How to control H2S Stripper 1

[jcee]

Dear all,
I am new to this forum and dont know if this is the right place to get some ideas/answers/help?! So I will just start to ask a question ;-)

I work for a german refinery and my new job here is to rework all the Advance Process Control Systems. As I am more skilled in programming/IT than in process engineering I have problems fully understanding the chemistry/physics behind. So could anyone give me a hint where and how to start...?! Thanks a million.

My current problem: A HS2 stripper which recieves desuplhurized distillate from a CHD. Its objective is to stripp out the H2S and at the same time allow to control to Flashpoint.

I can mainly adjust the: reflux, head_pressure, head_temp, bottom_temp (via reboiler), level (via rundown)

What I do monitor (online): Flashpoint,H2S

I have to asume: that composition and desulphurization level varies a lot. often product and throughput changes.

What I am faced with:
1) Flashpoint limit not reached (=non optimal operation)
2) Sometimes H2S breakthrough in the bottom rundown happens.

I am really scared when 2) happens and that has to be avoided as much as possible but on the other hand FP has to be optimized.

A had a look at the process data for many days but can not figure out the complex relationship between pressure/temp and timing.. Seems that H2S breakthrough happens from time to time and mainly when shifting to higher throughput!!

What is the best approach? And ideas for me? We are also interested in vendors delivering suitable controls to implement... our current APC is offline 90% aas it seems not fast/intelligent to handle the process safely...

:-(

Please help, any tipp is welcomed..
Jens

 

[25362]

Do you use direct stripping steam injection and/or a reboiler? What is the product distillation range ? Have a look at Process control systems (Mc Graw-Hill) and Controlling multivariable processes (The Instrument Society of America) both by F.G. Shinskey. Also Elements of Petroleum Processing by D.S.J. Jones (Wiley).
You may find ideas of worth.

I once had an exchange of correspondence with a Jens working in the Hamburg BP refinery on the subject of lube solvent dewaxing. Is it by chance you or was it another Jens?

Good luck.

 

[jcee]

Distillation range is 160 to 340..360°C, blending components come from (straight-run) crude distillation tower and/or intermediate product tankfarm.
Sulphur, fash and distillation requirements are changing once..twice a day.
We are only stripping with reboiler, no steam. Its really difficult to operate the stripper to reduce flashpoint giveaway and make sure that no H2S breakthrough happens.
Currently we are running the unit with a (poor) APC, consisting of some PIDs and feedforward but that seems unefficient.. :-( .. the common problems..
As far as I can see, we need some advanced control strategies.. hear a lot about neuronal networks, reasoning, multi variable etc..
But dont know what a good approach would be.. Will have a look at the books you mentioned.

;-) I am not the Jens from Hamburg BP, we are in Wilhelmshaven. But we have another Jens here. He is working in technical department and as I remember, he is from Hamburg.. Maybe its him ?!
Was his name "Jens Br..."(?!);-)

regards

 

[25362]

No, his name was Jens Mer.... Anyway, I wish you good luck in your endeavours.

 

[RogerH]

jcee
It is very difficult to get the flash down to levels closer to the specification (56/60°C) in a Diesel Hydrotreater. The amount of heavy naphtha/Kerosene is very small produced in the hydrotreater and in principal there isnt any of this material available. This is the reason that you most probably won't be able to control your flash point. You have to accept that your flahs will come out in the 90°C region.
The only way to control flash would be to feed your hydrotreater with some Kerosene or heavy naphtha, e.g. 5-10 vol% on feed. Then you have a chance to control your flash or to have some catalyst with cracking activity (going towards a MHC). The latter might make sense (if you have H2 enough) for selective (back end) MHC. Criterion did have such cat's on their "cat-meny", but I dont know how well they are working. Somebody else out on the

http://www perhaps

have experience of this?