Pipes, not me.!

I have a system which has had two trunnions (on a letdown station) anchored down as opposed to a stop and guide as requested on the calculation. this system is currently in operation (hot)

Caesar gives overstress of 300%(610n/mm2)
Kelloggs hand calcs suggest stress levels of 3600n/mm2 on the trunnion.

(JCL, Anchor nodes 20400 & 20500 on the CAESAR II file i sent)

I am thinking these stress levels are indicative only, and that local yielding, anchor slippage (base plates are bolted) and some minor buckling will reduce the actual stress levels somewhat.
Ideally, i would like to use FEA to asses what is going on, but i dont have it so my question is, at what (over)stress levels do i raise the alarm?. I have requested that the anchors are removed and the spools local are replaced with immediate effect for safety, but at what point/stress level can you expect rupture or catastrophic failure given all of the factors of safety invloved?
The Min UTS as per App.A for API5L is 410n/mm2
The trunnions are way over this level (as is the Caesar file)

Or. can i actually give a technical answer due to the vagaries of local yielding and deflections.

answers on a postcard..........