How can we simulate a flare gas recovery system for a refinery?
Answers
|
21/07/2009
|
A:
|
keith bowers, B and B Consulting, kebowers47@gmail.com
There are at least two main 'types' of process simulations used.
1) is STEADY STATE (Hysim, Aspen,etc) assumes peak mass flows from each PSV at the same time to ascertain maximum built-up back pressure at each node and PSV in the system. This is to check that each PSV will pass the required mass at the absolute worst plant wide conditions.
A more limited steady state simulation looks only at the system in a given unit, with more judgment about how adjacent unit relieving events may interfere (load up the flare system)
2) DYNAMIC simulation is used to 'prove' the time staging of the different relief loads from an incident affecting either a single unit or multiple units. For instance, plant wide power failure may cause many PSVs to open and relieve excess pressure, but they open at different times and rates than 'the maximum' event. High pressure hydro-cracker reactors are usually de-pressured in a fire event as rapidly as possible to a safe, below temperature critical pressure, while large distillation columns will typically not relieve for 20-30 minutes after loss of top reflux. Dynamic simulation is the only method that can 'prove' the actual relieving loads are not coincident. Dynamic simulation results may enable flare header size reduction by 50% or more!
|