Take a look at thread744-266307: Mxy Moments in concrete floor design.
Essentially, it's just a way to acknowledge the Mxy (Twisting) moment in plate design. It forces you to include the Mxy moment when you come up with your "design" moments rather than just looking at Mx and My moments.
These "twisting" moments create flexural stresses in plate elements and it's not always easy to understand their direction or what to do with these stresses.
In my opinion these flexural stresses due to Mxy are somewhat analogous to torsional warping stresses in a Wide Flange steel beam. That analogy may not help other people as much. But, I say it because I can more easily visualize the warping stresses as they related to flexural design.
I usually don't use plate elements to model RC in order to avoid having to go through the additional effort of the whole Wood Armer method to resolve the Mxy component. I'm aware that a lot of engineers simply ignore the Mxy component and somehow consider that the slab must have some inherent torsional strength without providing additional reinforcement for it (which it probably does in reality, but hard to justify).
I prefer using a grillage of beam elements instead, with the torsional stiffness of the elements reduced to zero, so that any torsion effect is redistributed to bending in the orthogonal directions.
The problem is the stress resultants for a particular element are normally not in the x/y directions, unless the bending in the element is perfectly symmetrical. The Principal Stresses are at a different rotation to the X and Y axes.
But most times a slab is reinforced in the X and Y directions. When you rotate the axes to X/Y, from the Principal axes, you end up with Mx, My and Mxy.
Mxy has to be accounted for in determining the reinforcement requirements in the X,Y directions.
If you reinforced in the principal stress rotations, Mxy would be 0, but that rotation is different for every element so that is not practical.
An alternate solution would be to calculate the capacities in X and Y directions and then transfer those capacities to the principal stress rotation for each element. But that would be even less practical in most cases, but would remove any conservatism in the Wood Armer Method.
