Distillation Group, Inc. Home Services Newsletter Contact Distillation Technical Software Training Questions Links Search Navigation

Wash Zone Coking

Poor product quality from a main fractionator

Petroleum refinery visbreaker main fractionator

Color problems with visbreaker HVGO product

Subject: Visbreaker HVGO colour
Date: 07:40 PM 10/21/00 +0000


I need information about vacuum distillation. I was read your technical paper's and I am interested in improving vacuum distillation column performance and troubleshooting vacuum distillation columns. I have problems with colour of HVGO, where the colour is never good (always >D.8). Ratio for slopwax and HVGO reflux (wash) is already at a maximum. A little information for my vacuum unit, it is in a visbreaker plant with a visbreaker capacity of 8,600 ton/day.

A., Asian Refinery

<Information sent>

Subject: Visbreaker HVGO color
Date: Fri, 15 Dec 2000 15:28:16 -0600


Have you had any success improving your visbreaker product color?

Andrew Sloley.

Subject: Visbreaker HVGO colour
Date: 05:38 PM 12/16/00 +0000


Not yet, and thank for your follow-up. I want to give you a little information about my vacuum unit, sorry I can not e-mail you a figure for the unit. The visbreaker unit has a capacity of 8,600 t/d. For the vacuum unit, wash and HVGO sections are grid. Slopwax rate is 25 m3/h and HVGO pumped down as wash is 100-130 m3/h. Colour of HVGO is always bad. Will it help if I change the grid for packing? Could you give me some explanation about advantage grid versus packing?
Thanks for your information.


A., Asian Refinery

Subject: Visbreaker problem
Date: Thu, 21 Dec 2000 14:34:37 -0600


What diameter is the tower and what pressure does it run?

Andrew Sloley.

Subject: Re: Visbreaker problem
Date: Thu, 21 Dec 2000 13:59:50 -0800 (PST)


Sorry that the information was not complete!
About the tower :
- Height : 35800 mm
- Diameter : 4300 (bott), 11600, 3200 (top) mm
- Pressure : Top = 10 - 14 mmHg
Bott = 38 - 42 mmHg
- Trays : 3 draw trays
4 grid beds

A., Asian Refinery

Subject: Visbreaker problem
Date: Fri, 05 Jan 2001 11:06:05 -0600


Figure 1 shows a sketch of your unit with the data marked on it. I can make some specific comments for your situation plus some general comments on the subject of refinery main fractionator wash zones.

Figure 1
Visbreaker vacuum tower with operating data

Specific Comments

Current Operation

The main causes of poor quality HVGO products from refinery main fractionators are coked wash beds, poor liquid distribution to the wash bed, and poor vapor distribution to the wash bed. These causes are related. Poor distribution to a wash bed, of either vapor or liquid, leads to coking.

Coking partially blocks the wash bed. Vapor velocity increases due to the lower cross section area open. The higher vapor velocities increase entrainment. Entrainment carries black oil from the feed entry (flash zone) up to the HVGO product. Black HVGO results. This severely affects downstream operation and product quality. Your D8 color is typical for a coked wash zone.

Your data shows a pressure drop of 28 mm Hg across the tower. This is a very high pressure drop for a normally operating vacuum tower. Unless your collectors have exceptionally high pressure drops and the grid beds are very deep, a typical pressure drop in this service would be 12-18 mm Hg across the tower. High pressure drops occur across coked beds.
While incomplete, the available evidence supports the contention that the wash bed in your tower is badly coked. Coked wash beds lead to poor product qualities.

The process operation must also be checked. A commonly used number for grid or structured packing wash zone liquid rates is 0.15 gpm/ft2 (0.367 m3/hr) of tower cross-section area at the minimum liquid wetting point. For wash zones, the minimum liquid wetting rate is found on the bottom of the bed. The rate your data shows is 0.097 gpm/ft2 (0.237 m3/hr) of slop wax. This rate is low and is a probable contributor to coking. More details on liquid rates and data interpretation are in the general section comments.


The bed must be replaced. The wash bed must also be modified. Whatever caused the bed to coke in the first place must be identified and fixed. This may be a complex activity. The root cause of coking problems may not be apparent. Many units have been 'fixed' only to have coking occur again. Unit reliability and product quality depends upon identifying and fixing the real problem.

Both grid and packing can coke. No clear evidence exists on the superiority of either grid or structured packing in this service. Vapor and liquid distributor design, fabrication, and installation are so much more important that any minor differences between grid and structured packing can be ignored. Random packing should not be used in this service. Random packing inevitably has parts of the packing that hold liquid for long periods of time. Long residence times increase the risk of coked beds.

The process should be checked as well as the mechanical design. Current wash rates appear to be low for reliable operation.

Visbreaker Service Specific Comments

Visbreaker products are thermally unstable oils. At the high temperatures in wash zones they readily coke. Extreme attention to mechanical design details must be included to eliminate dead spots where coking can occur.

General Comments

Thermally Unstable Oils

Many refinery main fractionators process thermally unstable oils. Common services include:
1. Atmospheric crude columns
2. FCC main fractionators
3. Gas oil crude columns
4. Vacuum preflash columns
5. Vacuum crude columns
6. Delayed coker main fractionators
7. Fluid coker main fractionators
8. Visbreaker atmospheric columns
9. Visbreaker vacuum columns
10. Residue hydrocracker atmospheric columns
11. Residue hydrocracker vacuum columns
While services differ between units and plants, the list has been sorted into a generally least severe to generally most severe order.

Reliable operation with thermally unstable oils requires a great care with mechanical details. Coking is a product of time, temperature, and thermal instability. Mechanical details that create small liquid pockets or films with long residence times initiate coke formation. Once started, coking may continue until major problems develop.

Grid versus Packing in Wash Services

Both grid and packing can coke. No clear evidence exists on the superiority of either grid or structured packing in this service. Vapor and liquid distributor design, fabrication, and installation are so much more important that any minor differences between grid and structured packing can be ignored.

In general, grid will require a deeper bed than structured packing for the same de-entrainment effectiveness. For a given degree of wash bed effectiveness, pressure drop across grid or structured packing will be approximately equal. Grid has been used more often in this service because it has been available for a longer time than structured packing.

Random packing should not be used in refinery main fractionator wash service. Random packing inevitably has parts of the packing that hold liquid for long periods of time. Long residence times increase the risk of coked beds. Figure 2 and Figure 3 show drawings of typical, modern random packing. Random packing fills a vessel randomly. Some packing will always lie with spots where liquid can have long residence times. This is especially true in low liquid rate services. Wash zones are low liquid rate services. Long residence times, high temperatures, and thermal instability of the oil lead to coking.

Figure 2 [1]
Side and end view of ring

Figure 3 [2]
View of a ring

In contrast, Figure 4 shows a drawing with elements of typical, modern structured packing. The surface can drain freely. Coking tendency is reduced.

Figure 4 [3]
Structured packing elements

Wash Zone Liquid Rates

To keep terminology clear, we will use the following terms as shown in Figure 5:

The obvious question looking at Figure 5 is why do we make a distinction between overflash and slop wax?

Figure 5
Flash zone terminology

Figure 5 is not complete. We have not considered entrainment. The purpose of the wash zone is to remove entrainment from the feed. Entrained oil makes the products black. A very small amount of entrainment can make a product D 8 color.
We normally consider entrainment a very small quantity. This is not always true. At very low wash rates and high vapor velocities in the column, entrainment is a significant fraction of the liquid on the collector tray. In some unit, entrainment may reach 100% of the slop wax. Figure 6 expands the definitions from Figure 5 to include entrainment.

Figure 6
Flash zone terminology with entrainment

Obviously, with a functioning wash zone the entrainment never reaches the product above the wash bed. Therefore, the slop wax liquid is:

Measured slop wax is never equal to overflash.

Care must be taken with data interpretation to make sure that the overflash rate is determined when deciding if a wash bed has liquid coming out the bottom of it. Common methods of determining overflash rates include metals and asphaltene balances around the feed and slop wax. Other techniques can be used as well. It is critical to avoid confusing slop wax rates with overflash.


Many plants have very good experience and routinely run for five years without wash bed coking. Others coke as often as every 18 months. Understanding the process and correct mechanical design, fabrication, and installation is required for reliable wash zone operation. The hotter the operation and the less stable the oil, the more important every detail becomes. Standard design approaches based on old-time, low severity services fail to perform acceptably in severe services.

Andrew Sloley.
The Distillation Group.


[1] From Strigle, R. F., Jr. and Porter, K. E. US Patent 4,303,599 Tower packing. 1 December 1981.
[2] From Nutter, D. E. US Patent 4,576,763 Packings for gas-liquid contact apparatus. 18 March 1986.
[3] From Chen, G.; Kitterman, B. L.; Axe, J. R. US Patent 4,604,247 Tower packing material and method. 5 August 1986.


No performance, suitability for use, or lack of suitability for use for any given process service is implied to any particular model or brand of packing by these comments. Figures used have been used as illustrative of generic classes of equipment.


Images have been sized for full screen display on an 800x600 monitor.

This page updated 05 January 2001.
© 2001 Andrew W. Sloley. All rights reserved.