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

Name    : 
E-Mail  : 
Subject : 

Comments:
: Tod & Brian already addressed most of the issues. However, there are a few other points to consider.

: Keeping a stable TC in closed-loop control will make a plant test easier to conduct and many times make the MPC controller easier to operate, but will increase the response time of the column, and thus will lengthen the plant test time. 

: One of the keys is how well your DCS top temperature controller works. To consider keeping this in closed-loop control with your MPC application, it needs to work quite well (be stable, hold temperature to the setpoint, follow setpoint changes properly, and compensate for moves on other independents). If you've got a high purity overhead product, a standard top temperature controller is not going to be stable because of the composition profile at the top of the tower. It will tend to move the top reflux around too much, upsetting the rest of the column. To hold the temperature to the setpoint, and follow setpoint moves, the TC needs a decent gain on temp with reflux moves, and it needs to be tuned properly. Same thing with compensating for other IND moves.

: Tuning should be fairly aggressive. Slow (integral dominant) tuning will increase the Tss of all the responses on the column, and will increase the plant testing time too much. Since a column temperature is relatively slow, you want proportional dominant tuning. A lot of times, some derivative action will help too, if the measurement isn’t noisy. I also like proportional kick action on setpoint changes with this, to speed up the TC response - since the MPC controller will be moving the setpoint around more than the operators would. Tuning should be verified with setpoint step changes and disturbance step changes.

: Another thing to consider is where are the constraints at the top of the tower. If you are overhead condenser limited, and you get into situations where the top TC can't keep up with temperature increases (and just keeps dumping reflux into the column) that's a bad situation. Also, are you typically reflux pump limited? That's another case where you may not want to keep the top TC in closed-loop.
:  
: Sorry it took so long to follow-up on your original posting.


Optional Link URL:  
Link Title:         
Optional Image URL: