9 replies to this topic

[joeblack]

Hi
I'm working in an EPC company. Since CDU/VDU is an open art design, I am assigned to look for design and simulation of a crude distillation unit. I made a study and search on this topic and just found the "Handbook of Petroleum Processing" as a good refrence for design. Although this book is published in 2006, the whole design calculations such as detemining number of trays, pumparound duties; specifying flash zone, overhead, residue and... conditions; number of draw-off tray and all other properties are just based on repeating manual calculations. I was wondering is it the only way to design a tower? Is there any design solution that at least is integrated with softwares? can I design a tower just using a software or not?
I will really appreciate if someone help me or introduce me a good refrence so I can overcome this complicated design.

Best Regards

[Zauberberg]

Joe the Black,

If you never had a chance to design and/or operate a CDU tower before, my best advice is - do not try to design one just by yourself, and especially by using simulation software only. Those are the most complex distillation systems, involving multiple product draw-off's, multiple pumparounds, varieties in overhead system design and operation, control challenges, flexibility issues for various crude oil processing opportunities in future, and many, many more.

If I was at your place - I assume that you already know what kind of crude oil is to be processed, and what products and their specifications are required - I would first evaluate the best configuration upstream of the tower: hot and cold preheat trains, preflash drum/tower (yes or no?), and then move to design of the tower itself.

Probably the best two approaches you can have in this complex task, are:

1. After developing preliminary results from process model, and after furnishing preliminary H&MB, talk to the vendor (Koch-Glitsch, Sulzer) who can supply you with quality answers, preliminary column sizing, and then move the things further. Being in close communication with equipment vendor is something I would strongly recommend since - at the bottom line - they will do the final design work. Make sure that you have properly characterized crude oil (TBP, density, viscosity curves, Light ends content, PNA for Naphtha cut, Cetane numbers and indexes for Kerosene/Diesel cuts etc. - obtain a typical Crude Assay if you don't have it already), and quality ranges for all your products.

2. Another option is to speak directly to various Technology Licensors. In spite of the fact that CDU's are, more or less, an open art, there are reputed Licensors who can do that work (or majority of it), similarly to catalytic processes (e.g. HT, FCC, Hydrocracking) or gas liquefaction processes (e.g. APCI, ConocoPhillips, Linde). One of these Licensors - if I remember well - is Axens (IFP group technologies) from France. Perhaps you can Google-around and see who else might be appropriate.

Good luck,

[joeblack]

Thank you Zauberberg for you reply. By the way I'm neither Joe nor black, its just a nick name

Yes absolutely you are right but maybe I was not clear enough on previous post. I didn't mean that I'm going to design a tower with all details; just process aspects are needed. So the flexibility of tower for future development or something like this, is not intended.
The crude assay is available and the specification of products are given as well. The whole tower configuration and side products are determined before:
-Overhead with one product drum and total condesing
-Three side products: Kerosene, LGO and VGO
-Two pumparounds
-Two of three side strippers working with steam and one of them has reboiler
and so forth.

In fact the goal is we are going to change the cut point slightly and predict the change in product distribution and specification. This may need to specify number of trays, the draw-off trays, flash zone and tower top operating conditions, calculating required PA duties, reflux ratio and so on.

The upstream configuration is typical heat exchanging with products and has a preflash drum. We are in contact with Axens not only for this project, but also for previous jobs. This is not going to be our final design to make the project basis. Even making a draft will be rechecked and revised with an experienced professional like Axens.

Let me know if you still insist I have to give up and appreciate any other comment from you.

[Zauberberg]

Your job now appears to be much more straightforward.

If you have given the tower configuration (diameter, type and number of trays, location of draw-off's and PA/s etc.), feed conditions, and main equipment design performance (exchanger/cooler duties, pump capacities, control valve flows/dP's etc.) then you can try to build the simulation model reflecting tower design performance. Then, you can use it to simulate what you would need in reality, i.e. changing the product cut points, for example. This is accomplished in several ways:

- Changing product draw-off rates, and
- Changing pumparound flows and duties, including overhead reflux as well (tower top temperature), and
- Changing feed conditions (enthalpy)

The model can reasonably predict what ranges of cut point change you can achieve without having the need to install additional trays (once when the tower is erected it is too late for that), and see if this matches your requirements and expectations.

Best regards,

[joeblack]

Thanks again
Although only the configuration is available but the results are not trustworthy. That's because the basic design of the tower was given to a new company and we tried to simulate that tower with whole available and given data but could not reach same results, that's why I needed to try re-calculate the tower and see if I can at least get close to the given results. Just using a simulating software tends to a try and error procedure and its not my favorite.
By the way thanks a lot for your constructive advices. I will reconsider the simulated tower and continue on simulation file.

Edited by joeblack, 15 March 2010 - 12:37 AM.

[Zauberberg]

You need to rely on the column internals vendor (I presume Koch-Glitsch or Sulzer) - their results are the most important ones. Can you explain how the company you mention in your post did the tower design work? What type of internals are used, and have the calculation been verified by the vendor?

If the company has sized the tower without contacting KG or Sulzer for confirmation of results, and if the exact type of internals was not specified, then the results are absolutely useless and you don't have anything to start with.

[joeblack]

The CDU/VDU PFDs and tower data sheets are available. Some of the internals, tray efficiency, vapor and liquid loads, tray type and so forth are mentioned in data sheet but I'm almost sure they are not confirmed with vendor because some items like weir hight are noted to be finalized by vendor. Since the company's simulation file is not given to us as well, I suspect the results maybe phony (it's just a guess). But using these information I cannot get same results from simulation. You said if it's not confirmed with vendors like Koch-Glitsch the whole design is useless, so I think we cannot verify their work and my diligence is going nowhere.

It is worth mentioning that their basic design was going to be sent to KBC for verification, but it is held and that's the reason I started to simulate their work.

[Zauberberg]

No reason to do so - you'll put a lot of efforts into the work, and the results you get will always be considered non-reliable. You would probably be the first person to suspect the model results, if the model was built based on someone's preliminary design work.

The suggestion is: wait for the final datasheet issued by the vendor, and then build a model that matches with the results closely. From that point onwards, you can try to play with it and see how many "degrees of freedom" you have in changing design draw-off's and product cut points.

Even better, the operating flexibility you want should also be discussed with the vendor. One way to solve all the problems during design stage, is to build several cases that reflect different operating conditions, different cut points etc. and then submit the initial datasheet containing all the options to equipment vendor. If it is possible to design the internals that will make all the options viable, you've got everything that you wanted.

[Zauberberg]

Here is one interesting document that can assist you in future work:

Attached Files

•  CDU Simulation using PROII.ppt   751KB   198 downloads

[joeblack]

Here is one interesting document that can assist you in future work:

Thaks a lot sorry that I reply too late. Didn't check forum for a long time