the creep curve

If you hang a constant weight on a hot metal bar and record how much it stretches over days, then plot strain against time, you get the creep curve — the story of a slow-motion failure in one line. It is the signature shape every high-temperature engineer learns to read, because its middle section sets how long the part will last.

The curve has three stages. Primary (transient) creep starts fast then slows, as the metal work-hardens and resists more. Secondary (steady-state) creep is the long, straight, middle stretch where hardening and softening balance and the strain rate is roughly constant and at its minimum — this steady creep rate is the single most useful design number, because it dominates the life. Tertiary creep is the final upward swing, where the rate accelerates as internal voids and cracks form, the section thins, and the bar heads toward rupture. A concrete picture: a blade might creep quickly for the first 100 hours (primary), crawl steadily for thousands of hours (secondary), then rapidly stretch and rupture over the last few hours (tertiary).

Engineers care most about the secondary (minimum) creep rate, because for a part that must last years it decides the life; a common design rule limits total creep strain to about 1 percent over the service life. Raising either the stress or the temperature lifts the whole curve, steepens the steady-state slope, and shortens the time to rupture — so creep test data are gathered at several stresses and temperatures and extrapolated (via parameters like Larson-Miller) to the long service times that cannot be tested directly.