Refinery configurations for maximising middle distillates A comparison of configurations for upgrading residual oil products for the maximum production of diesel
Arun Arora and Ujjal Mukherjee Chevron Lummus Global
efiners globally continue to face numerous challenges as 5.0 World environmental laws become 4.5 R ECA increasingly stringent. Principal 4.0 among them in the near future will 3.5 be the International Maritime 3.0 Organisation’s (IMO) proposed changes to bunker fuel oil sulphur 2.5 limits, from the current limit of over S, wt% 2.0 3.5% down to 0.5% globally and 1.5 from 1% to 0.1% in Emission 1.0 Control Areas (ECA, see Figure 1). 0.5 Global demand for high-sulphur 0.0 residual fuel oil (HSFO) is steadily 2005 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 declining too, by 35% since 1995. Both of these changes will signifi- cantly impact a refiner’s ability to Figure 1 Requirements of IMO regulations Source: International Maritime Organisation market any significant quantity of HSFO at a price that will maintain technology are just two of the diesel-driven vehicles in Europe refinery profitability. Refineries reasons why the demand for diesel and developing countries. Max two-column width currently making a significant continues to outpace that of gaso- Current refining investment is Blue background around amount of fuel oil and lacking the line. IMO regulations will indirectly predominantly made in Asia, the brown. complexity to upgrade the residual increase diesel demand further as Middle East, Russia and Latin oil to premium products (middle refiners are forced to blend in addi- America — regions with growing Always translate into UK distillates) will face two difficult tional low-sulphur diesel to meet demand for refined products. English (except images). options: either invest in commer- fuel oil sulphur specifications. Tightening of product quality speci- Leading cap on rst word cially proven and reliable solutions Worldwide, production of mid- fications will accelerate the only. to convert HSFO to more valuable distillates is projected to account for implementation of deep conversion liquid products such as Euro V 55% of the rise in oil demand units in existing refineries, but often diesel to greatly improve the refin- expected over the next 20 years. The these refineries are constrained by ery’s profitability, or face a threat to shift to diesel puts emphasis on plot space, hydrogen and other shut down the refinery as the opera- bottom-of-the barrel processing. infrastructural issues. Grassroots, tion becomes uneconomical to export-oriented refineries are all continue. Growing demand geared towards high conversion to Worldwide demand for refined mid-distillates. Shift in product demand products is projected to increase For the strategically oriented The IMO’s looming specification significantly in the next 20 years, refiner, stringent requirements for changes (see Figure 1) are likely to driven by population growth and high-quality products actually accelerate the decline in demand for the transition of emerging markets present an opportunity to invest in HSFO by the year 2020, if not earlier. into the global economy, with the the right technologies to signifi- Worldwide, including emerging majority of growth coming from cantly improve refinery margins. markets such as China, India and China in particular and Asia in Based on increasing product the Middle East, there is a shift in general. According to OPEC, global demand and the closure of multiple product demand from gasoline to demand for diesel fuels is expected non-performing refineries, refining diesel. Ethanol substitution in gaso- to grow by 10 million b/d by 2030, margins are expected to recover line and improvements in engine driven by an increased share of by 2015.
www.eptq.com PTQ Q3 2011 75 current trends only such refineries Coke Coke will survive in the future) drum drum • Ideally, should be part of a conversion platform encompassing Fractionator Reflux drum Wet gas complementary technologies. Technologies on the cusp of Unstablised commercialisation were excluded, naphtha because we did not want to prescribe Gas oil any solution without a reasonably stripper long operating history. For example, Light gas oil there are several slurry-phase resi- Residual oil due conversion processes on the Heavy gas oil verge of commercialisation, but Heater without a commercial operating history there is no data on reliability Figure 2 Schematic of a delayed coking unit and on-stream factor — a major consideration in any residue upgrad- A wider and more intense residue conversion technologies, ing process because of the difficult requirement for the deployment of keeping the intentions of a global nature of the feedstock. emissions reduction technologies refiner in mind. The conversion Major refinery processes included may also act as a catalyst for new technology: in this evaluation were: investments. Modern hydroprocess- • Should be commercially proven • Delayed coking ing technology will eliminate the and reliable with a good on-stream • LC-Fining (a high-conversion need for expensive downstream factor residue hydrocracking process) remediation technologies. • Should maximise the most valua- • Residue desulphurisation (RDS) It is our view that refining should ble product (diesel) while retaining • Solvent deasphalting (SDA) be viewed as an ongoing business, the capability to address niche • Combinations of the above, along where long-term average margins product demands for the foreseea- with secondary processes such as and product price differentials will ble horizon hydrocracking, residue catalytic support the investments that are • Should be flexible to handle more cracking (RFCC) and gasification needed. difficult feedstocks (VR and coke), FCC feed/product • Should be environmentally desulphurisation and various Residue upgrading technologies compliant to meet future stringent gasoline-producing processes. In view of the increasingly stricter specifications In the studies we conducted for regulations expected in the near • Should have enough complexity various clients, the residue conver- future, along with the emerging so that the refinery remains profita- sion technologies that rose to the trends in product demand, CLG ble when margins remain depressed forefront were delayed coking, LC- evaluated multiple combinations of for prolonged periods (based on Fining and RDS. The screening
www.ptqenquiry.com 76 PTQ Q3 2011 for further information www.eptq.com Advantages and disadvantages of delayed coking
Advantages Disadvantages Lower on plot capital investment compared to hydrogen addition processes Coke handling, plot area limitations, and transportation and logistics Can handle very poor-quality (high in contaminants) feeds Additional environmental health and safety (EHS) requirements Widely used, with many references Hydrogen addition still required to upgrade products and the process does not share the same process platform as other hydroprocessing units Favoured in low crude oil price environment Loss of liquid yield compared to hydrogen addition processes No residual liquid product to deal with Coke disposition is a major issue
Table 1 phase quickly ruled out several Installed LC-Fining units technologies as being too expensive, such as gasification, or not geared Start-up Client BPSD Unconverted oil towards maximising diesel, the 2011 Shell Canada 47 300 Stable HSSC product of choice. A brief descrip- 2010 GS Caltex 60 000 Stable FO tion of the primary upgrading 2007 Neste Oil 40 000 Stable FO 2003 Shell Canada 79 000 Stable HSSC processes follows. 2000 Slovnaft 23 000 Stable LSFO 1998 AGIP Petroli 25 000 Stable LSFO Delayed coking 1988 Syncrude Canada 40 000 Coker feed Delayed coking is the most widely 1984 BP-Amoco 75 000 Coker feed Total 8 units 389 300 used residue conversion technology and is particularly valuable when a long-term off-take arrangement for Table 2 coke exists. Almost every major grassroots refinery in the world has VGO. HCGO is either sent to a FCC considered it as a primary residue feed pretreater (in a gasoline- Catalyst addition conversion process, with the excep- oriented refinery) or a hydrocracking Effluent tion of locations such as Scandinavia, unit (in a diesel-oriented refinery). line Thermowell Western Europe and Eastern The coke produced by a standalone Density nozzle detector Canada, where coking units are not delayed coker (see Figure 2) is lower radiation preferred. Fuel-grade coke is used in value fuel-grade coke. If a hydro- source infrastructure projects (cement, processing unit such as an LC-Fining well power) and demand remains robust unit precedes the delayed coking Density detectors in developing countries. However, unit, the coke produced from the Normal with even more large coking units delayed coking unit can be of supe- bed level coming online, coke demand could rior anode-grade quality, suitable for come under pressure. use in the aluminium industry. Vacuum residue, normally Table 1 shows the main advantages Skin destined for fuel oil, is thermally and disadvantages of the delayed TCs cracked to obtain nearly 70% of coking process. distillate products. All distillate Catalyst products require further hydro- LC-Fining withdrawal processing to make finished The LC-Fining process is a resid- line Feed Recycle products. Coker naphtha requires uum conversion process that pump special and more severe hydro- hydrocracks the most difficult, processing compared to straight-run heavy, lower-value hydrocarbon naphtha. Light coker gas oil (LCGO) streams such as petroleum residua, that boils in the diesel boiling range heavy oils from tar sands and shale Reactor temperature 410–440°C has a much higher nitrogen content oils to lighter, more valuable prod- Reactor pressure 110–180 bar compared to straight-run diesel and ucts such as VGO, diesel and Resid conversion 55–80% operating pressures required for naphtha. The process involves an Hydrogen P. P. 75–125 bar hydroprocessing are relatively ebullated-bed reactor (see Figure 3) Chem H2 higher. Heavy coker gas oil (HCGO) that completely mixes oil and consumption 35–300 Nm3/m3 boils in the vacuum gas oil (VGO) hydrogen. Due to continuous addi- Desulphurisation 60–85% boiling range. HCGO has much tion and the withdrawal of small CCR reduction 40–70% Demetallisation 65–88% higher total aromatics, nitrogen, quantities of catalyst, the run polycyclic aromatics and asphaltenes, lengths between shutdowns are and requires more severe operating long. Unconverted oil from the LC- Figure 3 Schematic of a LC-Fining reactor conditions compared to straight-run Fining unit can be used as fuel oil with typical operating conditions
www.eptq.com PTQ Q3 2011 77 Make-up hydrogen
External gas oils HDT/MHC for MHC/HDT reactor
Wash H tower 2 purification
LP amine scrubber HP amine scrubber
Vacuum residue Hydro- and processed diluent distillates and gas oils To LVGO fractionation LC-Fining reactors with To HVGO to FCC fractionation inter-stage stripper Hydroprocessed residue
Figure 4 LC-Fining process with integrated hydroprocessing
Advantages and disadvantages of LC-Fining
Advantages Disadvantages Higher liquid gain compared to delayed coking On plot investment is higher than delayed coking units Can handle feeds higher in metals and other contaminants Residue stability may become a concern at high conversions (feed dependent) compared to fixed-bed processes Middle Eastern feeds, for example, have no stability concerns even at high conversions (proven commercially) Long run lengths More complex process compared to delayed coking and requires better operator training Can be integrated easily with other hydroprocessing units Not as much commercial experience as delayed coking units but adequate Ebullated-bed technology is a mature technology and over Spent catalyst disposal (trucks, rail car) has to be considered. Spent catalyst normally 30 years’ operating experience has led to many technological sent to metals reclaimer advances and made the process very reliable Requires less plot space compared to delayed coking units Unconverted oil disposition can become an issue depending on sulphur/stability specifications
Table 3
or as feed to power plants or a to the coking unit means less coke reduced if the feed quality delayed coking unit. Maximum make and thus a higher yield improves. conversion is dependent on feed- of liquid fractions that can There are only two ebullated-bed stock. Operating unit conversion subsequently be converted to trans- processes in the world that have ranges from 60% to over 80%. portation fuels. been proven by long commercial The LC-Fining unit operates at The LC-Fining unit has inherent history: LC-Fining and H-Oil. Table pressure levels similar to high- flexibility to meet variations in 2 is a list of operating LC-Fining pressure hydroprocessing and feed quality/throughput, product units, and Table 3 shows the main therefore offers excellent opportuni- quality and reaction operating advantages and disadvantages of ties for capital reduction by severities (temperature, space the process. permitting integration of either velocity, conversion and so on). The unconverted oil from the LC- hydrotreatment (Shell, Canada) or This flexibility is a direct result of Fining unit is normally used as fuel complete hydrocracking (Neste, the ebullated catalyst bed reactor oil. When combined with a delayed Finland, see Figure 4). system. In an ebullated-bed unit, if coking unit downstream, the uncon- The conversion of Conradson the metals or sulphur content of verted oil is converted to distillates carbon is economically important if the feed increases, the product and anode-grade coke, which LC-Fining vacuum bottoms are fed quality is maintained by increasing fetches a far higher price compared to a downstream coking unit. A catalyst consumption. Conversely, to fuels-grade coke. While LC- lower carbon-content resid product the catalyst consumption is Fining can handle a relatively high
78 PTQ Q3 2011 www.eptq.com become less severe and conversions !TMOSPHERIC can be pushed much higher. The ,OW 3. !TMOSPHERIC RESIDUE ,OW 3. yield slate shifts towards lighter DISTILLATES DISTILLATES RESIDUE AND 6'/ products and catalyst consumption VDU drops significantly. Without heavy VDU asphaltenes in the process, unit 0REMIUM operating factors improve as well. &##2&## The obvious disadvantage is the FEED loss of global conversion, as a Solvent significant volume of residue is LC-Fining Solvent deasphalting removed as pitch and, without a LC-Fining dedicated disposition of the large deasphalting volume of pitch (such as a gasifier), 0ITCHPELLETS ,3&/ the economics may not be favoura- ble. The option becomes very attractive in those situations where Figure 5 SDA upstream of an LC-Fining an SDA is already in operation and unit 0ITCHPELLETS there is a need to upgrade the DAO DELAYED COKING to diesel rather than routing to an metals content in the feed, the high FCC unit for conversion to level of nickel and vanadium in the gasoline. Figure 6 SDA downstream of an LC-Fining unconverted LC-Fining bottoms The SDA process can also be unit could limit the production of anode- integrated downstream, where grade coke in the downstream deasphalting removes the heaviest metals, nitrogen and sulphur from delayed coking unit. LC-Fining by asphaltenic residue from the petroleum residua in the presence of itself produces significantly more unconverted oil. The DAO can be hydrogen (see Figure 7). Conversion liquid yield compared to delayed recycled back to the LC-Fining results from the level of desulphuri- coking and improves the refiner’s process, while the pitch can be sation required and is not by itself a volume gain. blended in with incremental VR to target. The process is normally used The LC-Fining process is also easily an existing delayed coking unit to produce low-sulphur fuel oil or to integrated with a solvent deasphalt- (BP, Texas City). Conversion is produce a feed stream that is suita- ing unit either upstream (see Figure boosted and the volume of pitch ble for cracking in a residue FCC 5), downstream (see Figure 6) or as to be handled is reduced (RFCC) unit. an inter-stage process. significantly. The RDS process is by far the An upstream SDA significantly most widely used residue upgrad- reduces metals, CCR and Residue desulphurisation ing process. Catalyst and process asphaltenes. Operating conditions RDS is a fixed-bed process that has innovations for RDS include the required in the LC-Fining unit multiple beds of catalyst to remove upflow reactor (UFR) and onstream
Make-up hydrogen To gas recovery Reactors Recycle gas H S scrubbing 2
Cold HP Unstabilised separator Product naphtha H O 2 stripper
Fresh feed Filter Sour Product water Hot HP LP separator separator
Figure 7 Schematic of RDS process
www.eptq.com PTQ Q3 2011 79 Upgrading configurations
Catalytic CLG explored several configura- reforming tions, including some that are based on the residue upgrading platforms Gasoline Atmospheric described here, along with other tower HDT major processes such as hydroc- FCC racker, hydrotreater and FCC to maximise conversion to mid-distillates.
Vacuum Jet/diesel Hydrocracking Refinery with delayed coking as tower primary upgrader The configuration shown in Figure 8 is one of the most common refin- Delayed ery configurations and is a coking benchmark against which other configurations have been evaluated. The configuration is robust and, Coke to power plants depending on the crude slate, the and capacities of the hydrocracking and cement industry FCC unit vary to obtain the right balance between gasoline and diesel Figure 8 Simplified refinery with delayed coking as a residue conversion process production. In extreme situations, where gasoline production is to be catalyst replacement (OCR). UFR is RDS is a widely used technology, avoided, the configuration will have typically used in revamp situations especially in the Far East. It is the no catalytic reforming and no or when concerns about metals only technology that can produce FCC unit. levels and catalyst pore mouth <0.5 wt% sulphur fuel oil. The Figure 9 shows a refinery config- plugging might shorten downflow technology is used in this context uration where there is virtually reactor run lengths due to excessive in Japan, but the most prevalent no demand for gasoline and pressure drop. OCR is used when use of RDS is as a unit feeding a the refiner is only interested in the metals level in the feed is RFCC unit for the production of making middle distillates and excessive. gasoline. petrochemicals naphtha. Such a
LPG LPG Light naphtha Petro-chemical Heavy naphtha NHT Naphtha CDU
Jet A-1
Coker H2 plant Naphtha
Diesel VDU HCR DHT
HCGO UCO
Fuel oil LCGO DCU Fuel coke
CFB
CFB: Circulating fluidised bed boiler
Figure 9 A delayed coking-based refinery with no gasoline production
80 PTQ Q3 2011 www.eptq.com Catalytic reforming Gasoline
FCC Atmospheric tower HDT
*ETKEROSENE Vacuum Hydrocracking tower $IESEL Jet fuel Diesel
Delayed LC-Fining coking unit
Anode-grade coke
Figure 10 Optimised residue conversion using LC-Fining and delayed coking
configuration is likely to become the LC-Fining process to the coking unit, which also converts increasingly important in the next upgrading of residue. part of the UCO to distillates to be decade. The LC-Fining unit is the primary processed in downstream hydro- residue conversion process, where processing units. This configuration Further optimisation of the delayed conversion is pushed to the maxi- has no undesired or low-valued coking-based refinery mum because unconverted oil products and is therefore truly The overall profitability and return stability is not an issue. The uncon- “bottomless”. on investment in the configuration verted oil, low in sulphur and Furthermore, the configuration is depicted in Figure 10 improves metals, is converted to high-priced very amenable to phasing; the LC- significantly with the addition of anode-grade coke in the delayed Fining unit can be built first and
Catalytic reforming
Gasoline RFCC Atmospheric tower
,#/
Vacuum Hydrocracking *ETKEROSENE tower $IESEL Diesel
Residual desulphurisation (#/ Low sulphur (RDS) fuel oil
Figure 11 Refinery scheme with RDS as residue upgrader
82 PTQ Q3 2011 www.eptq.com Environment. The daily availability refinery’s current situation and to completion. These factors not only will be profitable until such time as coking, provides the maximum residue hydrocracking, Middle East Petrotech, of these values has been acknowl- propose operating improvements. guaranteed2010. that the project was there is a market for fuel oil. The return and the highest NPV, edged by ISO 14001 auditors. In other words, SIR has highlighted successful3 Sieli G, Gupta for both N, Delayed parties, coking but and also LC- delayed coking unit can be phased afollowed noticeable by improvement the “delayed in coking work- ensuredFining technology continuous — a winning improvement combination, E^llhgle^Zkg^]in after a few years. ingalone” productivity. option. The recommended after2008 ERTCthe Cokingproject and was Gasification completed. conference, TheThe implementation solution with RDS of an becomes EMS technologyActions taken platform by the is management proven and Rome, Italy. basedrelevant on when an advanced there isDVR a tech high- ofthe SIR solution are is steps not dependent towards the on Kh[^km
ppp'imj^gjnbkr'\hf ppp'^imj'\hf www._hk_nkma^kbg_hkfZmbhg ptq enquiry.com IMJJ++)**-. for further information www.eptq.com PTQ Q3 2011 83
belsim.indd 5 10/3/11 13:15:32