I agree that the design is troublesome, but when you say that I will not do the project by thinking so sensitively, you will also enable the poor architectural designs to be modeled in weak hands like you. Instead, I think it would be more accurate to try to change the parts that you have trouble with by talking to the architect, or to think about how to make the statics more convenient over the existing design, although it is not possible to change them. Currently, the elevator edges can be modeled as webbed. From the ends of these curtains, parallel beams to the 1st and 18th steps can be thrown and connected to the side beams. You will need to reduce the step widths by 1-2 cm to make the last step coincide with the beam. In addition, beams can be connected from the ends of the curtains to the floor landing in a horizontal direction. Overhanging beams can be covered with styrofoam later, or you can consider threading the landing at the modeling stage. This will strengthen your connection a little bit. In addition, it would be appropriate to solve the ladder together with the structure by performing a semi-rigid analysis. However; I'm not sure if it's the right approach but; it can be considered to compare the displacements of the stairwell and the main structure with the stairwell deleted on separate pages, and to approximate the resulting values by adding curtains to the main structure. Of course, it does not guarantee that the displacements will be at the same time and in the same direction, but I guess it will be better than the first situation. Good work.