Foundation bearing capacity testing determines whether soil and rock strata can safely support structural loads, and numerical analysis has become an indispensable tool for interpreting load test data, extrapolating results, and verifying design assumptions. This article explains the testing standards, sample requirements, laboratory and in-situ testing items, and the finite element and limit analysis techniques applied by accredited third-party laboratories and geotechnical engineers. A clear understanding of these workflows helps project owners control geotechnical risk, optimize foundation design, and satisfy code acceptance requirements.
Technical Overview
Foundation bearing capacity refers to the maximum load that a soil or rock stratum can sustain beneath a footing, pile, or mat foundation without undergoing shear failure or producing settlement that would compromise structural serviceability. Traditional assessment relies on empirical bearing capacity equations developed by Terzaghi, Meyerhof, Hansen, and Vesic, which express capacity as a function of soil unit weight, cohesion, internal friction angle, foundation width, embedment depth, and shape factors. While these closed-form solutions remain the basis of most design codes, they cannot fully capture complex soil stratification, layered deposits, inclined or eccentric loading, soil-structure interaction, or progressive failure mechanisms.
Numerical analysis of bearing capacity complements physical testing by constructing a mathematical model of the foundation-soil system and solving it under incremental loading. The dominant techniques include the finite element method (FEM), the finite difference method (FDM), and upper- and lower-bound limit analysis. When calibrated against field plate load tests, penetration tests, and laboratory strength data, these models allow engineers to predict ultimate bearing capacity, load-settlement behavior, and failure mechanisms with a level of detail that classical equations cannot provide. Modern practice, as encouraged by standards such as EN 1997-1 (Eurocode 7), GB 50007, and ASTM guidance documents, treats numerical analysis as a verification tool that must always be anchored in measured geotechnical parameters.
Three classical failure modes guide the analysis: general shear failure, characterized by a well-defined wedge and slip surfaces reaching the ground surface; local shear failure, in which slip surfaces remain localized beneath the footing; and punching shear failure, typical of loose sands and soft clays. Numerical models reproduce these mechanisms and identify the governing mode for a given geometry and soil profile, which directly influences how the measured or computed capacity is factored into design.
Testing Samples
Undisturbed Soil Samples
Undisturbed samples are recovered using thin-walled Shelby tubes, piston samplers, or block sampling techniques in accordance with ASTM D1587 and ISO 22476 series sampling guidance. These samples preserve the in-situ fabric, density, and stress history of the soil, making them essential for triaxial and oedometer testing that supplies the strength and stiffness parameters required for numerical modeling. Sample quality is classified by disturbance indices, and only high-quality specimens should be used to derive parameters for bearing capacity analysis.
Remolded and Compacted Samples
For engineered fill, embankment material, and compacted subgrades, remolded specimens are prepared at target moisture content and dry density following compaction standards such as ASTM D698 and ASTM D1557. These specimens are used to establish the relationship between placement conditions and achievable bearing capacity, which is critical for quality control during construction and for calibrating numerical models of compacted ground.
Rock Cores
Where foundations bear on rock, NX or larger diameter cores are extracted for unconfined compression testing per ASTM D7012 and point load testing per ASTM D5731. Core recovery, rock quality designation (RQD), and discontinuity spacing are recorded because jointing and weathering frequently control the effective bearing stratum more than intact rock strength. Numerical models for rock masses typically incorporate Hoek-Brown strength parameters derived from these measurements.
In-Situ Test Locations and Test Plates
Field bearing capacity verification requires representative test positions selected at the proposed foundation bearing level. Plate load tests use rigid circular or square steel plates, commonly 300 mm to 1000 mm in diameter, seated on a prepared bedding surface. The plate assembly, reaction system, and loading points constitute the physical "sample" for in-situ capacity evaluation, and their proper preparation directly determines the quality of data used for numerical back-analysis.
Testing Items
The scope of a bearing capacity investigation combines field measurements, laboratory strength testing, and derived numerical results. Core testing items include:
- Ultimate bearing capacity at the tested or modeled bearing stratum
- Allowable (characteristic) bearing capacity after applying code safety factors
- Cohesion and internal friction angle for effective stress and total stress analysis
- Deformation modulus and compressibility parameters for settlement prediction
- Subgrade reaction modulus for structural foundation interaction analysis
- Load-settlement curve characteristics, including yield point and plastic behavior
- Bearing capacity factors and shape, depth, and inclination corrections
- Factor of safety against shear failure under design load combinations
- Predicted total and differential settlement within the structural serviceability limits
Ultimate and Allowable Bearing Capacity
The ultimate bearing capacity is the pressure at which the soil reaches a shear failure mechanism, identified from a load test by the point where settlement continues rapidly under constant or decreasing load, or by an agreed settlement criterion such as a plate settlement equal to 10 percent of its width. The allowable capacity is obtained by dividing the ultimate value by a safety factor, typically 2.5 to 3.0 in limit-based design, or by verification within a partial factor framework as specified in Eurocode 7 and GB 50007.
Shear Strength Parameters
Cohesion (c) and friction angle (φ) are the primary inputs for both classical equations and numerical constitutive models. Consolidated undrained (CU), consolidated drained (CD), and unconsolidated undrained (UU) triaxial tests per ASTM D4767 and ISO 17892-9, direct shear tests per ASTM D3080, and unconfined compression tests per ASTM D2166 provide these parameters under drainage conditions matching the construction and loading scenario being analyzed.
Deformation and Settlement Characteristics
Because most modern foundations are governed by settlement rather than shear failure, compressibility parameters carry equal weight to strength. One-dimensional consolidation testing per ASTM D2435 and ISO 17892-5 yields compression and recompression indices, while plate load test curvature supplies a field-scale deformation modulus. Numerical analysis integrates these values to predict settlement-time histories, which are compared against serviceability limits defined by the structural designer.
Testing Methods
Static Plate Load Testing
The static plate load test, performed per ASTM D1194 for bearing capacity or ASTM D1196 for modulus determination, is the most direct field verification method. Load is applied in increments through a hydraulic jack reacting against kentledge or anchored frames, with settlement recorded by dial gauges or displacement transducers at each stage. The test proceeds until failure, an agreed settlement limit, or the maximum design load multiplied by a verification factor. The resulting load-settlement curve is the primary dataset for numerical model calibration.
In-Situ Penetration Testing
Cone penetration testing per ASTM D5778 and ISO 22476-1 provides continuous cone resistance and sleeve friction profiles, from which bearing capacity is estimated using established correlations. Standard penetration testing per ASTM D1586 and EN ISO 22476-3 supplies N-values that correlate with density and allowable pressure. Pressuremeter testing per ASTM D4719 and flat dilatometer testing per ASTM D6635 measure in-situ stiffness and limit pressure directly. These profiles define soil stratification and parameter variability for the numerical model geometry.
Laboratory Strength Testing
Laboratory programs combine triaxial compression, direct shear, unconfined compression, and consolidation testing on representative samples to establish the parameter set used in analysis. Multi-stage testing and stress-path testing may be employed where loading conditions deviate significantly from conventional axial compression. Index testing, including Atterberg limits and grain size distribution per ASTM D4318 and ASTM D6913, supports soil classification and correlation checks on the derived parameters.
Finite Element Analysis
Finite element analysis discretizes the foundation and surrounding soil into a mesh of elements, applying constitutive models such as Mohr-Coulomb, Drucker-Prager, modified Cam-Clay, or hardening soil models. Loading is applied incrementally, and the analysis tracks stress redistribution, plastic zone development, and settlement until a failure mechanism emerges or displacement criteria are met. Key modeling decisions include mesh refinement near the footing edge, boundary distance to eliminate artificial confinement, interface elements between footing and soil, and the simulation of construction sequence and embedment. Convergence studies verify that the computed ultimate load is mesh-independent.
Limit Equilibrium and Limit Analysis
Limit equilibrium methods assume an assumed failure surface and solve force or moment equilibrium to yield a capacity estimate, aligning closely with classical bearing capacity theory. Rigorous numerical limit analysis computes upper and lower bounds on the true collapse load using finite element formulations with plasticity constraints. The bracket between the bounds provides an objective measure of solution uncertainty and is increasingly used to verify both classical equations and elastoplastic FEM results.
Back-Analysis and Model Validation
In professional practice, the numerical model is validated by back-analyzing the measured plate load test curve: model parameters are adjusted within laboratory-determined ranges until the simulated load-settlement response matches field observations. The validated model is then used to extrapolate capacity to the actual foundation dimensions, embedment depths, and load combinations, since scale effects mean that a plate test result cannot be transferred directly to a full-scale footing without such analysis. This combined field-numerical workflow satisfies the verification principles embedded in GB 50007, Eurocode 7, and related national codes.
Applications
Numerical analysis of foundation bearing capacity testing supports a wide range of engineering and regulatory activities:
- Verification of allowable bearing pressure for shallow footings of buildings and industrial structures
- Design and acceptance of compacted subgrades for highways, airports, and rail corridors
- Assessment of bearing strata for bridge piers, abutments, and retaining wall foundations
- Scale extrapolation of plate load test results to full-scale mat and raft foundations
- Capacity evaluation for heavily loaded silos, cranes, and machine foundations with eccentric or dynamic loads
- Settlement prediction for foundations on layered soils and soft clay deposits
- Checking the effect of adjacent excavation, dewatering, or surcharge loading on existing foundations
- Optimization of foundation size and embedment to reduce construction cost while meeting code factors of safety
- Technical documentation for design review, construction approval, and dispute resolution
FAQ
What is the difference between ultimate and allowable bearing capacity?
Ultimate bearing capacity is the pressure at which the soil fails in shear or reaches a defined failure settlement criterion. Allowable bearing capacity is the ultimate value divided by a global safety factor, or reduced through partial factors in limit state design, so that adequate margins against both collapse and excessive deformation are maintained throughout the structure's service life.
How reliable is numerical analysis compared with field load tests?
Numerical analysis is a prediction tool whose reliability depends entirely on the quality of input parameters and model validation. When a model is calibrated against a measured plate load test curve and parameter values are constrained by laboratory testing, predicted capacities for the actual foundation typically fall within a narrow band of the verified values. Without such calibration, numerical results should be treated as indicative and verified by physical testing at critical locations.
Which constitutive model should be selected for bearing capacity analysis?
Mohr-Coulomb is the standard baseline for drained analysis of sands and normally consolidated clays. Soft and sensitive clays under undrained loading are better represented by total stress models or modified Cam-Clay, while heavily overconsolidated soils and staged excavation problems benefit from hardening soil models that capture stress-dependent stiffness. Rock foundations require Hoek-Brown or equivalent discontinuum approaches where jointing controls behavior.
What safety factors are required by design codes?
Conventional allowable stress design applies global safety factors of approximately 2.5 to 3.0 against shear failure. Eurocode 7 instead applies partial factors to soil parameters and actions in its design approaches, while GB 50007 specifies characteristic values adjusted by empirical modification factors with typical overall margins in the same range. The governing code for the project location determines the exact verification format.
Why can a plate load test result not be used directly for a large footing?
Plate load tests engage only a shallow and limited stress bulb, while a full-scale footing influences a much deeper soil volume that may include weaker layers. Scale effects on both strength mobilization and settlement mean the measured value must be corrected or, preferably, extended through validated numerical analysis that models the actual foundation geometry and the full soil profile.