I take it these are existing passenger cars which are starting to show cracks from normal operating conditions, tension and compression loads from the draft gear, and now you are being asked if they still meet the 800k buff loading criteria without catastrophic failure. It seems to me that you do have a fatigue problem or you wouldn’t have the cracking, although at a much lower stress level than the 800k buff load would cause. I think you have to look at where the cracks are forming and why, and how these cracks are oriented w.r.t. the loads and stress fields if this shear plate now sees the 800k buff loading. Do the cracks settle down once they form, or do they keep growing over time? What details are causing this cracking? I would at least drill small holes at the tips of the cracks. Without seeing the details and the end arrangement, outboard of (but including) the body bolster, I would imagine (expect, by design) that the shear plate would initially buckle under the compression loading, and it would them become two diagonal tension fields from the inner end of the stub sill, at the body bolster, to the outer ends of the two side sills. The body bolster and the end sill and any other cross car framing (cross ties, etc.) provide a cross car force system to hold the side sills and center sill apart during this tension field action. They also act in horiz. bending in concert with the shear pl. The shear pl. acts to do this over the stub sill/draft pocket. But, all of the welded joinery is acting to cause the existing cracks and may cause them to grow during this impact loading. If the cracks are oriented in the direction of the tension field stress lines, they are less of a problem on impact. If the cracks are diagonal to or cross the tension field stress lines, they are almost certainly an early failure mechanism. The weld details are all important to prevent the cracking from starting.
I’d look at these shear plates like buckled webs on a plate girder. They first start acting in shear, then buckle out of plane, and then start acting as a tension field of some depth, from the stub sill web to the side sill. And, I would certainly use FEA to study this problem, just not expect exact (certain, absolute) answers from it. I’d do the FEA model to represent the as built end frame, on two supports at the side sills, and loaded at the rear draft lugs, and run that through the normal draft and buff loads, and the 800k impact load. Then introduce the cracks one at a time (a linear discontinuity 1" long, for example), and only one at a time, to see what these did to the stress picture After the impact a slightly crumpled car end outboard of the body bolster is o.k. as long as the rest of the car is intact. Are these transit cars or over the road passenger cars, how old? I don’t think there is an exact answer to your problem, the next crack will be slightly different, starting at a slightly different defect. All you can do is offer some good engineering judgement about a range of failure loads and the nature of the failure. You might see that some cracks and the details that caused them should be welded up and cleaned up, while others are of little consequence, but I’d still drill the tips. I’d like to see an end arrangement drawing and some details.