Measurement uncertainty analysis is a fundamental requirement for the reliable determination of cement compressive and flexural strength, quantifying the confidence that can be placed in results obtained through standardized mortar prism testing. This article presents a systematic framework for identifying, quantifying, and combining uncertainty components in cement strength determination in accordance with ISO 679, EN 196-1, GB/T 17671, and the Guide to the Expression of Uncertainty in Measurement (ISO/IEC Guide 98-3). Accredited third-party laboratories apply these evaluations to support laboratory accreditation, proficiency testing, and conformity decisions against the strength classes defined in GB 175 and EN 197-1.
Technical Overview
Cement strength determination is the single most important test used to characterize the mechanical performance of hydraulic cements and to assign strength classes for commercial and structural use. The test measures the compressive and flexural strength of standardized mortar prisms prepared, cured, and loaded under strictly controlled conditions, and the reported result directly influences conformity assessment, acceptance decisions, and engineering design.
Because the measured value is shaped by a long chain of operations—sampling, proportioning, mixing, compaction, curing, specimen geometry, machine calibration, and loading—the reported strength is never an exact number. Measurement uncertainty, as defined in the GUM, is a parameter associated with the result that characterizes the dispersion of the values that could reasonably be attributed to the measurand. Quantifying this uncertainty is mandatory for laboratories accredited to ISO/IEC 17025 whenever statements of conformity are issued.
The evaluation follows four logical steps: formulation of a mathematical model, identification of all significant uncertainty sources, quantification of standard uncertainties by Type A (statistical) and Type B (other means) methods, and combination of the components into a combined and expanded uncertainty. The principal normative references are:
- ISO 679, EN 196-1, and GB/T 17671 — determination of strength of cement mortars using 40 mm × 40 mm × 160 mm prisms (ISO method)
- ASTM C109/C109M — compressive strength of hydraulic cement mortars using 50 mm cube specimens
- ISO/IEC Guide 98-3 (GUM) and JJF 1059.1 — evaluation and expression of uncertainty in measurement
- ISO/IEC 17025 — requirements for reporting results and documenting decision rules
- ISO 13528 — statistical methods supporting proficiency testing and interlaboratory comparisons
Testing Samples
Cement Types Covered
The uncertainty evaluation methodology applies to all hydraulic binders tested by the ISO mortar prism method, including:
- Portland cement and ordinary Portland cement (P·I, P·II, P·O; CEM I)
- Blended cements: slag, pozzolanic, fly ash, and composite cements (P·S, P·P, P·F, P·C; CEM II through CEM V)
- Special-purpose cements such as rapid-hardening, sulfate-resisting, and low-heat Portland cement
- Other binders tested with the same prism procedure, including masonry cement and white cement
Sampling and Specimen Preparation
Representative cement samples are taken and divided in accordance with GB/T 12573 or EN 196-7 to ensure the tested portion is characteristic of the consignment. The standard mortar consists of one part cement to three parts standard sand by mass at a water-to-cement ratio of 0.50, i.e., 450 g of cement, 1350 g of graded standard sand, and 225 g of water. The components are mixed in a planetary mixer, compacted in 40 mm × 40 mm × 160 mm steel molds on a jolting table, and cured in moist air at 20 ± 1 °C with relative humidity not below 90% for the first 20 to 24 hours, followed by immersion in water at 20 ± 1 °C until the test age. Three prisms are prepared for each test age.
Testing Items
Uncertainty analysis in cement strength determination typically covers the following deliverables:
- Compressive strength at specified ages (commonly 3 days and 28 days; also 1 day and 7 days where required) with complete uncertainty budgets
- Flexural strength at the corresponding ages, including the uncertainty associated with the smaller specimen population of three prisms
- Repeatability and reproducibility studies of the strength test method within a single laboratory or across laboratories
- Conformity assessment against the strength-class limits of GB 175 and EN 197-1, including documented decision rules
- Re-evaluation of the uncertainty budget after machine recalibration, a change of standard sand lot, relocation of equipment, or amendment of the test method
Testing Methods
Measurement Procedure and Data Reduction
At the specified test age, controlled within tight tolerances such as 24 h ± 15 min and 28 d ± 8 h, each prism is first broken in flexure at a loading rate of 50 ± 10 N/s over a 100 mm support span. The flexural strength is calculated as Rf = 1.5 · Ff · L / b3, where Ff is the peak force, L is the distance between supports, and b is the side of the square section. The resulting six half-prisms are then crushed in compression through 40 mm × 40 mm auxiliary platens at a loading rate of 2400 ± 200 N/s, and the compressive strength is Rc = Fc / A with a nominal bearing area A of 1600 mm2.
Data reduction per GB/T 17671 and ISO 679 incorporates robust averaging rules: if one flexural result among three deviates by more than 10% from their mean, it is discarded and the average of the remaining two is reported; for compression, if one of the six half-prism results deviates by more than 10% from the six-value mean, the mean of the remaining five is reported, and the rejection of more than one value invalidates the set.
Mathematical Model for Uncertainty Evaluation
The compressive strength model is expressed as Rc = (Fc/A) · kcal · krate · kgeo · kalign · kcure, where each k factor is a dimensionless correction term with an expected value of one, accounting respectively for force calibration error, loading-rate deviation, geometric and planeness imperfections, eccentricity of loading, and curing regime deviations. Sensitivity coefficients follow from partial differentiation; for the dominant force term, ∂R/∂F = 1/A, so relative standard uncertainties in force propagate one-to-one into relative uncertainties in strength.
Identification of Uncertainty Sources
- Repeatability (Type A): material heterogeneity, mixing efficiency, compaction consistency, and the intrinsic scatter among the six half-prism results
- Force calibration of the compression testing machine and load cell, verified to class 1 of ISO 7500-1, including long-term drift between calibrations
- Loading-rate deviation from the specified 2400 ± 200 N/s window, since strength is mildly rate-dependent
- Specimen geometry: mold dimensional tolerances, planeness and perpendicularity of loaded faces, surface defects, and the condition of the auxiliary platens
- Centering and alignment of the specimen between platens, including platen coaxiality and the resulting eccentric loading
- Curing regime: water temperature stability within 20 ± 1 °C, humidity during the initial 24 hours, and adherence to age tolerances
- Material factors: grading and moisture content of the standard sand, mixing water quality, and the representativeness of the cement sample
- Indication resolution of the force measuring system and rounding of the reported strength to 0.1 MPa
Type A Evaluation
The Type A component is derived from the experimental standard deviation of the replicate results: uA = s/√n, where s is the standard deviation of the n accepted compressive results (normally six half-prisms) or flexural results (normally three prisms). Precision studies published with the ISO method indicate relative repeatability standard deviations for 28-day compressive strength typically between about 2% and 4% for a competent laboratory, which makes Type A the dominant term in most budgets.
Type B Evaluation
Type B components are quantified from prior knowledge. Contributions from calibration certificates (expanded uncertainty U with k = 2) are treated as normally distributed and divided by the stated coverage factor. Contributions bounded by specification limits, such as loading-rate windows, temperature bands, and geometric tolerances, are treated as rectangular distributions and divided by √3. Each component is then converted into a strength-equivalent standard uncertainty through its sensitivity coefficient.
Worked Example: 28-Day Compressive Strength Budget
Assume a mean 28-day compressive strength of 52.4 MPa obtained from six half-prisms with a standard deviation s = 1.94 MPa, giving uA = 1.94/√6 = 0.79 MPa (1.50% relative). The Type B components are estimated as follows:
- Force calibration from the machine certificate (U = 0.8%, k = 2): u = 0.21 MPa (0.40%)
- Loading-rate deviation within ±200 N/s (rectangular): u = 0.15 MPa (0.29%)
- Specimen geometry and planeness tolerances (rectangular): u = 0.12 MPa (0.23%)
- Centering and alignment effects (rectangular): u = 0.10 MPa (0.19%)
The combined standard uncertainty is uc = √(0.792 + 0.212 + 0.152 + 0.122 + 0.102) = 0.85 MPa, corresponding to 1.6% relative. With a coverage factor k = 2 for approximately 95% confidence, the expanded uncertainty is U = 1.7 MPa (3.2% relative), and the result is reported as (52.4 ± 1.7) MPa, k = 2. Because the budget is dominated by the Type A term with a substantial number of degrees of freedom, the effective degrees of freedom computed by the Welch–Satterthwaite formula comfortably support k = 2.
Reporting Requirements
In line with the GUM and ISO/IEC 17025, the test report must state the measurand, the measurement procedure and standards used, the result, the expanded uncertainty with its coverage factor and level of confidence, and, where statements of conformity are made, the decision rule applied. The uncertainty budget should be retained as documented information so the evaluation can be audited and updated.
Applications
- Conformity assessment against strength-class limits of GB 175 and EN 197-1, applying a guard band equal to the expanded uncertainty so that pass decisions account for measurement doubt
- Acceptance testing and dispute resolution on construction projects, where the reliability of a marginal strength result must be demonstrated quantitatively
- Support of ISO/IEC 17025 accreditation scopes and participation in proficiency testing schemes, where uncertainty feeds En-number and zeta-score comparisons per ISO 13528
- Internal quality control at cement plants, allowing statistically defensible action limits and trending of strength results in statistical process control
- Research and product development, providing the statistical basis for comparing binder systems, admixtures, and alternative constituent materials
- Structural engineering applications, where the confidence level of strength data informs mix proportioning and safety verification
FAQ
What is measurement uncertainty in cement strength determination?
It is a quantified estimate, derived from statistical analysis of replicate specimens (Type A) and from calibration data, tolerances, and engineering judgment (Type B), of the range within which the true strength of the cement mortar is expected to lie at a stated level of confidence, typically 95%.
Why does repeatability usually dominate the uncertainty budget?
The scatter among the six half-prisms, driven by material heterogeneity and small variations in mixing, compaction, and curing, typically contributes 2% to 4% relative standard uncertainty, whereas the force calibration of a properly verified class 1 machine contributes only about 0.4%. As a result, specimen preparation discipline matters more to uncertainty than the calibration certificate itself.
What expanded uncertainty is typical for 28-day compressive strength?
In a well-controlled laboratory operating the ISO prism method, expanded uncertainties of approximately 2% to 4% relative (k = 2) are typical for 28-day compressive strength. The worked example above yields (52.4 ± 1.7) MPa, equivalent to 3.2% relative.
How does uncertainty differ from measurement error?
Error is the single signed difference between a measured value and a reference value, which cannot be known exactly. Uncertainty is a positive dispersion parameter that characterizes the doubt about the result; it does not correct the value but states how far the true value may reasonably lie from the reported one.
How can a laboratory reduce measurement uncertainty?
Effective measures include strict control of the curing water temperature at 20 ± 1 °C, timely calibration and preventive maintenance of the compression machine, use of certified standard sand with verified moisture condition, disciplined centering of specimens, standardized operator training, and control of the mixing and compaction sequence. Increasing the number of replicates reduces the Type A component in proportion to 1/√n, although the standardized method fixes the number of prisms per age, so improvements must come mainly from process control.
How is uncertainty used in conformity decisions?
Under ISO/IEC 17025, a laboratory must document its decision rule. With a guard-banded acceptance rule, a result such as 43.0 MPa against a 42.5 MPa class limit would fail to demonstrate conformity if the expanded uncertainty (e.g., ±1.7 MPa) overlaps the limit, whereas simple acceptance would pass it. The chosen rule must be communicated to the client and applied consistently.
