New residue process increases conversion, produces stable residue in Curacao refinery

Roger Marzin, Pedro Pereira
Pdvsa-Intevep SA
Los Teques, Venezuela

Michael J. McGrath, Howard M. Feintuch
Foster Wheeler USA Corp.
Clinton, N.J.

Gregory Thompson, Edward Houde
UOP LLC
Des Plaines, Ill.

Based on a presentation at the 1998 National Petrochemical & Refiners Association Annual Meeting, San Francisco, Mar. 15-17, 1998.

A new hydrovisbreaking technology uses a dual catalyst system to achieve higher conversion levels, lower asphaltene and Conradson carbon contents, and more stabilized residue than conventional visbreaking technologies.

Pilot plant tests by Intevep SA and a commercial application of the new Aquaconversion process at the Refineria Isla Carazao SA Emmastad, Curacao, refinery, confirm these advantages.

Demands for higher-quality transportation fuels, tighter capital restrictions, and higher refinery utilizations will be further magnified by declining crude quality, decreasing fuel oil demand, and more-severe product specifications. These factors have sparked renewed interest in converting the heavier fraction of the crude barrel.

Worldwide residue conversion capacity is about 12.5 million b/d. Although major technical advancements related to the catalytic conversion of residues have been made, conventional thermal processing technologies, such as delayed coking and visbreaking, represent more than 60% of the world's installed conversion capacity.

A recent internal UOP LLC survey indicates a continuing demand for these technologies, primarily because of their lower investment and operating costs, high reliability, and ability to economically produce both finished products and incremental feedstock for catalytic conversion processes.

Aquaconversion process

In 1996, UOP, Foster Wheeler USA Corp. (Fwusa), and Intevep entered into an alliance. The purpose of this new alliance is to promote the commercialization and the ongoing development of Intevep's novel Aquaconversion process. Although the Aquaconversion process has a wide range of refining applications, it is currently being promoted as either a replacement of, or a modification to, conventional visbreaking.

The Aquaconversion process reduces the viscosity of the heavier components of the refinery's fuel oil pool and reduces fuel oil production for a minimal incremental investment (Fig. 1 [62,836 bytes]). In addition, the process enables the refiner to profit from the additional yields of distillate and lighter products produced at conversion levels significantly higher than those attainable with conventional visbreaking technologies (Fig. 2 [54,232 bytes]).

In the late 1980s, Intevep started a program to investigate alternative catalysts and processing routes for upgrading and converting Pdvsa's heavy and extra-heavy crude oil reserves. Intevep modified the objectives of the original program in the early 1990s to focus on the application of these techniques to low-pressure, low-investment, heavy-oil upgrading approaches such as visbreaking. Since that time, Intevep's Aquaconversion development efforts have concentrated on the most-notable problem associated with conventional visbreakers: their tendency to form unstable products when operated at high conversion levels.

Thermal cracking mechanism

A basic understanding of the reaction mechanism involved in thermal cracking is fundamental to an understanding of Intevep's solution to thermal instability ( Fig. 3 [63,017 bytes]). Thermal cracking reactions progress at elevated temperatures through a free radical mechanism that enables various carbon and hydrogen components of the feedstock to dissociate and form hydrogen-based and hydrocarbon-based free radical intermediates. These thermally produced free radicals then enter into other reactions that eventually create a broad spectrum of products ranging from light gases to highly condensed coke.

Because hydrocarbon free radicals are highly susceptible to spontaneous cracking, the thermal cracking reaction mechanism also involves the cleavage of carbon-carbon bonds present within both the longer-chain paraffinic components of the feedstock and the larger alkyl side chains of the feedstock's aromatic components.

The cleavage of these carbon-carbon bonds creates smaller paraffins and olefins, which reduce the viscosity of the feedstock and decrease the refiner's fuel oil yield. As cleavage of the carbon-carbon bonds within the larger alkyl side chains of the aromatic structures proceeds, the alkyl groups that remain attached to the aromatic structure become shorter and less susceptible to further cleavage.

The hydrocarbon free radicals that are produced, however, begin to enter into competing reactions by combining with other free radicals and creating multiaromatic rings. These condensation reactions eventually produce highly condensed asphaltenic structures, as evidenced by the tendency of visbroken residues to have higher asphaltene and Conradson carbon contents than their original feedstocks.

The tendency of a visbreaker to produce asphaltenes results in an upper conversion limit beyond which product instability can occur. As thermal conversion increases, the molecular weight of the bulk phase of the visbroken product decreases.

At the same time, the higher thermal severity increases the rate of condensation reactions, resulting in the creation of additional asphaltenes. Eventually, a conversion level is reached at which the bulk phase of the visbroken product is unable to maintain the asphaltenes in solution, causing precipitation of the asphaltenes and the formation of an unstable visbroken residue.

Aquaconversion-reaction mechanism

Intevep's Aquaconversion development program focused on ways of improving the stability of visbroken residues so refiners benefit from the increased viscosity reduction associated with higher visbreaking conversion levels. Intevep theorized that the key to improved product stability, and therefore increased viscosity reduction, involved saturating the thermally produced aromatic free radicals before they could condense to form asphaltenes. Although saturation reactions are easily promoted by catalysts and elevated hydrogen pressures, neither of these conditions exists in a conventional visbreaking unit.

Intevep solved the aromatic condensation dilemma by developing a novel oil-soluble, dual-catalyst system, which, at conventional visbreaking processing conditions, converts water into hydrogen and then inserts the hydrogen at the critical point in the thermal cracking reaction sequence where the asphaltene-forming condensation reactions occur (Fig. 3).

This hydrogen-transfer mechanism inhibits aromatic condensation and produces a more-stable visbroken product, which has a higher hydrogen content and lower asphaltene and Conradson carbon contents than the product from a conventional visbreaking unit.

The Aquaconversion reaction mechanism proceeds by the unique interaction between the two non-noble metal catalysts. The first catalyst enhances the dissociation of water into hydrogen and oxygen free radicals. The highly reactive hydrogen free radicals that are formed accelerate the thermal cracking rates of the paraffinic components of the feedstock and stabilize the resulting thermal products by saturating olefinic free radicals.

The second catalyst minimizes the condensation reactions by promoting the addition of hydrogen to the aromatic free radical. The result is the formation of a smaller aromatic component as well as additional hydrogen free radicals and carbon dioxide.

This entire reaction sequence effectively terminates the undesirable aromatic-condensation reactions so the refiner can benefit from the viscosity reduction associated with higher visbreaking conversion and still produce a stable visbroken product.

Pilot plant test

The ability of the Aquaconversion process to terminate asphaltene formation by the addition of hydrogen is best illustrated by the results from one of Intevep's early pilot plant programs. This program, conducted in Intevep's 1 b/d pilot plant located at its research facilities in Los Teques, Venezuela, investigated the impact of operating temperature on Tia Juana Pesado vacuum residue when processed in both visbreaking and Aquaconversion process configurations. Physical properties of the vacuum residue feedstock used in the pilot plant program are summarized in Table 1 [ 64,100 bytes].

The configuration of Intevep's pilot plant is similar to that of a conventional visbreaker (Fig. 4 [55,697 bytes]) and consists of a five-zone electrical resistance heater followed by a reaction chamber and a hot separator.

The feedstock was heated to reaction temperature prior to its entering the pilot plant's reaction chamber. Effluent from the reaction chamber flowed to the separator vessel for the separation of liquid and vapors. Separator overheads, which consisted of primarily gas, steam, and light distillates, were then cooled.

The condensed steam and light condensates were recovered and weighed, and the gas stream was measured and analyzed. Finally, all of the liquid hydrocarbon products recovered from the pilot plant were combined and refractionated to determine the yield distributions and product qualities.

Initially, Intevep's pilot plant was operated as a conventional visbreaker to establish the baseline visbreaking performance (Table 1). The pilot plant was then operated in an Aquaconversion process configuration by mixing the incoming feedstock with water and the two oil-soluble catalysts.

Baseline visbreaking

As shown in Table 1, the conventional visbreaker configuration produced an overall conversion level, as measured by the production of material boiling below 330° F., of slightly less than 3 wt %. Similarly, the conversion of the feedstock to product that boiled below 932° F. was approximately 28 wt %.

At these conversion levels, the 662° F.+ visbroken product had a gravity of 3.7° API, a viscosity of 3,750 cSt at 210° F., and Conradson carbon and asphaltene contents of 23.6 and 25.7 wt %, respectively.

The ability of the visbreaking operation to produce a lower-viscosity product would result in a 9% reduction in fuel-oil yield relative to blending cutterstock directly into the vacuum residue feedstock. Unfortunately, any further reduction in the viscosity of the feedstock would be restricted by the stability of the 662° F.+ visbroken product, which, as measured by a Shell P-value of 1.15, indicated that the visbreaking operation was close to Intevep's definition of product instability (Shell P-value of less than 1.15).

Aquaconversion vs. visbreaking

After completion of the visbreaking test, Intevep simulated the Aquaconversion process by introducing catalyst and steam into the pilot plant. Initially, the operating temperature of the pilot plant was maintained at the same temperature as the prior visbreaking test to allow a direct comparison of the performance of the visbreaking and Aquaconversion processes. Subsequent Aquaconversion tests were conducted at progressively higher operating temperatures to determine the impact of processing severity on product quality and stability.

Table 1 illustrates that when both the visbreaking and Aquaconversion processes were operated at the same temperature, no significant difference in conversion to 932° F.- product occurred. However, a comparison of the quality of the 662° F.+ products from the two processes indicated that the Aquaconversion process produced a higher-quality product, as evidenced by its higher API gravity (4.9° API) and its lower viscosity (1,970 cSt at 210° F.). The reduced viscosity of the product from the Aquaconversion process results in a decrease in fuel oil yield of an additional 4% relative to the visbreaking operation.

The most-dramatic result of the first Aquaconversion test, however, was the effect on the product's asphaltene and Conradson carbon contents. When compared to the product from the visbreaking operation, the 662° F.+ product from the first Aquaconversion test contained 35% fewer asphaltenes and 12% less Conradson carbon.

In addition, the higher P-value of the product from the Aquaconversion process verified that a significantly more-stable product was produced.

Aquaconversion at higher severity

Intevep increased the severity of the test by raising the operating temperature of the pilot plant by 9° F. Information regarding this second higher-severity Aquaconversion test are also summarized in Table 1.

The results of the second Aquaconversion test indicate that the conversion to 932° F.- product increased to slightly more than 36 wt %, or approximately 30% higher than in the previous visbreaking test. Similarly, conversion to 330° F.- product increased to 7.5 wt %, or roughly 250% higher than from the visbreaking test.

Remarkably, the various product P-values indicate that the significantly higher conversions in the second Aquaconversion test were achieved while also producing a 662° F.+ product that was actually more stable than the product from the conventional visbreaker.

The higher conversion level of the second Aquaconversion process test produced a 662° F.+ product with a gravity of 5.4° API, a viscosity of 1,720 cSt at 210° F., and Conradson and asphaltene contents of 15.1 and 22.0 wt %, respectively.

After adjusting for differences in 662° F.+ product yields, the test indicated the asphaltene and Conradson carbon contents of the product from the second Aquaconversion test were approximately 40 and 20% lower, respectively, than from the conventional visbreaking operation.

In addition, the reduced viscosity of the product from the second Aquaconversion test would result in a 20% decrease in fuel oil yield relative to the visbreaking operation.

A comparison of the product qualities, product stabilities, and conversion levels from these pilot plant tests verified that the Aquaconversion process:

• Promotes the addition of hydrogen to the thermally produced products
• Produces a more-stable residue than the conventional visbreaking process
• Enables a higher degree of conversion to be achieved at similar product stability.

Ultimately, these benefits allow the refiner to profit from the lower fuel-oil cutterstock requirements of the lower-viscosity Aquaconversion product and the increased naphtha and distillate yields produced at the higher-conversion levels attainable in the Aquaconversion process (Fig. 2).

Commercial demonstration

Pdvsa conducted a commercial demonstration of the Aquaconversion process in the existing 36,000 b/d parallel-train soaker visbreaker at Isla's Curacao refinery. The objectives of this test were to:

• Commercially demonstrate the Aquaconversion process
• Confirm its ability to achieve higher conversions at acceptable product stability
• Generate sufficient data to determine the economic benefits of the technology.

The initial Aquaconversion commercial demonstration tests were conducted during the latter part of December 1996. Additional testing of the Aquaconversion process on other Venezuelan-based feedstocks was conducted in the Isla unit in mid-1997.

Modifications to Isla's visbreaker

The Isla visbreaking unit consists of two parallel visbreaking trains, each containing a heater, soaker chamber, and stripper ( Fig. 5 [54,801 bytes]). The overhead streams from each stripper are fed to a common fractionator. The two stripper bottoms streams are combined to form a fuel oil blending component.

The parallel train configuration of the Isla unit enabled the Aquaconversion demonstration tests to be structured so that one of the unit's two trains was operated in a visbreaking mode while the other train was operated in an Aquaconversion mode. This approach provided a direct comparison between visbreaking and the Aquaconversion process.

Minimal modifications were required to enable the commercial Aquaconversion test to be performed in Isla's existing visbreaking unit. These modifications primarily involved the installation of a skid-mounted catalyst-injection system.

The catalysts used in the demonstration test were preblended and used once-through to eliminate the need for catalyst-recovery facilities and to minimize disruptions to existing refinery operations.

Certain characteristics of the Isla visbreaking unit could not be modified and affected the ease of interpretation of the test data.

The most-significant factor that affected the results of the test involved the amount of coke that had been deposited within the visbreaker heater coils prior to the start of the Aquaconversion demonstration test.

Because the demonstration test was performed immediately before a scheduled visbreaker turnaround, existing coke deposits in the heater coils significantly limited the amount of steam that could be injected into the Aquaconversion train. As is discussed later, the reduced steam partial pressure caused by this limitation affected the quality of the products recovered at the highest-severity Aquaconversion test.

Results

The testing sequence used in the Isla Aquaconversion demonstration in December 1996 was similar to Intevep's pilot plant testing sequence described earlier. Initially, both trains of the commercial unit were operated as conventional visbreakers to establish the baseline visbreaking performance ( Table 2 [54,964 bytes]).

Testing then introduced catalyst and steam to the Aquaconversion process train. The second train continued to operate as a visbreaker at the baseline operating conditions. Based on the P-value of the 662° F.+ product from the Aquaconversion train, subsequent Aquaconversion tests were conducted at progressively higher operating temperatures to determine the impact of processing severity on product quality and stability.

The visbreaking baseline performance summarized in Table 2 indicates that when both of the trains were operated in the visbreaking mode, the overall conversion level, as measured by the production of 330° F. 2 products, was 5.6 wt %. Conversion to 662° F.- product was approximately 23 wt %, which resulted in the production of a 662° F.+ visbroken product having a gravity of 2.7° API, a P-value of 1.20, and a viscosity of 5,600 cSt at 210° F.

Once the visbreaking baseline performance of the commercial unit was established, catalysts and steam were introduced to the Aquaconversion process train. Because the heater firing remained constant, the introduction of steam to the Aquaconversion train lowered the operating temperature by 3° F. and reduced both the 662° F.- and 330° F.- conversion levels.

Interestingly, the observed increase in P-value of the 662° F.+ product was much greater than what would be expected based on the reduced operating temperature. Based on this higher P-value, the severity of the Aquaconversion process train was increased by raising its operating temperature by 8° F.

The results of the second Aquaconversion test conditions are also summarized in Table 2. The conversion to 662° F.- product for this second test increased to approximately 28 wt %, or slightly more than 20% higher than the 662° F.- conversion achieved in the baseline visbreaking operation.

Similarly, in the second Aquaconversion test, the conversion to 330° F.- product increased to 6.9 wt %. At these higher conversion levels, however, the 662° F.+ product from the Aquaconversion train had a higher stability (1.35 P-value), higher quality, and lower viscosity than the product from the visbreaking train. These results were consistent with the previous pilot plant tests and verified the commercial application of the Aquaconversion process.

To determine the maximum viscosity reduction and conversion that could be achieved in the Aquaconversion train, its operating temperature was increased by an additional 8° F. As Table 2 shows, conversion to 662° F.- products increased to approximately 31 wt %. Similarly, conversion to 330° F.- product increased to 7.8 wt %.

A comparison of the properties of the 662° F.+ products from the second and third Aquaconversion tests indicated that a reduction in the product's API and an increase in the product's viscosity occurred at the higher Aquaconversion severity. This decline in product quality was actually the result of the coke deposits that were present in the heater tubes prior to the start of the Aquaconversion demonstration test.

At the higher conversion levels of the third Aquaconversion test, these deposits significantly limited the amount of steam that could be injected into the heater.

A review of Intevep's Aquaconversion pilot plant database verified that the reduction in product quality observed during the third Aquaconversion test was indeed the direct result of the reduced steam partial pressure in the commercial unit. Interestingly, even at the higher conversion level of the third Aquaconversion test, the stability of the 662° F.+ product from the Aquaconversion train was still well above that of the product from the visbreaking train.

This improved stability indicated that it was possible to achieve even higher conversion levels in the Aquaconversion train. However, because of limitations in the heater's allowable pressure drop and the condensing capabilities of the fractionation system, no further attempts were made to increase the conversion level.

Economics

Based on the positive results of the commercial Aquaconversion demonstration test, the Isla refinery asked UOP to conduct a detailed engineering review of the existing visbreaking unit to determine the extent of the modifications required to revamp it to an Aquaconversion unit.

Although the engineering review is currently under way, initial results indicate that the existing heaters should be adequate for the revamp, although slight modifications would be required to allow for the additional steam requirements of the Aquaconversion process.

In addition, portions of the unit's stripping section might require retraying because of the increased conversion levels. Finally, catalyst addition and recovery systems would also be required.

Intevep is working with Isla's refinery planning group to evaluate the impact of various revamp options on the overall refinery product yields and economics. Preliminary economics indicate simple paybacks for the revamp could range from 12 to 18 months, depending on the assumptions used.

Future work

The three Aquaconversion process partners have identified several near-term development programs to further improve the economics of the Aquaconversion process and extend its current range of feedstocks and refining applications. These programs include studies related to alternative process configurations, new catalyst formulations, and the downstream processing of the products from the Aquaconversion process. These issues will be addressed in pilot plant testing programs to be conducted by Intevep, Fwusa, and UOP.

In parallel with the ongoing Aquaconversion process optimization work, Fwusa will focus on applying the technology directly at the wellhead for heavy-crude upgrading. In the recent past, these types of wellhead projects have considered using either delayed cokers or diluent addition followed by pipelining. Unfortunately, these traditional approaches have been hindered by coke-disposal issues or the large diluent and capital requirements associated with the construction of diluent and crude pipelines.

The ability of the Aquaconversion process to reduce the viscosity of its feedstock and also improve its gravity may allow low API, highly viscous crudes to be economically upgraded to transportable syncrudes directly at the wellhead, thus reducing both capital and transportation costs.

The conceptual design of a 100,000 b/d wellhead application of the Aquaconversion process for the upgrading of a heavy Venezuelan crude has recently been completed by Fwusa. The preliminary evaluation of the economics of this approach indicates a strong incentive to use Aquaconversion technology for wellhead applications.

Acknowledgment

The authors would like to acknowledge the valuable contributions made by both the management and operating personnel of Isla's Curacao refinery during the commercial demonstrations of the Aquaconversion process.

The Authors