Engineering task and calculation objective
The EN18 module performs the detailed fatigue assessment of pressure vessels and pressure vessel components to DIN EN 13445-3 clause 18. If you want to calculate fatigue life in detail, you assess cyclic loading from pressure and temperature fluctuations here as well as from varying external loads, such as piping forces or changing liquid levels. The basis is the standard's S-N curves for welded components, unwelded components and bolts, together with linear damage accumulation over all cycle types.
Unlike the simplified method of clause 17 (module EK17), the detailed assessment works with structural stresses: the stress ranges at the detail under consideration must be known, typically from a finite element analysis, from stress formulas with stress magnification factors, or from measurements. In return, the method exploits the load-bearing reserves far better and yields higher allowable cycle numbers, which often decides feasibility for heavily cycled equipment such as pressure-swing adsorbers, autoclaves or frequently started reactors.
For welded constructions, the geometry of the detail under consideration determines the fatigue class; thickness effect, temperature and, for unwelded parent material, the mean stress effect are captured via correction factors.
Standard and calculation basis: DIN EN 13445-3/18: 2021-12
Calculation workflow
- Compile the load spectrum: All cyclic loads occurring in operation are recorded as cycle types with frequencies: pressure cycles, temperature transients, external load cycles. Superimposed cycles of different magnitude are decomposed into individual stress cycles using a counting method such as the reservoir or rainflow method.
- Determine the structural stresses: The stress ranges are determined for each detail to be verified. For welds, the standard uses the structural stress at the weld toe (without the notch of the weld itself, which is contained in the fatigue class); for unwelded regions, the effective notch stress with a notch factor.
- Classify the detail: Welded details are assigned to the fatigue classes of the standard on the basis of their geometry; the class depends on the weld type, loading direction and extent of testing. Separate fatigue curves apply to unwelded components and bolts.
- Apply correction factors: The endurable stress ranges are adjusted with correction factors for wall thickness, temperature and, where applicable, mean stress. For elastic-plastic behaviour, the stress range is increased via a plasticity correction.
- Determine allowable cycle numbers: For each cycle type, the allowable number of load cycles at the corrected stress range is determined from the governing S-N curve.
- Form the damage sum: The partial damages of all cycle types are added according to Miner's rule. The verification is satisfied if the total damage does not exceed 1; otherwise the design, weld quality or operating regime must be adjusted.
Input quantities
| Quantity | Symbol | Unit |
|---|---|---|
| Calculation temperature ( ≤380°C ferritic, ≤500°C austenitic material, ≤450°C nickel / nickel alloy) | T | °C |
| Material | Wk | – |
| Design strength | K | MPa |
| Safety factor | S | - |
| Manufacturing allowance | c1 | - |
| Corrosion allowance | c2 | - |
| Offset of center-lines | A1 | - |
| Noncircularity factor | A2 | - |
| Angular offset | A3 | - |
| Local peaking | A4 | - |
| Exponent 1.5 for circumferential weld / 0.6 for longitudinal weld | x | - |
| Nominal thickness of component | en | mm |
| Nominal thickness (for A2) | e | mm |
| Nominal thickness of plate 1 (en1<en2) | en1 | mm |
| Nominal thickness of plate 2 | en2 | mm |
| Constants | C1 | - |
| Constants | C2 | - |
| Modulus of elasticity (at operating temperature) | E | MPa |
| For cylinders | Km | - |
| Plasticity correction factor for stress due to mechanical loading | ke | - |
| Plasticity correction factor for stress due to thermal loading | kv | - |
| Exponent | m1 | - |
| Exponent | m2 | - |
| Allowable number of load cycles | N | - |
Calculated results
| Quantity | Symbol | Unit |
|---|---|---|
| Coefficient | Fe | - |
| Coefficient | Fs | - |
| Thickness correction factor | fe | - |
| Thickness correction factor in welded components and bolts | few | - |
| Mean stress correction factor | fm | - |
| Surface finish correction factor | fs | - |
| Temperature correction factor | fT* | - |
| Overall correction factor for unwelded components | fu | - |
| Overall correction factor for welded components | fw | - |
| Mean stress sensitivity factor | M | - |
| Allowable number of load cycles | N | - |
| For given number of load cycles | n | - |
| Assumed mean cycle temperature | T* | °C |
| Stress range | Δσ | MPa |
| Reduced mean equivalent stress for elastic-plastic conditions | σeq,r | MPa |
| Effective total equivalent stress range | Δσf,all | MPa |
| Allowable number of load cycles | N | - |
| Allowable load cycles for bolts acc. 18.12 | N | - |
| Condition for stress range | sigma | – |
Calculation options
Type
Fatigue life of welded components · Fatigue life of unwelded components
Frequently asked questions
Where do I get the structural stresses for the assessment to clause 18?
Three routes are common: a finite element analysis with evaluation of the structural stress at the weld toe (for example by extrapolation over reference points ahead of the weld), analytical formulas with stress magnification factors for standard details such as nozzles or shape deviations, or strain measurements on the component. The quality of the fatigue assessment stands or falls with the correct determination of these stress ranges.
What distinguishes the assessment of welded and unwelded regions?
For welds, the notch effect of the seam is already contained in the class-specific S-N curves; the structural stress is applied, and the mean stress effect is largely omitted, because residual stresses from welding are assumed. Unwelded parent material, by contrast, is assessed with the effective total stress including a notch factor and benefits from favourable mean stresses. The same stress range can therefore lead to very different fatigue lives depending on the location.
Must temperature cycles be considered even if the pressure remains constant?
Yes. Temperature transients generate thermal stresses through temperature differences across the wall and between connected components, which can form a full stress cycle at every start-up and shutdown. For thick-walled components and fast transients, these thermal stress ranges frequently dominate the fatigue assessment over the pure pressure cycles.
When is it worth switching from the simplified assessment (clause 17) to clause 18?
Whenever the simplified assessment does not confirm the required number of load cycles, or when loads are present that clause 17 covers only roughly, such as temperature transients or external load cycles. The extra effort lies in determining the structural stresses; in return you obtain more realistic, usually considerably higher allowable cycle numbers and the ability to improve individual details in a targeted way.