1. It's probably not wrong. But a potentially better design would be to put the support under the elbow itself, rather than upstream of it.
2. a. API 520, ASME B31.1, for open system relief.
2. b. L.C. Peng's "Pipe Stress Engineering" for closed system relief.

I'll note a known limitation of CAESAR is that it doesn't handle second order effects, and directionality of force isn't variable.

I.E. the initial relief force is downwards at the end of the pipe. If this force causes the pipe to bend around the support, the force's direction would be changed. Because the force's direction changed, it's now doing something different to the pipe. So on and so forth until you meet an equilibrium in force vs force response from the pipe. CAESAR only allows you to input the initial downward force statically, and if you wished to do so dynamically, it'd take an enormous manual effort to get CAESAR to emulate this.