LPG constraint within an FCC DMC controller

Name    : 
E-Mail  : 
Subject : 

Comments:
: I currently have two separate DMC controllers that span our FCC complex.  The first covers the reactor/regen/main fractionator and the second the gas plant.  Unfortunately the LPG flow rate (overhead of the debutaniser) is a constraint, due to a pump amps limit and very frequently the reactor severity needs to be set, to limit the LPG production.  Currently this is done by the operator setting the reactor temperature high limit.  This has one of two results:

: 1. The temperature is set too low, leading to us not fully utilising the capacity that we have.
: 2. The temperature is set too high and the debutaniser gives up on the light naphtha RVP constraint, so as not to violate the LPG high limit (the LPG flow is a constraint within the gas plant controller).

: Clearly there is an opportunity here to try to incorporate the LPG constraint within the reactor/regen controller, but the question is how best to do this?  I see that I have 3 options

: 1.  Our upcoming RTO project will span the entire complex.  I could simply allow the RTO to set external targets in the reactor/regen to try to ensure that we are within the LPG constraint.  For me this is the easiest answer, but will it work?  With RTO running infrequently what happens if these external targets result in the LPG rate being too high or too low?  I see this as a step in the right direction, but I believe that we will still see the same two problems as before, ie a give away of capacity or a give up in qualities.

: 2.  Bring together the two applications such that the DMC controller can try to control the LPG constraint with the reactor temperature handles (and potentially feed rate).  We are in a reasonably good position to try this, in that the LPG rate is an MV within the gas plant controller and the debutaniser overhead drum level is a CV.  We would then have to break a level at the front of the gas plant.  All of the feeds to the deethaniser (fractionator distillate product and liquid condensed after the wet gas compressor) come together in one drum and this drum level could be opened.  The concern I have with this, is that the process gains to the debutaniser overhead level will be very small and any small variations in the gas plant will result in significant moves in the LP targets for the Reactor/Regen.  Then you add to this that all of the effects that the controller predicts are in bulk flow terms and not in changes in gas plant feed quality, I get a little nervous.  For example if you reduce reactor temperature I suspect that the LPG yield reduces more than the feed to the gas plant does, thereby leading the controller to underestimate the gain of reactor temperature to LPG.  

: 3.  Develop a model that predicts the LPG flow rate from reactor/regen variables and include this in the front end controller.  The benefit of this, is that hopefully the model understands the effects of reactor temperature, catalyst circulation and feed rate, more correctly than the DMC models that would be developed in option 2. Potentially the RTO model can be used to develop this model. The difficulty is trying to include the constraint in both the reactor/regen controller and the gas plant controller.

: I am sure that there must be several other applications out there that must be similar to this, if anyone has experience of such an application, I would be delighted to hear about it.

: Thanks.
:    


Optional Link URL:  
Link Title:         
Optional Image URL: