Investigation on Crack Propagation Mechanisms in Surrounding

Rock Induced by Excavation Unloading of Deep-Buried Caverns

Donghan Wang

2,3

, Kaiwen Song

2,3

, Qian Dong

and Junhong Huang

2,3

Hubei Key Laboratory of Blasting Engineering, Wuhan, China

Sanya Science and Education Innovation Park, Wuhan University of Technology, Sanya Hainan, China

School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan Hubei, China

Keywords: Rock.Mechanics, Numerical.Simulation, Roadway.Surrounding.Rock, Transient.Unloading, Small-Scale

Surrounding Rock Specimens.

Abstract: To investigate the deformation patterns and failure mechanisms of roadway surrounding rock under transient

excavation unloading, and to simulate the roadway excavation unloading process, a model test system for

roadway excavation and unloading was developed. Multiple sets of jointed rock mass model specimens were

fabricated using high-strength gypsum materials. Numerical simulations were employed to explore the

influences of joint quantity, length, stiffness, and spatial configuration on the failure characteristics of

surrounding rock during excavation unloading. The results indicate that under transient unloading conditions:

Jointed rock masses exhibit a higher degree of failure compared to intact rock masses. Rock masses containing

longer joints demonstrate more pronounced failure phenomena than those with shorter joints. Joint stiffness

exerts relatively minor influence on both peripheral displacement and damage extent of the excavation. Rock

masses with mixed-length joints show greater susceptibility to failure compared to those with uniform-length

joints. Multi-jointed rock masses are more prone to crack formation during unloading, potentially leading to

more significant rock deformation and crack propagation. In contrast, rock masses with fewer joints

experience less impact under such transient unloading conditions, consequently demonstrating enhanced

stability and safety of the surrounding rock.

1 INTRODUCTION

To prevent geotechnical hazards in deep rock mass

engineering, such as roadway instability and

rockbursts, a deeper understanding of the interaction

mechanisms between surrounding rock and support

structures is imperative. Previous studies

predominantly focused on interpreting rock

unloading phenomena through loading theories.

However, stress release induced by roadway

excavation is the primary cause of the loosened zone

in surrounding rock, necessitating consideration of

the mechanical properties of both rock mass and

support structures under varying stress states during

excavation unloading. International research on

excavation unloading commenced earlier. For

instance:(Cai et al., 2007)simulated damage

distribution characteristics in surrounding rock

during deep tunnel excavation using software like

PFC. In China:(Lu et al., 2008)proposed and

validated the concept of transient unloading during

excavation through mechanical analysis and duration

calculations of load release.(Lu et al., 2008) identified

excavation unloading as the primary cause of large-

scale damage zones in surrounding rock.

Research on deep rock mass excavation unloading

characteristics initially focused on classical

theoretical mechanics and numerical simulations,

supplemented by theoretical analyses of practical

engineering issues. Key findings include:(Luo et al.,

2023)analyzed the post-unloading mechanical states

and deformation behaviors of rock masses using

classical mechanics.(Dong et al., 2017)investigated

the impacts of initial in-situ stress, excavation radius,

and dynamic rock strength on unloading-induced

failure and stability.(Dong et al., 2017)demonstrated

that the extent of surrounding rock failure correlates

strongly with unloading duration, with shorter

durations inducing greater disturbance magnitudes.

Instantaneous excavation unloading induces

vibrations and failure in rock masses. Through case

studies, (Fan et al., 2015) and(Lu et al., 2007)found

that larger unloading volumes during excavation

result in stronger vibrations, and transient unloading

366

Wang, D., Song, K., Dong, Q. and Huang, J.

Investigation on Crack Propagation Mechanisms in Surrounding Rock Induced by Excavation Unloading of Deep-Buried Caverns.

DOI: 10.5220/0013604600003970

In Proceedings of the 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2025), pages 366-373

ISBN: 978-989-758-759-7; ISSN: 2184-2841

may amplify the overall vibration response of

surrounding rock.

Current research on dynamic excavation

unloading primarily focuses on the dynamic response

of rock masses. To investigate crack propagation

mechanisms in both jointed and intact rock masses

under excavation unloading, this study designed

laboratory model tests and numerical simulations,

followed by systematic analysis of the results.

2 SIMULATION TEST SYSTEM

FOR LOOSENING OF

UNDERGROUND CAVERN

STRUCTURAL PLANES

UNDER TRANSIENT

EXCAVATION UNLOADING

This section analyzes surrounding rock stress

distribution, introduces a self-developed testing

system, and integrates theoretical-experimental

methodology to establish foundations for subsequent

lab tests.

2.1 Stress Distribution of Roadway

Surrounding Rock

To elucidate the stress distribution of surrounding

rock in semi-circular arched roadways with vertical

walls and predict their failure patterns, the following

analytical approaches are conducted:

The stress distribution of roadway surrounding

rock constitutes a plane strain problem, which can be

solved using the complex variable method (2018; Zhu

et al., 2014; Dai & Zhang, 2012). The in-situ stress

field of semi-circular arched roadways with vertical

walls comprises vertical stress σ



, horizontal

stress σ



, and shear stress τ. The surrounding rock

stress can be expressed by two complex potential

functions:φ(z)，and ψ(z).

002

13 0

() () ()

() ( ) () ()

σσ

ϕωξϕξ

σσ ω

ξψξ

−









=− − +











(1)

Parameters 𝜑



(𝜉), 𝜓



(𝜉), and 𝜔(𝜉) in Equation

(1) are determined, while 𝜎





and 𝜎





are assigned

according to Equation (2):

()

arctan 1

VH VH

HV HV

σσ σσ

στ

σσ σσ

στ

σσ σσ

ττ



+−





=+ +









+−





=− +











−−



=++















(2)

Consequently, the stress distribution of roadway

surrounding rock is derived as:

()

2''

4Re

()

2()() ()

()

() ()

ρθ

ρθ ρθ

ϕξ

σσ

ωξ

ξϕξ ψξ

σσ τ

ρωξ























−+ = +

















(3)

2.2 Introduction to the Test System

To intuitively investigate the deformation behavior of

jointed rock masses during instantaneous excavation

unloading, a simulation test system for transient

unloading-induced loosening of structural planes in

underground caverns was designed. The system

operates as follows:

(I) Mechanical Configuration

The front end of a lever-type loading/unloading

assembly is embedded into the pre-excavated slot

within the surrounding rock model (containing

structural planes) and tightly contacts the slot

wall.The rear end is connected to a suction-cup

electromagnetic actuator via a welded cylindrical

steel tube.

(II)Unloading Simulation Process

Under axial tensile loading, the lever mechanism

applies controlled interfacial pressure to the slot wall

in the chamber assembly. Deactivating the

electromagnetic actuator triggers instantaneous

pressure release, simulating transient stress-field

unloading (Figure 1). This system facilitates lab-scale

simulation of excavation loading/unloading cycles,

with the lever mechanism enabling instantaneous

structural plane unloading in pre-excavated models

through controlled energy storage/release.

In terms of material selection, gypsum is adopted

for rock mass modeling due to its ease of processing,

customizable composition, and cost-effectiveness.

Gypsum-based materials remain widely utilized in

dynamic rock experiments. Dimensions and physical

characteristics of the surrounding rock specimens are

detailed in Figure 2.

The model incorporates centrally positioned pre-

existing cracks with mica sheets embedded along

vertically penetrating surfaces to simulate primary

Investigation on Crack Propagation Mechanisms in Surrounding Rock Induced by Excavation Unloading of Deep-Buried Caverns

367

structural planes in surrounding rock. A roadway

excavation of standardized geometry is implemented

at the specimen's center to improve real-world

scenario simulation fidelity.

Figure.1: Schematic diagram of simulation test system.

(a)Gypsum specimen model

size drawing

(b)Physical diagram of

the plaster specimen

igure.2: Size drawing and physical drawing of surroundin

ock specimen.

3 DESIGN OF LABORATORY

TESTS FOR TRANSIENT

UNLOADING DURING

DEEP-BURIED CAVERN

EXCAVATION

3.1 Uniaxial Compressive Strength

Testing of Gypsum

Gypsum was selected as the surrounding rock

simulation material for subsequent tests. To

determine the compressive strength of the gypsum

material used in the tests, uniaxial compression tests

were conducted using a 50 kN microcomputer-

controlled electronic universal testing machine

(Figure 3). The standard cylindrical specimens

(Figure 4) measured 50 mm in diameter and 100 mm

in height. A loading rate of 5 mm/min was applied to

ensure complete acquisition of stress-strain curves.

Photographs of five specimens before and after

failure are summarized in Table 1, with

corresponding stress-strain curves presented in Fig. 5.

Accounting for material heterogeneity and

experimental errors, the average uniaxial

compressive strength and elastic modulus of the

gypsum material were determined as 3.92 MPa and

5.77 GPa, respectively. This section must be in two

columns.

igure 3: Physical diagram

of experimental equipment.

Figure 4: Dimensions o

test standard parts.

Table 1: Typical photographs of standard gypsum

specimens before and after uniaxial compression failure..

Pre-failure Post-failure Pre-failure Post-failure

Figure 5: Stress-strain curves of gypsum specimens.

3.2 Experimental Methodology

The simulation test system (Section 2.1) applied 2

MPa confining pressure with instantaneous unloading

(≤0.5s). Mica-embedded gypsum specimens with

variable joint geometries (quantity/length/

orientation) simulated excavation-induced structural

planes. Fourteen test groups (Table 2) captured

systematic crack propagation patterns near

discontinuities, providing empirical validation for

structural plane evolution mechanisms under

excavation unloading disturbances.

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

368

Table 2: Schematic diagram of test conditions.

Schematic

Diagram

Test Case 1 (Without

mica sheets)

Test Case 2 (Two 1

cm mica sheets)

Test Case 3 (Four 1

cm mica sheets)

Test Case 4 (Two 1

cm mica sheets)

Test Case 5 (Four 1

cm mica sheets)

Schematic

Diagram

Test Case 6 (Two 2

cm mica sheets)

Test Case 7 (Four 2

cm mica sheets)

Test Case 8

(Two 1cm and two

2cm mica sheets)

Test Case 9 (Two 1

cm and two 2 cm

mica sheets)

Test Case 10

(Two 2 cm mica

sheets)

Schematic

Diagram

Test Case 11 (Four 2

cm mica sheets)

Test Case 12(Two 1

cm mica sheets)

Test Case 13(Four

1 cm mica sheets)

Test Case 14(Two 1

cm mica sheets)

3.3 Analysis of Recorded Experimental

Phenomena

Post-test documentation and analysis of gypsum

failure characteristics were conducted as follows:

3.3.1 Failure Characteristics of Intact and

Jointed Rock Masses

A comparative analysis of intact rock masses (Test

Case 1) and jointed rock masses (represented by Test

Cases 6 and 12) before and after testing is presented

in Table 3. To enhance visibility of experimental

phenomena, crack trajectories in the images were

manually highlighted with black lines to accentuate

fine fractures that were otherwise difficult to discern.

Experimental observations showed intact rock

under high confining pressure exhibited plastic

deformation with gradual microcrack development

and eventual fracture, while sudden unloading

induced elastic recovery that reduced plasticity and

accelerated crack dilation. In contrast, jointed rock

masses demonstrated reduced integrity due to pre-

existing discontinuities, causing localized stress

concentrations and accelerated crack propagation

under equivalent loading, resulting in extended cracks

and more pronounced failure phenomena.

Table 3: Phenomena of unjointed rock mass and partially

jointed rock mass after failure.

Test Case 1

Test Case 6

Test Case 12

8 16

Investigation on Crack Propagation Mechanisms in Surrounding Rock Induced by Excavation Unloading of Deep-Buried Caverns

369

3.3.2 Failure Characteristics of Jointed

Rock Masses with Varying Properties

(I)Different Joint Lengths

Comparative analysis indicates 2 cm-jointed rock

masses exhibit longer post-unloading cracks than 1

cm-jointed counterparts due to reduced transient

stress concentrations near joint planes in shorter joints,

enhancing stability and limiting crack propagation.

Table 4: Diagram of post-failure phenomena at different

joint lengths.

Test Case 7

Test Case 8

Test Case 9

Table 5: Diagram of post-failure phenomena at different

joint lengths.

Test Case 6

Test Case 12

Mixed-length joints exhibit compound transient

failure: longer joints initiate fracturing as shorter

joints amplify damage, reducing safety margins

versus uniform joints during dynamic excavation (

Table 4, Table 5).

(II)Different Joint Dip Angles

Table 6 documents failure progression in specimens

with 1 cm mica sheets at 30°,45°,60° dip angles

(Cases 2/4/14) under transient unloading. Mica sheets

are spatially indexed per Figure 6 for precise damage

characterization.

Comparative analysis of Test Cases 2, 4, and 14

revealed that under identical joint lengths and

transient unloading conditions: Cracks exhibited

comparable lengths and predominantly propagated

along joint planes. Increasing joint dip angles resulted

in steeper crack trajectories (i.e., greater angles

relative to the horizontal).

Figure 6: Schematic diagram of the naming of mica sheets.

Table 6: Diagram of the phenomenon after failure at

different joint dip angles.

Test Case 2

Test Case 4

Test Case 14

Test Cases 3/5/13 (Table 7) with four mica sheets

each demonstrated that steeper dip angles of No.②

mica sheet (45º/60º/0º) inversely correlated with No.

① mica sheet proximity to excavation boundaries

(Case 13>3>5). Post-test analysis revealed crack

coalescence between mica sheets and excavation

boundaries in Case 5 but no crack initiation along No.

① sheet in Case 13. Proximity to excavations reduces

confinement, enabling transient unloading-induced

crack formation, while distant fractures under higher

confinement exhibit suppressed crack initiation/

propagation.

Table 7: Diagram of the post-failure phenomenon of

different joint dip angles.

Test Case 3

Test Case 5

Test Case 13

(III)Different Joint Numbers

Table 8: Phenomena after failure of different joint quantities.

Test Case 10

Test Case 11

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

370

Table 8 demonstrates crack-joint quantity

proportionality: Multi-jointed systems near

excavation boundaries undergo simultaneous joint

propagation-deformation-cracking processes under

transient unloading, where adjacent joint interactions

amplify rock deformation/crack coalescence,

substantially destabilizing rock masses. Fewer-joint

configurations exhibit diminished joint interplay and

unloading sensitivity, yielding superior

stability/safety through restricted crack development.

Figure 7: Correspondence Between Joint and Crack

Dimensions: Lengths and Quantities.

Figure 8: Relationship between Joint Dip Angles and Crack

Trajectories.

Experimental results (Figures 7, 8) demonstrate

intact rock masses generate fewer cracks than jointed

counterparts under transient unloading. Increased

joint quantity directly elevates crack numbers in

single-length jointed systems. Equivalent joint

quantities show steeper Joint #② dip angles position

Joint #① closer to excavation boundaries, promoting

crack-boundary intersection and potential crack

multiplication. Single-length joints exhibit

proportional crack length-to-joint size relationships,

while mixed-length systems develop complex stress

redistribution enabling extended crack propagation.

Joint quantity, geometric orientation, and spatial

arrangement collectively govern transient unloading-

induced crack evolution and instability mechanisms.

4 NUMERICAL SIMULATION OF

TRANSIENT UNLOADING

DURING DEEP-BURIED

CAVERN EXCAVATION

Existing studies on jointed rock transient unloading

primarily addressed joint geometry/spatial

configurations, overlooking joint material

heterogeneity. This chapter employs finite element

analysis with an implicit-explicit-implicit sequential

method to investigate deformation-damage

mechanisms in surrounding rock during transient

unloading, including systematic analysis of unloading

rate effects.

4.1 Model Establishment

Figure 9: Schematic Diagram of Transient Unloading

Model for Jointed Rock Masses.

The jointed rock mass model (100 m × 100 m)

features a central excavation cavity (a=5 m) for

simulating excavation-induced transient unloading.

Plane strain conditions are enforced through

displacement constraints on front/rear surfaces, with

peripheral non-reflective boundaries for wave

reflection suppression. A 40 MPa global load

simulates in-situ stress. Initial equilibrium is

established by applying radial confinement

equivalent to in-situ stress at cavity periphery (t=0),

followed by linear unloading to 0 MPa to replicate

blasting-induced transient stress release. The

geometric configuration and staged loading protocol

are illustrated in Figure 9.

Investigation on Crack Propagation Mechanisms in Surrounding Rock Induced by Excavation Unloading of Deep-Buried Caverns

371

4.2 Influence Mechanisms of Joint

Stiffness on Surrounding Rock

Deformation and Damage During

Transient Unloading

Subsequent numerical simulations focus on the

damage and deformation characteristics of

surrounding rock under varying joint stiffness

conditions. In practical engineering, rock masses

typically contain both persistent and non-persistent

joints. However, this study exclusively addresses

non-persistent joints, as under high in-situ stress

conditions, joints are predominantly closed, with

normal stiffness values generally ranging between

kn=0.05E and kn=0.2E (2018; Zhu et al., 2014; Dai

& Zhang, 2012). Accordingly, joint normal stiffness

values are set to kn =0.05E, 0.10E, 0.15E, 0.20E, and

0.25E, while tangential stiffness is fixed at 20% of the

normal stiffness (ks=0.2kn).

Figure 10 clearly illustrates the influence of joint

stiffness on cavern convergence displacement in

short-joint configurations, showing that under short-

joint conditions, the maximum convergence

displacements occur within the 70 º -80 º sector

relative to the cavern, with a slight reduction (0.27

cm) as joint stiffness increases from kn =0.05E to

kn=0.25E. Notably, an anomalous displacement

behavior is observed along the horizontal axis closest

to the joints, where convergence displacement

increases by 0.70 cm during the same stiffness

escalation, indicating a counterintuitive stiffness-

dependent response.

Figure 11: Peripheral Displacement Diagrams of Cavern

under Different Joint Normal Stiffness.

The influence of joint stiffness on damage

evolution is also significant, as demonstrated in

Figure 1, which presents damage contours of short

joints under varying joint stiffness conditions. As

joint stiffness increases, stress wave reflection

diminishes, leading to a reduction in the displacement

differential between reflection and incidence zones.

Consequently, the damage severity at joint tips

decreases markedly, with the damage zone extent

shrinking significantly and gradually disconnecting

from the cavern periphery. Similarly, the damage

intensity near the cavern, particularly on the side

adjacent to the joints, is notably reduced with

increasing joint stiffness due to weakened reflection

effects.

Joint stiffness exerts relatively minor influence on

peripheral displacement and damage. When kn

increases from 0.05E to 0.25E, the maximum

convergence displacement decreases by only 0.27cm,

while damage zones at joint tips transition from

persistent to non-persistent states. In joint support

design, a "zoning control strategy" should be adopted:

implementing stiffness reinforcement in main

displacement zones while coordinating flexible

supports in near-field horizontal regions to create

mechanical buffer belts.

kn=0.05E kn=0.10E kn=0.15E kn=0.20E kn=0.20E

Figure 12: Damage Contours of Surrounding Rock under

Different Joint Stiffness Conditions.

5 CONCLUSIONS

The self-developed roadway surrounding rock

excavation unloading model test system was

employed to simulate instantaneous failure of

surrounding rock during tunnel excavation, yielding

the following conclusions:

(1) Model tests revealed that both intact and jointed

rock masses develop inverted “八 ”-shaped cracks

under excavation unloading. Compared to intact rock

masses, prefabricated jointed rock models exhibit

more pronounced crack propagation.

(2) When multiple joints of equal length exist in a

single model, the number of formed cracks aligns

closely with the original prefabricated joint count.

Longer prefabricated joints result in longer final crack

lengths.

(3) Rock mass models containing prefabricated joints

with varying lengths and dip angles demonstrate

significant uncertainty in the quantity, length, and

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

372

orientation of final cracks, necessitating further

systematic investigation.

(4) The coupled "damage suppression-displacement

excitation" effect induced by joint stiffness on

surrounding displacement and damage of

underground chambers reveals the dual attributes of

rock mass structural stiffness parameters.

ACKNOWLEDGEMENTS

This work is supported by the Foundation of Hubei

Key Laboratory of Blasting Engineering

(No.BL2021-13), the National Natural Science

Foundation of China (Grant No. 52108368, No.

52109165), the Fundamental Research Funds for the

Central Universities (WUT: 2024IVA028).

REFERENCES

2018, Некоторые основные задачи математической

теории упругости. In H. Zhao, T. Fan, & C. Wang

(Eds.), Beijing: China Science Publishing& Media Ltd.

Cai, M., Kaiser, P. K., Morioka, H., Minami, M., Maejima,

T., Tasaka, Y. & Kurose, H. (2007). FLAC/PFC

Coupled Numerical Simulation of AE in Large-scale

Underground Excavations. International Journal of

Rock Mechanics and Mining Sciences, 44(4), 550-564.

http://doi.org/ 10.1016/j.ijrmms.2006.09.013

Dai, J. & Zhang, L. (2012). Shuxue Wuli Fangcheng

[Shuxue Wuli Fangcheng]. Southeast University Press.

Dong, C., Zhao, G., Li, Y., Meng, X., Lu, X. & Zhang, L.

(2017). Mechanical properties and failure mechanism

of Surrounding Rocks in Deep Circular Tunnel under

Excavation Unloading [Mechanical properties and

failure mechanism of surrounding rocks in deep circular

tunnel under excavation unloading]. Journal of Mining

and Safety Engineering, 34(3), 511-518, 526

Fan, Y., Lu, W., Yang, J., Yan, P. & Chen, M. (2015).

Attenuation Law of Vibration Induced by Transient

Unloading During Excavation of deep Caverns

[Attenuation law of vibration induced by transient

unloading during excavation of deep caverns]. Rock

and Soil Mechanics(2), 541-549. http://doi.org/10.

16285/j.rsm.2015.02.033

Lu, W. (1996). A Study on Interaction between Stress wave

and Slipping rock Interface [A Study on Interaction

between stress wave and slipping rock interface]. Rock

and Soil Mechanics(3), 70-75.

Lu, W, Zhou, C., Chen, M., Jin, L. & Yan, L. (2008).

Research on Transient Characteristics of Excavation

Unloading [research on transient characteristics of

excavation unloading]. Chinese journal of rock

mechanics and engineering, 27(11), 2184-2192. http://

doi.org/10.3321/j.issn:1000-6915.2008.11.003

Luo, S., Yan, P., Lu, W., Dong, Z., Zhou, C., Yang, Z. &

Hu, Y. (2023). Stability Index of Surrounding Rock

during Deep Rock Excavation Considering Energy

Release Speed. Applied sciences, 13(5), 3000. http://

doi.org/10.3390/app13053000

W. Lu, M. Chen, P. Yan, & C. Zhou. (2007). Study On

Vibration Characteristics of Surrounding rock Induced

by Tunnel Excavation under high in-situ stress [study

on Vibration Characteristics of Surrounding rock

induced by Tunnel Excavation under High in-situ

stress]. chinese journal of rock mechanics and

engineering, 26(z1), 3329-3334. http://doi.org/10.

3321/j.issn:1000-6915.2007.z1.111

Yan, P., Lu, P., & Xu, H. (2007). A Primary Study to

Damage Mechanism of Initial Stress Dynamic

unloading when Excavating Under high geostress

condition [A primary study to damage mechanism of

initial stress dynamic unloading when excavating under

high geostress condition]. Explosion and Shock

WAVES, 27(3), 283-288. http://doi.org/10.3321/j.issn:

1001-1455.2007.03.016

Yang, J. (2014). Coupling Effect of Blasting and Transient

Release of In-situ Stress during Deep Rock Mass

Excavation [ Doctor, Wuhan University].

Zhu, J., Yang, J., Shi, G., Wang, J. & Cai, J. (2014).

Calculating method for conformal mapping from

exterior of unit circle to exterior of cavern with arbitrary

excavation cross-section [Calculating method for

conformal mapping from exterior of unit circle to

exterior of cavern with arbitrary excavation cross-

section]. Rock and Soil Mechanics(1), 175-183.

Investigation on Crack Propagation Mechanisms in Surrounding Rock Induced by Excavation Unloading of Deep-Buried Caverns

373