Re: FCC severity control

Name    : 
E-Mail  : 
Subject : 

Comments:
: I have never included "severity" as a controlled variable in an application.  I have included apparent conversion (defined as [feed-LCO-bottoms] / feed) as a constraint variable.  This calculation has relatively few inputs and is not terribly susceptible to measurement noise (especially if SP's are used instead of PV's).  Some people define severity = conversion / [100 - conversion].  So this type of severity calc. might also work.  I always try to avoid controlling calculated variables that include a lot of inputs because measurement noise is additive.

: The real question here is how is the best way to "fill up" the FCC - with feed or severity?  Usually the answer is both - i.e. increase the feed as long as it can be cracked.  In general, I would rather see real constraints setting the feed / severity trade-off (e.g. refinery fuel gas balance, C3/C4 make, LCO make, gas recovery unit flooding, etc.).  However, it may sometimes be useful for the planners to be able to set conversion in a narrow band and then maximize feed.  The advantage of this approach is that riser temperature can move up and down depending on how refractive the feed is.  The disadvantage is that octane may move up and down as well since it is more a function of riser temperature than conversion.

: So, as usual, the answer to all of your questions is “It depends…”  I try to avoid including “artificial” constraints such as conversion or severity in the controller but sometimes it’s necessary.  However, I would definitely take a long hard look at how inherently noisy any calculated variable may be.


Optional Link URL:  
Link Title:         
Optional Image URL: