Re: Column overhead temp as an mv vesus reflux as an mv

Name    : 
E-Mail  : 
Subject : 

Comments:
: You bring up an interesting question.  My comments below are a result of partially my own experience, but also largely from the experience of those engineers with whom I have worked in the past.

: First thing, as a general rule, you like to nail down one side (only) of the distillation for a two-cut splitter.  This could mean a temperature controller on a bottom tray cascaded to bottom reboiler or an overhead tray temperature cascaded to reflux.  This keeps the column from "drifting" in response to changes in rate, composition, etc.  Sometimes just a FOT on the bottom is also sufficient for these purposes.  Once you have fixed one side of the column, it is best to control the composition on the other side using the DMC controller.  How do I choose whether to fix the top or the bottom?  It usually just depends on where the most sensitive thermocouple is.  Probably slightly more than half of the applications I do are fixed on the bottom.

: There are always variants and exceptions depending on the particular circumstances that you face.  Options for "fixing" one side of the distillation may take a different form for "mass-balance" type control schemes (where the controller MV is top and/or bottom flow rate out).  When the ratio of internal reflux/bottom flow out is high, it is often better to cascade bottom level control to reboiler heat input.  If the overhead reflux ratio is very high, or total reflux, then the reflux flow should be cascaded to the overhead accumulator level.  I consider either of these configurations as having one end of the column "fixed".  I would then not try to close a temperature controller on the other end.

: There are some complicating factors in closing an overhead temperature loop that you should consider carefully.  For example, if you have an FOT on the bottom as an MV in the controller, closing the overhead temperature will approximately double the time to steady-state of the controller model.  You will have to wait for the vapor to travel up the column and then get pushed down again by the overhead temperature controller.  Besides doubling your testing time, it can create other difficulties, especially if you have side streams.  Compositions in the middle of the column may have gains that are very close to zero (as the temperature goes up and then back down).  This has the potential to confuse the overall optimization function.  

: Good luck in your application.

: Tod


Optional Link URL:  
Link Title:         
Optional Image URL: