One of the frequent causes of extended coke drum cycles is the difficulty in filling the drum with cooling water. Once feed is switched out of a coke drum, the coke tends to settle back and partially plug the feed nozzle in the bottom head. This effect is especially pronounced when shot (i.e., BB-size pellet) coke is produced because of charging a low-gravity feed. Not only is the time required to fill the drum with water extended, but it becomes difficult to drain the drum.
To keep the feed nozzle open during a switch, a small amount of steam must be cracked into the drum that is being taken off-line. By the time the feed has been diverted from a full drum, the flow of steam should be several thousand pounds per hour. In practice, the steam flow is adjusted to maintain the pressure in the feed line several psi above the pressure in the bottom of the coke drum.
Coke drum overhead vapor valves are massive affairs. Switching a pair of 20-ft diameter drums may require the manipulation of four 18-in. and four 12-in. valves. Size aside, the vapor valves are difficult to open and close because of coke deposits on valve internal parts. Removing the insulation on piping upstream of the vapor valves or injecting a small liquid quench (typically 3 vol%) into the vapor line at the first 90° turn will help.
To minimize the amount of hard physical labor required to switch drums every 12 hours, the vapor valves should be fitted with locally controlled air operators. Such a device is little more than an air gun permanently mounted to the stem of each valve. The speed of the valve movement is controlled with a Win. air valve.
Figure 2-5 shows the valves used during the coking cycle. Prior to switching out of a coke drum, the process operators must be sure that the empty coke drum is truly empty. Trying to speed up the cycle by cutting hot feed (900°F) into a coke drum that is incompletely drained will cause a foamover.
The quick way to be sure that a coke drum is clear of liquid is to use a “tailpipe” thermocouple (Fig. 2-6). This temperature point is located between the coke drum bottom head and the condensate collection drum. When the coke drum is draining freely, the tailpipe temperature will be close to the coke drum temperature.
A cold tailpipe indicates a coke drum with liquid accumulated in the bottom head. About 30 minutes before switching time, the operating crew should verify that the tailpipe thermocouple is reading at least 370°F to 400°F.
Turning around a coke drum is a cooperative effort between the process operators and the decoking crew. It is labor intensive and normally accomplished without the intervention of supervisors. The work itself is best characterized as dirty, difficult, and if not done with care, dangerous.
To get a 12-hour cycle, every facet of the complex task of turning around a drum must be carried out at maximum speed. This means that both the decoking crew and the operators must be personally committed to the objective of short cycles.
One demonstrated method of obtaining this commitment is to put the decoking crew on a piece-work basis. Then the decoking crew cuts one coke drum and goes home. Typically, when the 12-hour cycle is being achieved, the decoking crew will spend four hours on the job site.
But it is somewhat unsettling to pay a person for eight hours of work and only have him on the job for half that time. In one refinery, management attempted to rectify this matter by assigning the decokers to general maintenance work after they completed cutting a full coke drum. Suddenly, the time to cut a drum of coke increased from four to seven hours.
Motivating operating personnel can be difficult. Adding labor-saving devices such as air operators on the large vapor valves will demonstrate to the process operators that refinery management is placing great emphasis on short cycles. A careful and honest review by first-line supervision explaining how increased coke production will enhance refinery profitability is also helpful.
How long should it take to turn a coke drum around? The elapsed time from when one switches out of a full drum until it is empty and ready to receive resid again is called the cycle time. Most delayed coking units are designed for a 24-hour cycle time. Shorter cycles can be used in refineries where coke drum volume limits capacity. Each of the steps detailed in the Troubleshooting Checklist following this chapter can be contracted to save time.
The decided trend toward poor-quality crudes means that the tons of coke produced per barrel of resid are, on the average, increasing. Coke-drum size is therefore a more frequent bottleneck than is the capacity of the heater, combination tower, or wet gas compressor. With this in mind, a common assignment for an engineer is to determine how the operating crew can reduce cycle time.
As a basis for discussion, the details of operating on a 12-hour coke drum cycle will be reviewed. The discussion that follows applies to a 20-ft diameter coke drum producing about 1,000 T/D of anode-grade coke. The volatile combustible matter averages 1 2%. In really large-diameter coke drums (26 ft in diameter) producing hard coke with 8%-10% VCM, it is unlikely that a 12-hour coke drum filling cycle could ever be obtained. However, 16-hour cycles in 27-ft drums have become a common industry practice.
Increasing the recycle ratio is supposed to stop shot coke formation. However, in some cases, increased recycle ratio has no effect. Why?
Whether or not increased recycle ratio reduces shot coke formation is dependent on the composition of the incremental recycle.
Recycling 950°F-1,000°F coker gas oil will increase the formation of needle coke.
Recycling 600°F-700°F coker gas oil will reduce the formation of needle coke by raising the coke drum outlet vapor temperature. Heavy recycle is necessary to effectively suppress shot coke formation.
An additional area of concern relates to the quality of the coke produced from hydrotreated vacuum resid feeds. For a variety of reasons, premature furnace fouling, inclusion of semi-coked liquids in the quenched coke drum, and carry-over of coke fines into the fractionator, the delayed coking of 100% hydrotreated feed has not been successfully achieved in commercial operations.
Attempts lo do so have resulted in the production of high VCM coke and/or commingling of hydrotreated and virgin coker feedstocks.
Sulfur causes puffing of the calcined needle coke. Increasing the delayed coker recycle ratio by cutting the heavy coker gas oil end point will make more needle coke and reduce the sulfur content of it.
Other ways to improve needle coke quality include minimizing steam in the heater passes, reducing the gravitity of the recycle gas oil, and raising the coke drum pressure.
Needle-grade coke contains no shot coke. Commercial needle coke is typically produced by charging FCCU slurry oil to the delayed coker.
All of the components in the slurry oil coker charge are produced as a vapor from the FCCU reactor fresh feed. Even if vacuum tower bottoms are charged to the FCCU, none of the asphaltenes in the resid will vaporize, and therefore, the asphaltenes will not be in the coker charge.
The outside surface of shot coke spheres is coated with a layer of needle coke. The coating gives shot coke its polished-surface appearance.
The formation of cross-linked aromatic rings can be reduced by lowering the coke drum pressure, raising the coke drum outlet temperature, improving the vacuum tower operation, and by adding steam to the coke drum.