Here’s another suggestion, which hasn’t been suggested, quite the same way, on getting the adapter loose enough to “torque” off the bar.
Heating the mated combination or freezing the combination doesn’t seem to have much effect, since the SS materials just respond the same way, and shrink or expand together too slowly. Then they return to the same temperature even if you act quickly. But, how about “differential heating”
of the two parts ?
The trick is to keep the “bar as cool as possible”, relative to the adapter's “elevated” temperature.
Perhaps the cwt. bar could be (safely) secured in a large vice, and just heat the adapter itself, from an air-flow directed nozzle adapter of very hot HEAT GUN. A hair dryer might work, set at its highest level, but not quite as effectively as the bench tool.
Before inserting in the bench vice, shield the bar nearest to the adapter, away from the hot air stream as best you can, perhaps even with just several wrapped layers of fibreglass wool walll insulation. The glass will keep the heat gun’s air stream from hitting and warming the bar too quickly, as the “much smaller mass” of the adapter’s threaded hole expands “radially outward” much faster, hopefully releasing its grip on the cwt. threads.
At that point - acting as quickly as you can - grip the adapter with heavy work gloves, something like “welders gloves”, and see if it will budge.
If that still fails, devise an “insulated lever bar” – say, a 2-foot long piece of 2x4 , drilled at its length centre , with hole being slightly larger than the cwt. threads diameter. Slip it over the end of the bar and bolt the 2x4 to the adapter’s two holes, one above and below the adapter ring centre.
Finally, you can slam the 2x4 at its top end, with a sledge hammer, giving that added levered torque the 2x4 extension provides to the adapter, to break the thread’s bind. The 2x4 being wood, won’t affect the continual heat being delivered to the adapter ring’s “opposite face” side, so the adapter continues to expand, while the heat “isolated” cwt. bar still remains relatively cooler, for a while.
If it doesn’t break the seal with repeated strikes the first time, have a bucket of ice-water handy, to pour ONLY over the bar’s far end itself, and give the 2x4 lever another slam, since the ice water should then have even shrunk the bar diameter somewhat, while the adapter is still quite hot, and its centre diameter still expanding from the continual heat gun air stream.
I think that the triple combination of the “differential heating” of the adapter outer area, with something like a 1,200 watt heat gun, while additionally super cooling the bar “nearby ” (but not on), the adapter, should respond well, to a strong slam at the extreme end of the 2x4 lever arm - with say a 5 lb. sledge hammer.
Desperate times ... call for desperate measures !