Creep Fundamentals
The fundamentals of creep — time-dependent deformation under sustained load, the three stages of creep, the variables that govern creep rate, and the engineering significance of creep in structural analysis.
What is creep?
Creep is the time-dependent inelastic deformation of a material under a sustained load at elevated temperature. Unlike elastic deformation (which is instantaneous and recoverable) and plastic deformation (which is instantaneous and permanent), creep deformation accumulates over time and is permanent (with a small recoverable anelastic component). Creep occurs because the material's microstructure evolves under load at high temperature — dislocations climb, grain boundaries slide, and diffusion occurs. The creep rate depends on the stress, the temperature, the accumulated strain and the material microstructure.
The three stages of creep
A creep test at constant stress and temperature produces a strain-time curve with three distinct stages. In primary creep, the strain rate decreases — the material strain-hardens as it deforms. In secondary creep, the strain rate is approximately constant — the strain hardening and the thermal recovery (softening) are in balance. In tertiary creep, the strain rate increases — damage (voids, cracks) accumulates and the effective load-bearing area decreases. The three stages are not always equally prominent; in some materials or conditions, one stage may dominate or be absent.
The creep curve
The creep curve plots strain against time at constant stress and temperature. The shape is characteristic: a rapid initial strain (elastic + plastic), followed by primary creep (decreasing slope), secondary creep (constant slope), and tertiary creep (increasing slope) ending in rupture. The relative duration of each stage depends on the stress and temperature. At high stresses or temperatures, the tertiary stage may dominate; at low stresses or temperatures, the secondary stage may dominate and the tertiary stage may be short. The creep curve is the primary experimental data for calibrating creep constitutive models.
Variables governing creep
- Stress: higher stress increases the creep rate, typically as a power law (Norton law) at moderate stresses
- Temperature: higher temperature increases the creep rate, typically following an Arrhenius (exponential) relationship
- Time: the accumulated creep strain affects the rate through strain hardening (primary) and damage (tertiary)
- Material microstructure: grain size, precipitates, prior deformation and heat treatment all affect creep resistance
- Environment: oxidising or corrosive environments may degrade the microstructure and accelerate creep damage
Engineering significance
Creep has three engineering consequences. First, excessive deformation: creep strain may cause loss of clearances, loss of alignment or dimensional changes that impair function. Second, stress relaxation: in constrained components, creep strain replaces elastic strain, reducing the stress — this causes loss of bolt preload, loss of interference fits and load redistribution. Third, creep rupture: the accumulated damage in tertiary creep eventually causes fracture. The structural analysis must address all three: the deformation, the stress changes and the life consumption.
Creep is not a secondary effect to be added as an afterthought — at elevated temperature, it is often the primary life-limiting mechanism. The analysis must treat creep as a first-order phenomenon with a time-dependent constitutive model and a life assessment.
Creep mechanisms and why they matter
The macroscopic creep curve can arise from different microscopic mechanisms, including dislocation climb, diffusion and grain-boundary processes. The dominant mechanism changes with temperature, stress and microstructure, which is one reason a single fitted power law should not be extrapolated without evidence. The stress exponent and apparent activation energy inferred from test data can provide clues that the controlling mechanism is changing. For engineering analysis, this matters because a model calibrated over one regime may predict the wrong stress or temperature sensitivity in another. Qualification of the constitutive law should therefore consider the range of stress, temperature and time represented by the service condition rather than only the visual quality of a curve fit.
From test specimen to structural response
Uniaxial creep tests provide the material basis, but structural behaviour also depends on redistribution and constraint. A local peak stress may relax rapidly while neighbouring material picks up load; a nominally uniform bar may be well represented by a simple creep curve, whereas a thick wall, welded joint or notched component may not be. The analyst should distinguish between the material creep law, which describes local strain rate, and the structural response, which emerges when that law acts throughout a constrained geometry. This distinction is central to interpreting FEA: the highest initial stress is not necessarily the location with the highest long-term creep strain or damage.