Rock Mechanics and Underground Engineering

Introduction and Context

Rock mechanics is the study of the mechanical behaviour of rock and rock masses, including their strength, deformation, permeability, discontinuities, and response to excavation or loading. Underground engineering applies this knowledge to tunnels, caverns, shafts, mines, foundations, slopes, and underground infrastructure.

Rock engineering differs fundamentally from soil engineering. Intact rock may be extremely strong, yet a rock mass can be weak because of joints, faults, bedding, weathering, or groundwater. Consequently, design must distinguish between the properties of intact rock specimens and the behaviour of the rock mass as a geological structure. The central engineering task is to develop a defensible ground model, select suitable analytical or numerical methods, and manage uncertainty through investigation, monitoring, and observational construction.

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Distinguish intact rock properties from rock-mass properties.

2. Characterize discontinuities and explain their influence on stability.

3. Apply major rock-mass classification systems critically.

4. Use Mohr–Coulomb and Hoek–Brown criteria within their limitations.

5. Assess foundations, slopes, tunnels, and underground excavations in rock.

6. Select excavation methods and support systems.

7. Explain observational construction and monitoring-based design.

8. Evaluate uncertainty, risk, sustainability, and professional responsibilities in rock engineering.

1. Geological and Mechanical Description of Rock

1.1 Intact rock and rock mass

Intact rock is the relatively unfractured material between discontinuities. A rock mass consists of intact rock blocks plus discontinuities and the infilling, weathering, groundwater, and stress conditions associated with them. Laboratory tests generally measure intact rock, whereas engineering performance is usually governed by the rock mass.

Important rock types include igneous rocks such as granite and basalt, sedimentary rocks such as sandstone and limestone, and metamorphic rocks such as schist, gneiss, and marble. Mineral composition, grain size, fabric, cementation, porosity, degree of alteration, and geological history influence mechanical behaviour.

1.2 Essential intact-rock properties

The principal properties are:

• Unit weight and density: relevant to overburden stress and gravitational loading.
• Porosity and permeability: control storage and groundwater movement.
• Uniaxial compressive strength (UCS): the peak axial stress sustained by an unconfined cylindrical specimen.
• Point-load strength index: a rapid index test often correlated cautiously with UCS.
• Tensile strength: commonly obtained from the Brazilian indirect tensile test.
• Triaxial strength: describes the effect of confining pressure.
• Elastic modulus and Poisson’s ratio: describe deformation under approximately elastic loading.
• Durability and slake durability: indicate resistance to deterioration during wetting and drying.
• Anisotropy: directional variation caused by foliation, bedding, schistosity, or preferred mineral orientation.

Laboratory results must be interpreted with attention to specimen size, saturation, loading rate, end restraint, sampling disturbance, and orientation relative to fabric. Small, high-quality specimens commonly provide higher strengths than the equivalent rock mass.

1.3 Stress and deformation around excavations

The in-situ stress state includes vertical stress, horizontal stresses, tectonic stresses, and local stress concentrations. A preliminary estimate of vertical stress is:

σv = γH

where σv is vertical stress, γ is unit weight, and H is depth below ground surface. This relationship assumes approximately uniform overburden and does not include tectonic or topographic effects. Horizontal stress cannot reliably be inferred from depth alone; it should be measured or estimated from geological and regional evidence.

Excavation removes confinement and redistributes stress. Depending on rock structure and stress level, the response may include elastic deformation, joint opening, block movement, brittle spalling, squeezing, swelling, or rock burst. The excavation geometry, orientation, size, sequence, and rate of advance are therefore design variables.

2. Discontinuities and Rock-Mass Structure

Discontinuities include joints, faults, bedding planes, foliation, cleavage, shear zones, and contacts. Their engineering description should include:

• orientation: dip and dip direction;
• spacing and persistence;
• aperture and roughness;
• wall strength and weathering;
• infilling material and thickness;
• seepage and water pressure;
• number of joint sets and block size; and
• evidence of movement or mineralization.

2.1 Orientation and kinematic significance

A discontinuity’s orientation must be compared with the excavation or slope geometry. In slopes, planar failure is possible when a persistent plane daylights in the slope face and has insufficient shear resistance. Wedge failure may occur where two discontinuity planes intersect and the intersection line plunges out of the face. Toppling is associated with steep discontinuities dipping back into or approximately parallel to a slope.

In tunnels, adverse joint orientations can produce wedges at the crown, walls, or portal. The stereographic projection is a useful tool for identifying kinematic feasibility, but it does not by itself establish stability. Shear strength, persistence, groundwater, block size, and support must also be evaluated.

2.2 Shear strength of discontinuities

For an idealized planar discontinuity, the Mohr–Coulomb form is:

τ = c + σ′n tan φ

where τ is shear strength, c is cohesion, σ′n is effective normal stress, and φ is the friction angle. Effective normal stress is approximately:

σ′n = σn − u

where σn is total normal stress and u is water pressure. The equation assumes a linear strength envelope over the stress range considered. It is not automatically appropriate for rough, interlocked, dilatant joints.

Rough joints may exhibit peak strength greater than residual strength because asperities interlock and must be sheared or overridden. A commonly used conceptual description is the Barton–Bandis approach, in which joint strength depends on joint roughness coefficient, joint wall strength, normal stress, and residual friction. Roughness decreases with weathering, displacement, and scale. Laboratory-scale roughness should not be transferred directly to a large fault or bedding plane.

Groundwater reduces effective normal stress, may soften infilling, and can create seepage forces. Drainage, grouting, lining, or water-pressure control may therefore be as important as mechanical support.

3. Rock-Mass Classification Systems

Classification systems organize field observations into empirical estimates of ground quality and support requirements. They are valuable for communication and preliminary design but must not replace engineering analysis.

3.1 Rock Quality Designation

Rock Quality Designation (RQD) is the percentage of recovered intact core pieces longer than 100 mm, measured along the core run:

RQD = 100 × (length of qualifying core pieces / total core-run length)

RQD is sensitive to drilling quality, core orientation, and the selected run length. It describes fracture frequency along the borehole, not the full three-dimensional block structure. It should therefore be interpreted with joint surveys and logging of weathering and infilling.

3.2 Rock Mass Rating

The Rock Mass Rating (RMR) system combines ratings for UCS, RQD, discontinuity spacing, discontinuity condition, groundwater, and orientation adjustment. The total rating classifies rock mass quality and supports preliminary estimates of stand-up conditions and support. Its usefulness depends on consistent logging and correct interpretation of discontinuity conditions. RMR is not a universal factor of safety and should not be used outside the range of conditions for which its correlations were developed without engineering judgment.

3.3 Q-system

The Q-system expresses rock quality using block size, joint condition, and water or stress effects. In its classical form:

Q = (RQD/Jn)(Jr/Ja)(Jw/SRF)

where Jn represents joint-set number, Jr joint roughness, Ja joint alteration, Jw water or stress-related water reduction, and SRF the stress reduction factor. The system is particularly associated with tunnel support selection, but its empirical correlations require appropriate geological and excavation contexts.

3.4 Geological Strength Index

The Geological Strength Index (GSI) estimates rock-mass structure and discontinuity surface condition. It is used with the Hoek–Brown criterion to derive rock-mass strength and deformation parameters. GSI is most reliable when based on a clear description of blockiness and surface condition rather than a mechanically adjusted number chosen to produce a desired design outcome.

4. Failure Criteria and Rock-Mass Strength

4.1 Mohr–Coulomb criterion

For effective stresses, the Mohr–Coulomb failure condition is:

τf = c′ + σ′n tan φ′

where τf is shear strength at failure, c′ is effective cohesion, σ′n is effective normal stress, and φ′ is effective friction angle. The criterion is simple and convenient for limit-equilibrium analysis, but real rock strength envelopes are often nonlinear, particularly over a broad range of confining stress.

4.2 Hoek–Brown criterion

For intact rock and rock masses, the generalized Hoek–Brown criterion is commonly written:

σ′1 = σ′3 + σci (mb σ′3/σci + s)^a

where σ′1 and σ′3 are major and minor effective principal stresses at failure, σci is intact-rock UCS, and mb, s, and a are rock-mass parameters. These parameters depend on the intact material constant mi, GSI, and disturbance factor D. The parameters must be selected consistently with the adopted version of the criterion and the quality of the site data.

The criterion is useful for nonlinear rock-mass strength and numerical modelling. It is not a substitute for structural analysis where failure is controlled by persistent discontinuities. A heavily jointed rock mass may fail by block release at stresses far below the intact-rock strength.

4.3 Brittle failure and squeezing

At high stress relative to rock strength, brittle spalling may develop around an excavation. In very weak or highly deformable rock, squeezing can impose large convergence and support loads. Swelling may occur when clay-bearing minerals absorb water or when anhydrite and other minerals undergo volume change. These mechanisms require different support and construction responses; stronger bolts alone may be ineffective in squeezing ground if the support cannot accommodate deformation.

5. Foundations on Rock

Foundations on rock require verification of bearing resistance, settlement, sliding, overturning, discontinuity-controlled failure, and groundwater effects. The apparent high strength of rock must not lead to uncritical use of high allowable pressures.

Critical questions include:

1. Is the founding material intact, weathered, fractured, or sheared?

2. Are adverse bedding or joint planes present beneath the foundation?

3. Can blocks slide, rotate, or topple under combined loading?

4. Is the rock susceptible to dissolution, scour, swelling, or degradation?

5. Will excavation, blasting, or grouting alter the foundation conditions?

For a block resting on an inclined discontinuity, a simplified sliding check compares driving shear force with available resistance. If the resultant normal force is N and the driving shear force is T, a basic factor of safety is:

FS = [cA + (N − uA) tan φ] / T

where c is discontinuity cohesion, A is potential sliding area, u is average water pressure, and φ is friction angle. This expression assumes a planar sliding surface and does not include three-dimensional block geometry, bolts, roughness-induced dilation, or earthquake forces unless separately incorporated.

Foundation design should use geological mapping, oriented core logging, geophysical methods where appropriate, boreholes, discontinuity shear testing, permeability testing, and direct observation during excavation. Dental concrete, rock removal, grouting, dowels, shear keys, drainage, and load redistribution may be required.

6. Rock Slopes

Rock-slope design combines geological structure, slope geometry, groundwater, seismicity, weathering, and construction effects. Potential mechanisms include planar sliding, wedge sliding, flexural or block toppling, ravelling, and circular or compound failure in heavily weathered or soil-like rock.

A kinematic analysis identifies whether a mechanism is geometrically possible. Stability analysis then incorporates shear strength, water pressure, block weight, seismic loading, and support. Limit-equilibrium methods are useful for planar and wedge failures, while numerical methods may be required for complex interaction, deformation, progressive failure, or coupled groundwater effects.

Slope stabilization measures include scaling, trimming, drainage holes, surface protection, rock bolts, cable anchors, shotcrete, mesh, buttresses, retaining structures, and controlled blasting. Drainage is often the most cost-effective intervention because water pressure directly reduces effective normal stress.

7. Tunnels and Underground Excavations

7.1 Ground behaviour and excavation response

Tunnel behaviour is controlled by the interaction among rock mass, stress field, excavation geometry, excavation method, and support. Possible ground classes include stable self-supporting ground, structurally controlled blocky ground, squeezing ground, swelling ground, flowing ground, and rock-burst-prone ground.

The excavation creates a disturbed or yielded zone around the opening. Support must control loosening and deformation without assuming that the support carries all overburden weight. In many designs, the rock mass itself contributes to load bearing if deformation is controlled and a stable ground-support interaction is developed.

7.2 Excavation methods

Drill-and-blast is flexible and suitable for hard or variable rock. It requires blast design, perimeter control, ventilation, scaling, and management of vibration and overbreak. Poor blasting can damage the remaining rock mass and increase support demand.

Mechanical excavation includes roadheaders, continuous miners, tunnel-boring machines, and excavators with breakers. Roadheaders are effective in weak to medium-strength and variable ground; hard-rock TBMs provide high production in relatively uniform conditions but are less flexible during severe geological change. Selection should consider strength, abrasivity, jointing, groundwater, tunnel diameter, alignment, access, spoil handling, and project duration.

7.3 Support systems

Common support elements are:

• rock bolts and fully grouted anchors;
• cable bolts for longer reinforcement;
• steel fibre or mesh-reinforced shotcrete;
• steel ribs or lattice girders;
• forepoling, spiles, pipe umbrellas, and face reinforcement;
• steel sets and lagging; and
• cast-in-place or segmental concrete linings.

Bolts reinforce and clamp blocks, mobilize confinement, and improve load transfer. Shotcrete provides surface retention and helps form a composite support shell. Drainage may reduce water pressure but must be designed to prevent erosion, clogging, or uncontrolled inflow.

Temporary support and permanent lining have different functions. The design should state required service life, fire and durability performance, waterproofing, inspection access, seismic demands, and consequences of failure.

8. Observational Construction and Monitoring

Rock engineering contains unavoidable uncertainty. The observational method provides a structured approach rather than an informal decision to “watch and wait.” It requires:

1. a ground model and preliminary design;

2. identification of uncertainties and plausible behaviours;

3. predicted ranges of deformation, loading, and groundwater response;

4. measurable action levels;

5. instrumentation and reporting procedures;

6. preplanned responses to threshold exceedance; and

7. competent authority to modify construction.

Typical instruments include convergence stations, multipoint extensometers, load cells, piezometers, inclinometers, microseismic systems, prism surveys, radar, and fibre-optic sensors. Monitoring must be frequent enough to detect trends before instability develops. Data should be interpreted using rates, acceleration, spatial patterns, and correlation with excavation stages—not isolated readings alone.

The method is valid only where intervention remains feasible and safe. It is not a justification for inadequate investigation, absent contingency planning, or continued excavation after warning signs.

9. Applied Case: Illustrative Mountain Tunnel

Background

Consider an illustrative 6.2 m diameter road tunnel through alternating schist and quartzite at a mountain site. The tunnel alignment intersects two joint sets and a weathered fault zone. Seasonal groundwater inflow is expected.

Problem

The initial design proposed uniform rock bolts and shotcrete based on an average classification rating. Probe drilling during excavation revealed that the fault zone was more persistent than indicated by surface mapping. Crown convergence increased rapidly, and seepage softened clay-rich infilling.

Analysis

Mapping identified a wedge-shaped crown block formed by the intersection of the two joint sets. Oriented core and face mapping revised the structural model. Piezometer readings showed elevated water pressure near the fault. The original average classification obscured the local adverse structure; therefore, a discrete wedge analysis and deformation monitoring were added.

Approach and solution

Excavation rounds were shortened, perimeter blasting was controlled, and probe drilling was extended ahead of the face. The fault zone was pregrouted where inflow threatened face stability. Systematic fully grouted bolts, fibre-reinforced shotcrete, steel ribs, and forepoling were installed in the weakest zone. Drainage was introduced in a controlled manner, and convergence action levels were defined: increased inspection at the first threshold, reduced advance and additional reinforcement at the second, and evacuation or stabilization at the critical threshold.

Outcome

Convergence stabilized after shorter rounds and installation of the modified support. The permanent lining was designed after deformation had substantially reduced and groundwater conditions had been confirmed. The project experienced additional support cost but avoided uncontrolled collapse and major delay.

Learning points

The case demonstrates that average rock-mass ratings cannot replace structural mapping; groundwater can transform a stable block problem into a progressive failure; probe drilling reduces geological uncertainty; and monitoring is effective only when linked to predetermined decisions.

10. Technology, Sustainability, and Professional Governance

Three-dimensional geological modelling, LiDAR, photogrammetry, digital core logging, automated convergence monitoring, and numerical modelling improve data integration. Machine-learning systems can assist fracture detection, image classification, and anomaly identification, but they require representative training data, validation, uncertainty reporting, and professional review. An algorithm must not silently replace geological judgment.

Sustainable rock engineering seeks to reduce overbreak, spoil generation, cement and steel consumption, groundwater disturbance, noise, vibration, and energy use. Reuse of excavated rock should be assessed for mineralogy, durability, contamination, and grading rather than assumed suitable.

Professional responsibilities include protection of workers and the public, accurate reporting of uncertainty, independent review for high-consequence works, compliance with applicable regulations, blast and groundwater controls, emergency planning, and preservation of auditable design records. Geological information should not be selectively presented to support an unsafe or predetermined construction decision.

Challenges and Limitations

Rock engineering remains difficult because:

• geological conditions vary sharply over short distances;
• boreholes sample only a small fraction of the rock mass;
• discontinuity persistence is difficult to establish;
• scale effects invalidate direct transfer of laboratory strength;
• groundwater conditions change seasonally and during excavation;
• classification correlations are empirical and context-dependent;
• numerical models depend strongly on constitutive assumptions and input quality; and
• construction can alter the ground before its properties are fully understood.

Accordingly, designs should use multiple lines of evidence, sensitivity analyses, conservative assumptions where consequences are severe, and continuous updating as new ground is exposed.

Best-Practice Design Workflow

A defensible workflow is:

1. define the performance requirements and failure consequences;

2. develop a geological and hydrogeological model;

3. plan investigations to test the most consequential uncertainties;

4. characterize intact rock, discontinuities, stress, and groundwater;

5. identify credible failure mechanisms;

6. select suitable analytical, empirical, and numerical methods;

7. design excavation sequence, support, drainage, and contingency measures;

8. review constructability, safety, environmental effects, and durability;

9. monitor and compare actual behaviour with predictions; and

10. update the design and record the basis for all changes.

Chapter Summary

Rock mechanics requires simultaneous consideration of intact material, discontinuities, rock-mass structure, stress, groundwater, excavation method, and support interaction. Intact-rock laboratory strength is not equivalent to rock-mass strength. Discontinuity orientation, persistence, roughness, infilling, and water pressure commonly govern foundations, slopes, and tunnels. RQD, RMR, Q, and GSI provide useful frameworks but must be applied within their limitations.

Mohr–Coulomb analysis is convenient for linear shear resistance, while Hoek–Brown represents nonlinear rock-mass strength; neither replaces structural or kinematic assessment. Safe underground construction depends on suitable excavation methods, staged support, groundwater management, monitoring, and a disciplined observational method. Modern digital tools and AI can improve characterization, but professional judgment, transparent uncertainty, and safety governance remain essential.

Key Terms

Intact rock: Relatively unfractured rock material between discontinuities.

Rock mass: Intact rock together with its discontinuities, weathering, groundwater, and structure.

Discontinuity: A fracture, joint, fault, bedding plane, foliation, or other mechanical break.

RQD: Percentage of a core run consisting of intact pieces longer than 100 mm.

RMR: Empirical Rock Mass Rating classification system.

Q-system: Empirical tunnel classification based on block size, joint condition, and stress or water effects.

GSI: Geological Strength Index describing rock-mass structure and surface condition.

UCS: Uniaxial compressive strength of intact rock.

Effective stress: Total stress minus pore-water pressure in the relevant normal-stress context.

Mohr–Coulomb criterion: Linear relationship between shear strength, cohesion, and effective normal stress.

Hoek–Brown criterion: Nonlinear failure criterion for intact and jointed rock masses.

Persistence: The extent to which a discontinuity continues through the rock mass.

Roughness: Geometric irregularity of a discontinuity surface.

Squeezing: Time-dependent or stress-induced closure of an excavation in deformable ground.

Rock burst: Violent brittle failure and ejection of rock caused by high stored stress.

Overbreak: Excavation beyond the intended profile.

Forepoling: Forward-projected reinforcement installed ahead of a tunnel face.

Observational method: Design approach using predictions, monitoring, action levels, and planned responses.

Convergence: Closure or relative inward movement of an underground opening.

Review Questions

Short Answer Questions

1. Distinguish between intact rock and a rock mass.

2. What information should be recorded when describing a discontinuity?

3. State the principal limitations of RQD.

4. Why does groundwater reduce discontinuity stability?

5. What is the purpose of convergence monitoring in a tunnel?

Long Answer Questions

1. Explain how discontinuity orientation and persistence influence rock-slope failure mechanisms.

2. Compare RMR, Q, and GSI in terms of purpose, inputs, and limitations.

3. Explain the assumptions and applications of the Mohr–Coulomb and Hoek–Brown criteria.

4. Discuss the selection of excavation and support methods for a tunnel in variable rock.

5. Describe a complete observational-construction framework for underground engineering.

Analytical / Application Questions

1. A foundation block has potential sliding area 20 m², cohesion 50 kPa, effective normal force 2,000 kN, average water pressure 30 kPa, friction angle 30°, and driving shear force 900 kN. Using the simplified planar-sliding equation, calculate the factor of safety and comment on its meaning.

2. A proposed tunnel crosses a persistent fault with clay infilling and high seasonal groundwater. Develop an investigation and support strategy, identifying at least three uncertainties and corresponding monitoring measures.

3. A rock slope contains two joint sets whose intersection daylights in the slope face. Explain the sequence of kinematic, strength, groundwater, and support analyses needed before construction.

Multiple Choice Questions with Answers

Question 1

Which feature most directly distinguishes a rock mass from intact rock?

A. Mineral colour

B. Discontinuities and their properties

C. Specimen diameter

D. Laboratory temperature

Correct Answer: B. Discontinuities and their properties

Question 2

In the RQD calculation, which core pieces are normally counted?

A. Pieces longer than 50 mm

B. All recovered pieces

C. Intact pieces longer than 100 mm

D. Only pieces containing joints

Correct Answer: C. Intact pieces longer than 100 mm

Question 3

In the Mohr–Coulomb equation, effective normal stress is approximately:

A. Total normal stress plus water pressure

B. Total normal stress minus water pressure

C. Water pressure minus cohesion

D. Shear stress minus friction angle

Correct Answer: B. Total normal stress minus water pressure

Question 4

Which failure mechanism is associated with the intersection of two discontinuity planes?

A. Wedge failure

B. Pure compression

C. Uniform settlement

D. Hydraulic heave only

Correct Answer: A. Wedge failure

Question 5

Which parameter is included in the classical Q-system expression?

A. Joint roughness coefficient Jr

B. Poisson’s ratio only

C. Concrete slump

D. Blast powder factor only

Correct Answer: A. Joint roughness coefficient Jr

Question 6

The Hoek–Brown criterion is particularly useful because it represents:

A. Only groundwater flow

B. A nonlinear rock-mass strength envelope

C. Only foundation settlement

D. Tunnel ventilation demand

Correct Answer: B. A nonlinear rock-mass strength envelope

Question 7

For a planar discontinuity, increasing water pressure generally:

A. Increases effective normal stress

B. Reduces effective normal stress and shear resistance

C. Eliminates cohesion effects in all cases

D. Increases intact-rock UCS

Correct Answer: B. Reduces effective normal stress and shear resistance

Question 8

Which method is most closely associated with flexible excavation in variable hard rock?

A. Drill-and-blast

B. Uncontrolled hydraulic filling

C. Surface paving

D. Passive ventilation only

Correct Answer: A. Drill-and-blast

Question 9

A valid observational method requires:

A. Monitoring without action levels

B. Predictions, measurements, and preplanned responses

C. No preliminary design

D. Use of one instrument only

Correct Answer: B. Predictions, measurements, and preplanned responses

Question 10

If c = 50 kPa, A = 20 m², N = 2,000 kN, u = 30 kPa, φ = 30°, and T = 900 kN, the simplified planar-sliding factor of safety is closest to:

A. 0.95

B. 1.35

C. 1.87

D. 2.75

Correct Answer: C. 1.87