The supposed to be linear region of my stress strain curve is concave shaped (see attached image). As a result, linear regression was resulting in bad slope value and erroneous yield strength. Anyone knows how to evaluate the yield strength in this case?
I tried taking only the linear portion from the middle part of the curve and then did the regression(see attached image). Is that okay ?
Thank you for your answer Timothy A Ebert . Looks like it is okay to do piecewise regression. However, I am now confused about the slope.
Do I need to take an average of the all the slopes taken for each segmented portion?
Furthermore, when extrapolating the data, is it okay taking just the slope of the last linear portion?
Thanks a lot for your help.
The average of the slopes would be a meaningless value. You need a good justification for focusing on one critical area. From the graph it looks like the material starts to deform at about 800 MPa, and once it starts to fail it takes less effort to continue the process. So you just want to focus on performance before the point of no return (or something like that).
The decision must be made after the sample has been unloaded. Meaning of 0.2% - material operation criterion in linear area and absence of residual deformations. Concave curve means, likely, that the deformations of the steel are inelastic
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression