EN 1990 is not the Eurocode for reinforced concrete, steel, or timber. It is the base standard that sets the common design principles for the whole Eurocode system: safety, serviceability, durability, robustness, design situations, limit states, and the method of verification. In the current second generation, the document describes the basis of structural and geotechnical design, and the main verification approach remains the limit state concept combined with the partial factor method.
That is exactly why EN 1990 matters more than its title might suggest. It does not size a section for the designer and it does not replace the material-specific Eurocodes, but it decides what correct design is supposed to look like in the first place: what has to be checked, in which situations, at what reliability level, and according to what logic for combining actions. Without this layer, the other Eurocodes would be a set of separate rules. With it, they form one system.
Where does EN 1990 sit in the Eurocode system?
The simplest way to put it is this: EN 1990 is the overarching document, while EN 1991 to EN 1999 develop the different parts of design in detail. EN 1991 describes actions on structures, and the following Eurocodes move into materials and detailed calculation rules. EN 1990 does not compete with them. It creates the common basis for buildings and civil engineering works, together with geotechnical, fire, seismic, execution, and temporary structure aspects.
This distinction has a very practical purpose. A designer may know the material rules well, but if the design situations, action combinations, or the correct type of limit state are set up wrongly, the calculations can still be formally correct while addressing the wrong problem. EN 1990 is there to organise that earlier stage, meaning the logic of the whole design process.
What does the standard cover, and what does it not cover?
In its current form, EN 1990 establishes principles and requirements for the safety, serviceability, robustness, and durability of structures, including geotechnical structures, in line with the consequences of failure. The document is intended to be used together with the other Eurocodes when designing buildings and civil engineering works, and the basis of verification remains the limit state approach.
At the same time, it should be stated clearly what the standard does not do. EN 1990 does not tell you how to calculate the resistance of a specific steel section, how to design reinforcement in a slab, or which timber grade to choose for a load-bearing element. Those are tasks for the later Eurocodes. The role of EN 1990 is different: to establish a common level of requirements and a common verification format for all those calculations.
The four pillars of design under EN 1990
The core of the standard is built around four ideas: safety, serviceability, durability, and robustness. In the first generation of the Eurocodes, the emphasis was usually placed on safety, serviceability, and durability. The second generation puts stronger emphasis on robustness and on a wider set of requirements tied to the real performance of structures in use. One of the key directions of change in the second generation is a stronger focus on robustness together with better practical usability.
This has concrete consequences. A structure may satisfy its resistance requirement and still fail to meet the code requirements if it deflects too much, develops unacceptable vibrations, performs poorly in service, or is too vulnerable to disproportionate damage after an initial local event. EN 1990 therefore does not allow design to be viewed only through the lens of “will it collapse”. It requires a broader view: will it perform properly over the intended design working life and at the assumed level of reliability?
Limit states – the real heart of the whole standard
The most important layer of EN 1990 is limit state design. This is where the standard moves from general requirements to the practical mechanism for checking a structure. Instead of one general “margin of safety”, the process is divided into different types of limit states and different sources of uncertainty, with verification carried out mainly through the partial factor method.
That creates two basic orders of verification: ultimate limit states and serviceability limit states. The first relates to structural safety, the second to normal performance in use. This division may look obvious, but it is what organises most design decisions. Without it, it becomes very easy to confuse a structure that is “safe” with one that is “well designed”.
ULS – when the issue is safety
ULS covers situations in which the structure may lose its ability to carry load, lose stability, or lose its overall ability to perform safely. This includes section failure, buckling, loss of equilibrium, excessive deformation leading to failure, and fatigue-related issues. This is the layer in which the designer is responsible for ensuring that the structure does not enter a dangerous state.
SLS – when the structure still stands, but stops working as it should
SLS concerns normal use. This includes deflections, vibrations, displacements, damage to finishes, loss of comfort, or other unacceptable service effects. The second generation of EN 1990 develops this side of the standard more clearly, especially in relation to limits on deflection, vibration, and foundation movements, because these are the issues that often do not cause failure, but quickly lead to a poor assessment of the whole design.
| Type of limit state | What it concerns | Typical consequence of exceedance | Example checks |
|---|---|---|---|
| ULS | safety and resistance | failure, loss of stability, collapse | section resistance, stability, buckling, fatigue |
| SLS | normal use | excessive deflections, vibrations, cracking, poor structural performance in service | displacements, vibration comfort, limitation of service-related damage |
The main point from this table is simple: design under EN 1990 does not end when the structure “holds”. If the structure does not meet serviceability requirements, the design is still incomplete. In day-to-day practice, this is often where disputes begin, because the client usually judges a structure by how it behaves in use, not by the fact that it has not reached its ultimate resistance.
The second important point is that ULS and SLS are not two competing methods. They are two parallel filters through which the design has to pass. EN 1990 does not allow one to be chosen instead of the other. It requires both, because only then is the structure both safe and fit for use.
Design situations – not one calculation, but several different scenarios
EN 1990 requires the designer not to analyse the structure in one abstract situation, but in the relevant design situations. Persistent, transient, accidental, seismic, and, in the broader view, fatigue-related situations – each of them organises a different type of hazard and a different set of checks. This is not a formal exercise. A structure behaves differently in normal use, differently during execution, differently under an accidental event, and differently under repeated loading over time.
This is the point at which EN 1990 forces scenario-based thinking. Instead of one “model for everything”, the designer has to ask what may really happen over the life cycle of the structure and which limit states are critical in each case. Only then do action combinations and factors make sense. If this step is done badly, the later calculations lose value even if the equations themselves are correct.
Actions and combinations – where EN 1990 meets EN 1991
One of the most frequently misunderstood issues is the relationship between EN 1990 and EN 1991. EN 1991 provides the catalogue of actions, but it is EN 1990 that determines how those actions are to be represented and combined into the combinations used for limit state verification. That is why characteristic, combination, frequent, and quasi-permanent values, together with ψ factors for variable actions, matter so much.
Design is not about mechanically “throwing everything into one basket”. It is necessary to consider which actions can really occur together and which cannot. This matters both for safety and for economy. Over-simplified action combinations can lead to unnecessary oversizing, but combinations that are too optimistic are equally problematic. EN 1990 is there to organise that boundary.
| Standard | What it is responsible for | What it does not determine on its own |
|---|---|---|
| EN 1990 | basis of design, reliability, limit states, design situations, combinations, and verification | detailed material resistance rules and member sizing details |
| EN 1991 | actions on structures | the rules for verifying all limit states |
| EN 1992–EN 1999 | material-specific and calculation rules for particular types of structures | the common design philosophy of the whole system |
This comparison shows clearly that EN 1990 is not an add-on to the “real” Eurocodes. It is the top layer. Without it, EN 1991 gives only a set of loads and the material standards only rules for specific structures. It is EN 1990 that turns them into one design method.
From a practical point of view, this means one more thing: knowing load values alone is not enough. It is just as important to set the design case and the verification logic correctly. That is the step that separates software output from actual structural design.
Reliability, consequence classes, and partial factors
One of the most technical, but at the same time most important, layers of EN 1990 concerns reliability. The standard links the required level of design to the consequences of failure. In the background material for the second generation, it is stated directly that the consequence class is used, among other things, to determine a consequence factor, select reliability management measures, modify allowable failure probability or target reliability levels, and choose methods for improving robustness. The Eurocode system broadly covers classes CC1 to CC3, while CC4 may require additional provisions.
This shows that partial factors are not arbitrary corrections added to equations. They sit on top of a defined way of working with uncertainty. The probabilistic background to this logic shows that, for a reference class such as CC2, the traditional target level for ULS has been associated with a reliability index of about β = 3.8 over 50 years, and for SLS about β = 1.5. The point is not that every daily design should be carried out probabilistically. The point is that the numbers in the tables did not come from nowhere.
In practice, this is one of the most misunderstood aspects of the Eurocodes. Partial factors are often treated as a formal obligation rather than as a carrier of the chosen reliability level. EN 1990 is built precisely on that logic: to separate uncertainty on the side of actions, resistance, modelling, and consequences, and then bring it together in one coherent design format.
What did the second generation of EN 1990 really change?
The most important changes in the second generation do not overturn the whole philosophy of the standard. The core remains the same: limit states, partial factors, and a common level of reliability. The change is more about wider scope, better structure, and greater usability in everyday work. The main directions of change include better practical usability, provisions for assessment, reuse, and modification of existing structures, a stronger treatment of robustness, and an expanded scope covering new methods, materials, and market expectations.
Several more detailed shifts are also visible. There is a more consistent approach to ULS verification, stronger alignment with EN 1997, more guidance on serviceability of buildings in terms of deflection, vibration, and foundation movements, improved wording for nonlinear analysis, further development of fatigue-related provisions, changes in the annexes dealing with reliability management and reliability analysis, and a clearer emphasis on robustness and sustainability. These are not cosmetic changes. They move the standard closer to real design problems.
2026 update
By 2026, the editorial direction of the second generation is also much clearer. EN 1990-1 is being implemented for new structures, while EN 1990-2 is being developed and published for the assessment of existing structures. This shows that the system is maturing toward a clearer separation between the design of new works and the assessment of existing ones.
| Area | First generation | Second generation |
|---|---|---|
| role of the standard | base standard for the Eurocode system | the same role, but with a wider and better organised scope |
| geotechnics | present, but less tightly integrated | stronger alignment with EN 1997 |
| serviceability | more general | more guidance on deflection, vibration, and foundation movement |
| robustness | present, but less developed | much stronger emphasis |
| practical use | more framework-oriented | greater emphasis on ease of use and implementation |
The most important conclusion from this table is simple: the second generation does not change the meaning of EN 1990, but makes it more operational. The standard remains the base document, but it looks less like an abstract design manifesto and more like a tool for real calculations and structural assessment.
This is especially visible in the areas where the first generation sometimes felt too dispersed for users. The new direction moves toward greater clarity, a stronger link to the actual life cycle of the structure, and better organisation of the relationship between structure, geotechnics, robustness, and service performance.
National Annexes – where the common standard meets national practice
It also needs to be said clearly that EN 1990 does not operate in a national vacuum. The Eurocode system has always included Nationally Determined Parameters, meaning points at which the standard allows national choice. In practice, this means the National Annex contains the decisions adopted in a given country for the areas left open by the European text.
That has very practical consequences. The text of EN 1990 on its own is not always enough to make a design decision. The designer still has to check the National Annex and the way the provision is implemented in the country of use. In the UK, for example, that means working not only with BS EN 1990, but also with the UK National Annex where relevant. A fair amount of misunderstanding comes from treating the European standard as if it were detached from that national layer.
What does EN 1990 really mean for day-to-day design?
The most important point is that EN 1990 organises not only calculations, but the whole way of thinking about a project. It forces a series of simple but fundamental questions: which limit state governs, which design situation has to be checked, which actions can really occur together, what level of reliability is needed, and whether the structure will be not only safe, but also serviceable and durable. That is the real practical value of the standard.
The second point is even simpler. EN 1990 does not design the structure for the engineer, but it protects the engineer from arbitrariness. It sets a common logic for design, execution, and assessment of structural behaviour. Because of that, concrete, steel, timber, geotechnics, and actions do not operate as separate worlds, but as parts of one system. And that is exactly why this standard is a foundation, not an add-on.
Sources and reference materials:
https://eurocodes.jrc.ec.europa.eu/EN-Eurocodes/eurocode-basis-structural-design
https://publications.jrc.ec.europa.eu/repository/handle/JRC139110
https://knowledge.bsigroup.com/products/eurocode-basis-of-structural-and-geotechnical-design-new-structures
