Based on up-to- date thermo-mechanical simulators like GLEEBLE, the SR is often used to register in real time the REC and REX of austentite that should then disappear in the following phase transformation in cooling. After the austenite deformation stops (sample size is fixed), REC and REX are presumed to display linear segments of different slopes on the SEMILOG stress-time plot. However, to my knowledge, nobody tried to exclude evident issues resulting in inevitable uncertain errors as follows.
1) After the stop command at a programmed time t0, the ACTIVE deformation still goes on with decreasing rate (and hence stress) until actually stops at some t1>t0. Just this t1 rather than t0 should be taken for the start of following RELAXATION process (and related softening), however no accepted criterion exists to determine t1.
2) Physically, at small enough time lapses t-t1, S(t) function should be linear and approaches Log type only later on. Thus, at the least times (say 0.0001 to 0.001 s) of relaxation, the plot will show a HORIZONTAL (constant S) segment with a length which does not visibly differ (!!!!!) from that of much longer time intervals (e.g. 1 to 10 s) where the softening slope become clear. This PURELY NUMERICALl effect hinders one to draw proper tangent lines and find deflection points. Meanwhile, nobody clearly recommended from which time t2>t1>t0 to begin the Log time axis.
3) The somewhat stressed sample under a high temperature exposure can undergo creep that weakens constraints by the FIXED clamps, and this geometric effect accelerates an APPARENT stress reduction regardless of actual material softening. This often results in very unrealistic softening ratios of about 90%!! It is unclear how to exclude this error. The later is more or less tolearable when only temporal parameters do matter, but it is absolutely inacceptable when activation energies or similar charateristics are of interest..
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size