Saturday, March 24, 2012

Debottleneck crude-unit preheat exchanger network inefficiencies

In this case history, a crude distillation unit (CDU) preheat train network in a Saudi Aramco refinery was simulated and analyzed for anticipated modifications to the network. This analysis helped eliminate inefficiencies in the network, and, based on the insights from the analysis, various options were generated and the existing network was reconfigured. The reconfiguration allowed the temperature of the crude preheat network, which processes Arab Light crude oil, to be increased to the maximum of 277°C from a previous temperature of 261°C.

Existing configuration.

Desalted crude from the tank is heated by the crude column top pumparound, light gasoil (LGO) product, heavy gasoil (HGO) product, LGO pumparound (LGO PA), HGO pumparound (HGO PA), heavy vacuum gasoil (HVGO) pumparound and vacuum residue (VR) product, as shown in Fig. 1 in exchangers E1 to E7, respectively. The current crude preheat temperature entering the CDU furnace is around 261°C. This exchanger network is validated using heat exchanger design software and by adjusting the fouling coefficients.
Modifications required.
The base-case network was altered for anticipated modifications in the future. The reasons for the modifications are listed below:
Vacuum slop circuit. In the current configuration (Fig. 2), the vacuum slop is recycled to the vacuum tower through the vacuum furnace. The purpose of this recycle is to recover the VGO components and send the VGO to the hydrocracker; however, this is not achieved in the current operation due to vacuum furnace limitations and insufficient separation in the wash section. As a result, this vacuum slop stream (which is lower in viscosity) goes with the vacuum tower bottoms. The mingling of streams deteriorates the feed to the asphalt oxidizer and creates operational problems in meeting the penetration property of the asphalt.
To address this concern, the vacuum slop stream from the vacuum tower is available at a temperature of 380°C, which is withdrawn as a separate cut and is used to increase the preheat temperature of the crude. This proposed new exchanger is configured to be in parallel with the existing heat exchanger E4 in Fig. 1. Fig. 3 shows the rerouting of the vacuum slop.
Future splitter configuration. To meet the clean gasoline specification of 1% benzene in gasoline, the existing naphtha splitter must remove the benzene precursors in the catalytic reformer feed by increasing the initial boiling point of the heavy naphtha. This process requires a higher reboiler duty. In addition, the heavy naphtha from the hydrocracker needs to be processed in the naphtha splitter, as this feed also contains benzene precursors.
Currently, hydrocracker heavy naphtha is not part of the naphtha splitter feed. The hydrocracker heavy naphtha feed volume is 12,500 barrels per day (bpd), and the existing naphtha splitter capacity is 23,000 bpd. Figs. 4 and 5 show the naphtha system’s current and planned configurations, respectively. As the current naphtha splitter cannot handle this higher throughput with higher reboiler requirement, the existing naphtha splitter will be mothballed. The existing reboiler, which uses HGO PA flow and gives a duty of 10.4 million kilocalories per hour (MMkcal/hr), will also be mothballed.
High-pressure steam will be used in the reboiler of the new naphtha splitter to meet the higher reboiler requirements. For the column to be in heat balance, this 10.4 MMkcal/hr of heat removal is required. In the proposed exchanger network, this stream (HGO CR) will be used to preheat the crude. 
 Synthesis of crude preheat train.
A new, preliminary heat exchanger network (Fig. 6) was synthesized to accommodate the above modifications. While modifying the crude preheat train network, the following impact on the equipment was kept in mind:
·         Prevention of vaporizations in the furnace pass-control valves, as it is difficult to control two-phase flows across pass-control valves. Inadequate flow in the furnace pass flows will also lead to coking
·         Column heat balance.
·         Column hydraulics.
·         Impact of hot streams going directly to the other unit.
The changes made in the base-case network are listed below:
·         Exchanger N1 was added parallel to E4 (see Fig. 6) using vacuum slop (vacslop) and vacuum residue ex-E7 as the hot fluid. This modification is required to improve the viscosity of the vacuum residue to the asphalt oxidizer. The current viscosity of the feed to the asphalt oxidizer is 1,500 centistokes (cst), and the required viscosity is 2,000 cst.
·         Another exchanger N2 (E5-2, similar to E2) was added parallel to E2 using HGO PA fluid ex-E5 (hereafter referred to as E5-1) as the hot fluid. This modification is performed to accommodate the 10.4-MMkcal/hr duty in the HGO PA circuit.
·         Increased area in E4 from the 2-parallel-1-series arrangement to a 2-parallel-2-series design and added cooler N3 downstream of E4.
Due to the first two modifications, the inlet temperature to E4 has increased, which decreases the logarithmic mean temperature difference (LMTD) available across the unit. Since E4 is the LGO PA exchanger, the column will not be in heat balance if the required heat removal is not performed. The required duty was 18.8 MMkcal/hr, and the available duty was 12.7 MMkcal/hr (see Table 1). Therefore, additional area and a cooler were added in the LGO PA circuit to meet the duty requirement of the column.
The required HGO PA duty is 26.8 MMkcal/hr, and the available duty is 29.8 MMkcal/hr. As the heat removed in HGO PA is higher by 3 MMkcal/hr, the requirement of LGO PA duty will come down by 3 MMkcal/hr. As both LGO and HGO are mixed outside of the column and go to the diesel hydrotreater (DHT), the splitting of the duty between LGO and HGO pumparound is not a concern from a separation point of view. However, it does impact the column draw temperature, which will slightly reduce the LMTD across E3 (HGO product/crude exchanger) and E5 (HGO PA/crude exchanger).

Results of network modification.

In the modified network, the obtained preheat temperature was 266°C. The duty, LMTD and area of each exchanger in the network are presented in Table 1. From Table 1, it can be observed that:
·         Exchanger E6, which has a higher area, is experiencing the lowest LMTD; therefore, any modification that increases the LMTD will significantly increase the heat recovered from E6.
·         The exchanger preceding exchanger E6 is heated by HGO circulating reflux (CR), which is at 337°C; this is higher than the hot stream (HVGO CR) temperature of E6, which has decreased the LMTD in E6.
This preliminary network was analyzed for possible improvement in the preheat temperature. The analysis indicated that heat recovery can be increased by 45% by boosting the area by 56% (see Table 2).
The analysis also indicated that the driving force across exchanger E7 further limited the heat recovery. Fig. 7 displays the driving-force plot. The figure indicates that the driving force in E7 can be increased by decreasing the inlet temperature in E7. This temperature adjustment can be achieved by operating E5 in parallel with E7.
Case 1. Based on the insights derived from Table 1 and Fig. 7, to improve the heat recovery, the crude stream in E7 and E5 was split by operating E5 in parallel with E7. The objective of this modification is to increase the LMTD across E7 and E6. However, it also decreases the LMTD across E5-1. The net effect is shown in Table 3, and the modified network is shown in Fig. 8. With this arrangement, the preheat temperature has increased from 266°C to 269°C.
Case 2. From LMTD and approach data in Table 3, it can be inferred that heat recovery in E5-1 can still be improved by increasing the area. Hence, another case study was performed by adding two similar exchangers in a series in E5-1. The results are tabulated in Table 4. The preheat was found to be increased to 277°C.
The HGO PA is now providing an extra 4.2 MMkcal/hr more than required, which will reduce the LGO PA duty requirement by the same amount for the column to be in heat balance. Then, the required LGO PA cooler duty comes down to 2.6 MMkcal/hr.

Share |

Friday, March 23, 2012

Optimize hydrogen management for distillate production

Moving more distillate streams to diesel production and/or fuel oil (FO) production is a major refining activity. The distillate qualities and quantities, hydrogen price and consumption, product demand and prices, and refinery configuration constraints are key factors for these decisions.
Producing higher-quality products from poorer-quality crudes requires more hydroprocessing. These distillate streams, e.g., heavy kerosine (HK), gasoils (GOs) and light cycle oils (LCOs), etc., are also used as cutter stocks to upgrade vacuum residues (VRs) and downgraded as low-value FO products. FO and VRs are major constrains if no resid-upgrading facilities are available.
The heavy distillate streams from the fluid catalytic cracking unit (FCCU) must be hydrotreated to upgrade todiesel specifications. Hydrotreating consumes substantial hydrogen quantities, and this process adds more costs. When adding new hydrogen-consumer streams, hydrogen demand can exceed available refinery supplies. In such scenarios, optimizing existing hydrogen supplies is the key to improving the total profitability of the refinery.
A simple reliable optimization tool for routing intermediate distillate streams maximizes benefits while meeting all constraints of hydrotreating capacity, hydrogen, VR utilization, product specifications and prices. The Excel-based calculator is generic for hydrogen management, and it also fits into various refinery configurations where different resid-evacuation options are practiced.

Hydrogen management.

Processing opportunity crude oils and meeting critical product specifications of EURO III, IV and V are the real challenges for refiners. The quality of straight-run (SR) distillate streams obtained from the processing of high sulfur (S) and low API crude oils is considered inferior. Thus, refining higher-quality products from poorer-quality crudes has increased hydrogen addition. Conversely, LCOs (from the FCCU) are such intermediate distillate streams; they are also routed through the diesel hydrodesulfurization (DHDS) unit to meet product specifications. These streams (especially LCO) consume substantial hydrogen quantities, thus increasing processing costs to meet final diesel specifications. Under these conditions, optimizing hydrogen consumption is the key to total profitability.
Planning tool. To fully optimize the routing of intermediate distillate streams and hydrogen management, additional requirements for these streams must be added to existing planning and optimization tools. These tools must not only determine hydrogen and hydrocarbon routings, but also accommodate individual unit capacities and refinery configuration constraints.
These intermediate distillate streams, in general, are also routed to residue evacuation when no modern resid-upgrading facilities such as coker/visbreaker/solvent deasphalting (SDA) are available. Adding intermediate distillate streams as cutter stock with residues produces FO, when capital investment may not be required. In the presented study, an optimization tool evaluates the routing of intermediate distillate streams to the diesel pool and/or resid-upgrading while meeting the refinery configuration constraints.
More hydrogen demand. Hydrogen consumption for intermediate distillate streams has grown significantly. Hydrogen-addition processes, in general, are preferred due to two factors. First, new environmental regulations over transportation fuels require higher-quality refinery products. Second, the differential prices for light- and heavy-crude oils continue to increase as light-crude reserves are declining, and supplies of heavy-crude oils are increasing. Refiners are taking advantage of these spreads; they are incorporating more lower-cost, heavier, sour, opportunity crudes into the feedslate.
Under these conditions, it is essential to understand the crude oils and their hydrogen content. Increased hydrogen consumption is an additional cost to process these crude oils. Therefore, to produce the same yields of transportation fuels either carbon rejection and/or hydrogen-addition processes must be selected. In actuality, even with incremental new carbon-rejection process capacity, additional hydrogen consumptions and, thus, their enhanced process capacities are preferred. This processing scheme enables optimizing hydrogen management for the refinery. The presented optimization tool can help facilitate efficient hydrogen usage in various resid-upgrading scenarios.

Problem definition.

While processing high-sulfur crude oils at a crude distillation unit, the processed distillate qualities were considered inferior. These intermediate distillates and LCO (from the FCCU) streams are routed through the DHDS unit to meet diesel-product specifications. These streams (especially, LCO) consume additional hydrogen to meet final diesel specifications. However, there are capacity limitations for hydrogen (maximum of 35 tpd) and hydraulic limits for DHDS capacity (maximum of 6,000 tpd). This is a common problem for any refinery. The presented study can be replicated to any other refineries with many commonalities and/or additional constraints.
In the presented case, the VR is being evacuated as FO, where it consumes distillates as cutter stock to meet the final product specifications. Two grades of FO (180 cst and 380 cst) are produced when the cutter profiles are different. Thus, the available distillate streams are being used either for diesel production, which has a higher value and/or routed to FO production, which is needed for upgrading VR.
The minimum VR production is approximately 3,500 tpd while processing 18,000 tpd of crude oils (6 MM tpy crude oil processing basis). The distillate streams available at this refinery are HK (high sulfur), GO (high sulfur) and LCO (high sulfur), etc. Fig. 1 shows the processing flow diagram for routing distillate streams. The cutter requirement depends on the quality of the produced distillate products. However, while optimizing the overall FO production, one must consider the total hydrogen consumption for upgrading both SR and cracked feedstocks.

A simple optimization tool was developed for optimal routing of the distillate streams to diesel and/or FO production. The study considered all constraints of meeting DHDS capacity, hydrogen capacity, VR utilization and product specifications, prices, etc.


Diesel production is always the first choice due to its higher value over FO production. When distillate streams are routed through the DHDS unit, hydrogen is consumed to meet diesel-product specifications and, of course, increases processing costs. However, FO production needs no additional cost, but FO demand is declining. In this scenario, VR upgrading is one of the limits when equivalent distillates are downgraded. Thus, an optimal decision must be made between additional hydrogen consumption vs. producing more low-value FO products.

Hydrogen consumption.

In diesel hydrotreating, hydrogen consumption is governed by feed properties and product specifications. The affecting variables are carbon/hydrogen (C/H) ratio, S, basic nitrogen (N) and metal content, etc. To estimate the C/H ratio, a correlation as a function of specific gravity is applied. In this study, C, H and impurities (I) are evaluated in a balanced approach, especially across the DHDS unit to determine H2 consumption while upgrading distillates. The estimated H consumption is based on these assumptions:
·         Data for distillates are from refinery test runs
·         H2 consumptions are based on the distillate quality
·         All components, other than C and H, are considered as impurities for the calculations
·         Estimated cost for H2 is $2,150/ton.
In this approach, the Excel-based optimization tool was developed to maximize benefits by optimum routing of intermediate streams to diesel (via DHDS) and/or FO production. The spreadsheet is enabled with macro, where input and output are linked with a single click button, as shown in Fig. 2. The model has provision to enter all inputs for the total distillate quantities available to routing and qualities, product specifications and prices, cutter profiles to meet product specifications and all process limits, e.g., DHDS capacity, hydrogen and VR upgrading, etc. These input data are treated as the Base Case, which is normally being practiced. The output data are reported as H2-consumption profiles for each stream, capacity utilization, optimum routing of intermediate distillates, final product profiles and overall benefits.

Share |


Twitter Delicious Facebook Digg Stumbleupon Favorites