On your Q1 (aside from the issues pointed out by Mr. Luff), in terms of the application of CAESAR II, a rupture disk burst causes impulse type loads that are very similar to those caused by relief valves. In reality, the key difference is the speed of application of the unbalanced loads (rupture disks being faster to open than relief valves), but I've yet to find a PRV manufaturer who could tell you how fast their valves open. For that reason, a conservatively-quick opening rate is often presumed; along the lines of a rupture disk. The faster the load is applied, the higher the DLF, to a maximum of 2.

On your Q2, the equation offered by the manufacturer must be considering the suddenly-applied unbalanced pressure force (P*A), multiplied by the conservative DLF value of 2. In many cases, applying this 2PA force in a static load will provide good 'ballpark' results. There are cases though, where presuming DLF of 2 is overly conservative.
If you are not sure, I would suggest developing your own force-time profiles based on your particular circumstances and employing time history analysis.

On your Q3, there are several variables that come into play. If you have a large pressure and/or area, your P*A load can be significant, and in these cases, lesser load durations can still develop a significant system response.

The DLF generator in CAESAR's dynamic input processor will give you an idea of how various natural frequencies will respond to your particular force-time profile. That, in conjunction with the magnitude of the unbalanced pressure load may help you decide when it is necessary to consider some loads and ignore others.