Bridge Analysis Workflow

Learning Objectives

  • Convert a physical bridge into an analysis model with explicit assumptions, supports, member properties, and load paths.
  • Distinguish global bridge analysis from local component analysis.
  • Generate dead-load, moving live-load, environmental, and extreme-event load cases.
  • Use influence lines or automated vehicle sweeps to obtain response envelopes.
  • Apply appropriate transverse load distribution or refined analysis.
  • Combine force effects at the correct limit states without double counting.
  • Perform equilibrium, compatibility, sensitivity, and reasonableness checks before design.
  • Produce traceable design actions for prestressed-concrete and steel plate-girder design.

Bridge analysis is a workflow, not a single calculation. The engineer must decide what physical behavior the model needs to capture, generate the correct loading cases, determine critical vehicle positions, extract envelopes, distribute demand to components, combine effects, and verify that the numerical results remain consistent with structural mechanics.

Select Model Fidelity Deliberately

Escalation ladder

  1. Start with hand checks or a line-girder model when the response is primarily longitudinal and load sharing is well understood.
  2. Use a grid or frame model when transverse distribution, cross-frame action, support skew, or system interaction matters.
  3. Escalate to shell or refined finite-element modeling when deck participation, torsion, distortion, local stiffness, or unusual geometry controls.
  4. Use local submodels only where detail-level stress or deformation is required; do not make the entire bridge needlessly complex.

Fidelity is response-specific

The best model is the simplest model that represents the stiffness, load path, and boundary conditions governing the response under study. More elements do not automatically mean more accuracy.

1. Define the Analysis Question

Global Analysis

Global analysis determines support reactions, span moments, shears, torsion, displacements, and overall load distribution. Typical models include line-girder, grillage, frame, or finite-element models.

Local Analysis

Local analysis resolves behavior that the global model intentionally idealizes: deck-strip effects, bearing zones, diaphragms, cross-frames, anchorage zones, local plate buckling, connections, or stress concentrations.

Model only what the decision requires

Greater model complexity does not guarantee greater accuracy. Each degree of freedom, release, stiffness, and element must correspond to a physical assumption that can be explained and checked.

2. Establish Geometry and Structural Idealization

Model-definition checklist

  1. Establish span lengths, skew, curvature, deck width, girder spacing, support elevations, and continuity.
  2. Define material properties and section properties appropriate to the analysis stage.
  3. Assign support restraints that match the actual articulation system.
  4. Represent diaphragms, cross-frames, bracing, or transverse stiffness at the fidelity required by the question.
  5. Define composite and non-composite stages separately where stiffness changes during construction.
  6. Document eccentricities, rigid offsets, releases, and any intentionally omitted behavior.

Boundary-Condition Audit

Checklist

3. Establish Analysis Stages

Construction Stage

Before the deck develops composite action, steel or precast girders may carry self-weight, wet concrete, formwork, equipment, and erection forces using a different stiffness and bracing condition from the final bridge.

Final Service Stage

The hardened deck, continuity details, barriers, wearing surfaces, utilities, and final articulation define the service model. Time-dependent effects may need a separate treatment for prestressed or staged concrete systems.

4. Build Load Cases

Keep fundamental actions separate so the model remains auditable. Typical categories include structural dead load, superimposed dead load, wearing surface and utilities, vehicular live load, pedestrian load, braking, centrifugal action, wind, temperature, water, earth pressure, seismic action, collision, and construction loads where applicable.

Do not combine too early

Store unfactored component load cases independently. Apply code-prescribed load factors and combinations only after the response from each action can be inspected and verified.

5. Solve Moving-Load Response

Moving-load envelope workflow

  1. Select response locations along each span and at critical supports.
  2. Sweep each required concentrated vehicle model across all relevant positions.
  3. Apply the distributed lane component to the response-sign regions required by the adopted loading rules.
  4. Calculate response for every vehicle position.
  5. Store maximum and minimum reaction, shear, moment, torsion, and displacement envelopes as required.
  6. Record the vehicle position that generated each governing envelope point.

6. Account for Transverse Distribution

Approximate Distribution

Code-based distribution factors or simplified lever-rule methods convert lane-level live loading into individual girder effects within stated applicability limits.

Refined Analysis

Grillage or finite-element analysis explicitly represents transverse stiffness and can be required or beneficial for significant skew, curvature, unusual spacing, complex cross-sections, or other conditions outside simplified-method assumptions.

Skew, curvature, and torsion are not cosmetic

Significant skew or curvature can change reaction distribution, cross-frame forces, torsion, and local flange behavior. If these effects can alter the governing load path, the model must represent them explicitly rather than treating the bridge as a straight orthogonal system.

7. Form Limit-State Combinations

Generic factored load effect

Represents a code-defined linear load combination while preserving the origin of each component effect.

U=∑iγiQiU=\sum_i \gamma_i Q_i

Variables

SymbolDescriptionUnit
UUFactored force effect for the limit state-
γi\gamma_iLoad factor assigned by the governing specification-
QiQ_iUnfactored response caused by action i-

For each design location, combinations should be assembled from response components—not from guessed total loads. Sign conventions matter: maximum positive moment, minimum negative moment, maximum shear, minimum reaction, and extreme displacement can be controlled by different load arrangements.

8. Verify the Model Before Designing Members

Minimum engineering checks

  1. Equilibrium: Sum reactions and compare with applied vertical, longitudinal, and transverse actions.
  2. Symmetry: For symmetric geometry and loading, confirm symmetric response where expected.
  3. Simple benchmark: Compare at least one model result with a hand solution for a simplified case.
  4. Deflected shape: Confirm displacement direction and continuity are physically plausible.
  5. Sensitivity: Vary uncertain stiffness or restraint assumptions and determine whether conclusions change materially.
  6. Envelope continuity: Check for numerical spikes caused by mesh, release, or moving-load discretization errors.
  7. Units and signs: Confirm force, length, stress, temperature, and axis conventions before exporting design actions.

9. Design Handoff

The output of bridge analysis should be a documented set of force effects for each component and limit state, not a screenshot of one software plot. For a girder, the handoff typically includes maximum/minimum moment and shear envelopes, reactions, torsion where relevant, fatigue stress-range demand, construction-stage effects, and the load combination producing each result.

Prestressed Concrete Handoff

Include construction/transfer conditions, service moments, factored flexure and shear, reactions, deflection/camber inputs, and stage-specific section properties.

Steel Plate-Girder Handoff

Include non-composite erection demand, composite service demand, positive/negative moment regions, shear envelopes, fatigue-sensitive stress ranges, bracing forces where modeled, and support reactions for bearing/stiffener design.

10. Reporting and Traceability

Checklist

Analysis Acceptance Gates

Release model results for member design only after these gates

  1. Geometry gate: dimensions, offsets, section assignments, skew/curvature, and construction stages match the design basis.
  2. Boundary-condition gate: support restraints, releases, springs, bearing directions, and eccentricities match the physical articulation.
  3. Loading gate: self-weight, DC/DW, live load, temperature, wind, braking, seismic, and construction loads are complete and not duplicated.
  4. Numerical gate: moving-load step size, response-station spacing, mesh density, and solver settings are sufficiently converged for the response of interest.
  5. Physical gate: equilibrium, symmetry where applicable, deflected shape, reaction direction, and load path are physically credible.
  6. Independent-check gate: at least one benchmark or hand solution reproduces a simplified subset of the model within an explainable tolerance.
  7. Traceability gate: every reported design action retains station, sign, stage, load combination, and controlling load arrangement.

Minimum Analysis QA Record

Checklist

Interactive visualization

The workflow simulator below lets you move through model definition, loading, vehicle sweep, envelope extraction, distribution, combinations, and verification while watching the analysis outputs become progressively design-ready.

Bridge analysis workflow

Concept and model scope

Trace how span geometry, support articulation, construction stage, and moving-load position become reactions, envelopes, and a checked design handoff.

This laboratory uses one representative simply supported span to connect a physical bridge to a line-girder or grillage idealization.

The selected view changes which physical assumptions are visible; every response value is calculated from the same span, stage, load, and restraint state.

The equations are an educational statics model, not a project design check. They omit nonlinear materials, local connection effects, wind, seismic action, and code load factors.

The scalar controls expose their own focused definitions and ranges from their highlighted labels. Physical bridge, line-girder, and grillage are alternative teaching views of the same deterministic span state.

Controls

Learning objective: manipulate the physical idealization and watch the same model drive geometry, response envelopes, and verification status.

Service stage active
120 m
12.0 m
2.4 m
6 girders
50%
700 kN
15%
250 kN
12×
Overlays

Shared model view

Camera presets expose the load path without unrestricted orbit controls.

A deterministic representative-span view showing service stage, 120 metre span, 12.0 metre deck width, 6 girders, and the selected response overlays. The deformed shape uses a labeled visual exaggeration.700 kN truckR₁ 4939 kNguided bearingR₂ 4939 kNguided bearingdeformed shape · visual scale 12×M envelope · scaled response overlay120 m representative span · 12.0 m deck width
structural modeldeformed response, scaledsupport reactionsmoving-load envelope

Direct results

Values update from the same deterministic model shown at left.

physical

Max positive moment

158664 kN·m

Moving-load maximum over 41 span stations and 41 truck positions.

Peak support shear

5289 kN

Largest absolute reaction-side shear in the same sweep.

Peak deflection

39.4 mm

Teaching screen: span / 800 = 150.0 mm.

Critical girder moment

24328 kN·m

Distribution factor: 0.153 of envelope.

Governing state

Illustrative response within limit

The representative span remains below the teaching deflection screen, and the normalized support flexibility is not the governing concern.

Service stage: the truck is active, and the selected restraint transfers the displayed braking action.

Left reaction
4939 kN
Right reaction
4939 kN
Longitudinal
45 kN
Effective inertia
29.4 m⁴

Workflow trace

Geometry → supports → loads → envelopes

The current stage and view keep the physical load path tied to the analytical response.

Equilibrium check

0.000 kN residual

The displayed vertical reactions sum to dead plus active live load.

Envelope control

41 response stations · 41 load positions

Use the response-charts preset to inspect positive, minimum, and selected-position traces.

Key Takeaways
  • The analysis model is a controlled abstraction of the physical bridge.
  • Keep load cases independent until limit-state combinations are deliberately formed.
  • Moving-load analysis must search vehicle position rather than assume it.
  • Approximate distribution methods have applicability limits; refined analysis is not automatically superior unless it is correctly modeled and checked.
  • Equilibrium, benchmark, deflected-shape, sensitivity, and sign/unit checks are mandatory before member design.
  • Good bridge analysis ends with traceable design actions that can be audited independently of the software used.