Flare

It is clear from you question you do not have the experience or knowledge to perform this complicated analysis. API RP 520 provides general requirements. However, the subject matter is both broad AND very detail oriented. ALL the specific details of the physical facility must be evaluated for designing a safe pressure relieving SYSTEM. Overlooking even a single 'simple' detail may result in a safety system that does not work!

The basic rules are pressure vessels AND piping must be protected from pressures exceeding 110% of their maximum design values at the maximm temperature of even abnormal operation. The design challenge is ensuring the 'back pressure' in the relief SYSTEM , when added to the relief device opening setting, does not exceed the maximum allowable value. Considerable experience and good judgement is required to assess the probability of co-incident relieving events (two or more relief devices opening near the same time) contributing to a high back pressure in the flare header to which the relief device(s) vent.

In one particularly large load ( 1.4 MM#/hr at MW of 18) , the flare TIP (54" ID) had 90 psig back pressure!. The vessel pressure was 3300 psig. We installed a bank of computer controlled 'drag valves' , with redundant processers and power supplies, and opened enough drag valves to keep the vessel pressure below allowable, and at the same time, depressure the vessel to below 50% as fast as the catalyst bed supports would allow to protect against fire impingement heating of the vessel wall resulting in failure. This design allowed this very large and dangerous relief load to be dealt with before the distillation columns (other large loads) would need relief, thus allowing a single relief header and flare instead of two costly ones. Considerable dynamic process simulation effort of the entire SYSTEM was required to ensure the intended performance would be achieved. We STILL installed conventional pilot operated pressure relief valves on the reactors as redundant 'insurance' against reactor over-pressure and potential failure, but did not have to consider those loads in sizing the header and flare since those were smaller than the depressuring rate! Calculated flame length was greater than 2000 feet!