RANDOM PACKING DEBOTTLENECKS REFINERY DE-ETHANIZING STRIPPER

Stephen J. Deley
BP Oil Co.
Toledo, Ohio

Kenneth Graf
Norton Chemical Process Products Corp.
Stow, Ohio

BP Oil Co. successfully packed a de-ethanizing stripper at its Lima, Ohio, refinery to improve capacity and ethane removal. Design capacity increased from 76,000 b/d to 92,000 b/d, and ethane in the LPG product decreased from 6-7 LV %% to 3.5 LV %.

The improved performance has been evident in 2 years of operation since the revamp.

Critical project decisions included:

Feed preheat
Liquid distributor design
Water decanting capability
Weld-free internals supports
Materials of construction
Support-ledge removal.

STRIPPER TOWER

The 150,000 b/d BP refinery in Lima, Ohio, includes a saturates gas plant to separate saturated light ends into fuel gas, propane, isobutane, butane, isopentane, and naphtha. The C2/C3 separation is performed by an absorber/stripper combination, as is common in fluid catalytic cracking unit (unsaturated) gas plants.

The absorber/stripper system includes a primary absorber, high-pressure condenser and separator, and de-ethanizing stripper. A simplified process flow is shown in Fig. 1.

The saturates gas plant is very heavily loaded, particularly during the summer months. Compounding the problem is the increase in light ends over a period of several hours during a reformer reactor swing. The frequency of reactor swings depends on reforming severity, which is greater in the summer gasoline season.

The stripper tower is 9 ft in diameter. Before the revamp, it contained 20 three-pass valve trays, most at 24in. tray spacing.

The stripper has both a bottom reboiler and a side reboiler, with a draw and return between the fourth and fifth trays from the bottom. Built in 1969, the stripper was retrayed for additional capacity in 1984.

Before the revamp, the overloaded stripper exhibited poor ethane rejection from the bottoms. A gamma scan of the tower indicated downcomer flooding.

As a result of the overloaded stripper, the downstream LPG production was off-specification during the summer, so a portion of the LPG was flared to reduce the ethane content.

There also was a loss of propane to fuel gas, estimated to be 250 b/d. This loss occurred because the lean oil rate to the absorber had to be reduced to help unload the stripper.

PROJECT SCOPE

An engineering study resulted in the recommendation that the stripper be packed to debottleneck the saturates gas plant.

BP planned to engineer this project for implementation during the next major turnaround. The project proceeded quickly, however, when a sudden window of opportunity opened.

Trays above and below the side reboiler were evaluated independently for replacement with packing. Because the stripper tower's highest internal loadings occur above the side reboiler, packing was required for this section of the tower.

The trays below the side reboiler, however, experienced lower loadings, and their replacement would have been costly and time-consuming. Norton Chemical Process Products Corp. recommended against changing these trays.

In the engineering study, both structured and random packing had been considered for the stripper tower.

In recent years, structured packings have generated much interest. These packings now have the reputation of being universally more efficient than random packing.

Norton recommended against structured packing, because there is little experience with it in this service, at this liquid rate (52 gpm/sq ft). Experience with structured packing is limited to 45 gpm/sq ft in distillation service.

Norton recommended No. 50 IMTP packing - a nominal, 2-in., high-performance random packing - for the stripper tower. The IMTP packing had been used in at least 10 de-ethanizing strippers before this revamp. At design loadings of 3,308 gpm liquid and C, (capacity factor) of 0.110 fps vapor, Norton rated the No. 50 IMTP packing at 82% of its capacity in this system.

BP agreed to design for a 25 ft, 10 in. deep bed of this packing to replace the 16 valve trays above the side reboiler.

DESIGN CHOICES

After the basic project was identified, many decisions needed to be addressed:

Feed vaporization. Two phase feeds are more difficult to handle in packed towers than in trayed towers. De-ethanizing strippers often use feed preheaters to add inexpensive heat in place of reboiler heat, and to relieve stripper internal loadings.

In raising the feed temperature by 10-150 F., BP's preheater caused slight vaporization of the stripper feed. Feed vaporization would not affect the trayed tower, but it would be important in the packed tower.

Norton could design for the possibility of two-phase feed, but this would require expensive feed equipment and the sacrifice of 2 ft of packing height. Instead, BP chose to design for afl-liquid feed. Thus, if the feed began to vaporize and the tower efficiency dropped, BP would reduce the preheat to correct the tower operation.

Liquid distributor. BP preferred a trough-type liquid distributor at the top of the stripper tower. In addition to its ease of installation and of sealing, the trough-type distributor lends itself well to post-fabrication testing with water.

Norton recommended a deck-type distributor, as shown in Fig. 2. This distributor can better handle the 50 gpm/sq ft of liquid flow without allowing hydraulic gradients that would cause liquid maldistribution.

Because this type of distributor does not completely contain its liquid (the tower wall contains the liquid), it cannot be water-tested in a fabrication shop unless a piece of tower shell, including support ring, is also fabricated.

The aggressive project schedule (4 weeks from order until shipment) did not allow Norton time to fabricate a piece of tower shell and conduct a water test. Based on Norton's experience and recommendation, BP elected to accept the deck-type distributor, without water test, for the stripper tower.

Water decanting. The trayed stripper tower had three trays with inlet sumps for water decanting. A small, continuous flow from the sumps back to the high-pressure separator may have been removing free water, but it was not clear whether the flow was effective or necessary.

The packed tower would have no possibility of decanting water, as packing height could not be sacrificed to make room for a tall water decant tray. After some deliberation, the loss of decanting capability was accepted, for three reasons:

Unlike absorber/de-ethanizers without high-pressure separators, the presence of the high-pressure separator with effective water removal keeps most of the water out of the stripper.
The stripout can remove a small amount of water from the tower.
The old water "decanters" from the tray inlet sumps probably were very ineffective anyway, as the liquid velocities and turbulence were much too high in the tiny sumps.

Weld-free supports. The packing-support plate and liquid distributor required complete support ledges at the location of, respectively, former Trays 6 and 19. Tray support rings were available at these locations, but each ring was incomplete in three spots where the down-comers had been.

BP preferred not to weld ledge pieces in to complete these rings because welding to the shell would require extensive stress-relieving. Norton was able to provide ledge segments that bolted in, to complete the two rings without welding. The completion of Ring 19 had to be rope-gasketted along the tower wall to seal the liquid distributor.

Materials of construction. The choice of 410 stainless steel (SS) for the fabricated tower internals (support plate, liquid distributor, etc.) was straightforward. It matched the existing trays, and it is resistant to chloride attack. It also is inexpensive and readily available in 10-16 gauge thickness.

The decision to use 410 SS for the packing, however, was not automatic, as the material is neither inexpensive nor commonly available in thickness suitable for packing manufacture. BP considered stainless steel types 409, 410, 430, and 304, but chose the 409 SS, as it was technically preferred over 430 and 304, and is less expensive and more readily available than 410 SS.

Support ledge removal. The elevation layout of the packed stripper tower put the tower packing into a shell area containing 12 support rings and 72 downcomer bolting bars. The possible removal of these old attachments was an important issue, as it would add much time to the revamp schedule.

Norton immediately recommended removal of the downcomer bolting bars, as old bolting bars interrupt packing uniformity and produce vertical channels for vapor and liquid bypassing. The sloped portions of the downcomer bolting bars also would collect liquid and move it sideways in the tower. The need to remove the horizontal rings, however, was more questionable.

By consuming about 6% of the tower cross-sectional area, the presence of the 2.5-in. wide rings would have a small detrimental effect on both tower capacity and efficiency. Norton recommended removal of the 12 support rings.

REVAMP EXECUTION

A reinstallation meeting was held at the refinery with a Norton representative, the BP maintenance and engineering team, and the installation contractor to discuss the specifics of the packing job. This advance preparation proved to be very beneficial when the work began.

Another benefit was having Norton personnel on site during the entire packing installation. This round-the-clock coverage enabled immediate response to any issue, as well as close observation to ensure installation according to design.

The project execution plan was to remove the tray support rings and downcomer bolting bars. As the stripper revamp was the critical path item during this brief plant shutdown, removing the tray support rings and downcomer bolting bars became a significant part of the overall schedule. Although BP agreed to remove the downcomer bolting bars, the support ring removal issue was reopened.

Because the stripper was designed for a significant capacity increase, the minimal loss in capacity created by not removing the tray support rings was not deemed a problem. BP also believed that the packing revamp would gain efficiency relative to the overloaded trays. If so, a risk of slight efficiency loss caused by the old tray support rings would be tolerable.

Because of shutdown timing constraints, BP decided to not remove the tray support rings at that time. If the expected efficiency was not achieved, the rings were to be removed during the next scheduled shutdown. (Because no efficiency loss has been perceived, there now are no such plans.)

The installation of the packing and internals was uneventful. The packing was loaded from the 2 cu ft shipping boxes into the tower through a makeshift chute.

The packing loading was completed in less than 8 hr. The entire installation was completed within the anticipated time frame; the time from tower opening to tower closing was 7 days.

START-UP, OPERATION

Start-up of the stripper with the new packing went smoothly. The tower operated well at the initial flow rates (about 45,000 b/d, compared to a design capacity of 92,000 b/d). The ethane removal was improved significantly over operation with trays.

During this time, the stripper feed preheater was bypassed to minimize the likelihood of feed vaporization. The cooler stripper feed was not a problem at low rates because the reboiler duty was sufficient to release the ethane.

When the stripper feed rate was increased, however, the operation of the tower deteriorated rapidly. With the low feed temperature at higher rates, there was not enough heat input to the tower to achieve the desired ethane removal.

The lower feed temperature caused a buildup of ethane in the stripper. After several hours, the buildup would release large amount of vapor for a few minutes. This problem was eliminated by using the preheater to increase the feed temperature. The higher feed temperature immediately resulted in the smooth, stable operation of the stripper.

Although there could be a small amount of vaporization of the feed, there has been no observed loss of efficiency. The average preheat is now about 100 F.; the typical feed temperature is 1200 F.

After the tower operation became steady, the stripping severity was adjusted to meet the vapor pressure specification of the LPG product. The LPG vapor pressure specification of 208 psig is equivalent to about 3.5 LV % ethane in the LPG. Before the revamp, the LPG product typically contained 6-7 LV % ethane.

in addition, the lean oil flow rate to the absorber was increased for additional propane recovery from the fuel gas. Because of the increased packed tower capacity, the absorber lean oil flow rate can be maintained higher than it was before packing the stripper.

The elimination of the three water draws from the stripper has not been a problem. The small amount of water that enters the stripper exits with either the "stripout" or the bottoms; it does not build up in the tower.

As was the case with the tray operation, ammonium chloride salts tend to build up in the stripper tower. This buildup manifests itself via somewhat erratic operation of the stripper.

To remedy this situation, BP occasionally washes the stripper on-line, for about an hour, using high-pressure steam condensate. In every instance, this apparently has eliminated the salt buildup.

Analysis of operating data from the saturates gas plant indicates that the stripper packing is generating 10 theoretical stages of stripping. This implies an HETP of 31 in. for the No. 50 IMTP packing, which is about the expected efficiency for the service and system involved.