Premcor implements fast-track schedule for heavy oil upgrade project

May 27, 2002
A recent conversion of Premcor Refining Co.'s refinery at Port Arthur, Tex., allows it to process highly cracked stocks derived from Maya crude.

A recent conversion of Premcor Refining Co.'s refinery at Port Arthur, Tex., allows it to process highly cracked stocks derived from Maya crude. Premcor addressed inherent project challenges within a fast-track schedule and the project was completed in 31 months.

Since 1990, more than 10% of US refining capacity has been reconfigured to process large quantities of heavy crude oil, with delayed coking as the critical process for maximizing liquid product yield. In recent years, hydro cracking is associated with delayed coking technology in many projects further to enhance economic incentives for converting refineries to heavier, less-expensive crude oil.

Project challenges

Premcor named the crude conversion and expansion project the heavy oil upgrade project (HOUP). Conceptual design determined the optimal process configuration:

  • Delayed coking unit with gas plant and naphtha pretreater.
  • Hydrocracker, which produces feed to an existing FCC unit.
  • Sulfur-removal unit with a sour water stripper and amine-recovery unit.
  • Flare and associated offsites and utilities for the new units.

The change in crude slate and capacity increase also required revamping the existing refinery crude unit and three hydrotreaters and upgrading and expanding the refinery offsites and utility systems.

Click here to enlarge image

In 1998, Premcor selected the Foster Wheeler Corp., Clinton, NJ, selective yield delayed coking (SYDEC) technology for the coker unit and Chevron Lummus Global LLC's (CLG) optimized partial conversion (OPC) technology for the hydrocracker unit.

Premcor selected Foster Wheeler USA Corp. for engineering services. Foster Wheeler defined the project through front-end engineering and design (FEED), performed early procurement, and converted the work-on an open-book estimating basis-to lump-sum turnkey (LSTK) execution through project completion.

Premcor's objectives for HOUP included:

  • Streamlined project cost through value engineering.
  • A fast-track schedule of 31 months from conceptual design to mechanical completion.

The accompanying box shows the HOUP overall project scope, which included new units, offsites, and revamps.

OPC concept

Feedstock for the Premcor hydro cracker project is one of the more difficult feeds that can be processed through a hydrocracker unit.

Click here to enlarge image

Premcor wanted to process a feedstock derived from Maya crude with a majority of the blend from previously cracked stocks. Table 1 presents unit feed specifications.

The table shows that the design feedstock blend is 60% heavy coker gas oil, 20% light cycle oil (LCO), and 20% straight-run heavy vacuum gas oil. Cracked stocks derived from Maya crude are difficult to process in hydroprocessors due to high aromatics and nitrogen.

Click here to enlarge image

The hydrocracker design basis was a 35,000-b/sd unit that would achieve 50% actual conversion of 650°+ F. material in the feed, leaving 13,000 b/sd of 650°+ F. material for the FCC unit to process.

Premcor initially envisioned the plant to be a single-stage, once-through (SSOT) unit to achieve the desired conversion levels. The proposed feedstock, however, represented a severe processing challenge because of high aromatics and nitrogen. The unit required a liquid hourly space velocity (LHSV) of 0.3 hr-1, resulting in a reactor volume equivalent to two large-diameter reactors operating in series.

A two-stage, recycle system offers investment economy and contributes other benefits such as lower hydrogen consumption and improved overall product qualities.

Click here to enlarge image

This project represented the first commercial application of the OPC design concept and was meant as an alternative to an SSOT partial-conversion configuration. Fig. 1 presents a simplified process flow diagram of the unit.

The OPC configuration consists of a two-stage unit with the second stage operating in a "clean" environment. Both reaction stages are in the same recycle gas loop, which offers substantial savings in capital investment because the unit uses a single recycle gas compressor and product separator system.

The fractionation system is between the two stages. Only a portion, therefore, of the 650°+ F. feed material is processed in the second stage. The remainder is sent to the FCC and provides good-quality feed.

The "clean" environment, in which the second stage operates, is defined as relatively free of nitrogen and ammonia, which are strong inhibitors to cracking activity. Cracking reaction rates substantially increase in the clean second-stage environment, thus allowing lower reactor temperatures and higher LHSVs than would be required at first-stage conditions to achieve the same conversion targets.

Click here to enlarge image

Fig. 2 highlights the benefits of a clean second-stage environment in terms of catalyst activity and selectivity.

Implementing the OPC configuration reduced the amount of hydrocracking catalyst and improved overall product qualities.

Premcor installed one large 15-ft diameter, 92-ft length reactor for the first stage and a 9-ft diameter, 58-ft long second-stage reactor.

LCO upgrading in hydroprocessing units requires relatively high pressures and low space velocities and is associated with high amounts of heat release, high hydrogen consumption, and accelerated catalyst deactivation.

In addition, due to the LCO stream's boiling range, most of it is recovered as kerosine in the fractionator without being processed in the clean second-stage reactor. There, additional product quality enhancement (aromatic saturation) would occur.

The unit could recycle kerosine recovered off the fractionator to the second-stage reactor (Fig. 1). Pilot-plant laboratory testing of the Premcor feeds in the kerosine recycle mode demonstrated that a large portion of the kerosine could be cracked to naphtha products.

This is a benefit from an operating flexibility standpoint because it allows increased naphtha yield during the gasoline season.

Foster Wheeler, therefore, added this feature, which required only an additional jumpover line and control valve back to the second-stage feed drum, to the design.

Another important aspect was reactor internals, which play a vital role in ensuring proper mixing and avoiding bypassing, leading to better operability and longer catalyst life.

This was especially critical in the Premcor hydrocracker, which processes high heat-release cracked stocks that can cause serious temperature maldistribution problems. CLG included a new generation of reactor internals technology in the design.

The key to this design is how the gas and liquid are circulated, leading to thorough mixing with minimal pressure drop.

Basic engineering design

Process simulations and process design began in May 1998 and were finished in August 1998.

At the project onset, the parties agreed that Foster Wheeler and Premcor representatives would be resident in CLG's offices during development of the process design package.

Having a process design engineer participate in the entire process design package effort allowed the project to move rapidly into the detailed design phase and was critical for the project's final success.

Major benefits of having a resident process design engineer included:

  • A single-point contact between Premcor, Foster Wheeler, and CLG. All communications were routed through one person, which avoided confusion and contradictory instructions to CLG.
  • An understanding of design intent. Working closely with CLG engineers, Foster Wheeler's engineer gained a thorough understanding of the design, enabling transfer upon package completion.
  • Quick response on urgent items.
  • Prioritizing process data issues. If, for overall schedule reasons, the project required a particular equipment data sheet out of the general sequence, the engineer could communicate this fact and assess effects on other equipment.
  • Data sheet preparation review of preliminary CLG data to ensure that all the data required from Foster Wheeler design groups were included.
  • A hydraulic review of the performance of preliminary hydraulic calculations based on simulations. This minimized inconsistency between equipment nozzle sizes and line sizes.
  • Design criteria. On-site team discussion on essential (i.e., would affect process guarantees) and optional process design items.

A key element of team success in achieving the project objectives was a conscious effort to achieve thorough scope definition through FEED, and to secure design buy-in from Premcor's representatives, project management, engineering and construction lead engineers.

Team participants awarded the project a top score (more than 900 out of 1,000 using the Construction Industry Institute's project definition rating index), which indicated the readiness to complete the design and build the LSTK project.

Detail design; schedule

Although seemingly contradictory with the high project definition goal set in the FEED phase, the extreme fast-track schedule requirements forced the team to think "outside-of-the-box" and use innovative approaches in the project's engineering, detail design, and construction phases. Table 2 outlines HOUP's major milestones.

Innovative execution approaches

Immediately upon the contract award, Foster Wheeler and CLG began process design of the delayed coker and hydrocracker, respectively. The 31-month project schedule goal, however, did not allow the companies to execute FEED before ordering critical equipment and before initiating engineering flow sheet work.

The firms, therefore, placed orders for long-lead equipment items within 6-8 weeks from the contract award and initiated flow sheet preparation concurrent with front-end work.

P&ID preparation

It was not unusual for Foster Wheeler to initiate P&ID work on the delayed coker and hydrocracker units.

This approach, however, was more daring and risky applied to CLG's first commercial application of its OPC design concept for a hydrocracker. Through team cooperation during that phase, CLG produced flow sheets on this schedule:

  • Contract award, April 1998.
  • P&ID issue for review by Premcor (both units), August 1998.
  • Hydrocracker PDP Issue by CLG, August 1998.
  • P&ID issue for design (both units), October 1998.

This approach allowed detailed unit design to commence in October 1998, fewer than 5 months after the contract award.

Critical equipment purchases

To support the overall project schedule, Foster Wheeler issued inquiries for the long-lead equipment items immediately after contract award. This meant that, for the OPC hydrocracker, Foster Wheeler and Premcor committed orders for high-pressure, stainless steel reactors at about 50% completion of CLG's basic process design.

This concerted effort was carried through the entire project, which allowed Foster Wheeler and CLG to place all long-lead equipment and material purchase orders early enough in the project schedule to support the construction effort.

Construction

As a schedule-saving concept, Foster Wheeler planned and implemented a construction sequence that involved preerection of coker structural steel, early staging of critical equipment for preerection dress out, and off-site fabrication of 38 modules.

Preinsulated coke drums were lowered into the steel framework, saving several weeks vs. traditional coker schedules.

Foster Wheeler introduced other innovations to streamline coker cost, scheduling, and operability. These innovations increased efficiency by using structural steel and concrete, and improved the coker switch deck and overhead valve operating deck layout designs.

Project management included supervision of a fully integrated team of 56 subcontractors, represented by 3,553 craft laborers.

The team successfully sequenced long-lead procurement components and erection of individual LSTK process plants, offsites and utilities, and reimbursable revamps.

The construction scope involved 25,000 cu yards of concrete, 7,500 tons of structural steel, and 98 miles of pipe.

Project start-up

Over a 5-month period before mechanical completion, Premcor's 45 person start-up and operations team trained as an extension of the integrated involvement of its plant operators during design.

Plant operators participated in a special 1-week training session to cover start-up activities and to discuss emergency scenarios that could occur in the hydrocracker.

The start-up required 2 months for sequencing through the delayed coking unit (89 systems), the hydrocracking unit (60 systems), and the sulfur recovery, amine, and sour-water stripper units (40 systems).

The Premcor hydrocracker was commissioned in late 2000 and early 2001.

Premcor introduced atmospheric gas oil feed on Dec. 29, 2000, in preparation for liquid-phase catalyst sulfiding. Premcor sulfided and titrated the second stage shortly thereafter. After catalyst sulfiding, the company introduced heavy vacuum gas oil feed on Jan. 2, 2001.

After a few weeks, it started feeding cracked stocks to the unit.

This allowed adequate time for the unit to line out and overcome start-up challenges with equipment and instrumentation.

On Jan. 27, 2001, Premcor introduced heavy coker gas oil feed to the unit in gradual increments, displacing gas oil feeds. Adding heavy coker gas oil was uneventful and did not cause any unit upsets.

The unit started processing LCO on Mar. 28, 2001. Again, this addition was uneventful, other than an expected increase in first-stage top bed reactor temperatures.

A licensor guarantee 3-day test run, completed in June 2001, showed that the unit exceeded C5+ liquid yield estimates at lower-than-expected H2 consumption. The only outstanding issue from the test run was kerosine product quality.

The data scatter included points equal to and less than the expected value due to the severity of LCO feed processing requirements.

HOUP was successfully completed in 31 months (eclipsing the previous 35-month industry benchmark for Maya crude conversion of Shell Oil Co.'s refinery, Deer Park, Tex.).

Through the first year of operation, the fouling rates of the first and second-stage catalysts indicate the unit will exceed the catalyst life predictions.

The catalyst exhibits better-than-expected hydrotreating activity, which allows for better FCC feed quality (lower sulfur and nitrogen).

Premcor used recycle kerosine to increase naphtha production in spring 2001.

Click here to enlarge image

Premcor will use it in the future when justified by refinery economics. Fig. 3 shows that reactor stability, as indicated by the radial temperature differences across an axial position, is excellent.

The plant operates at 110% of design feed and has converted over 75% of the 650°+ F. material in the feed (compared to a design value of 50%).

Despite the higher operating severity, catalyst stability continues to be excellent.

A 60-day reliability performance test of the entire HOUP facilities, run from July 26 to Sept. 23, 2001, demonstrated that the overall HOUP project achieved contractual guaranteed reliability requirements.

The authors

Click here to enlarge image

Jean Marc Papon is a project director at Foster Wheeler USA Corp., Clinton, NJ, responsible for all activities on the HOUP from process design, construction, and plant commissioning to start-up. He has worked for Foster Wheeler for more than 25 years. A project manager since 1989, he became project director for the Premcor Port Arthur HOUP project in 1998. Papon holds a BS in chemical engineering from the National Institute of Industrial Chemistry, Rouen, France, 1973.

Click here to enlarge image

Jay Parekh is lead process engineer at Chevron Lummus Global LLC, Richmond, Calif., and has been involved in licensing activities for 8 years. His specialty is hydrocracking in which he has participated in process design, plant commissioning, and technical service activities. Parekh was the lead start-up engineer for the Premcor Port Arthur hydrocracker unit in 2001. He holds a BS in chemical engineering from the University of California, Davis, 1991.