Dear Ihsan

I am just wondering if you can use correlation coefficient  to check if their is significant relationship between them. I think can help you to see if both techniques following the variation instead to focus on just the differences between them.

All the best

Ammar

Dear Ammar

Thank you for your replay. I am not sure if this technique can be adopted in this case or not. This is because I do not have a huge amount of data. For example, If I got an experimental moment of 1200 N.mm and an analytical value of 950 N.mm, so can I say that the analytical result agreed well with the experimental one? 

You can try to build the confidence interval for the experimental value to see the analytical one falls to that interval or not. But in general the accuracy requirements are defined by certain application. The error 250 N.mm can be significant for the one application and can be insignificant for another.

Hi Ihsan

I believe your comparison should be concentrated before the failure of the structure not after this point, i.e. the 250 N.mm is significant if it is before the failure of the structure.

Good luck