Abstract
Chimney failure (crown displacement) in tunnel construction presents a critical concern of tunnelling projects in weaker rock formations at shallow depth of cover. Understanding the mechanisms and implications of chimney failure is of paramount importance for ensuring safe and efficient tunnel excavation, preventing potential hazards, and implementing effective remedial measures to maintain tunnel stability and structural integrity. An attempt has been made in this paper to simulate chimney failure in a tunnel and measured the vertical displacements and stresses in both unsupported and supported tunnels with cavities and compared with the field observations. This paper also discusses the importance of appropriate support strategies and remediation techniques to ensure the stability and structural integrity of tunnels while addressing challenges associated with crown displacement and cavity formation.
1 Introduction
Tunnels have been excavated for a wide range of purposes, serving the needs of road construction and water transportation for mining and civil construction projects. A conventional and time-tested technique, drilling and blasting, continues to excavate tunnels of various shapes and sizes. However, the Himalayas stand out as one of the most unstable regions on the planet, presenting immense challenges for underground excavations. This region’s fragile geology, frequent tectonic activities and intricate geological structures make tunnelling endeavours exceptionally demanding. When undertaking excavation projects in the Himalayas, numerous hurdles are encountered, including the risk of rock bursts, difficult squeezing ground conditions, potential cavity failures, and water ingress. The phenomenon known as cavity failure, or chimney failure (Figure 1), emerges as a particularly significant issue during tunnel construction among these challenges, warrants an extensive research and investigation. There are a number of reasons leading towards cavity formation during tunnel construction like groundwater inflow (Figure 2), weak or unstable ground conditions (Figure 2), inadequate support system, construction-induced vibrations and geological features.
FIGURE 1
FIGURE 2
Schmidt et al. (2010) identified two distinct types of stability problems during tunnel construction: failures that affect already supported sections of the tunnel, and failures that occur at the tunnel face, where excavation and support installation are actively in progress. Occurrence of cavity failure behind the lining is frequent cause of instability during tunnel drivage (Liu et al., 2024). Cavity failure has become an unavoidable and serious problem during tunnel excavation in special geological conditions (Huang et al., 2017; Liu et al., 2022; Nguyen et al., 2017; Liu et al., 2013). The nature of rockmasses modify due to submerging in groundwater causing expansion, softening, muddy and changes in their geo-mechanical properties (Chen et al., 2022). Inappropriate support system under such conditions lead to tunnel cavity failure and collapse. Chen et al. (2022) combined in situ monitoring with numerical simulations to investigate cavity failure mechanisms in a tunnel excavated through phyllitic rock. Their study showed that tunnel excavation disturbed the groundwater regime, leading to continuous seepage toward the cavity, which in turn caused phyllite dissolution and progressive weakening of the mechanical properties of the surrounding rock mass, grouting material, and steel arch support. It was found that the stresses around the surrounding rock exceeded the tensile strength to cause cavity failure. Further, they proposed grouting by replacing steel arch as remedial measures for prevention of treatment of tunnel collapse in phyllite strata. It enriched the strength of the rockmass and acted like a water barrier. It reduced the rock pressure near the collapse by 60.8% replacement.
Groundwater inflow into the tunnel void is one of the most common reasons for cavity formation. As the tunnel is excavated, it can intersect or disturb underground water sources, increasing water pressure in the surrounding soil or rock. If the pressure becomes excessive, it can force water to flow into the tunnel, causing erosion and cavity formation. Cavity formation becomes more likely if the ground surrounding the tunnel has inherently weak or unstable soil or rock formations. Excavating in such conditions can cause the surrounding ground to collapse and create voids, particularly for loose soils, fractured rock masses, or areas with high clay content susceptible to subsidence and ground instability.
Inadequate or ineffective support systems within the tunnel can contribute to cavity formation. When the support system, such as shotcrete, rock bolts, steel ribs, or tunnel linings, is not designed or installed correctly, it may fail to provide sufficient reinforcement, leading to the surrounding ground’s collapse and the collapse of cavities’ formation. Vibrations generated during tunnel construction, mainly from blasting or heavy machinery, can cause ground disturbances and contribute to cavity formation. The vibrations can loosen the surrounding soil or rock, making it more susceptible to collapse and developing voids.
Geological features such as underground streams, aquifers, or water-filled voids and sheared zone can increase the likelihood of cavity formation. If the tunnel intersects or disturbs these features, water can enter the tunnel void, causing erosion and cavities. In areas with significant tectonic activity or geological faults, natural fractures or fault zones can exist in the ground. These fractures can provide pathways for water to infiltrate and erode the surrounding soil or rock, leading to cavity formation during tunnel excavation. It is essential to recognize that these reasons can often overlap and interact. For example, groundwater inflow into the weak or unstable ground can exacerbate cavity formation, while insufficient support systems can further contribute to ground collapse and cavity development.
Rajput et al. (2023) suggested some of the remedial measures adopted by the tunnel excavating firm. These include installation of pipe-roofing of an appropriate length and diameter in crown area of tunnel which forms an umbrella roof to prevent the flow of muck. Grout material is injected through the pipe to consolidate the loose muck. After setting of injected grout material, fallen mucks from the roof are removed to create space for installation of steel ribs at an appropriate spacing along with application of lagging and bulkhead in it. Ribs are backfilled with shotcrete material and these are repeated up to the tunnel face. In case of roof cavities extending for longer distance in the vertical direction, proper lapping of the installed pipes is maintained. Application of forepoles and shotcrete are subjected to the behaviour of the loose muck in each cycle of removal of fallen roof muck for creation of space.
Therefore, it is required to develop comprehensive understanding of the geological conditions and appropriate engineering measures, essential to minimize the risk of cavity formation in tunnel construction. Numerical simulation study on calibrated models will help in estimation of vertical displacements and vertical stresses after exposure of freshly excavated unsupported roof. Further, study of displacements and stresses after application of different supports on the numerical models will help in testing the efficacy of the support system. Accordingly, it will help to identify and implement the additional support measures and remediation techniques to ensure the stability and structural integrity of tunnels while effectively addressing the challenges posed by cavity formation.
An attempt has been made in this paper to understand the mechanism of cavity formation in numerical models and proposed a support design as remedial measures based on analysis of vertical displacements and stresses to deal with such incidences.
2 Literature review
Tunnelling in the Himalayan region is often complicated by the presence of weak geological formations, high overburden stresses, and active tectonics. One of the major geotechnical challenges during tunnel excavation in this region is the formation of cavities in the crown area, often due to undetected weak zones, weathered rock mass, and sheared strata. These cavity formations pose significant risks to tunnel stability and construction timelines. Accurate and reliable predictions of rock deformation are essential in both civil and mining engineering projects involving rock masses (Abbas et al., 2024). Ensuring the precision of these forecasts is critical for optimizing structural design, enhancing safety, and enabling efficient planning and execution of excavation activities (Koopialipoor et al., 2022). In the beginning, tunnel stability was assessed using empirical approaches (Yang et al., 2021) and rock mass classification systems (Bieniawski, 1993). Nowadays, the deformation of the rock mass around the tunnel crown and walls are being analysed through numerical modelling techniques (Leu et al., 2001; Schmidt et al., 2010; Kumar, 2023).
Ground reaction curve (Figure 3) gives an idea of the support load capacity to be designed along with the permissible level of roof deformation (Rabcewicz, 1973). Schmidt et al. (2010) studied the factors influencing tunnel safety and behaviour of micropile umbrella and carried out 3D numerical simulation and analysis to investigate the effects of a micropile umbrella on the stability of tunnel of Burgos (Spain) constructed in miocenic sand clay. Schmidt et al. (2010) presented a 3D numerical analysis to investigate the effect of a micropile umbrella on the stability of the tunnel face. It was found that the micropiles should be installed as parallel as possible to the tunnel axis, minimising the pile angle. The length of the pile should not be too long in order to avoid high bending moments, which can exceed material resistance and produce failure. The separation of the piles should not permit soil collapse between the pile tubes. As well as the overlapping of the single umbrellas with each other in order to secure a forerunning support: it should be chosen in a matter to grant a mutual support between the umbrella segments. Beg et al. (2025) highlighted that Himalayan tunnel projects face severe stability issues due to complex geology such as thrust and shear zones, folding, and intense in situ stresses along with challenging climatic and hydrological conditions. These factors drive common problems like rock bursts in deep overburden, squeezing in weak schistose mass, and significant water inflows, all of which escalate project cost and timeline. It emphasized that successful tunnelling in this region requires a holistic approach combining detailed geological investigations, adaptive construction techniques, and engineered support systems.
FIGURE 3
Under most conditions, tunnelling causes a transfer of the ground load by arching to sides of the opening, termed the ground-arch effect. The horseshoe shape is generally preferred for all but the weakest rocks, since the flat bottom facilitates hauling. By contrast, the stronger and more structurally efficient circular shape is generally required to support the greater loads from soft ground. Ability of roof to remain standing after excavation in underground mine is necessary to install a temporary or permanent support and remove the material. It is affected by the nature of rock mass and its parameters and also excavation technique. A number of scientists (Lauffer, 1958; Bieniawski, 1993; Barton et al., 1975) have worked over the stand-up time of roof and proposed the width of unsupported excavation during construction of tunnel. Chimney failure has been found to be mainly occurring due to tunnelling operations in weak ground at shallow depth of cover. It is very difficult to predict such circumstances unless a comprehensive mapping of the site is carried out in advance. It is not related to stand-up time of roof (Lauffer, 1958) (Figure 4). Ground reaction provides an idea of the supporting of the roof.
FIGURE 4
Upadhyay et al. (2023) reported the formation of a cavity between RD 312–320 m during the drivage of Adit-2 at the Rammam III Hydroelectric Project in the Sikkim Himalayas, attributed to a weathered, weak, or sheared zone located above the crown level, which remained undetected during excavation. By systematically applying mitigation measures guided by the DRESS (Drainage–Reinforcement–Excavation–Support–Solution) philosophy, the project team successfully navigated through the problematic stretch and reached the rock face at RD 320 m. The adopted approach effectively mitigated loose rock falls and cavity formation triggered by high pore water pressure and a sheared rock mass. Stabilization in the cavity-prone section was achieved using drainage holes, steel ribs, shotcrete, and forepoling. According to Diwakar et al. (2022), the presence of complex geological features such as faults, folds, and shear zones makes the Himalayas one of the most challenging regions in the world for underground excavation projects. Ahmed et al. (2025) studied influencing factors affecting ground deformation during construction of Khari-Banihal Railway Tunnel (KBRT) in the geologically disturbed Western Himalayas and developed a predictive model using real-time 3D monitoring data and multiple linear regression (MLR) techniques. Li et al. (2020) conducted numerical simulations to assess the performance of tunnel linings in the presence of surrounding rock cavities. Their findings revealed that such cavities significantly impact tunnel stability and can lead to structural damage, with cavity width exerting a more pronounced effect on the tunnel’s stress state than cavity depth.
Spyridis and Bergmeister (2024) studied the environmental impacts of tunnel construction which can be significantly reduced by optimizing the design and materials of the primary and secondary linings. Shen et al. (2014) discussed the 2008 Wenchuan earthquake in China which destroyed a number of tunnels to different degrees of the vulnerability, which stopped the emergency lifeline supply from the safety places. Huang et al. (2022) carried out the resilience analysis modelling of the underground infrastructure both at structural and system level due to man-made or natural hazards arising from climate change.
Cavity formation at the crown during Himalayan tunnel drivage is a multifaceted problem stemming from geological complexity, hydrological pressure, stress redistribution, and dynamic loading. A robust engineering response integrates detection, adaptive support, and excavation sequencing tailored to each tunnel. This integrative framework with proven success in real-world Himalayan projects, provides a blueprint for future works in similar settings.
3 Field study
The work of four lane road from the end of Pandoh bypass to the Takoli section of National Highway-21 from 221.305 km to 242.000 km in the State of Himachal Pradesh, started in 2017 (Figure 5). Around 21.4 km of underground highway tunnels (4 twin tunnel, 2 single tube and cross-passage between LHS and RHS Tunnel), 3 major bridges, 10 minor bridges, 1 elevated bridge, 48 culverts, and 13 km of highways (including 290 m cut and cover section) has been constructed in the most challenging Pandoh-Takoli section of NH-21 (Figure 5). The tunnel is horseshoe shaped having width of 13.20 m, height of 7.90 m and length of 2053 m (Figure 6). The distance between the two twin tube tunnels is 34 m (Figure 7). The length of the RHS and LHS tube is 2,144.452 m and 2,142.663 m respectively. Peak overburden is around 250 m (RHS) and there are no peak overburden data available for LHS due to missing topographical surveys. The estimated peak overburden varying in the range of 300–350 m. The tunnels are being constructed in an alteration of meta-sediments belonging to the Haimanta Formation (Precambrian to Cambrian age). The average thickness of single layers is in the range of centimetres to a few meters. The discontinuities of the rock mass are mainly widely spaced (range of spacing: close to very wide). Details of wedge-cut drilling and blasting pattern adopted for tunnel excavation is shown in Figure 8. Rock support system for different nature of rock during tunnel excavation is mentioned in Table 1. The selected tunnel section represents a typical Himalayan tunnelling scenario characterised by shallow overburden, weak and sheared zones above the crown, and drill-and-blast excavation. These site-specific conditions directly contributed to cavity formation, making this project a representative case for the analysis of chimney failures.
FIGURE 5
FIGURE 6
FIGURE 7
FIGURE 8
TABLE 1
| Rock class | Method of excavation | Blasting round length (m) | Shotcrete thickness (mm) | SN rock bolt length (m) | Drainage holes length and numbers | Spacing of lattice girder/ribs (m) |
|---|---|---|---|---|---|---|
| Fair | Full face | 3–5 | SFRS 100 | 3 (11 nos.) | 20 m (6 nos.) (5 m overlapping) | - |
| Poor | Full face | 2.5–3 | SFRS 150 | 2.5 (12 nos.) | 20 m (6 nos.) (5 m overlapping) | - |
| Very poor | Full face | 1.5–2 | SFRS 200 | SN: 2 m (14 nos.) SDA: 6 m (8 nos.) Forepole: 32 mm dia bar 9 m- 12 months | 20 m (6 nos.) (5 m overlapping) | 1.5 (lattice girder) |
| Extremely poor | Full face | 1.25–1.5 | SFRS 250 | SN: 2 m (14 nos.) SDA: 6 m (8 nos.) Forepole: 32 mm dia bar 9 m- 12 months | 20 m (6 nos.) (5 m overlapping) | 0.75 (rib with RCC lagging and backfilling) |
Rock support system for different nature of rock during tunnel excavation.
4 Numerical simulation
Numerical modelling is an important and popular tool to study the geotechnical problems. A model is created for developing a tunnel length of 50 m (Figure 9a) and the geological profile of the GT 3 and GT13 rock characteristics (Tables 2, 3) has been modelled (Figure 9b). The overburden above the tunnel is assumed to be constant. A draw of 1.25 m is considered by rock class of GT 3 and GT13. Drilling and blasting achieved a pull of 2.5 m in each round; therefore, tunnel is developed in the model at a distance of 2.5 m sequentially. In-situ stresses are initialised in the model as per the measurements of the site available. The primary aim of this study is to measure displacement on the crown, stresses, share failure around the tunnel. Observations were made along the entire length of the tunnel without support. After observing the condition of sequentially developed tunnel, rock bolts (Table 4), steel ribs (Table 5) and a shotcrete liner (Table 6) were installed. After installation of supports, vertical stresses and displacement were observed on the crown around the entire length of the tunnel.
FIGURE 9
TABLE 2
| Key parameters | Ground type | |||
|---|---|---|---|---|
| GT 1 | GT 2 | GT 3 | GT 13 | |
| Rock type | Bedrock | Bedrock | Bedrock | Fault rock |
| UCS - intact rock (MPa) | 30–60 | 30–60 | 30–60 | Variable |
| Discontinuity spacing (cm) | >60 | <20 | <20 | NA |
| Lithology | Alternation of slate to phyllites and metasilt to metasandstone; quartzite; locally thin interlayers of shale | Alternation of slate to phyllites and metasilt to metasandstone; quartzite; locally thin interlayers of shale | Alternation of slate to phyllites and metasilt to metasandstone; quartzite; locally thin interlayers of shale | Fault rock originating from various rock types |
| Description | Moderately to poor interlocking rockmass, medium to high strength, closely jointed; spacing: >60 cm; large prismatic to tabular rock blocks (>60 cm) | Moderately to poor interlocking rockmass, medium to high strength, closely jointed; spacing: 20–60 cm; small tabular rock blocks (20–60 cm) | Moderately to poor interlocking rockmass, medium to high strength, closely jointed; spacing: >60 cm; large prismatic to tabular rock blocks (>60 cm) | Heterogeneous fault rock, very closely to extremely fractured and/or sheared rock, without any or poor interlocking of blocks. Size of rock fragments from cobbles to gravel; occasionally with sandy to clayey matrix |
| Additional parameters — intact rock | ||||
| Specific gravity (kN/m3) | 27 | 27 | 27 | - |
| Young’s modulus (GPa) | 20–30 | 20–30 | 20–30 | - |
| Coefficient of elasticity | 0.25 | 0.25 | 0.25 | - |
| Indirect tensile strength (by brazilian test) (MPa) | 3–18 | 3–18 | 3–18 | - |
| Mi (Hoek-Brown constant) | 5–10 | 5–10 | 5–10 | - |
| Abrasivity (CAI) | 1–4 | 1–4 | 1–4 | - |
| Additional parameters — joints, cleavage | ||||
| Friction angle | 27–37 | 27–37 | 25–35 | Variable |
| Residual friction angle | 23–35 | 23–35 | 20–30 | Generally low |
| Roughness | Smooth-rough, planar | Smooth-rough, planar | Smooth-rough, planar | Generally smooth |
| Additional parameters — rockmass | ||||
| RQD | >70 | >30 | <40 | NA |
| GSI | 55–70 | 45–55 | 35–50 | 26 |
| UCS (rockmass) | 7.5–11 | 6–7.5 | 4.5–6 | 0.8–1.5 |
| Cohesion | 2–3 | 1.7–2.3 | 1.5–2 | 18–23 |
| Angle of internal friction (0) | 29–33 | 26–29 | 23–26 | 0.7–2.5 |
| Young’s modulus (GPa) | 9–16 | 5–9 | 2.5–5 | 0.3 |
Rockmass properties for GT3 and GT13 rock types.
TABLE 3
| Basic categories of behaviour types | Description of potential failure modes/mechanisms during excavation of the unsupported ground | |
|---|---|---|
| GT 1 | Stable | Stable ground with the potential of small local gravity induced falling or sliding of blocks |
| GT 2 | Potential of discontinuity-controlled block all | Voluminous discontinuity controlled, gravity induced falling and sliding of blocks, occasional local shear failure on discontinuities |
| GT 3 | Shallow failure | Shallow stress induced failure in combination with discontinuity and gravity-controlled failure |
| GT 4 | Voluminous stress induced failure | Stress induced failure involving large ground volumes and large deformation |
| GT 5 | Rock burst | Sudden and violent failure of the rock mass, caused by highly stressed brittle rocks and the rapid release of accumulated strain energy |
| GT 6 | Buckling | Buckling of rocks with a narrowly spaced discontinuity set, frequently associated with shear failure |
| GT 7 | Crown failure | Voluminous overbreak in the crown with progressive shear failure |
| GT 8 | Raveling ground | Flow of dry or moist, intensely fractured, poorly interlocked rocks or soil with low cohesion |
| GT 9 | Flowing ground | Flow of intensely fractured, poorly interlocked rocks or soil with high water content |
| GT 10 | Swelling ground | Time dependent volume increase of the ground caused by physical chemical reaction of rock and water in combination with stress relief, leading to inward movement of the tunnel perimeter |
| GT 11 | Ground with frequently changing deformation characteristics | Combination of several behaviours with strong local variations of stresses and deformations over longer sections due to heterogeneous ground (i.e., in heterogeneous fault zones; block-in matrix rock, tectonic mélanges) |
General categories of different ground behaviour types.
TABLE 4
| Yielding load (kN) | Length (m) | Diameter (mm) | Area (m2) | Young’s modulus (kN/m2) | Axial stiffness (kN/m) | Spacing (m) | Bond strength (MN/m) | Perimeter bolt length (m) |
|---|---|---|---|---|---|---|---|---|
| 250 | 6 | 32 | 4.7 x 10−4 | 2.1 x 108 | 9.87 x 104 | 1.25 (roof); 2.5 (perimeter) | 0.1 | 6 |
Roof bolts details used as the support system.
TABLE 5
| Steel ribs spacing (m) | Section depth (m) | Area (m2) | Moment of inertia (kg-m2) | Young’s modulus (MPa) | Poisson’s ratio | Compressive strength (MPa) | Tensile strength (MPa) |
|---|---|---|---|---|---|---|---|
| 1 | 0.136 | 1.02 x 10−3 | 3.56 x 10−6 | 2.1 x 105 | 0.30 | 400 | 250 |
Steel ribs details used as the support system.
TABLE 6
| Thickness (cm) | Young’s modulus (GPa) | Compressive strength (MPa) | Tensile strength (MPa) | Strength (MPa) with time |
|---|---|---|---|---|
| 10 | 10–50 | 30–70 | 3–7 | 0.1–0.5 (6 min); 2-20 (24 h) 30-35 (7 days); 30-70 (28 days) |
Shotcrete properties as the support system.
The steps of modelling consisted of.
Generation of radial cylindrical brick of size with 84 m and 50 m along x (−42 to 42) and y (0–50)- axes with 50 m roof and 20 m floor.
Generation of boundary of the semi-circular tunnel (radius = 7.0) to be excavated.
Group the developed geometry into various zones, i.e., heading 1, heading 2, heading 3, benching 1, benching 2, benching 3, benching 4, benching 5, benching 6, floor 1 and floor 2.
Assigning Mohr-Coulomb constitutive material model properties to these zones. Applying the rock properties GT3 (bulk modulus- 2.083 GPa, shear modulus- 960 MPa, density- 2,750 kg/m3, cohesion- 0.224 MPa, angle of friction- 42°, tension- 0.6 MPa). Applying the rock properties GT13 to a weaker chimney zone (bulk modulus- 0.750 GPa, shear modulus- 346 MPa, density- 2,750 kg/m3, cohesion- 0.90 MPa, angle of friction- 23°, Tension- 0.6 MPa).
Applying boundary conditions (left, right, front, back, and bottom faces).
Initializing the vertical and horizontal stresses with required gradient.
Applying the gravity and solving the system for attaining the equilibrium state.
Initializing the velocity and displacement as zero for recording the deformation after excavation.
Development of 2.5 m length of tunnel (considering 2.5 m pull in each blast cycle) (Figure 10).
Recording the vertical displacement of the points in the crown of the excavation (Figure 11a).
The above two steps are repeated till the complete excavation has taken place.
Application of the rock bolt support (1.5 m × 1.5 m spacing) in staggered pattern and the shotcrete liner (2.5 m in one go) (Figure 12).
Recording the vertical displacement of the points in the crown of the excavation (Figure 11b).
The above two steps are repeated till the complete excavation has taken place and vertical stresses and displacement are observed at each sequential development of 2.5 m (Figure 13).
Continuous crown displacement causing chimney formation during tunnel sequential excavation in the numerical model is shown in Figure 14.
FIGURE 10
FIGURE 11
FIGURE 12
FIGURE 13
FIGURE 14
In the numerical simulations, the support system was represented using built-in structural elements available in the simulation software package, namely, roof bolts (cable elements), steel ribs (beam elements), and shotcrete liners (liner/shell elements). These support elements are not governed by rock mass constitutive models such as Mohr–Coulomb or Hoek–Brown. Instead, they are simulated using dedicated structural constitutive laws specifically developed to represent the mechanical behaviour of support systems. Roof bolts were modelled using cable elements, which follow a one-dimensional axial elastic–plastic constitutive law combined with a bond–slip interaction at the grout–rock interface. The steel bar behaves linearly elastic up to the yield force and exhibits perfectly plastic behaviour beyond yielding. Load transfer between the bolt and surrounding rock is represented by shear coupling springs, which follow a linear elastic bond law and fail once the bond strength is exceeded, allowing for debonding and pull-out behaviour. Steel ribs were simulated using beam elements governed by elastic or elastic–perfectly plastic beam constitutive laws. In conditions where the ribs are sufficiently stiff and yielding is not expected, a linear elastic beam model was adopted. For yielding steel ribs in squeezing ground conditions, an elastic–perfectly plastic structural beam model was used, enabling the development of plastic hinges and redistribution of internal forces. Shotcrete support was modelled using liner (shell) elements, which behave as a two-dimensional continuum rather than a one-dimensional structural member. The liner was primarily simulated using a linear elastic constitutive model, which is commonly adopted in practice. The formulation allows the capture of key failure mechanisms such as tensile cracking, compressive yielding, and bending failure, depending on the stress state developed in the liner. Overall, the support system was simulated using structural elements governed by axial and bending elastic–plastic constitutive laws with appropriate interface coupling, providing a realistic representation of support–rock interaction without employing rock mass constitutive models for the supports. The mechanical properties and installation sequence of these support components were assigned based on field specifications (Tables 4–6) to realistically capture rock–support interaction during sequential excavation.
Although cavity formation during Himalayan tunnel excavation is influenced by hydrological pressure and blast-induced dynamic loading, the present numerical simulations adopt a quasi-static framework to isolate the dominant role of geological weakness and stress redistribution in progressive crown cavity development. The weakening effects of groundwater are incorporated through reduced rock mass parameters, while blasting-induced disturbance is represented by staged excavation and stress release. Explicit hydro-mechanical and dynamic analyses are identified as important extensions for future work.
5 Analysis
The monitoring of the crown was carried out using extensometers at different point along the section of the tunnel (Figure 15) and these observations were used to calibrate the numerical model and predict the roof/side deformation during different stages of tunnel excavation. The displacement graph (Figure 16a) illustrates the initial stages of excavation, specifically the displacement of the crown’s top surface around the tunnel portal. In the beginning of tunnel excavation, there is a noticeable rise in vertical displacement, indicating ground movement. However, as the excavation advances to a certain distance, the amount of vertical displacement saturated depicting equilibrium or stabilizes. The introduction of support during this phase effectively controlled the crown displacement, demonstrating the effectiveness of the support system in maintaining stability. However, the analysis of the displacement graph reveals that the existing support system, as depicted in Figure 16a does not effectively control the displacement of the crown. Consequently, additional support measures are necessary to ensure the stability of the tunnel. This can be attributed to the presence of weaker rock, which demands additional support to maintain stability. The graph (Figure 16a) clearly demonstrates that, in the absence of enhanced support, crown displacement continues, increasing the risk of tunnel instability or potential failure.
FIGURE 15
FIGURE 16
The examination of the displacement graphs (Figures 16b,c) reveals a significant increase in the value of crown deformation as compared to previous graphs (Figure 16a). The existing support system proved to be inadequate in controlling the displacement, ultimately resulting in cavity failure approximately 26 m from the tunnel portal. This indicates that either the ground conditions or the implemented support measures are insufficient to withstand the forces and pressures exerted on the tunnel walls. The occurrence of cavity failure (Figure 16c) signifies a collapse or deformation of the surrounding rock, posing a considerable risk to the overall stability of the tunnel. It is important to note that the interpretation of the graphs may vary without specific numerical values or additional context. When assessing tunnel stability, it is crucial to consider other factors such as geological conditions, engineering design, and construction methods.
A clear correlation is observed between field-monitored crown displacement (Figure 15), weak geological conditions identified above the crown (Tables 2 and 3), and the limitations of the installed support system (Tables 4–6). Despite support installation, continued displacement was recorded, indicating insufficient confinement in the weak GT13 zone. Numerical results corroborate these observations, showing progressive deformation leading to cavity formation, thereby confirming the inadequacy of the existing support system under the prevailing geological conditions.
6 Discussion
The present study offers valuable insights into the mechanism of chimney (cavity) failure in tunnels excavated under weak and heterogeneous Himalayan geological conditions, by integrating field observations with calibrated numerical simulations. The discussion below interprets the results in the context of geological controls, displacement behaviour, support system performance, and broader implications for tunnel design and construction.
The numerical simulation results demonstrate that crown displacement is strongly influenced by the presence of weak and sheared rock mass zones above the tunnel crown, particularly under shallow to moderate overburden conditions. The progressive increase in vertical displacement observed during sequential excavation confirms that stress redistribution and loss of confinement play a dominant role in initiating chimney failure. These findings are consistent with earlier observations that cavity formation is not solely governed by stand-up time concepts but is controlled by localised geological weakness, groundwater conditions, and excavation-induced stress relief.
Comparison between unsupported and supported excavation stages highlights the critical role of timely and adequate support installation. The unsupported tunnel condition exhibited a continuous increase in crown displacement with excavation advance, ultimately leading to the formation of a cavity. In contrast, the supported condition showed a marked reduction in displacement magnitude and deformation rate, indicating that support systems significantly enhance tunnel stability. However, the occurrence of cavity failure despite the installation of support in weak zones suggests that conventional support measures may be insufficient when applied uniformly, without accounting for localised geological variability.
Field monitoring data obtained from extensometers provided valuable calibration for the numerical model, enabling objective validation of the simulated displacement trends. The close agreement between measured and simulated crown displacements reinforces the reliability of the numerical approach adopted in this study. Importantly, the identification of displacement thresholds beyond which deformation accelerates offers a practical indicator for early warning and adaptive support design during tunnel construction.
The results also emphasise the influence of groundwater conditions on chimney failure. The presence of perched aquifers and water inflow above the crown reduced effective stress and weakened the rock mass, thereby exacerbating deformation and cavity growth. This underscores the necessity of incorporating groundwater effects into both numerical modelling and support design, particularly in Himalayan tunnelling projects where hydrological uncertainty is high.
From a design perspective, the study highlights the need for site-specific and performance-based support strategies rather than reliance on generalised ground classifications alone. The integration of field monitoring, numerical simulation, and geological characterisation provides a robust framework for assessing tunnel stability and optimising support systems in weak ground conditions. Such an approach aligns with modern sustainable tunnelling practices, as it enables minimisation of overbreak, efficient use of support materials, and improved construction safety.
Overall, the findings of this study contribute to a better understanding of chimney failure mechanisms in Himalayan tunnels and demonstrate the effectiveness of combining numerical modelling with field data for evaluating support performance. The proposed approach can be applied to similar tunnelling projects in complex geological environments to enhance stability assessment, support design, and risk mitigation strategies.
7 Conclusion
This study focused on the analysis and remediation of cavity failure during tunnelling in the Himalayas and also assess the efficacy of the existing support system. Numerical modelling of unsupported and supported tunnels in the simulation software, the performance of the support system and its ability to mitigate deformation and stresses was evaluated. Based on the observations, the proposed support system proved to be effective in controlling crown displacement and vertical stresses, indicating its efficiency in maintaining tunnel stability. It was also evident that the support requirements varied for different rock classes, highlighting the importance of adapting the support system to suit the specific geological conditions encountered during tunnelling in the Himalayas. The analysed support system comprising rock bolts, shotcrete, and steel ribs demonstrated effective control of crown displacement and vertical stress concentration, reducing deformation by approximately 40%–55% compared to unsupported conditions, as found from field and simulation results. This study provides valuable insights into the analysis and remediation of cavity failure during tunnelling in the challenging geological conditions of the Himalayas. The findings emphasize the significance of an appropriate support system design and the consideration of pore water pressure effects to ensure safe and successful tunnel construction in the region as per the changing ground conditions.
Statements
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Author contributions
RK: Conceptualization, Data curation, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – review and editing, Formal Analysis. AsK: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – review and editing, Supervision, Writing – original draft. AbK: Conceptualization, Data curation, Investigation, Methodology, Resources, Visualization, Writing – original draft, Writing – review and editing, Software, Validation. BP: Methodology, Formal analysis, Validation, Writing – review and editing. SR (5th author): Data curation, Formal Analysis, Writing – review and editing, Conceptualization, Investigation, Methodology, Resources, Visualization, Writing – original draft. SR (6th author): Methodology, Formal analysis, Validation, Writing – review and editing. KS: Data curation, Formal Analysis, Supervision, Validation, Writing – review and editing, Funding acquisition, Project administration. KZ: Data curation, Formal Analysis, Funding acquisition, Supervision, Validation, Writing – review and editing. AZ: Data curation, Formal Analysis, Funding acquisition, Supervision, Validation, Writing – review and editing. JS: Data curation, Formal Analysis, Supervision, Validation, Writing – review and editing. ZR: Data curation, Formal Analysis, Supervision, Validation, Writing – review and editing.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
1
AbbasN.LiK.FisshaY.LeiW.EmadM. Z.ChandrahasN. S.et al (2024). Stress-deformation and stability challenges in himalayan tunnels: impact of geological discontinuities. Discov. Mater.4, 72. 10.1007/s43939-024-00144-z
2
AhmedA.MishraS. K.AzadM. A.AlquamarC.SinghT. N.AnsariA.et al (2025). Prediction of tunnel ground deformation – a case study from Western himalaya, India. Results in. Earth Sci.3, 100052. 10.1016/j.rines.2024.100052
3
BartonN.LienR.LundeJ. (1975). “Estimation of support requirements for underground excavations,” in Proc. Sixteenth symp. On rock mechanics (Minnesota, U.S.A.: Univ. of Minnesota), 163–177.
4
BegY.SinghV.KunduS.BhowmikR. (2025). “Tunneling challenges in Himalayan region—A review,” in Proceedings of 9IYGEC 2023 (Singapore: Springer Nature), 2, 207–217. 10.1007/978-981-97-6988-9_21
5
BieniawskiZ. T. (1993). “Classification of rock masses for engineering: the RMR system and future trends,” in Rock testing and site characterization. Editor HudsonJ. A. (Oxford: Pergamon), 553–573.
6
ChenL.WangZ. F.WangY.XitongB.LaiJ. (2022). Characteristics and failure analysis of a railway tunnel collapse influenced by cavity in phyllite strata. Eng. Fail. Anal.142, 106794. 10.1016/j.engfailanal.2022.106794
7
DiwakarK. C.GautamK.DangiD.KadelS.HuL. (2022). Challenges in tunneling in the himalayas: a survey of several prominent excavation projects in the himalayan Mountain range of south Asia. Geotechnics2, 802–824.
8
HuangF.ZhaoL. H.LingT. H.YangX. L. (2017). Rock mass collapse mechanism of concealed karst cave beneath deep tunnel. Int. J. Rock Mech. Min. Sci.91, 133–138. 10.1016/j.ijrmms.2016.11.017
9
HuangH.ZhangD.HuangZ. (2022). Resilience of city underground infrastructure under multi-hazards impact: from structural level to network level. Resilient Cities Struct.1 (2), 76–86. 10.1016/j.rcns.2022.07.003
10
KoopialipoorM.AsterisG.SalihM. A.AlexakisD. E.MamouA.ArmaghaniD. J. (2022). Introducing stacking machine learning approaches for the prediction of rock deformation. Transp. Geotech.34, 100756.
11
KumarR. (2023). Analysis and remediation of cavity failure during tunnelling in the himalayas.MTech thesis. Dhanbad, India: Indian Institute of Technology (Indian School of Mines).
12
LaufferH. (1958). Gebirgsklassifizierung fur den stollenbau. Geol. Bauwesan24, 46–51.
13
LeuS. S.ChenC. N.ChangS. L. (2001). Data mining for tunnel support stability: neural network approach. Automation Constr.10, 429–431. 10.1016/S0926-5805(00)00078-9
14
LiJ.FangY.LiuC.ZhangY.LuW. (2020). Performance investigation of tunnel lining with cavities around surrounding rocks. Adv. Civ. Eng.2020, 1364984. 10.1155/2020/1364984
15
LiuX. R.LiD.ZhongZ. L.WuQ. (2013). Research on collapse mechanism and disaster treatment of mayakou tunnel by the action of fluid-structure interaction (FSI). Disaster Adv.5 (4), 903–908.
16
LiuX. Z.LiuF.SongK. Z. (2022). Mechanism analysis of tunnel collapse in a soft-hard interbedded surrounding rock mass: a case study of the yangshan tunnel in China. Eng. Fail. Anal.138, 106304. 10.1016/j.engfailanal.2022.106304
17
LiuJ.XieQ.HanB.WangJ.LvG.LuoX. (2024). Experimental and numerical study of the failure mechanism and stability of tunnel linings under different cavity conditions. Eng. Fail. Anal.162, 108412. 10.1016/j.engfailanal.2024.108412
18
NguyenN. H. T.BuiH. H.NguyenG. D.KodikaraJ. (2017). A cohesive damage-plasticity model for DEM and its application for numerical investigation of soft rock fracture properties. Int. J. Plast.98, 175–196. 10.1016/j.ijplas.2017.07.008
19
RabcewiczL. (1973). Principles of dimensioning the supporting system for the new Austrian tunnelling method. Water Power25, 88–93.
20
RajputA.KaushikA.IqbalM. A.GuptaN. K. (2023). Non-linear FE investigation of subsurface tunnel with GFRP protection against internal blast. Int. J. Impact Eng.172, 104423. 10.1016/j.ijimpeng.2022.104423
21
SchmidtF.SagasetaC.KonietzkyH. (2010). “3D analysis of a micropile umbrella for stabilizing the tunnel face of a NATM tunnel,” in Numerical methods in geotechnical engineering – Benz and nordal, 749–753.
22
ShenY. S.GaoB.YangX. M.TaoS. (2014). Seismic damage mechanism and dynamic deformation characteristic analysis of mountain tunnel after wenchuan earthquake. Eng. Geol.180 (8), 85–98. 10.1016/j.enggeo.2014.07.017
23
SpyridisP.BergmeisterK. (2024). Sustainable tunnel design: concepts and examples of reducing greenhouse gas emissions through basic engineering assumptions. Tunn. Undergr. Space Technol.152, 105886. 10.1016/j.tust.2024.105886
24
UpadhyayS.KaushalV.BaishtaD.BhakatS. K. (2023). Mitigating loose rock fall and cavity formation in Adit-2 tunnel of rammam III hydroelectric project in himalayas, India: challenges and solutions. IOP Conf. Ser. Earth Environ. Sci.1249, 012014. 10.1088/1755-1315/1249/1/012014
25
YangW.ZhengJ.ZhangR.LiuH. (2021). An empirical model for characterizing 3D deformation at the face of shield tunnel in soft clay. Tunn. Undergr. Space Technol.112, 103862. 10.1016/j.tust.2021.103862
Summary
Keywords
chimney failure, numerical simulation, roof bolt, shotcrete, support, tunnel
Citation
Kumar R, Kumar A, Kumar A, Pandit B, Ram S, Ray S, Skrzypkowski K, Zagórski K, Zagórska A, Stasica J and Rak Z (2026) Assessment of cavity formation in the crown during tunnel excavation in the Himalayan region. Front. Earth Sci. 14:1745569. doi: 10.3389/feart.2026.1745569
Received
13 November 2025
Revised
02 February 2026
Accepted
12 February 2026
Published
13 March 2026
Volume
14 - 2026
Edited by
Manoj Khandelwal, Federation University Australia, Australia
Reviewed by
Alfrendo Satyanaga, Nazarbayev University, Kazakhstan
Ying Yuan, Hebei GEO University, China
Updates
Copyright
© 2026 Kumar, Kumar, Kumar, Pandit, Ram, Ray, Skrzypkowski, Zagórski, Zagórska, Stasica and Rak.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Ashok Kumar, ashokmin@iitism.ac.in; Krzysztof Skrzypkowski, skrzypko@agh.edu.pl
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.