September 2018

Special Focus: Refining Technology

Technical considerations for the Heydar Aliyev refinery revamp

The following case study describes the early development phases of SOCAR’s ongoing major revamp project at the Heydar Aliyev Refinery (HAR) in Baku, Azerbaijan, and its interface with the associated Azerikimya (AZK) steam cracking complex near Sumgait. References are made to the related SOCAR polymer project in Sumgayit, as well.

Alkhasli, E., SOCAR Turkey Enerji; Baerends, M., Baars, F., Fluor Energy and Chemicals; Khidirov, B., SOCAR; Asgar-zadeh, S., Azerbaijan National Academy of Sciences

The following case study describes the early development phases of SOCAR’s ongoing major revamp project at the Heydar Aliyev Refinery (HAR) in Baku, Azerbaijan, and its interface with the associated Azerikimya (AZK) steam cracking complex near Sumgait. References are made to the related SOCAR polymer project in Sumgayit, as well.

Present project situation

To fully understand the project background, a brief overview of SOCAR’s downstream assets in Azerbaijan is needed. SOCAR operates two refineries, both located in the capital city of Baku, that are connected by pipelines. The sweet crude oil feedstock to both refineries is domestically sourced.

FIG. 1. HAR block scheme.
FIG. 1. HAR block scheme.

The older Azerneftyag refinery is the smaller of the two refineries and has no conversion facilities. Conversely, it does include a significant lube plant, along with bitumen blowing for asphalt production. This facility is located in a fairly central location in Baku on the Caspian Sea. As this location is slated for urban development, the refinery will need to be closed in the near future.

The HAR is located in the northeast of Baku. It is a complex refinery that processes approximately 6 MMtpy, or about 130,000 bpd, of crude oil. The facility includes delayed coking (DC) and fluid catalytic cracking (FCC) units for high conversion of crude to transportation fuels. Additional facilities include a distillate hydrotreater (originally operated as an FCC feed pretreater), kerosine caustic treatment, a naphtha hydrotreater and a continuous catalytic reformer (NHT/CCR). The HAR process scheme is shown in FIG. 1.

The HAR produces LPG, gasoline, kerosine and diesel, primarily for the domestic market, as well as low-sulfur coke. Together with the Azerneftyag refinery, a substantial amount of light cycle oil (LCO), vacuum gasoil (VGO) and fuel oil is exported. The overall product slate (including Azerneftyag) includes lubes and asphalt.

FIG. 2. The locations of SOCAR’s Azerneftyag, HAR, AZK and OGPC.
FIG. 2. The locations of SOCAR’s Azerneftyag, HAR, AZK and OGPC.

The HAR also provides feedstocks to the AZK steam cracking complex. AZK feedstocks include light naphtha, FCC C3, FCC C4 and FCC dry gas (C2 gas). The AZK facility, which was started up in 1988, is located at the other end of the Absheron Peninsula, approximately 40 km north of the HAR refinery. FIG. 2 shows the locations of Azerneftyag, HAR, AZK and SOCAR’s oil-gas processing and petrochemicals complex (OGPC).

The AZK facility includes an EP-300 naphtha cracker with a standard design for 300,000 tpy of ethylene production capacity based on naphtha cracking. In addition to naphtha, the AZK facility receives other feedstocks (LPG-range materials and C2dry gas) from the HAR, transferred mainly by pipeline. Ethylene is converted into low-density polyethylene (LDPE) for local consumption, whereas propylene, C4, pyrolysis naphtha and pyrolysis fuel oil are largely exported.

The OGPC project

In looking at the longer-term development of SOCAR’s downstream assets, several factors drive toward a solution, dubbed the OGPC project:

  • The imminent closure of the Azerneftyag refinery for city development reasons
  • The long-term closure of the HAR due to its age and to urban development reasons
  • The long-term closure of the AZK steam cracker facility due to its age and its relatively small capacity
  • The availability of substantial amounts of natural gas from which NGL can be extracted as low-cost feedstock to a petrochemical facility centered on steam cracking
  • The economic and social advantages to the entire country of a major industrial development, which includes a move up the value chain from crude export to the export of petrochemical products
  • The export of motor fuels.
FIG. 3. Overall scheme of the OGPC project.
FIG. 3. Overall scheme of the OGPC project.

The OGPC project will be located near Garadagh (FIG. 2). This location is where large quantities of natural gas are imported, and it is well away from the Baku metropolitan area development. The overall scheme was studied and optimized in detail, followed by an extensive licensor evaluation/selection exercise for each of the units making up the entire complex. The envisioned complex would include a large (more than 10 B Nm3/yr) gas processing plant, which would feed NGL (C2+) to the petrochemical complex. The petrochemical complex is centered on a steam cracker and includes low-linear-density polyethylene (LLDPE), high-density polyethylene (HDPE), polypropylene (PP) and comonomer production facilities.

The 8-MMtpy–10-MMtpy refinery (FIG. 3) would be highly flexible and include FCC and hydrocracking to swing between diesel and gasoline production, as required. It will produce fuels that comply with the most stringent of worldwide quality standards, as well as lube oil base stock from the hydrocracker. The complex would make use of the various synergies available for integration between the refinery and the steam cracker (TABLE 1). However, the large demands of the OGCP project on capital and human resources necessitated that the project be delayed for some time, especially since SOCAR is undertaking several projects, including the SOCAR Turkey Aegean Refinery (STAR) project in Turkey, along with other major investments in infrastructure.

HAR revamp project

The shift of the OGPC project to a later date, in combination with the development of new polyethylene (PE) and PP plants at SOCAR Polymers’ Sumgayit plant (which required increased production of high-quality, polymer-grade ethylene and propylene by AZK) required various investment measures at both the HAR and AZK. These investments aim to secure the following objectives:

  • Accommodate the closure of the Azerneftyag refinery by expanding crude processing capacity and adding bitumen blowing capacity to maintain asphalt production at the HAR
  • Improve HAR economics by minimizing or eliminating low-value exports (VGO, LCO, fuel oil export), while maximizing the production of high-value transportation fuels (gasoline, jet, diesel) that comply with Euro 5 quality standards, and increasing production of RON 95 gasoline from RON 92
  • Supply AZK with the required feedstocks to supply SOCAR Polymer with sufficient ethylene and propylene; feedstocks from the HAR to the AZK complex are largely C3, C4 and dry gas so as to maximize conversion of naphtha into gasoline
  • Improve the HAR environmental footprint (e.g., by adding sulfur recovery)
  • Enable the HAR to continue safe operations by means of targeted equipment replacements
  • Minimize downtime caused by project execution
  • Minimize capital investment.
FIG. 4. The future configuration of the HAR.
FIG. 4. The future configuration of the HAR.

Studies were commissioned to determine the best way to achieve these objectives. The following are the main outcomes of these studies (they are also shown in FIG. 4, which displays the post-revamp refinery scheme):

  • Revamp the existing crude distillation unit/vacuum distillation unit (CDU/VDU) from 6 MMtpy to 7.5 MMtpy, while improving VGO cutpoint (to 560°C) and energy efficiency.
  • Add a new 400,000-tpy bitumen blowing unit,
    which will substantially reduce DC throughput.
  • Revamp the existing FCCU from 2 MMtpy to 2.5 MMtpy to eliminate VGO export, while maximizing naphtha and C3/C4 production—both for gasoline production and for AZK feedstock.
  • Add a new diesel hydrotreating unit to enable the production of Euro 5 diesel from a mixed feedstock of SR diesel, FCC LCO and coker light-cycle gasoil (LCGO). This measure also substitutes the export of low-value LCO for high-value Euro 5 diesel.
  • Convert the existing diesel hydrotreater (DHT)—formerly the FCC feed pretreater—to an NHT to process straight-run and coker naphtha. The existing DHT is unsuitable for Euro 5 diesel production (insufficient pressure), but is more than capable of ensuring a hydrotreated naphtha product with very low sulfur and nitrogen content, even with coker naphtha in the feed to the unit.
  • Add a new sharp-cut naphtha splitter to split NHT product into feedstock for the isomerization unit (low C7+ content) and CCR feed (low content of benzene and benzene precursors).
  • Add a new isom unit (including a deisohexanizer to achieve 89 isomerate RON) as required to achieve Euro 5 gasoline specs (benzene/octane/aromatics) and substantial RON 95 production.
  • Make limited modifications to CCR net gas recontacting to maximize LPG recovery and net gas H2 purity, while co-processing isom offgas—and replace the CCR reactor heaters due to age.
  • Add a new FCC naphtha desulfurization unit to ensure compliance to Euro 5 gasoline specs (sulfur); while the reuse of the existing NHT was investigated, it was found that reuse proved to be of limited value.
  • Add a new amine and mercaptan removal treatment of unsaturated C3–C4.
  • Add a new MTBE unit to achieve Euro 5 gasoline specs (octane/aromatics); the unit includes a splitter column for the MTBE raffinate to produce high-octane C4 raffinate for use in automotive LPG.
  • Recommission the delayed coker unit (DCU) gas plant to unload the FCC unsaturated gas plant, which processes coker light ends.
  • Add a new sulfur block consisting of an amine regeneration unit, sour water stripper (SWS) unit and sulfur plant to drastically reduce sulfur dioxide (SO2) emissions. At present, the refinery has no operational amine treatment and most hydrogen sulfide (H2S) ends up in fuel gas.
  • Add a new H2 plant and a new pressure swing adsorption (PSA) unit for CCR net gas to satisfy high-purity H2 demand, especially for the new DHT.
  • Provide amine treatment of the combined FCC and coker dry gas for export to Azerikimya.

The licensor/technology selection for the new scheme made extensive use of results from the OGPC licensor selection exercise. This reduction in effort made sense, given the high similarity of feedstocks (OGPC had a similar crude feed makeup to HAR) and unit capacities.

HAR revamp technical considerations

The CDU/VDU revamp will add a substantial number of heat exchangers to enhance thermal efficiency, while retaining the existing desalters and distillation columns (with new internals, however). The fired heaters will be replaced. The CDU heater will be replaced due to economic reasons, rather than for capacity reasons. The increase in VGO cutpoint will be achieved not only with a higher VDU heater outlet temperature, but also with steam injection (at present, the VDU is operated dry) and lower operating pressure, necessitating replacement of the ejector system.

FIG. 5. Liquid-driven ejector (left) vs. steam-driven ejector (right).
FIG. 5. Liquid-driven ejector (left) vs. steam-driven ejector (right).

The existing VDU is equipped with a vacuum system using liquid-driven ejectors, rather than steam-driven ejectors. The liquid-driven system uses more power (large, high-head circulation pumps) but no steam, and with much less cooling water (no need to condense motive steam) than a standard steam-driven ejector system. The system can be located at grade, whereas the steam-ejector system condensers must be located at an elevated platform to allow proper drainage to the atmospheric hotwell drum. A schematic comparing both systems is shown in FIG. 5.

With a liquid-driven ejector system and wet VDU operation in the future, there is a concern of water/oil emulsion formation at the ejector outlet (given the intense mixing of the driving oil with the vapor and condensation of the steam exiting the vacuum column by the colder motive oil). Previous experience at SOCAR with wet VDU operation, in combination with a liquid-driven ejector system, has shown that this emulsion can be routed to the desalter where the emulsion is broken into water and oil.

FIG. 6. View of the single-stage SWS.
FIG. 6. View of the single-stage SWS.

Another aspect of this project is the provision of a two-stage SWS system, necessitated by the low ratio of sulfur to nitrogen in the sweet domestic crude. With a single-stage SWS, the sulfur recovery unit (SRU) feed gas would be too rich in ammonia (NH3) for proper operation of the SRU (FIG. 6).

With a two-stage SWS (FIG. 7), the first stripper strips out only the H2S (directed to the SRU combustion chamber), while the second stripper releases NH3 (directed to the incinerator downstream of the SRU). The first column operates at an elevated pressure, facilitating the separation between H2S and NH3. It also uses an external reflux from the second column bottoms product. The second column operates under more typical conditions and produces an NH3-rich overhead gas. Any remaining H2S in the NH3 gas is removed by water washing in a small packed column, which is very effective in the presence of high concentrations of NH3. Although the two-stage SWS is more expensive and uses more utilities (steam mainly) than a single-stage SWS, the savings in the SRU and tail gas treating unit (TGTU)—lower throughput as the NH3 bypasses these sections—directionally compensates for this additional expenditure.

FIG. 7. Schematic for the two-stage SWS.
FIG. 7. Schematic for the two-stage SWS.

HAR-AZK interactions

One of the interesting challenges is the handling of HAR FCC-originated dry gas containing residual O2, NOx and diolefins in the AZK steam cracker. Due to incidences that occurred in France and the US,1,2 a known risk exists when introducing gas containing these contaminants into the steam cracker cold section. In the presence of oxygen, some of the nitrogen oxide (NO) will be oxidized into nitrogen dioxide (NO2). NO and NO2 will form a blue liquid called dinitrogen trioxide (N2O3) in the coldest sections (operating below –125°C) of the steam cracker separation system. Under abnormal operating conditions, the N2O3 may react with heavier diolefins (butadiene and heavier) and form a potentially explosive gum. Although this gum is normally considered inert at the low temperatures, when the cracker is shut down, this gum is prone to exploding when it is not adequately removed prior to heating up the cold box.

FIG. 8. A view of the steam cracker cold box, cold-end separators and exchangers.
FIG. 8. A view of the steam cracker cold box, cold-end separators and exchangers.

However, there are several ways to manage this risk. The simplest safeguard is to monitor the temperatures in the cold box; a rise in the temperature of the cracked gas entering the first plate-fin exchanger (FIG. 8) would be indicative of diolefin vapor entering the coldest separators. The safeguard can be combined with washing the cold equipment (plate-fin exchangers, vessels) with methanol prior to warmup during every shutdown to ensure that any gum that may have accumulated during operation is washed out.3,4 Additional protection can be provided by catalytic removal of the relevant impurities combined in a common train with the removal of other product specification relevant impurities (e.g., arsine, phosphine, COS).

An ultimate protection resides in processing the dry gas in a dedicated cold box, operating at sufficiently high temperatures (e.g., above –120°C) to prevent the formation of N2O3. The lighter components, including NO and NO2, end up in a common tail gas stream directed to fuel. This additional layer of safety adds to capital cost. Conversely, it unloads the steam cracker compressor and (part of the) cold box section as the C2+ liquid(s) are directed to the back end of the cold box (FIG. 9). Typically, the dry gas cold box refrigeration duty is supplied from the steam cracker C2 and C3 refrigeration systems.

FIG. 9. Dry gas processing via a dedicated cold box (top) or directly to the cracked gas compressor.
FIG. 9. Dry gas processing via a dedicated cold box (top) or directly to the cracked gas compressor.

While the schemes in FIG. 9 include trace removal of contaminants from the FCC dry gas, bulk H2S and CO2 removal will take place upstream. At present, the dry gas is caustic-treated at the HAR and at AZK (in the cracked gas compressor train, together with cracked gas from the furnaces). In the future, the dry gas will be amine treated at the HAR. The HAR FCC dry gas contains fairly high quantities of CO2 compared to H2S, related to the low sulfur content of the crudes being processed. Using a selective amine (removing H2S while slipping CO2) to treat the FCC dry gas would result in an acid gas rich in H2S to ensure proper operation of the SRU; however, AZK would continue to have a high operating expense due to the caustic needed for removal of the CO2 slipped from the upstream HAR amine absorber.

FIG. 10. Acid gas enrichment in the SRU/TGTU.
FIG. 10. Acid gas enrichment in the SRU/TGTU.

In the final analysis, it was decided to employ a non-selective amine (DEA) to minimize CO2 slip to the AZK caustic scrubber. The inevitable dilution of acid gas to the SRU with CO2 proved to be manageable, especially because the SRU reaction furnace does not see significant NH3 now that a two-stage SWS is employed, resulting in the sour gas stream to the SRU being low in NH3. For scenarios of very low H2S production (at more or less constant CO2 production), the SRU/TGTU is equipped with a bypass line to the TGTU amine absorber/regenerator (which necessarily operates with a selective amine solvent) as an enrichment facility, rejecting CO2 to the incinerator and providing a more concentrated H2S acid gas stream back to the SRU furnace (FIG. 10).

A final interaction between the HAR and AZK is related to both hydrogen and C4. The original intent of the project was to hydrogenate the excess refinery C4 (MTBE raffinate after subtraction of C4 blended into the gasoline and/or sent to the sales LPG pool) at AZK. Processing hydrogenated C4 in the steam cracker enables longer run lengths and optimum olefin yields.

As a result of unrelated developments, the AZK complex would have insufficient hydrogen to saturate the incoming HAR FCC C4, making it more logical to increase the size of the H2 plant at HAR. In this context, it also made sense to install the new C4 hydrogenation unit at HAR, as well, creating the possibility to allow blending of hydrogenated C4—having a higher LPG blend motor octane number (MON) value (TABLE 2).

FIG. 11. Original concept design of the MTBE raffinate splitting and C<sub>4</sub> hydrogenation.
FIG. 11. Original concept design of the MTBE raffinate splitting and C4 hydrogenation.

Splitting the C4 raffinate into a light and heavy fraction is no longer necessary. The raffinate splitter originally envisaged (FIG. 11) to produce a light raffinate rich in high-MON isobutene—a tall column, given the relatively close boiling points of the components to be separated—can be removed from the scope, resulting in the final lineup (FIG. 12).

Project execution and status. The HAR revamp project kicked off in 2015. During that year, all conceptual study work was concluded. Licensor proposals were obtained and evaluated, ensuring that the licensor process design package (PDP) preparation could start in H2 2015. The main non-licensed unit (CDU/VDU) underwent a pre-front-end engineering design (pre-FEED) study to define its main revamp scope.

FIG. 12. Final lineup design of the MTBE raffinate splitting and C<sub>4<sub> hydrogenation.
FIG. 12. Final lineup design of the MTBE raffinate splitting and C4 hydrogenation.

During 2016, the licensor PDPs were completed. The author’s company, which was the contracted project management consultant (PMC), reviewed the PDPs and managed licensor interfaces. The authors believe that the schedule for this complex revamp project, with numerous interfaces to simultaneous revamp work at another site, was fairly ambitious. Having achieved this schedule without significant setbacks can be attributed to a few important success factors.

Due to a high level of mutual trust between all parties, many of whom were involved in SOCAR’s projects for some time, various potential issues could always be resolved in a timely manner, which resulted in negligible interruption of the work process. Significant preparatory work was done to ensure that the various licensor kickoff meetings were productive, and technical open ends were identified early on, enabling early closure.

Often, in complex revamp projects, the devil is not only in the details, but specifically in the interface details. Recognizing and addressing these details before they can do any harm requires solid technical knowledge and mutual trust in the project team. Due to these elements, the project is in good shape at the time of this publication. FEED was largely completed at the end of 2016, with targeted mechanical completion in 2019/2020. HP


  1. “Shell Berre steam cracker blast cause cited,” Oil & Gas Journal,, December 1990.
  2. Hack, W. R., “Investigation of NOx related cold box incident,” AiChE Spring Meeting and Global Congress on Process Safety, April 2009.
  3. Grenoble, D., “Report on activities of the NOx in ethylene recovery task group,” AiChE Spring Meeting and Global Congress on Process Safety, May 2013.
  4. Bancroft, J., “Practices to mitigate NOx risk in the cryogenic section of ethylene plants,” AiChE Spring Meeting and Global Congress on Process Safety, March 2011.

The Authors

Related Articles

From the Archive



{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}