FCC REVAMP-Conclusion

Aug. 12, 1996
Bud Dahlstrom, Kevin Ham, Myron E. Becker, Tom P. Hum Consumers' Co-operative Refineries Ltd. Regina, Sask. Larry Lacijan, Tom Lorsbach UOP Des Plaines, Ill. Consumers' Co-operative Refineries Ltd. (CCRL) installed an elevated feed-injection system and a vortex-separation-based riser termination device on the fluid catalytic cracking (FCC) unit at its Regina, Sask., refinery.

Canadian refiner gets 2-year payout with FCC revamp

Bud Dahlstrom, Kevin Ham, Myron E. Becker, Tom P. Hum
Consumers' Co-operative Refineries Ltd.
Regina, Sask.

Larry Lacijan, Tom Lorsbach
UOP
Des Plaines, Ill.

Consumers' Co-operative Refineries Ltd. (CCRL) installed an elevated feed-injection system and a vortex-separation-based riser termination device on the fluid catalytic cracking (FCC) unit at its Regina, Sask., refinery.

The modifications have increased yield of high-value liquid products in the gasoline-to-distillate range and decreased production of low-value products such as dry gas, LPG, fractionator bottoms, and coke. These yield benefits have exceeded projections and increased refinery profitability by an estimated US$920,000/year.

The first article in this two-part series described the project objectives and the technologies incorporated (OGJ, July 29, p. 83). This one provides details of the project execution schedule and the economic and operational benefits realized as a result of the revamp.

Project execution

On Feb. 14, 1995, CCRL and UOP met to discuss the scope and economics of the project. Considering the economic benefits, CCRL decided to install the new technologies on a fast-track schedule during its June 1995 turnaround.

By not delaying the project until 1996, CCRL would realize an extra year's worth of yield benefits.

Engineering

On Feb. 21, UOP gave CCRL enough engineering information to prepare a cost estimate and submit a "Request for Expenditure." The FCC revamp was approved on Feb. 27.

Because of the tight schedule for June completion, CCRL elected to act as general contractor. This saved the time required to select a general contractor and relate the project details to the contractor's employees.

The company assigned the revamp project to a team comprising: one engineer working full time, two engineers working part time, and two draftsmen. A construction coordinator also was assigned to the project to supervise installation.

Fig. 1 [30618 bytes] shows the schedule of the major tasks accomplished before FCCU shutdown.

In a cooperative effort between CCRL and UOP, all long-delivery items were specified and tendered for procurement as quickly as possible. Only 12 weeks before the June turnaround, the main hurdle to completing the project on schedule was the ability to obtain critical components. Critical delivery items were: expansion joints, fabricated vessel components, 5-Cr piping and valves, control valves, instrumentation, strainers, spring hangers, and the Optimix feed distributors.

The CCRL/UOP engineering team completed the detailed design of the entire project at the end of May 1995. To reduce the schedule, components and piping were prefabricated to the maximum extent possible. Prior to shutdown, foundations were completed for new supports, and control wiring was installed on the FCC structure to minimize the work required during the turnaround.

Construction

The FCCU was scheduled to be shut down for 4 weeks to accommodate this revamp. Fig. 2 [49421 bytes] shows the schedule of the major tasks to be completed during this period.

The workers cut and removed the reactor head, along with the cyclones and dust bowls. The top head section of the reactor was lowered and placed on a special stand at ground level. The head section was then retrofitted with a new reactor-shell extension piece, new cyclone dust-bowl transitions, and new cyclone inlets and expansion joints.

While one construction crew worked on the reactor head section, another worked on the reactor internals. This team removed the existing riser extension with the T-disengager and installed the vortex chamber supports, vortex chamber, and new riser extension with swirl-type outlet arms.

When work at ground level on the reactor head section and the installation of the reactor internals were completed, the head was lifted back onto the reactor and rewelded. The reactor internals and refractory linings were then completed.

The Optimix distributor nozzles were prefabricated and delivered before the shutdown. Four cold-wall riser nozzles with 5-Cr internal sleeves were installed on the riser early in the shutdown to accommodate the new Optimix feed distributors and new piping.

A number of new nozzles were installed on the reactor shell and overhead vapor line to provide the instrumentation, purge steam, and stripping steam connections associated with the new equipment. Piping, instrumentation, and electrical modifications were ongoing throughout the shutdown.

Two new platforms were installed on the reactor for access to the new instrumentation. The platform below the regenerator was modified to permit access to the new feed distributor piping.

Start-up and operation

The revamped FCCU was brought on-line on July 4. The only notable change to normal start-up procedure was that purge-steam flows to the new feed nozzles and stripping-steam nozzles were maintained to prevent plugging.

CCRL and UOP believed that the hydrocarbon containment aspects of the vortex separation system (VSS) design also would affect the hot-air refractory cure, causing the hot air to bypass the upper sections of the reactor. For this reason, a vent to atmosphere was provided on the reactor dome to ensure adequate curing of refractory in the top section of the reactor.

During the initial phases of start-up, the unit experienced high reactor-side catalyst losses. The problem was traced to two points where purge-steam flow was entering the reactor just below the flapper valves on the cyclone diplegs. CCRL removed the purge steam flow, which immediately corrected the catalyst loss problem.

After this, catalyst containment with the new reactor internals was comparable to operation before the revamp. Reactor-side losses were about 0.04 lb/bbl, based on slurry bs&w results, while total unit catalyst losses averaged about 0.10 lb/bbl.

During this first start-up of the revamped unit, feed was introduced to the unit through the lower feed nozzle (Premix type) at the base of the wye. The catalyst level in the reactor section was controlled in the side stripper during initial operation.

On Day 3 of operation, the feed was routed to the elevated Optimix feed distributors. This change resulted in a marked reduction in regenerator temperature (lower delta coke) and a reduction in dry gas make. On Day 4, the bottom opening of the vortex chamber was sealed by raising the catalyst level in the reactor. Dry gas yield decreased further.

Subsequent unit start-ups have proceeded smoothly. CCRL's routine start-up procedure now calls for introducing feed through the elevated feed nozzles. On all start-ups, the catalyst level in the reactor section is maintained in the side stripper until the unit is lined out. The catalyst level is then increased to seal the bottom of the vortex chamber.

The operators have remarked that start-ups are more stable when feed is introduced at the elevated nozzles rather than at the lower feed injection point.

When the feed was directed to the elevated feed nozzles the first time, they discovered that the oil flow to one of the four nozzles was restricted. To date, this flow restriction remains unresolved. The restriction is specific to the oil side of the nozzle and is probably caused by foreign material left in the feed piping during construction.

Another feed nozzle became partially plugged during an extended feed outage in October 1995. Although steam flow was maintained to the nozzles during this outage, it was insufficient to prevent catalyst from backing up into one of the nozzles.

This problem stresses the importance of ensuring not only that the purge-steam flow is adequate during unit downtime, but also that the steam is evenly distributed to each nozzle. The flow restriction in this nozzle has decreased slightly over time, but obtaining balanced oil flow to the nozzle requires that the steam rate be somewhat reduced.

One notable consequence of installing the new equipment is about a 60° F. reduction in regenerator dense-bed temperature. This reduction causes a significant increase in the catalyst-to-oil ratio. Additionally, in order to increase delta coke sufficiently to raise the regenerator temperature high enough to burn the coke efficiently, the feed preheat was increased about 80° F. and the slurry recycle rate was increased. In addition, the vacuum unit cut point was adjusted to provide heavier feed to the FCCU.

The operability and reliability of the revamped unit is fully satisfactory. The unit operates smoothly, and no feed interruptions have been attributable to the new equipment.

Operator acceptance of the new systems has been favorable. A number of operator training sessions were conducted before the revamp. These sessions provided an appropriate forum for introducing the new technology and soliciting operator feedback while the project was still in the design stage.

Performance

CCRL normally operates the FCCU in distillate mode most of the year. To assess the benefits of the revamp, gasoline and distillate (LCO + HCO) are the main product yields of interest for this mode of operation.

To document the benefits of the revamp, CCRL measured unit performance by conducting a series of well-controlled test runs, before and after the revamp, over a range of reactor severities. Three tests were conducted before the revamp in May 1995. These tests spanned reactor temperatures of 965 to 980° F., which correspond to CCRL's typical operating conditions for distillate and gasoline severities.

Table 1 [37762 bytes] summarizes the operating conditions and product yields during and before revamp testing.

Following a period of initial operation and stabilization after the revamp, tests over a similar range of reactor temperatures were conducted during August 1995. Table 2 [41366 bytes] shows the operating conditions and product yields for these tests.

A comparison of Tables 1 and 2 shows that the revamp significantly increased yields of gasoline and gasoline-plus-distillate. Additionally, dry gas production decreased. One immediate benefit of the revamp was that CCRL was able to take off line one of the four reciprocating wet gas compressors.

Fig. 3 [23858 bytes] shows the shift in yields of dry gas and gasoline as a function of conversion, before and after the revamp. Fig. 4 [25625 bytes] similarly shows the shift in gasoline-plus-distillate yield before and after the revamp, as a function of conversion. The gasoline yield and conversion are not shown at the traditional ASTM 90% distillation point of 380° F. Instead, a reference boiling point of 360° F. was chosen to minimize data manipulation because this temperature is close to the "as-produced" gasoline cut.

It should be noted that, although the reported catalyst microactivity tests (MATs) are much lower in the after-revamp tests compared with the before-revamp testing, the surface area and metals-on-catalyst are similar.

Ordinarily, during the data normalization process, postrevamp operation would be credited with a significant yield advantage because of the lower catalyst MAT. The postrevamp operation, in this case, was not credited with a yield advantage during data normalization because UOP did not think that would produce a fair comparison.

UOP believes the actual postrevamp catalyst activities are higher than the values measured in equilibrium catalyst testing. Postrevamp equilibrium-catalyst MAT that is equal to or lower than that of before-revamp tests limits the conversion levels achieved in a revamped unit equipped with a more-contained riser termination system.

The product yields shown in Tables 1 and 2 are normalized to 100 wt % recovery. The split between gasoline and LCO is corrected to a 360° F. ASTM 90% point.

The HCO and main column bottoms (MCB) yields shown in Tables 1 and 2 are on an as-produced basis, with only small corrections for weight recovery. The two most important product yield deltas obtained from the revamp at constant reactor temperature are dry gas reduction and gasoline-plus-distillate increase, shown in Table 3 [20325 bytes].

Fig. 4 [25625 bytes] clearly shows that the revamp improved selectivities to gasoline and gasoline plus distillate at constant conversion. At constant reactor temperature, conversion decreased, gasoline yields decreased moderately, and essentially all of the desired increase in product volume was accounted for by distillate products.

This result is consistent with the decrease in post riser residence time achieved by the revamp. The increase in distillate yield is desirable, given CCRL's processing objectives and market demands. For an FCCU operating in a maximum-gasoline mode, a revamp to this technology would enable optimization of the reactor to a higher severity.

Economics, payout

The costs for the project were allocated into several categories: design, material, labor, contract, and license fees.

For the feed distributor part of the revamp, the total 1995 material and construction cost was US$675,000. For the VSS riser disengager part of the revamp, the 1995 material and construction cost was $1,050,000. The total project cost, including all fees, engineering design, material, labor, crane rental, contract inspection, and testing of the reactor head and reweld, was $2,600,000.

These modifications to the feed distributor system and the riser disengager were predicted to produce more high-value liquid products in the gasoline-to-distillate range and less low-value products, such as dry gas, LPG, MCB, and coke. The estimated annual benefit in product value was US$920,000.

The average annual operating cost was estimated to increase slightly because of increased steam consumption, but the decrease in dry gas production has enabled CCRL to idle one reciprocating gas compressor. This compressor used to operate about 25% of the year; therefore, taking into account the resultant savings of $90,000/year, unit operating costs have been reduced by $40,000/year.

The payout, in simple terms, was predicted to be about 3 years, based on the predicted yields provided in the original proposal. Initial performance testing has shown that actual gasoline-plus-distillate yields are significantly higher than predicted, resulting in increased product value. The payout is now estimated to be less than 2 years.

This payout is based on the product values in the local Canadian market served by the CCRL refinery. If the delta-yield economics were valued at typical U.S. Gulf Coast prices, operation at higher reactor severity to increase wet gas compressor loading to the previous gas compressor limit would produce the most profitable yield pattern.

Cooperative efforts

The revamp of the CCRL FCCU was successful because of the close working relationship between CCRL and UOP during the engineering and construction phases of the project. Revamp planning and execution were accomplished in less than 6 months.

The yield benefits obtained from the new Optimix feed distributor and VSS riser disengager exceeded projections and have increased refinery profitability.

Based on a paper presented at the National Petroleum Refiners Association annual meeting, Mar. 17-19, in San Antonio.

Copyright 1996 Oil & Gas Journal. All Rights Reserved.