A lot of piston aircraft engines don't have the same problem we have. The pipes comes straight down from the cylinders to the muffler. The bending loads, if any, are low. There are four or more pipes supporting the weight of the muffler in tension and not bending under the engine. Our rotary engine mufflers are cantilevered off the side of the engine and the exhaust temperatures are higher. In the case of professionally made or certified aircraft mufflers systems the parts may be heat treated to minimize the effects of welding.
The main vibration mode of the rotary is torsional about the e-shaft axis. That whips the side mounted mufflers up and down. Perhaps at some multiple of the muffler resonant cantilever bending frequency mode when the engine is running at a constant RPM. Consequently it fatigues the tubes where they are welded to the flange. I think what we need is a rigid connection and let the straps and tubes bend slightly.
It is well known the tubes always cracks just on the edge of the weld where the material properties are obviously affected by the welding process weakening the material. This heat affected zone could be as little as 1/8 th of an inch wide. I have seen dozens of pictures of this cracking phenomena. If you have been around exhaust system or steel tube space frames you too may have seen it. If the heat affected zone is inside the 3/8th thick exhaust flange it will be well supported. If it is outside the flange it will have no support what so ever. If inside the support will be applied to the tube in a place on the tube where the material strength will be at a maximum unaffected by the welding process
Loads and stress distribution on a typical exhaust flange.
Everett Hatch style exhaust flange.
3D drawing of Everett Hatch type exhaust flange.
Consequences of not welding the flange correctly.