Partial factors in Eurocode 7 convert characteristic or representative inputs into design values by scaling actions (γF), soil parameters (γM) or resistances (γR). In UK practice, you take the values from the National Annex and use Design Approach 1 for structural and geotechnical checks. The second-generation EN 1997-1:2024 formalises both input and output factoring for numerical analyses and expects you to document which one you chose.
TL;DR:
- For each structural or ground ultimate limit state check, run A1+M1+R1 and A2+M2+R1, then use whichever combination governs to control the design.
- The UK National Annex for the 2024 standard took effect in 2026; the 2004 edition remains current until March 30, 2028, so match tables carefully.
- Choose representative soil values after deriving parameters from tests, measurements, or correlations; weak site data calls for better investigation or greater conservatism, not altered factors.
- Input factoring reduces soil strength before analysis, while output factoring scales calculated effects; geometry and plastic redistribution can cause the results to diverge.
- Under the newer standard, designers must document how factor choices and design assumptions will be checked during construction and service life.
Table of Contents
- What partial factors do within limit state design
- The three factor types and how combinations are formed
- Design approaches and the UK National Annex
- Applying partial factors in numerical modelling
- Practical steps and pitfalls in partial-factor design
- Where partial factors come from: the theory behind them
- How EN 1997-1:2026 differs from earlier versions
- Choosing factors from ground investigation quality and variability
- Linking partial factors to the wider Structural Eurocode design
- A worked example: factoring a spread foundation
- An editorial view: factors are only as good as your documentation
- Remote engineering support for Eurocode 7 partial-factor work
- FAQ
- Sources
What partial factors do within limit state design
Partial factors belong to the partial-factor method, one verification route inside the limit state framework. They are not arbitrary safety margins. The JRC guidance describes the Eurocodes as fundamentally reliability-based, with partial factors calibrated against reliability concepts, and EN 1997-1:2024 also offers reliability-based design as a formal verification option alongside the partial-factor method.
The order of work matters more than most calculation sheets admit:
- Obtain derived values from tests, measurements or correlations, and record them in the ground investigation report, as the ground model guidance explains.
- Select representative values transparently, with the uncertainties stated.
- Only then apply partial factors to reach design values. Experts flag the confusion of these steps as a common error.
- For serviceability limit states, take the factors as 1.0 in the normal case.
EN 1997-1:2024 is the second-generation code, while the first-generation EN 1997-1:2004 stays current until its withdrawal on 30 March 2028, so both generations co-exist until then. Check which one your contract specifies before you start.
The three factor types and how combinations are formed
Each factor type acts at a different point in the calculation, as the Eurocode guidance on reliability-based verification sets out:
- γF scales actions or the effects of actions, such as permanent and variable loads.
- γM scales soil parameters, such as the tangent of the friction angle or undrained shear strength.
- γR scales resistances, such as pile base and shaft capacity or sliding resistance.
Exact values are nationally determined. Annex A of the UK National Annex lists partial and correlation factors for ultimate limit states and replaces the recommended values in the main Eurocode text where it specifies them.
Factors are grouped into sets and used as combinations. Combination 1 is A1+M1+R1 and Combination 2 is A2+M2+R1. For each structural (STR) or ground (GEO) check you run both and design to whichever governs. As a rule of thumb, Combination 1 tends to govern where structural actions dominate, and Combination 2 tends to govern where ground strength dominates, though you should confirm this case by case rather than assume it.
Pro Tip: Write the set label (for example A2+M2+R1) beside every design value in your calculation sheet, so a checker can see at a glance which factors were applied and why.
Design approaches and the UK National Annex
EN 1997 offers three design approaches. They differ in where the factors are applied: DA1 uses two combinations as described above, DA2 factors actions and resistances rather than soil parameters, and DA3 factors structural actions and soil parameters together. Technical summaries note that DA2 also has variants, depending on whether factoring is applied at the start or the end of the calculation.
The UK position is simpler. The National Annex prescribes Design Approach 1 for STR and GEO checks, and its tables are decisive for design values. The National Annex text also gives factors for different situations, including piles and slopes, and explains when model factors may apply to test-verified resistances.
Practical points for the UK:
- The NA to BS EN 1997-1:2024 took effect in 2026 and replaces the older NA to EN 1997-1:2004, so confirm which edition your project cites.
- Check the Annex A tables for the specific case: spread foundations, piles, anchorages, slopes, uplift and hydrodynamic conditions.
- Record any departure from the tabulated values and the reason for it.
Applying partial factors in numerical modelling
Numerical methods raise a question hand calculations hide: where exactly does the factor go? Practitioner work on action effect factoring shows that EN 1997-1:2024 specifies both input factoring (IFA) and output factoring (OFA), and that the two can give different results depending on geometry, plastic redistribution and modelling choices.
- IFA: factor the soil strength inputs (for example reduce tan φ′ and c′ by γM) and run the model with the reduced values.
- OFA: run the model with unfactored parameters, then factor the resulting action effects, such as bending moments or prop forces.
- Compare and judge: in many geometries the two land close together, but where plastic redistribution matters they can diverge, and OFA can avoid over-factoring working-state values.
- Be consistent: the chapter on formalised IFA and OFA stresses that the choice affects stiffness, redistribution and design outcomes.
Pro Tip: State the approach in the model report header, and show how the assumptions will be checked during execution, as the JRC guidance on implementation during execution requires.
Practical steps and pitfalls in partial-factor design
A defensible workflow runs in one direction: ground investigation, derived values, representative values, factor set and design approach, analysis, then documentation of assumptions and how they will be checked on site. The guidance on determining representative values includes worked examples and covers limited samples and depth-dependent properties.
Common pitfalls we see in review:
- Treating characteristic and representative values as the same thing.
- Not confirming which National Annex tables apply to the case.
- Mixing factoring approaches within one finite element model.
- Omitting the implementation checks that EN 1997-1:2024 now expects.
Pro Tip: Put a short assumptions table in the calculation note listing each parameter, its representative value, the factor set and the site check that will confirm it.
Backfill and compaction assumptions are a good example of a ground property that needs a site check, and this contractor guide to backfill and compaction covers how those specifications are verified in practice.
Where partial factors come from: the theory behind them
Partial factors grew out of limit state design, where the aim is to show that a design value of a load effect never exceeds the design value of the corresponding resistance. Rather than applying one global safety factor to everything, the method gives each source of uncertainty its own factor: loads vary, soil properties vary, and models are imperfect, and each deserves separate treatment.
The reliability basis is explicit. The Eurocode guidance explains that EN 1997 works alongside EN 1990, and the partial factors are calibrated so that structures reach a target reliability level, with characteristic, frequent and quasi-permanent combinations used for serviceability and ultimate checks. In other words, a partial factor is a calibrated shortcut to a probability-based outcome, which is why designers are not free to pick values they find convenient.
Geotechnical practice added a wrinkle structural codes do not have. Soil is a natural material with strong spatial variability, so the choice of representative value carries much of the uncertainty before any factor is applied. That is why the second generation puts so much weight on documenting the ground model and the route from derived to representative values, and why a factor cannot rescue a poorly characterised site.
How EN 1997-1:2026 differs from earlier versions
The first-generation EN 1997-1:2004 and the second-generation EN 1997-1:2024 co-exist until 30 March 2028, so many offices are working with both. The main changes are in how the code asks you to work, rather than in a simple list of replaced numbers.
- Reliability-based design is now an explicit verification option next to the partial-factor method.
- Numerical methods receive formal procedures, with both input and output factoring specified, where the earlier edition left more to interpretation.
- Design during execution is stronger: you must show how design assumptions, including partial-factor assumptions, will be checked during construction and service life, through implementation plans.
- Representative values receive more guidance, with the derivation route documented.
For UK users, the practical consequence is to read the 2026 National Annex alongside the main standard. The older National Annex tables belong with the 2004 code, so do not carry values across between generations without checking them against the edition cited in your project. The draft text of the second-generation standard also gives example tables and formula references for γF, γM and γR.
Choosing factors from ground investigation quality and variability
The factor tables do not change with the quality of your site data, so the adjustment happens upstream, in the representative value. A thin investigation with few tests leads to a more cautious representative value than a dense, well-targeted one, and the JRC guidance on representative values addresses how to treat limited samples and depth-dependent properties.
A practical approach:
- Start from the derived values recorded in the ground investigation report and the uncertainties noted in the ground model guidance.
- Judge variability by layer and by depth rather than quoting one site-wide number.
- Where correlations rather than direct tests supply a parameter, reflect that in how conservative the representative value is.
- Where temporary or transient situations justify it, technical summaries note that reduced partial factors may be used, provided the consequences justify it and the reasoning is documented.
Keep the roles separate: the representative value absorbs ground uncertainty, and the National Annex factor then covers the remaining calibrated safety margin. If the ground model feels weak, the right response is more investigation or a more cautious representative value, not a quiet change to the factor.
Linking partial factors to the wider Structural Eurocode design
Geotechnical design rarely stands alone. EN 1997 is used together with EN 1990, which sets the combination rules and the formulae for γF, and the foundation or retaining structure usually has to be checked against structural codes such as EN 1992 for reinforced concrete. The same project can therefore carry factors from several parts.
The link that catches people out is the interface. The action effects produced by a geotechnical model, such as wall bending moments or pile loads, feed straight into structural resistance checks. Keep a clear record of which factored or unfactored effects you pass across, so the structural designer does not apply the same factor twice or omit it.
Our Eurocode 2 checklist for reinforced concrete drawings shows where these assumptions surface in reinforcement design, and our retaining wall design guide applies the same thinking to earth pressures and wall verifications. The overview of Eurocode 7 is a useful reference when you need to confirm how the parts fit together.
A worked example: factoring a spread foundation
The inputs below are illustrative, not market data. Consider a strip foundation carrying permanent and variable loads on sand with a representative friction angle.
For the action side of Combination 1 (A1), apply the relevant factors to the permanent and variable actions to obtain the design action. Check the actual A1 values in the National Annex tables before using them in a real design.
For the ground side of Combination 2 (M2), apply the relevant factor to tan φ′ to obtain the design friction angle. The reduced design angle affects bearing capacity, which is then compared with the Combination 2 design action.
The lesson is in the method. Run both combinations, record the set label beside each value, and let the governing one control the design. Always replace these illustrative factors with the values the National Annex gives for your case.
An editorial view: factors are only as good as your documentation
The weak link in most partial-factor work is not the choice of γ. It is the unrecorded reasoning upstream of it. A perfectly selected factor sitting on an unexplained representative value gives a design that looks rigorous and cannot be defended.
Our view is that the second generation is right to push designers towards documented assumptions and implementation plans. It turns the factor from a number you look up into a claim you have to support, and that is a healthier culture for geotechnical design than treating the tables as the whole job.
— Mohammed
Remote engineering support for Eurocode 7 partial-factor work
When your team is stretched on EN 1997-1:2024 work, we embed experienced engineers directly into your workflow, with deliverables aligned to UK standards and Eurocodes, so you keep full control of the client relationship and skip permanent hiring.
Typical ways our Remote Engineering Support helps consultancies:
- Urgent numerical-modelling support, with the factoring approach stated in the report.
- Calculation packages that show the National Annex factor sets applied, ready for your review.
- Implementation plans and site-verification checklists that show how design assumptions will be checked.
Tell us about your project through the Remote Engineering Support page and we will scope the help you need, with a first conversation to agree the work.
FAQ
What are partial factors in Eurocode 7?
Partial factors are numbers that scale actions (γF), soil parameters (γM) or resistances (γR) to turn representative values into design values. They form part of the limit state design method and are calibrated against reliability concepts.
Which design approach does the UK use for Eurocode 7?
The UK National Annex prescribes Design Approach 1 for structural and geotechnical checks, using two combinations: A1+M1+R1 and A2+M2+R1. The National Annex values replace the recommended values in the main standard where it specifies them.
What is the difference between input and output factoring?
Input factoring applies the partial factor to soil strength inputs before the model runs, while output factoring applies it to the resulting action effects. The two can give different results depending on geometry and plastic redistribution, so state which one you used.
Do serviceability checks use partial factors?
Serviceability limit states normally take partial factors of 1.0, as the Eurocode guidance summarises. The ultimate limit state checks carry the factors from the National Annex tables.
What must designers document under EN 1997-1:2026?
You need to show how design assumptions, including partial-factor choices, will be checked during construction and service life. The JRC guidance describes this as implementation plans and site-check procedures.
Leave a Reply