Read
this thread, found using the [Search] facility.
The CNODE between 200 and 1200 defines an internal restraint. You should not define the same restraint (association) again, because now you have two restraints there. So this makes sense that the loads would drop by a factor of 2.
Jacketed systems are typically composed of two materials, so you can generate huge forces and moments in a jacketed system if you (erroneously) lock-up (or attempt to lock-up) the thermal growth. The problem isn't the CNODE, and assuming nothing else has been modeled at 200 and 1200, eliminate the restraint/CNODE here and change node 1200 to node 200. (So 1190-1200 becomes 1190-200, and 1200-1210 becomes 200-1210.) Your results shouldn't change. (Make sure you remove both restraint/CNODE associations at 200/1200.