Crude Distillation
Crude Distillation
(OP)
Hello Gentlemen,
We are working on tuning our DMC (Dynamic Martix Control) models for Crude tower and we are looking for better optimization methods to control the Naptha 90% cut point, Kero Flash and Kero freeze point. We usually control our Naptha 90% by manipulating the reflux to maintain the top temperature and our model was built based on this. However, after we commissioned the model the Naptha 90% target was on target, whereas, the Kero Flash point went off spec.
Several attempt methods were considered but it didn’t work, we finally took our model offline and we went back to traditional ways.
We do have an on line analyzer for Naptha, Kero specs. What DMC doses is it take the lab result and compare it with the anaylazer results and correct for results differences.
My question is, what is the typical way to control both Naptha and Kero spec if we consider dynamic control system?
We are working on tuning our DMC (Dynamic Martix Control) models for Crude tower and we are looking for better optimization methods to control the Naptha 90% cut point, Kero Flash and Kero freeze point. We usually control our Naptha 90% by manipulating the reflux to maintain the top temperature and our model was built based on this. However, after we commissioned the model the Naptha 90% target was on target, whereas, the Kero Flash point went off spec.
Several attempt methods were considered but it didn’t work, we finally took our model offline and we went back to traditional ways.
We do have an on line analyzer for Naptha, Kero specs. What DMC doses is it take the lab result and compare it with the anaylazer results and correct for results differences.
My question is, what is the typical way to control both Naptha and Kero spec if we consider dynamic control system?





RE: Crude Distillation
In a traditional refinery flash points are taken care off in the side strippers,whereas pour points are related to end points,for kerosene the higher the end point,the higher the pour(freeze) point..and also depends on the crude type and paraffin content.
Having said ,the model you use is as good as the information you provide ..
I guess in this case you have to re-evaluate the input to your model..before relying on it
RE: Crude Distillation
I believe using the reflux to control the Naphtha 90% point is correct, and the Kero flash would be controlled by adjusting steam to the side stripper.
The Kero freeze point would be adjusted by the Kero draw rate.
An interesting thing is that you can increase stripping steam to the top draw product (Kero in your case) for INCREASED Kero production at the same product specs. This is because you will move some of the very light initial components from the Kero into the naphtha (which will be swapped by some heavier components from the Naphtha). Simply an improvement in separation due to the reflux increasing as you increase the stipping steam.
I found that a good tower simulation can really help the Applications Engineer set up the model based predictive controller.
Regards,
Bob
RE: Crude Distillation