Quench column bottom level cascade to top reflux

[dagur777]

Does anyone has experience in bottoms level cascade to top reflux in distillation column control? The current plant setup is the top reflux drum level cascade to distillate product flow. Column bottom level cascade to top reflux. I haven't seen this kind of setup. I think it can cause the column unstable using bottom level to control top reflux. Any comment on this? Thanks

 

[DickRussell]

Your suspicion is reasonable, but tell us more. What controls the rate of bottoms pumpout - downstream needs? What feeds the column, and how is that flow controlled. How is reboiler heat input controlled? Then there is the fundamental question: is the column presently running in stable fashion and performing the desired separation, despite concerns? Normally, reflux would be controlled and perhaps reset by something like a temperature so as to maintain separation in the column. Levels are used for inventory control and reset either feed or pumpout rates.

 

[UmeshMathur]

This looks like a very bad control scheme to me, because the entire lag (dynamic delay) between column top and bottom trays, caused by holdup on the trays, might easily make the bottoms level controller erratic and difficult to tune. All major disturbances would really hurt such a control scheme. Personally, I have never seen an example of such a cascade in 39 years as a ChE.

If bottoms flow MUST be on flow control, I would consider controlling bottoms level with reboiler heat. This works only when you have an adequately high boilup to bottoms flow ratio, however, as in a tower which also has a high reflux ratio. Otherwise, bottoms level should be cascaded to bottoms flow.

 

[dagur777]

The bottom is on flow control with very small amount. That's why level controller cannot cascade to it.

 

[UmeshMathur]

Hence, you ought to cascade the bottoms level to the reboiler heating medium flow. The overhead reflux should then be on flow control, and the overhead draw on reflux drum level. This is a perfectly stable and workable scheme in such situations. See the following for an example of a control scheme with a very high boilup ratio (butane removal column):

U. Mathur and R.J. Conroy: "Successful multivariable control without plant tests", pages 55-65, Hydrocarbon Processing (June, 2003).