HYDROTREATING OPTIMIZATION-CONCLUSION OPTIMIZATION OF CATALYST SYSTEM REAPS ECONOMIC BENEFITS

Charles F. LeRoy, Michael J. Hanshaw, Steven M. Fischer, Tariq Malik, Randall R. Kooiman
Champlin Refining & Chemicals Inc.
Corpus Christi, Tex.

Champlin Refining & Chemicals Inc. is learning to optimize its catalyst systems for hydrotreating Venezuelan gas oils through a program of research, pilot plant testing, and commercial unit operation.

The economic results of this project have been evaluated, and the benefits are most evident in improvements in product yields and qualities.

CATALYST SYSTEM

The project has involved six commercial test runs, to date (Runs 10-15), with a seventh run planned. A summary of the different types of catalyst systems used in the test runs, and the catalyst philosophy that developed, is found in Table 5, Part 1 (see OGJ, May 27, p. 49).

Runs 10 and 11 used standard CoMo and NiMo catalysts for heavy gas oils hydrotreating. These catalysts had small pore sizes and suffered high deactivation rates because of metals contamination. When it was discovered that metals contamination was a problem, catalyst options were reviewed.

The catalyst for Run 12 was selected in early 1987, before a full understanding of the metals problem was acquired. At this point, a small amount (1.2%) of demetallization catalyst was added to the load for Run 12.

Directionally, this was correct, but the amount of demetallization catalyst was not sufficient to extend the run to the desired length.

During Run 12, metals analysis was improved, and the sources of the metals were determined. In light of this information, the catalyst system for Run 13 was loaded with 7.8 and 6.9% demetallization catalyst in Reactor Trains A and B, respectively.

These percentages were calculated based on a 12-month run at 60,000 b/d charge and 6 ppm (Ni+V). The metals loading capability of the demetallization catalyst was supplied by the catalyst vendor, and the total metals to be charged to the unit during the run was calculated. These factors were used to estimate the amount of demetallization catalyst needed.

Run 13 had a much longer run length and showed the benefits of the demetallization catalyst. This run was performed at low severity to favor run length.

Run 13 was conducted with catalysts from two different suppliers. One catalyst was loaded in Reactor Train A and the other in Reactor Train B. Champlin is able to run these two reactor trains independent of each other and evaluate each catalyst system separately.

Both suppliers' catalysts had a successful run. One showed better performance with lower stability, and the other showed higher stability, but with lesser performance.

Run 14 was conducted with the catalyst that showed better performance with lower stability, and Run 15 was conducted with the catalyst that showed higher stability, but lesser performance. Both Runs 14 and 15 were good runs because the demetallization catalyst protected the main bed from rapid metals deactivation.

Table 5, Part 1, shows the continued evolution of the catalyst system. As the feed to the unit was better understood, the catalyst system was modified accordingly. The substantial benefits derived from this catalyst optimization process are only possible when a consistent feedstock is run in the refinery.

This table also shows that, as new catalysts were introduced into the market, they were tried and improvements were seen.

Another troubleshooting approach used was spent catalyst analysis. Table 4, Part 1 (see OGJ, May 27, p. 49) shows the results of the spent catalyst analysis from Runs 10-14. There is a great deal of scatter in these data and a wide variability in the results, which were caused by inaccuracies in sampling technique and analytical test methods.

Still, the analysis has provided good insight into the catalyst system philosophy and temperature management approach used to operate the reactor system and optimize the overall system.

The spent catalyst analysis helped to show whether the amount of demetallization catalyst used was adequate. It also showed that the condition of the second reactor catalyst was very good in regard to both carbon and metals, and led to the modification of the temperature management approach used in Run 15.

TEMPERATURE MANAGEMENT

After Runs 13 and 14, it became apparent that significant progress had been made toward understanding the Unibon unit operation. The analysis of metals had been improved, the sources of those metals had been determined, and the selection of a catalyst system had been improved to allow the Unibon to reach the desired run length.

The focus of attention in Run 15 has been turned to increasing severity by changing the temperature management approach. In Runs 10-14, the WABTs (weighted average bed temperatures) in both reactor trains were controlled to be equal. This was done by setting the inlet temperature to the first reactor and then injecting quench hydrogen into the inlet of the tail reactor to drop its inlet temperature. Champlin believed that this was the best way to balance deactivation in the reactor system and optimize catalyst usage.

This approach has been changed in Run 15. The first reactor WABT is set to attain the desired desulfurization level. By reducing the quench, the second reactor is running at a WABT that is 20-25 F. higher than that of the first reactor. This delta WABT of 20-25 F. increases severity on the second reactor, because the second reactor is believed to have been under-utilized.

In Runs 13 and 14, the first reactor dictated the run length because of metals and coke deactivation, and the second reactor shut down with a great deal of catalyst activity remaining.

The objective in Run 15 is to increase severity in the second reactor so that both reactors are deactivated at the same time. This simultaneous deactivation will provide the highest activity for the longest run length.

Another area that is being studied is the reduction of aromatics to improve the quality of the FCCU (fluid catalytic cracking unit) feedstock.

It is desirable to operate in the optimum aromatics reduction temperature range. This can sacrifice some run length, but it may be economically beneficial in the FCCU.

Aromatics levels are being monitored and start-of-run temperatures are being raised to the range of maximum aromatics reduction. At end of run, the delta WABT will be reduced to maximize total aromatics saturation.

The analytical techniques for aromatics analysis are still being reviewed, and confidence in aromatics values is continuing to develop.

The refinery has recently purchased a gas chromatograph/mass spectrometer (GC/MS).

This analytical tool should allow better analysis of aromatic components. It will open up a whole new area of development.

This area of optimum aromatics reduction may provide a better method of assessing severity, from an economic viewpoint, than do desulfurization or denitrification.

Denitrification levels in the unit are also being watched. This variable has a big impact on the FCCU feedstock value. Denitrification tracks with desulfurization and WABT.

Because it is easier to analyze for desulfurization, it is used as a severity measure, but denitrification is still calculated and monitored.

FUTURE PLANS

During the past 5 years of work on the Unibon unit, a revamp project was developed to increase catalyst volume. The revamp will provide improved FCC yields, higher product values, additional Unibon capacity, and increased run length flexibility. Reactor volume will be effectively doubled with the addition of two reactors in series with the existing trains.

Cost of the revamp is estimated at $19 million. Project income varies between $4.5 million/year and $7.7 million/year, depending on the price basis and marketing assumptions used. Typical income is estimated at $6.0 million/year, with a discounted cash flow return on investment of 23.6%. Project start-up is planned for September 1991.

The original Unibon unit was designed and commissioned in 1976 to hydrotreat 48,700 b/d of 1.54 wt % sulfur feedstock to a 0.2 wt % sulfur product, while achieving a 1-year catalyst cycle length.

The Unibon rate and sulfur loading has increased substantially because of the construction of the coker in 1983, the implementation of deepcut operations in 1986, and the increased charge of high sulfur and nitrogen Venezuelan crude in 1986.

The average rate in 1989 was 53,600 b/sd, with a feed sulfur of 1.73 wt % and a product sulfur of 0.37 wt %. This is an increase of 10% in volume and 24% in sulfur load over design.

Catalyst cycle length is still estimated at 1 year. However, cycle length has been maintained at the expense of lower gas oil quality and FCCU yields.

Improvements in hydrotreating catalysts have also helped maintain the 1-year catalyst cycle length. Gas oil desulfurization has been reduced from the design value of 87% to about 79%.

Because of the higher load on the Unibon, the present reactor volume is about half the current design standard for new units. The current design standard for this unit is on the order of 1.0 liquid hourly space velocity (LHSV). This value compares to the original design value of 1.43 LHSV and current operating level of 1.76 LHSV. The revamp will lower the LHSV to about 0.9.

The project objectives are to raise the Unibon product quality and downstream FCCU yields by doubling the unit catalyst volume. The project will upgrade the FCC slurry product from the 3 wt % sulfur No. 6 fuel oil market to the 1 wt % sulfur market.

It will also provide additional gas oil hydrotreating capacity to adjust for future changes in crude slate and environmental regulations. It will allow the refinery to maintain a minimum catalyst cycle length of one year.

ECONOMIC IMPACT

The total benefits for project justification are estimated at $6.0 million/year. The downstream FCCU yields will be improved, with average FCC conversion increasing about 1.2%. Total liquid recovery will increase about 0.6% on charge. FCC catalyst consumption will decrease by about 1.5 tons/day. These benefits are valued at $4.5 million/year.

The FCC slurry product value will increase by $1.2 million/year by upgrading it to the 1 wt % sulfur No. 6 fuel oil market. The Unibon catalyst cycle length will be extended from 1 year to 18 months, which is a $300,000/year benefit.

The economics for this project were developed by conducting pilot plant tests and by using a computer simulation model for the Unibon and FCCU. The pilot plant tests were designed to show the potential improvements in Unibon yields and product qualities made possible by doubling Unibon catalyst volume.

Yields were obtained before and after an increase in reactor volume for two scenarios: first, at constant sulfur removal; then, at constant reactor temperature.

The first scenario corresponds to a strategy of maintaining constant gas oil quality while increasing catalyst cycle length. The second scenario corresponds to an increase in gas oil treating severity and an improvement in gas oil quality, at a shorter cycle length.

The pilot plant results were used to develop the reaction temperature coefficients and rate constants for the hydrotreater model. The computer simulation model was then calibrated using pilot plant data and actual run data from the commercial unit.

This model was then used to analyze cycle length issues, reactor volume impacts, and Unibon/FCC yield effects. All of these results are summarized in Table 1.

The results indicate that the optimum catalyst cycle length prior to the revamp is 9 months. Because a 9-month cycle length is less than the 12-month cycle typically run in the past, the 12-month cycle length case (Case 2, Table 1) was used as the basis for the economic evaluation.

Following the revamp, a 12-month cycle length will be the economic optimum. However, the hydrogen consumption required at this cycle length exceeds the current capacity of the Unibon makeup compressors. As a result, the post-revamp 18-month cycle length case (Case 5) was used as the subject case.

There are several other projects being studied in the Unibon area as a continuation of this effort. An increase in make-up compressor capacity is being reviewed as an option. This has an apparent benefit of $3 million/year in yield benefits from the FCCU, created by higher-severity operation of the Unibon.

The revamp of the existing fractionation towers would allow the production of an on-spec low sulfur (0.05%) No. 2 fuel oil product. This would recover this valuable product and avoid its downgrade to light cycle oil (LCO) in the FCCU. It would also free up FCCU capacity. This project has an apparent benefit of $3.5 million/year.

These two projects also allow for the consideration of mild hydrocracking catalyst usage in the reactor system.

ASSESSMENT

The program to improve the hydrotreating operation at the refinery has been very successful. It shows how continuous improvements can be made when all the refinery resources work together toward a common goal. The following list shows the major conclusions that were drawn from the work performed on the hydrotreater:

Key performance indicators show a 158% increase in run length from Run 10 (19.1 bbl/lb) to Run 15 (50.0 bbl/lb), with only a 2.5% reduction in severity, as indicated by desulfurization (from 86.6% in Run 10 to 84.1 % in Run 15).
The economic impact of this effort can best be seen in increased product values with higher severity operation. An increase in overall severity, as measured by desulfurization, from 80% to 90% is worth about $10.5 million/year, assuming a 1-year run length as a minimum.
The amount of demetallization catalyst used will continue to be studied and optimized with the vacuum unit yields.
The Champlin spent catalyst analyses were very difficult to interpret. Alone, they were not sufficient to allow proper conclusions to be drawn. This emphasized the need for multiple approaches to the same problem and the fallacy of relying too heavily on any one data point.