Brian Dyson (https://www.researchgate.net/scientific-contributions/BF-Dyson-2003401809) used to joke that there were many times more models for predicting creep-fatigue than there were researchers working on the topic.
Yet time fraction remains overwhelmingly the most frequently used method for calculating the effects of creep, even though many papers (https://www.researchgate.net/profile/Yukio-Takahashi-2, https://www.researchgate.net/scientific-contributions/J-Wareing-2041878341, https://www.researchgate.net/profile/Michael-Spindler) have been written that highlight time fraction's limitations and proposing demonstrably better models.
Why is it so difficult to change the methods to calculate the effects of creep that are routinely applied to both research and industrial applications?
It occurs to me that it is difficult simply because creep-fatigue is a two parameter problem (rather than a 1 parameter problem). Furthermore, I believe it is because the effect of fatigue is known to a significantly smaller uncertainty than we know the effects of creep. I am confident that we know the effect of fatigue on creep-fatigue because the following experimental considerations are the same for both fatigue tests and creep-fatigue tests (but are different for both creep tests and creep-fatigue tests):
1. The test specimens (fatigue specimens used for both)
2. The test machine (servo controlled)
3. The control mode used to conduct the test (typically extensometer or strain control)
4. The data that is logged during testing
5. The failure criterion used (drop in maximum load or stress)
I am interested to hear alternative points of view and will be happy to reply with further evidence and my own thoughts and answer any questions that arise.
Really, two-parameter (for example, creep-fatigue) phenomena are difficult for interpretation. But I do not believe that fatigue process is characterized a significantly smaller uncertainty than creep.
I would like to answer the question from actual operating condition point of view.
In lab, we use a certain combination of frequency, loads, wave-forms, thermal cycle etc. to get creep-fatigue test results within a reasonable time.
However, in actual operating conditions, the temperature gradient/thermal cycle, changes in loads, time of exposure to creep or fatigue conditions are vastly different.
Moreover, there is no standard pre-fixed or assumed cyclic wave-forms in actual conditions.
I think the above factors makes creep-fatigue problems difficult to solve.
With regards to the reply by Serge V. Shil’ko this reply is very important and leads into the observations that I want to discuss. I apologise for not having my response ready so that I could have posted it immediately. Please be patient I will respond in full soon.