Geotechnical Modelling and Analysis

Table of contents

Geotechnical modelling and analysis involve the mathematical representation of soil and rock behaviour to predict how the ground will respond to applied loads, excavations, and environmental changes. It bridges the gap between site investigation data and engineering design — transforming borehole logs and lab test results into actionable design parameters.

What Is Geotechnical Modelling?

A geotechnical model (or ground model) is a simplified representation of the subsurface conditions at a site. It synthesises geological mapping, borehole logs, geophysical surveys, and laboratory test data into a coherent understanding of:

  • Stratigraphy — layer geometry and continuity
  • Material properties — strength, stiffness, permeability
  • Groundwater conditions — water levels, flow regime, pore pressures
  • Geological structure — faults, joints, bedding planes
  • Geohazards — reactive soils, acid sulfate soils, fill, mine voids

Levels of Modelling

Level Description When Used
Conceptual model Block diagram or cross-section showing interpreted geology Early project stages, feasibility
1D model Borehole-based soil profile with depth-dependent properties Simple foundation design, site classification
2D model Cross-sectional model along a specific alignment Slopes, embankments, retaining walls, tunnels
3D model Volumetric representation of the full site Complex geology, major infrastructure, BIM integration

The Ground Model Development Process

Step 1: Data Compilation

All available data is assembled:

  • Borehole logs and CPT traces
  • Geophysical survey interpretations
  • Laboratory test results
  • Topographic and LiDAR data
  • Geological maps and memoirs
  • Aerial photography and satellite imagery
  • Previous investigation reports

Step 2: Geological Interpretation

The data is interpreted to define:

  • Layer boundaries and thickness variations
  • Geological structures and discontinuities
  • Weathering profiles and alteration zones
  • Groundwater regimes

Step 3: Parameter Assignment

Characteristic design parameters are assigned to each material unit:

Parameter Source Application
Unit weight (γ) Lab measurement All stability and deformation analyses
Friction angle (φ') Triaxial / direct shear Strength-based analyses
Cohesion (c') Triaxial testing Strength-based analyses
Undrained shear strength (su) UU triaxial / shear vane Short-term stability in clays
Young's modulus (E) Triaxial / oedometer Settlement and deformation analysis
Poisson's ratio (ν) Triaxial / empirical Settlement and deformation analysis
Permeability (k) Lab / in-situ testing Seepage and dewatering analysis
Coefficient of consolidation (cv) Oedometer Time-rate of settlement

Step 4: Model Validation

The model is checked against:

  • Adjacent site data and published geological maps
  • Historical performance of nearby structures
  • Sensitivity analysis to key parameters

Common Analysis Methods

Limit Equilibrium Analysis

Used for slope stability, bearing capacity, and retaining wall design.

Method Description Best For
Bishop Simplified Circular slip surface, method of slices Homogeneous or layered slopes
Janbu Simplified Non-circular slip surfaces Irregular slopes, weak layers
Spencer Satisfies moment and force equilibrium All slope geometries
Morgenstern-Price Non-circular, arbitrary interslice force function Complex slope geometries
Sarma Non-vertical slice boundaries Blocky rock slopes

Output: Factor of Safety (FoS)

Finite Element Analysis (FEM)

Provides stress-strain analysis of the full soil-structure system.

Software Developer Key Features
Plaxis 2D / 3D Bentley Industry standard, soft soil model, hardening soil
RS2 / RS3 Rocscience 2D/3D, slope stability, tunnel support
ABAQUS Dassault Advanced constitutive models, coupled analysis
Z_SOIL Zace Services Unsaturated soil, dynamics
DIANA DIANA FEA Tunnel lining, soil-structure interaction

Typical outputs:

  • Displacement contours and vectors
  • Stress distributions (effective stress, pore pressure)
  • Plastic strain zones (failure progression)
  • Structural forces in retaining walls, piles, tunnel linings
  • Ground settlement troughs

Finite Difference Analysis

Software Developer Key Features
FLAC 2D / 3D Itasca Explicit solution, large deformation, dynamic analysis

Discrete Element Method (DEM)

For granular materials and jointed rock masses.

Software Developer Key Features
PFC 2D / 3D Itasca Particle-based, fracture propagation
UDEC / 3DEC Itasca Distinct element, jointed rock, blocky systems

Seepage Analysis

Modelling groundwater flow and pore pressure distribution.

Software Developer Key Features
SEEP/W GeoStudio Steady-state and transient, unsaturated flow
Plaxis (seepage) Bentley Coupled deformation-seepage
MODFLOW USGS Regional groundwater modelling
SVFlux SoilVision Unsaturated flow, climate boundary conditions

Consolidation and Settlement Analysis

Software Key Features
Settle3 1D and 3D consolidation, staged construction, PVDs
Plaxis 2D/3D Coupled consolidation, soft soil creep
GEO5 Settlement Immediate and consolidation settlement
CDS (Consolidation Design System) PVD design, staged preloading

Dynamic / Seismic Analysis

Software Key Features
DEEPSOIL 1D site response analysis, liquefaction triggering
Plaxis / FLAC Dynamic time-history, earthquake loading
OpenSees Open-source, research-grade seismic analysis
QUAKE/W Tailings dam seismic assessment

Constitutive Soil Models

Model Parameters Best For
Mohr-Coulomb E, ν, c', φ', ψ Simple elastic-perfectly plastic, general use
Drucker-Prager E, ν, α, k Simplified 3D Mohr-Coulomb
Modified Cam Clay λ, κ, M, e₀ Soft clay consolidation
Hardening Soil (HS) E₅₀, E_ur, E_oed, m, c', φ' All soils — best general-purpose model
HS Small (HSS) HS + G₀, γ₀.₇ Small-strain stiffness (excavations, tunnels)
Soft Soil Creep (SSC) λ*, κ*, μ*, M Long-term creep in soft soils
Hypoplastic Granular parameters Sands, cyclic loading
Barcelona Basic Unsaturated parameters Expansive and collapsible soils

Model Selection Guide

Problem Type Recommended Model
Simple bearing capacity Mohr-Coulomb
Retaining wall deflection HS Small
Deep excavation HS Small
Tunnel-induced settlement HS Small
Soft clay embankment Soft Soil Creep or Modified Cam Clay
Slope stability (FoS only) Mohr-Coulomb (with phi-c reduction)
Liquefaction assessment UBC3D-PLM or PM4Sand
Unsaturated soil Barcelona Basic or extended Mohr-Coulomb
Dynamic / cyclic Hypoplastic or PM4 models

Applications

Application Analysis Type Key Outputs
Shallow foundation Settlement + bearing capacity Allowable bearing pressure, settlement
Deep foundation (pile) Axial and lateral capacity Pile capacity, load-settlement curve
Retaining wall Earth pressure + structural Wall deflection, bending moment, anchor loads
Slope stability Limit equilibrium or FEM Factor of safety, failure surface
Excavation support FEM, soil-structure interaction Wall deflection, ground settlement
Tunnel design FEM or DEM Ground loss, lining forces, settlement
Embankment / dam Consolidation + stability Settlement, pore pressure, stability
Liquefaction Cyclic stress + site response CRR, CSR, liquefaction potential, settlement
Ground improvement FEM Composite ground behaviour, effectiveness
Seepage / dewatering Seepage analysis Flow rates, drawdown, pore pressure

Australian Standards

Standard / Guideline Relevance
AS 1726-2017 Geotechnical site investigations (data for modelling)
AS 4678-2002 Earth retaining structures — analysis requirements
AS 2870-2011 Residential slabs (design soil parameters)
AS 2159 Piling — analysis methods
AS 1170.4 Earthquake loads — site response analysis
AS 3798 Earthworks — compaction and fill parameters
TfNSW QA Specification Modelling requirements for road projects

Frequently Asked Questions

What is the difference between a ground model and a geotechnical model?

A ground model describes the geological conditions (layers, structure, groundwater). A geotechnical model assigns engineering properties to the ground model units and is used directly in analysis.

How many boreholes are needed for a good 3D model?

It depends on geological complexity. Simple sites: 1 borehole per 1,000–2,000 m². Complex sites (e.g., variable alluvium, faulted rock): 1 per 500–1,000 m². Geophysics can supplement sparse borehole coverage.

Can numerical modelling replace site investigation?

No. Site investigation provides the input parameters for modelling. A model built on poor data will produce unreliable results — "garbage in, garbage out."

What is the most common numerical modelling mistake?

Using over-simplified constitutive models with generic parameters. Always calibrate models against site-specific laboratory test data and use the simplest model that captures the essential behaviour.