Posted by Dave Hoffman on May 16, 2005 at 15:39:15:
In Reply to: Reflux Drum Level posted by Graeme Taylor on May 13, 2005 at 08:17:10:
The easy answer is: No, it is not the case for all types of flooded condensors that have a drum. Often there is no valve between the flooded condensor and the drum in which case the drum is operated "full" at > 100% level. In this case, I've seen the DCS PC loop configured to manipulate either distillate rate or reboil heat input. In either case I usually end up opening the DCS PC because the system is very interactive and the PC dynamics are slow for all MV's.
For your configuration (with the valve between condensor and reflux drum), it depends on how sensitive the hydraulics are to changes in the drum level. It sounds like it is sensitive. The basic problem is that if you open the level controller you will create two ramp variables (the level and the PC output) for essentially one underlying phenomenom. DMC will have a problem with this (because they are fundamentally colinear in most variables but also ramps which must always be balanced). You do have the hot vapour bypass which could be used to independently control the PC output without having much effect on the level but I assume that you would normally want to keep this closed.
The other answer that could work would be to open both the LC and the PC. However, this might be difficult to test depending on disturbances, surge capacity, etc. You may also introduce inter-dependent non-linearities with this approach so be careful.
And finally, the last option leave both DCS controllers closed. If you choose this option, the level controller needs to be tuned quickly (i.e do not tune the DCS controller for surge volume level control) or else DMC will "fight" with the level controller when the distillate flow constraint is active.
Hope this helps a little - the correct approach may also depend on whether we're talking about a high constraint on the distillate flow or whether you are trying to minimize it to zero.