Pressure Cycling & Fatigue
Fatigue assessment of cyclic pressure systems, from pressure histories and structural stress ranges through cycle counting, local discontinuities, mean stress, cumulative damage and verification.
Why pressure cycling becomes a fatigue problem
A pressure boundary can be comfortably below its static strength limit and still accumulate fatigue damage if the pressure, temperature or attached-system loads vary repeatedly. Each pressure change modifies membrane stress in the shell and can create much larger local stress ranges at nozzles, welds, head junctions, supports and geometric discontinuities. Fatigue assessment is therefore based on stress or strain range between load states, not simply on the maximum operating stress. The relevant history may include startup and shutdown, normal control cycling, proof tests, pump or valve events, pressure transients and occasional upset conditions.
Define the cycle history before calculating life
The first engineering task is to establish the actual or design-basis sequence of pressure states. A vessel that sees one full startup/shutdown cycle per day and thousands of small control fluctuations should not be represented by one generic pressure range. Group events into physically meaningful cycle types and preserve their associated temperature and mechanical loads. When measured histories are available, cycle counting can identify repeated ranges; when only duty descriptions exist, a conservative operating envelope may be required. The assumed number of cycles should be traceable to the required design life and operating philosophy.
Nominal, structural and local stress range
Remote shell membrane stress often scales almost linearly with pressure and provides a useful baseline. Fatigue usually concentrates at discontinuities, where the local range contains bending and geometric concentration in addition to membrane stress. The correct stress measure depends on the chosen fatigue method: nominal stress with a detail category, structural or hot-spot stress, linearised membrane-plus-bending stress, or local notch stress. A singular FE peak at a sharp weld toe should not be paired directly with a nominal S-N curve. Stress extraction and fatigue data must form a consistent methodology.
Pressure plus thermal and external-load cycling
The largest fatigue range may not occur between minimum and maximum pressure. A hot pressurised state compared with a cold depressurised state can combine pressure membrane stress with thermal expansion, nozzle loads and support restraint. Startup may create transient through-wall temperature gradients that disappear at steady operation. Piping reactions can reverse as connected systems expand. The fatigue case should therefore be constructed from complete system states rather than pressure alone where coupled loads are significant. Time correlation matters: combining independent maximum pressure, thermal and piping stresses can create a range that never occurs physically.
Cycle counting and cumulative damage
For variable-amplitude histories, rainflow counting or an equivalent recognised method can reduce a stress-time record into ranges and cycle counts. Damage is then commonly accumulated using a linear summation such as Miner’s rule when that approach is permitted by the design basis. The apparent simplicity should not hide the assumptions: the fatigue curve, mean-stress treatment, environmental effects, weld class, thickness correction and sequence sensitivity can all influence life. Where only a few defined transients exist, direct event-by-event accounting is often clearer than applying generic counting software.
Mean stress, residual stress and welds
Mean stress can affect fatigue resistance in unwelded components because tensile mean stress reduces the available cyclic margin. Welded details are often treated differently because high tensile residual stress is already implicit in the fatigue class or assessment procedure. Do not apply a mean-stress correction automatically without checking the basis of the selected curve. Surface finish, weld quality, misalignment, residual stress and environment can be more influential than nominal mean stress. The fatigue model should reflect the manufacturing detail actually being qualified.
FEA workflow for pressure fatigue
Use the global pressure model to identify critical regions and establish converged structural stress ranges. Refine or submodel only where the selected fatigue method needs more local resolution. Evaluate all relevant state pairs and retain signs and local directions so that the true range is recovered. If plastic strain develops each cycle, an elastic S-N approach may no longer be sufficient and strain-life or nonlinear cyclic analysis may be needed. For transient thermal fatigue, map the temperature history into a structural transient solution and extract the complete stress-time response.
Verification and reporting
Verify the pressure contribution against analytical hoop and longitudinal stress changes, check local mesh convergence using the stress measure required by the fatigue method, and confirm that cycle counts match the duty definition. Report the governing location, cycle type, stress range, fatigue curve, modifiers, calculated usage and uncertainty. Include proof and test cycles if they consume life. A defensible fatigue result explains why the chosen stress measure and curve are compatible and why the assumed operating history envelopes credible service.
Fatigue is governed by stress or strain range and cycle count. The location with the highest static stress is not automatically the location with the highest fatigue usage.
Local weld and attachment fatigue
Pressure cycling can drive significant local ranges at attachment welds even when the attachment carries little external load. The shell expands and contracts beneath a relatively stiff pad, clip, support or reinforcement, producing local bending at the weld toe. These details should be included in the fatigue screening if they sit on a highly cycled pressure boundary. Use a fatigue method compatible with the weld classification or structural-stress approach and avoid interpreting a singular sharp-toe FE peak as a converged local stress.
Operating-envelope changes
A pressure system may accumulate a very different fatigue usage after an uprate, revised control philosophy or increase in startup frequency. Reassessment should therefore compare the new duty with the cycle basis used originally, not only compare the new maximum pressure with the old design pressure. Small additional stress ranges can become significant when repeated many times. Preserve a cycle ledger or equivalent operating record so that future life updates can distinguish consumed life from remaining design capability.