honestly, i have no experiences in designing of rails due to impact load. mostly in structural building due to quasi-static loads, there’s also dynamic analysis but simplest one with lumped mass and rigid floor. another interesting me are in member performances due to cyclic and ground motion (earthquake). thanks to CalculiX development for the feature and possibilities.
regarding brittle material, shear capacity depends on normal stress and work triaxially. the material definition is not based on simple stress-strain curve rather using some parameter were it need calibrated before. also notes on Modified Mohr-Coulomb (MMC) it has tension cut-off (Rankine) controlled by cohesion values.
The best shear method used here in the US is the modified compression field method once cracked, or simply principal tensions prior to cracking. The former is not particularly easy to model as it depends on crack size. Fortunately it is dependent on strains, shears, shear and flexural reinforcement and reinforcing spacing.
Most of my bridge design work also uses quasi static loads, though some types of reinforcing can be used at higher stresses or strains (~0.005) than yeild. Arching is also a factor that can sometimes result in higher strengths as well, sometimes used in bridge decks, and I have discovered appropriate in some locations in concrete bridge rails that have appropriate restraint. Tension field effects can also be useful in metal bridge rails, though the deflection needed, if large, usually makes the design very difficult due to chaotic effects with the vehicle interaction. Both of these can be helpful arresting progressive collapse.
reading paper i previously posted by links. i do replacement of RedBlue with CoolWarm colormap based on recommendations.
or may the RedBlue colormap is keep and stay available? another option is doing reversed value list, so it can be comparable with another type (qualitative: Classic, Turbo, Jet)
the modification files is placed at GitHub as before (attachment link)
hopefully you can helping again to re-test before i submit a revision.
