening valve V30. This action halted material extraction and prevented further suction of soil and bentonite, thereby mitigating the risk of surface collapse. However, this sudden operational change generated a secondary hydraulic transient, resulting in a water hammer within the slurry line, with pressure peaks reaching up to 15 bar.

As indicated by the detailed monitoring data, the initiation of the main failure can be identified at 11:16:49, when a sudden increase in Δflow was observed, followed in less than one second by a significant drop in confinement pressure. This event was immediately accompanied by a rapid decrease in the bentonite level within the bubble chamber. Within approximately 10 s, nearly 20 m³ of bentonite flowed into the steel bell, corresponding to an estimated flow rate of about 6500 m³/h. The simultaneous occurrence of pressure loss and rapid flow increase suggests that the bentonite slurry entered an existing air gap within the steel bell, where it accelerated and impacted the steel bell door at high pressure (likely exceeding 10 bar). This violent hydraulic interaction is inferred to have generated a shockwave, leading to the rupture of the weakest point in the watertight sealing system of the bell door. Field observations confirmed that the hydro-swelling gasket located at the upper part of the steel bell was expelled during this event, supporting the hypothesis of a high-energy hydraulic impact. Once the watertight integrity of the steel bell door was compromised, control of the bentonite flow and, consequently, the confinement pressure could no longer be maintained until the leakage was contained. Immediately following the failure, TBM operators reacted within less than one minute by closing the suction valves to stop material extraction, thereby limiting further ground loss and mitigating potential surface damage. The potential existence of an օդ gap in the upper part of the steel bell, allowing free movement of bentonite, was further investigated. Unlike previous breakout (BO) steel bells, this bell had been installed on 08/06/2020, approximately 1.5 months prior to the TBM arrival. During this period, bleeding of the sand–cement backfill inevitably occurred, leading to the formation of a void at atmospheric pressure in the upper section of the steel bell. Although the bleeding percentage of sand–cement mixtures is not typically measured for such temporary works, reference values for mortar indicate a maximum bleeding of approximately 2%. Assuming a similar (and conservative) value, this corresponds to an estimated void volume of approximately 18 m³. This order of magnitude is consistent with the observed sudden loss of bentonite during diaphragm wall (D-wall) piercing, thereby supporting the proposed mechanism of failure involving rapid slurry inflow into a pre-existing օդ gap, subsequent hydraulic impact, and loss of confinement control.

Interferometric Synthetic Aperture Radar (InSAR) data were employed to monitor ground settlement and subsidence associated with structural distress in the study area. Annual maximum settlement values were extracted to examine the temporal evolution of ground deformation. The findings indicate a gradual increase in settlement between 2017 and 2019; however, a marked anomaly was observed on 26 July 2020, when subsidence values abruptly increased from approximately 10 mm to nearly 50 mm. This abrupt escalation signifies a critical failure event and raised substantial concerns regarding the structural stability of nearby buildings, particularly due to the onset of excessive differential settlement. Figure 13 shows the recorded data by InSAR.

Fig. 13.

InSAR data recording before the supporting.

The mathematical framework is:

To quantitatively interpret the observed deformation during the TBM break-out event, a physically consistent relationship is established linking the sudden pressure drop, rapid slurry loss, and induced ground settlement.

During the incident, field data indicated a sharp reduction in excavation chamber pressure () accompanied by an abrupt increase in slurry inflow into the steel bell. The instantaneous slurry discharge rate can be approximated as:

This rapid discharge is attributed to a sudden pressure imbalance between the excavation chamber and the steel bell, which can be conceptually described using a simplified orifice-type flow relationship:

where is the discharge coefficient, is the effective leakage area (e.g., failure zone at the gasket), and is the slurry density. This formulation highlights that the slurry loss rate is directly governed by the magnitude of the pressure drop.

The cumulative slurry loss volume is assumed to translate into an equivalent ground volume loss , particularly under undrained or short-term conditions:

This volume loss is then directly linked to surface settlement through a volume loss-based approach. For localized failure conditions such as the present case, the maximum settlement can be approximated as:

where represents the effective area over which the volume loss is distributed. Alternatively, for distributed settlement profiles, the Gaussian formulation can still be used in a modified form:

Importantly, in this case, the settlement is not purely due to conventional tunneling-induced volume loss but is dominated by a sudden loss of confinement, leading to a rapid ground response.

Accordingly, the complete quantitative relationship can be expressed as:

This chain demonstrates that the abrupt pressure drop triggered a high-rate slurry loss event, which induced significant ground volume loss and consequently led to the observed settlement and structural response. The proposed formulation provides a direct mathematical and physical interpretation of the failure mechanism observed during the TBM break-out process.

It should be noted that a direct one-to-one agreement between field measurements and the proposed mathematical framework is not expected in this case. This is primarily due to the non-standard nature of the deformation mechanism, which differs fundamentally from conventional tunnelling-induced settlement.

The proposed formulation is based on a simplified physical interpretation that links pressure drop, slurry loss, and ground settlement through the chain:

However, the field measurements reflect a highly transient and complex failure process, governed by sudden loss of confinement rather than controlled excavation.

The resulted damage

According to the previous, the construction of underground railway stations, and particularly the execution of the station break-out phase, presents substantial risks to the stability of adjacent buildings and infrastructure. This phase—during which the tunnel boring machine (TBM) transitions from the station box to the running tunnel, or conversely—constitutes a critical stage of construction. Excavation-induced stress redistribution, potential face instability, and insufficient temporary support systems can trigger significant ground deformations. Such disturbances may manifest as vertical settlements, differential displacements, or complex soil–structure interaction phenomena, thereby jeopardizing the structural performance of surrounding assets. Documented case studies and project reports highlight the occurrence of visible damage in buildings situated in proximity to station excavations, including cracking, tilting, and, in extreme cases, partial structural failure. These outcomes not only endanger public safety but also contribute to project delays, cost overruns, contractual disputes, and diminished public trust in large-scale urban infrastructure initiatives. Although recent advances in numerical modeling, geotechnical monitoring, and ground improvement technologies have improved predictive and mitigation capabilities, the break-out phase continues to present uncertainties. Unanticipated ground responses underscore the urgent need for a more comprehensive understanding of the mechanisms governing building distress under these conditions. This includes the development of robust predictive frameworks, enhanced monitoring protocols, and more effective protection strategies for vulnerable structures within densely built urban environments. Figure 14 shows the cracks in the surrounding buildings. While, Fig. 15 shows the monitored points. Furthermore, the observed damage and cracking in the building are attributed to differential settlement. Based on InSAR measurements, the edge of the building closest to the tunnel experienced a maximum settlement of approximately 50 mm, while the edge farther from the tunnel exhibited a settlement of about 5 mm. This results in a differential settlement of approximately 45 mm, which is considered the primary cause of the observed structural damage and cracking.

Fig. 14.

The cracks in the surrounding building walls.

Fig. 15.

The monitored points on the buildings.

Results

The results will be presented in the subsequent sections. They will include an analysis of the failure of two major surrounding buildings, along with a detailed examination of both settlement and horizontal displacement. These results were obtained following the implementation of reinforcement and grouting measures in the surrounding soil. Additional monitoring points were installed at the previously identified locations to assess the effectiveness of the support interventions implemented after the failure. Regarding the monitoring methodology, fixed surface point markers have been installed on the building at designated locations. These markers serve as reference benchmarks for settlement and movement measurements. Monitoring readings are obtained at regular, predefined intervals using calibrated surveying equipment, in accordance with the project monitoring plan, ensuring accuracy, consistency, and reliability of the collected data.

The movement of building Z-24

Settlement

The settlement monitoring results at section Z-24 illustrate the ground surface response during TBM excavation over the periods from 27 to 31 July 2020 and from 31 July to 3 August 2020. Multiple monitoring points (P01, P02, P03, P04, P05, and P10) record vertical displacements above the tunnel alignment, where negative values indicate settlement and positive values indicate heave. The excavation sequence from rings 1621 to 1626 provides a temporal reference linking TBM advancement to ground behavior. During the first period, settlement values range approximately between − 12 mm and − 2 mm. An increase in settlement is observed during the excavation of rings 1622–1624, corresponding to a temporary loss of pressure in the excavation chamber. After ring 1626, the settlement trends stabilize, indicating that ground movements were contained without progressive development. In the second period, settlement values remain within approximately − 10 mm to + 2 mm. The monitoring curves show generally stable behavior with minor fluctuations, confirming that ground deformation remained controlled over time. No progressive settlement is observed. A limited number of anomalous readings are present in the dataset, including an isolated extreme value on 01 August 2020. These outliers are attributed to external interferences such as sun reflection or obstruction of the monitoring line of sight by construction equipment. As they are isolated, they do not affect the overall settlement trend. Overall, the results indicate that settlement remained minor and stable, with a short-term increase associated with chamber pressure loss, while the excavation process was effectively controlled and did not lead to progressive ground deformation. Figure 16 shows the settlement for z-24 building.

Fig. 16.

The settlement monitoring points in building z-24.

Horizontal movement

The two figures present the horizontal ground displacements in the normal direction at section Z-24, recorded between 27 July 2020 and 05 August 2020 at four monitoring points (P01-N, P02-N, P03-N, and P04-N). Displacements are reported in millimeters versus time. uring the period from 27 to 30 July 2020, horizontal displacements remain small, generally ranging between –6 mm and + 2 mm. All monitoring points show similar trends, with minor fluctuations during excavation that remain within a narrow range, indicating controlled lateral movements. rom 31 July to 05 August 2020, the displacement behavior remains stable, confined between –5 mm and + 2 mm. The curves overlap closely across all monitoring points, showing uniform soil response without progressive deformation. verall, horizontal movements remain minimal (within ± 6 mm), consistent across monitoring points, and without cumulative growth. These results indicate effective control of excavation-induced lateral displacements and stable TBM operation. Figures 17 presents the horizontal movements over the periods 28/07–31/07 and 31/07–02/08, respectively.

Fig. 17.

The horizontal movement.in building Z-24 from 28/07 to 31/07.

The movement of Bahrain embassy

The figures illustrate the recorded settlement behavior at the Bahrain Embassy during TBM excavation between 27 July and early August 2020, with settlement values (mm) plotted against time and multiple monitoring points represented by different curves. Throughout the monitoring period, settlement remained small and stable, generally ranging between –5 mm and –1 mm, indicating well-controlled excavation conditions and uniform ground response. During the initial phase (27–31 July), deformations were minor and consistent, with a single isolated positive reading (~ + 2 mm) likely caused by measurement error or temporary interference. In the subsequent phase (31 July–early August), similar stability was observed, although two additional upward spikes (+ 3 to + 4 mm) appeared; these are attributed to external factors such as sunlight reflection or obstruction of the optical monitoring system rather than actual soil movement. Excluding these anomalies, no evidence of progressive or cumulative settlement was detected, and all monitoring points exhibited consistent behavior. Overall, the low settlement magnitudes (generally < 5 mm) confirm that TBM operations and ground support measures were effective, with no significant impact on the Bahrain Embassy or surrounding structures. These results highlight the importance of continuous monitoring and careful interpretation of data to distinguish between true ground response and measurement artifacts, thereby supporting reliable engineering assessment and risk management. Figures 18 presents the settlement trends for the periods 29/07–31/07 and 31/07–08/08, respectively.

Fig. 18.

The settlement monitoring points.

Building repairing

Figures 19a–f collectively illustrate the staged construction sequence of the micropile-supported retaining system, demonstrating the integration of temporary stabilization measures with permanent structural elements. In Figure (a), the first row of micropiles is installed vertically through the fill and silt layers into the underlying clay stratum, forming a deep foundation system capable of transferring loads to competent soil layers. In Figure (b), inclined steel bracing is installed prior to excavation to provide temporary lateral support, effectively controlling ground deformation and protecting adjacent structures, including the cracked brick wall. Figure (c) shows the construction of the first reinforced concrete retaining wall, where starter bars are embedded to ensure structural continuity, while the temporary bracing remains in place to maintain stability during casting. In Figure (d), a second row of micropiles is installed to enhance the overall stiffness and load-bearing capacity of the system, further improving ground support conditions. Figure (e) illustrates the casting of the reinforced concrete slab, which integrates the micropiles into a unified structural system; at this stage, temporary bracing and steel elements are removed once the slab has attained sufficient strength to assume load-bearing functions. Finally, Figure (f) presents the construction of the second reinforced concrete wall, forming a rigid diaphragm that significantly increases the system’s resistance to lateral soil pressures and ensures efficient load transfer. This sequential methodology enables a controlled transition from temporary support systems to a fully integrated permanent structure, ensuring both construction safety and long-term stability.

Fig. 19.

The process for repairing the building.

Conclusions

This study has demonstrated that the station break-out stage in urban tunneling projects represents one of the most critical phases in terms of ground stability and adjacent building safety. Through the analysis of real monitoring data from the Zamalek Station project, it was shown that sudden loss of confinement pressure, leakage through the steel bell system, and subsequent hydraulic shocks were the primary triggers of ground deformation and localized building distress. The settlement and displacement measurements around sensitive structures, such as Building Z-24 and the Bahrain Embassy, confirmed that while ground movements were generally limited to within acceptable thresholds (–12 to + 2 mm), transient events caused by operational disturbances directly influenced settlement magnitudes and patterns.

The findings underscore the strong correlation between excavation chamber pressure stability and ground response, highlighting the necessity of rigorous face support management and continuous monitoring. The results also confirmed that anomalies in monitoring datasets—such as outlier settlement spikes—must be critically assessed to distinguish genuine soil behavior from measurement artifacts. Moreover, the underpinning and micropile reinforcement schemes implemented around adjacent buildings proved effective in mitigating risks, ensuring stability, and facilitating safe continuation of excavation works in a highly sensitive urban context.

From a broader perspective, this research contributes to advancing risk-mitigation strategies for tunneling projects in densely built areas. It demonstrates that adaptive operational measures—such as systematic mortar injection, targeted grouting, and steel door reinforcement—are essential in responding to unpredictable ground conditions and equipment malfunctions. Future underground construction projects should integrate real-time monitoring systems, probabilistic prediction models, and deep-learning frameworks to better anticipate soil–structure interaction during critical excavation stages. By adopting such integrated approaches, urban tunneling can achieve greater resilience, safeguarding both infrastructure development and the built heritage of historic city centers.

Future research will aim to extend the present findings toward practical tunneling applications, including the development of enhanced risk management frameworks, real-time monitoring approaches, and predictive models for assessing ground and structural responses in complex urban environments.

Beyond the specific case investigated, the present study provides broader implications for modern tunnelling monitoring and control strategies. The results demonstrate that critical deformation events are often governed not by gradual excavation processes, but by transient operational instabilities, particularly sudden loss of confinement pressure. Therefore, effective risk management in urban tunnelling requires a shift from conventional settlement-based assessment toward integrated monitoring frameworks that combine TBM operational data with ground response measurements in real time. The proposed interpretation framework highlights the importance of detecting early indicators such as pressure fluctuations and flow imbalance, which can serve as precursors to failure. In this regard, the study contributes to the advancement of next-generation tunnelling monitoring systems, supporting the development of data-driven early warning tools and enhancing the resilience and safety of underground construction in complex urban settings.

Author contributions

The whole authors (Zaid S Alajlan, Ahmad Abd-alma geed, Lojain Suliman, Mohamed Rabie, Hussein mahmoud,) contributed equally.

Funding

This study is supported via funding from Prince Sattam bin Abdulaziz University project number (PSAU/2026/R/1447).

Data availability

All data generated or analysed during this study are included in this published article.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

All borehole drilling and soil sampling conducted in 2017 were performed with full authorization from the relevant project authorities and the local municipal/regulatory bodies responsible for supervising geotechnical investigations in the study region. The fieldwork complied with all regulatory, safety, and technical requirements applicable at the time.