ZeroExperience, surely even given your screen name, you can do better than that! LOL, just giving you a hard time.
The references I listed go much deeper than that. The basic idea is that a researcher performed a whole bunch of computerized, nonlinear spread-of-plasticity analyses on a specific beam-column. Axial loads varied from very small to large. Plot the results and one gets a figure that looks about like the bilinear curve one would get from plotting Eq. H1-1a and b. The thing I don't remember is how one goes from that to being able to use H1-1 with LTB, etc.