Petróleo Brasileiro (Petrobras)

Transcription

1 Revamping crude towers for quality and yield A successful revamp of distillation towers for increased diesel quality and yield depended on reliable design and accurate assembly LEONARDO SOUZA, RAFAEL WAGNER, CLAUDIO ROCHA and HENRY GIRON Petróleo Brasileiro S.A. ANTONIO CORTINES Sulzer Chemtech Petróleo Brasileiro (Petrobras) has carried out a major project to increase its production of ultra-low sulphur diesel in order to reduce Brazil s dependence on foreign markets. This is in response to growth in internal demand for fuels and environmental restrictions limiting the market s tolerance to heavier diesel grades with higher sulphur content. As part of this effort the fractionation of several crude distillation towers, both atmospheric and vacuum, has been improved to increase the diesel yield while still meeting the quality requirements of the hydrotreating units. The project is presented as a case study of the successful revamp of a distillation unit where old multi-pass valve trays have been replaced by up-to-date structured packing and the associated internals, highlighting the importance of the chimney trays and their design to ensure proper liquid and vapour distribution throughout the packing bed. Need for more diesel In 2013, following impressive growth in the nation s internal Growth, % Figure 1 Growth (%) of ultra-low sulphur diesel internal sales relative to the results of 2013: Petrobras ( ) demand for environmentally friendly fuels with lower sulphur content (see Figure 1), the Refining Executive Management of Petrobras launched a corporate programme, called Promega, intended to increase the production of gasoline and medium distillates, including ultra-low sulphur diesel, also known as diesel S-10 due to the sulphur content specification of 10 ppm or less. Several actions have been employed so far, including fractionation improvement in some key sections of atmospheric and vacuum crude distillation towers, since diesel S-10 imposes a challenge beyond its ultra-low sulphur content: it requires an ASTM D-86 distillation curve with a T95% (temperature at which 95% of the product is vaporised) lower than 370 C, which is considerably lighter than the other automotive diesel grades commercialised by Petrobras (see Table 1). One of the refineries targeted for improvement of crude fractionation was REDUC (Duque Revamps

2 Diesel grades Sulphur content, ppm ASTM D-86 distillation curve point S max T95% = 370 C max S max T85% = 360 C max S max T85% = 370 C max Table 1 Specifications of automotive diesel grades commercialised by Petrobras for the internal market Scope of modifications in the atmospheric and primary vacuum towers The original design of the towers used only trays, with moving valves as contact devices for mass transfer between the phases. It is a well known revamping practice for crude distillation towers to replace trays with packing, usually structured, in order to increase capacity and improve fractionation, since all of the cross-sectional area of the tower is used for contact between the phases (there are no downcomers) and there is no need to have a vapour-liquid disengaging space as happens in trays. However, packing is very susceptible to liquid maldistribution issues, which can ruin fractionation if the necessary precautions are not taken at the design and assembly stages. In the particular case of the lubes distillation unit there were extra sources of concern regarding the atmospheric tower: the heavy diesel wash zone was intended to be replaced by structured packing due to insufficient quality in the cut. But the section above, which has four pass trays, would not be modified, since the quality and yield of the light diesel have been considered to be satisfactory. In addition, the atmospheric residue stripping section would keep its trays, with the originals replaced by fixed valve trays, due to fouling issues observed in previous turnarounds, and adding an extra tray to improve fractionation. Therefore, besides modifying the heavy diesel wash zone itself, replacing the trays with structured packing and its assode Caxias refinery). The refinery was built in the early 1960s and two of its three crude distillation units had never been modified for fractionation improvement until 2015; these two units were built in the 1970s and have been operating for lubes production since then. As a major turnaround had been previously scheduled for the last quarter of 2015 and included one of the lubes distillation units, the Promega staff at the refinery decided, in 2013, to use this turnaround to revamp the atmospheric and vacuum towers for increased diesel quality and yield. The lubes distillation units have the same capacity and are very similar to each other, Feed Atmospheric residue Atmospheric Kerosene Light diesel Primary vacuum Sweetening Vacuum gas oil Primary vacuum residue having an atmospheric tower that produces kerosene for jet fuel, light diesel, heavy diesel and atmospheric residue. This last cut is sent to a primary vacuum tower and produces gas oil (VGO), spindle oil and two other heavier cuts for lubes production. There is also a secondary vacuum tower that receives the primary vacuum residue and processes the heaviest cut for lubes and secondary vacuum residue, which is used for asphalt production. The atmospheric light and heavy diesel and primary vacuum gas oil cuts are mixed together before going to the hydrotreating units. Figure 2 shows a simplified refining scheme for the lubes distillation units. Lubes cut I (spindle oil) Lubes cut II Lubes cut III Hydrotreating Secondary vacuum residue Figure 2 Simplified refining scheme for the lubes distillation units Secondary vacuum Lubes cut IV Asphalt 2 Revamps

3 ciated internals (liquid distributor and bed support), there was a need to install two chimney trays, allowing transition between the three different sections: light diesel x heavy diesel (four pass trays), heavy diesel wash zone (structured packing) and atmospheric residue stripping (two pass trays). The modifications in the primary vacuum tower were simpler, but not free of concern: the gas oil x spindle oil section was targeted to have its fractionation performance improved by replacing its two pass trays with structured packing, but the sections below would keep their trays and, since the available free space was limited, the use of a chimney tray would excessively reduce the amount of packing to be installed. This issue was solved by keeping the lowest tray in the section, from which the spindle oil is drawn, and letting the liquid from the bed support drip directly onto this tray. The liquid distributor would use the same nozzle already in service for directing the internal reflux from the section above only the internal feed pipe was replaced by one fitted to the distributor. Design engineering With the scope of modifications defined, the next step was to design the new internals, an operation divided into two phases. In the first phase, a representative process simulation of the unit operating in the desired conditions (maximising heavy diesel to the limit of the ultralow sulphur diesel specification, and maximising vacuum gas oil to the limit of the viscosity Atmospheric tower specification for spindle oil) determined the flow rates and properties of the liquid and vapour internal fluxes. This information was required for the bidding process in which the structured packing, its associated internals, and the fixed valve trays would be selected from different suppliers. In addition for the bidding process, it was necessary to determine the desired number of theoretical stages for the packing beds, the limits of pressure loss under liquid load, percentage of flooding in turn-up conditions and up-lift resistance for the packings and trays and, finally, the materials to be used in their manufacture. Sketches describing the height restrictions of each section were also available (see Figure 3). When the bidding process Scope of supplier Vacuum tower Figure 3 Sketches showing the height restrictions considered for each section (values not shown) for the atmospheric (left) and vacuum (right) towers was completed, with one of the suppliers chosen, the process and mechanical details of the packing beds, their associated internals, and the fixed valve trays were exhaustively checked and discussed with the supplier, since the internals would also interact with the chimney trays, which were outside the supplying scope. The second phase involved design engineering the chimney trays themselves. The basic design defined the number and diameter of the downflow pipes, the number, arrangement, and dimensions of the chimneys and, in the specific case of the upper chimney tray, the dimensions of its sumps (see Figure 4). The arrangement of the downflow pipes in both trays was defined at the detailed Revamps

4 Chimneys nozzle Sump Down pipe design stage, along with all mechanical detail regarding supportation. A proper arrangement of the downflow pipes is essential to ensure that the distributor or tray below will be evenly fed and CFD simulations can be used to check the design. 1 Failure in properly designing chimney trays can lead to liquid maldistribution issues, which Seal pan nozzle Distributor Chimneys Down pipe Sump Figure 4 Sketch of the upper chimney tray with the relevant dimensions considered in the basic design (values not shown) can ruin fractionation 2 ; insufficient vapour disengaging in the sumps, which can cause pumps to cavitate if their suction pipes are not self-venting; and liquid entrainment. Assembly This stage is crucial to ensure proper installation of the internals. There are several reported cases where assembly problems Figure 5 Upper chimney tray, before shipping from the factory (left), and its partial assembly on the atmospheric tower (right) have led to maldistribution issues. 3 In the particular case of the lubes distillation unit, the new internals were pre-assembled outside the towers before installation, in order to check for missing parts and foresee possible assembly problems. In addition, the contractor responsible for the assembly was specifically instructed in the task and the installation was supervised by a process engineer. Finally, the trays were checked for leakage and the distributors were checked for leakage and performance. Figure 5 shows the upper chimney tray, before shipping from the factory, and its partial assembly on the tower; one of the sumps was already in place. There was literally no room for mistakes as can be seen by the proximity of one of the upper tray downcomers to the distributor. Operation with the new internals After a turnaround of one month, the unit started operating in mid-december The start-up went well and after some time the atmospheric cuts were all specified, while in the primary vacuum tower it took several weeks for the operators to find their way through the behaviour of the tower with the new internals, since the improvement in fractionation between gas oil and spindle oil affected the viscosity of the last cut and it was necessary to find a new operating point to get it specified. Gain evaluation The methodology used to evaluate the effectiveness of the 4 Revamps

5 new internals involved comparing the average yields of medium distillates in the five months after start-up with the values obtained in the same time period after the previous turnaround in This was done in order to eliminate possible variations in yields caused by operating time, for instance fouling or damage to the internals. In addition, medium distillates yields, including kerosene, were preferred as a comparison parameter since there may be seasonal maximisation of kerosene from diesel for the jet fuel market as a result of an increase in domestic flights for summer vacations. Finally, the ASTM D-86 distillation curves for diesel and gas oil were compared and analysed, to identify improvements in fractionation. Figure 6 shows an increase in kerosene yield of 2.2% of the crude feed rate against a loss of less than 0.5% in the diesel yield, including vacuum gas oil. Nevertheless, there was an increase in medium distillates of around 1.7% of the crude feed rate. This achievement was obtained with almost the same T95% for heavy diesel, as can be seen in Figure 7, which also shows an improvement in the fractionation of vacuum gas oil after the modifications: the ASTM D-86 distillation curve is more flat and its initial boiling point is 25 C higher, more or less. Finally, Figure 8 shows a lighter light diesel after the modifications, while keeping almost the same distillation curve for heavy diesel, indicating a fractionation Volumetric gain, % `1.0 Kerosene Diesel + VGO Kerosene + diesel + VGO Figure 6 Volumetric gain relative to crude feed rate 2016 vs 2012 Temperature, ºC Temperature, ºC VGO Figure 7 ASTM D-86 distillation curves for heavy diesel (top) and VGO (bottom) 2016 vs Revamps

6 Temperature, ºC Temperature, ºC Light diesel Light diesel Figure 8 ASTM D-86 distillation curves for light and heavy diesel 2012 (top) vs 2016 (bottom) improvement, since kerosene was maximised. In addition, a reduction in the overlap between T95% (lighter) and T5% (heavier) after the modifications can be seen, also suggesting an improvement in fractionation. The quality improvement observed indicates that diesel yield can be pushed harder, even during kerosene maximisation, in order to eliminate losses and further increase the yield of medium distillates. Conclusion This article has presented a successful case of a distillation towers revamp for increased diesel quality and yield. The project has shown that the key to success depends, essentially, on two objectives: reliable design in all aspects, and proper assembly. The first objective should be exhaustively discussed with the internals supplier and carefully checked: conventional design techniques seem to be satisfactory and CFD can be used as a design aid, if necessary. The second objective, proper assembly, should be very well planned and the contractor in charge of the assembly has to be specifically instructed in the task. Finally, the entire process, from design to assembly, should be supervised by an in-house coordination team comprising process and equipment engineers. References 1 Ropelato K, Rodrigues R, Bellote A, Schnaibel G, Torres G, Marsiglia M, Fonseca C, Castro A, CFD Study for chimney tray design in crude fractionators, 2007 AIChE Annual Meeting, Salt Lake City, Utah. 2 Ropelato K, Moraes A, Bellote A, Schnaibel G, Torres G, Marsiglia M, Fonseca C, Castro A, Feitosa F, The jet fuel problem strikes again, 2009 AIChE Spring Meeting, Tampa, Florida. 3 Kister H, Distillation Operation, 1990, McGraw-Hill, New York. Leonardo Leite Garcia de Souza is a Process Engineer in the Process Optimization Department of Duque de Caxias refinery (REDUC) with Petróleo Brasileiro S.A. (Petrobras S.A.), Rio de Janeiro, Brazil. He holds a BSc and a MSc in chemical engineering from Federal University of Rio de Janeiro and a Specialisation Course in Petroleum Refining from Petrobras Corporate University. llgs@petrobras.com.br Rafael Ramos Wagner is a Process Engineer in the Process Optimisation Department at Duque de Caxias refinery. He holds a BSc and a MSc in chemical engineering from Federal University of Rio Grande do Sul and a Specialisation Course in Petroleum Refining from Petrobras Corporate University. rafaelwagner@petrobras.com.br Claudio de Carvalho Rocha is a Process Engineer in the Design, Construction and Assembly Department of Duque de Caxias refinery. He holds a BSc in chemical engineering from Federal University of Rio de Janeiro and a 6 Revamps

7 Specialisation Course in Petroleum Refining from Petrobras Corporate University. Henry Giron Bonsaver is an Equipment Engineer in the Design, Construction and Assembly Department of Duque de Caxias refinery. He holds a degree in mechanical engineering. henry.giron@petrobras.com.br Antonio Carlos de Avelar Cortines is a Sales Manager with Sulzer Chemtech Brazil in São Paulo. He holds a BSc in metallurgical engineering from Fluminense Federal University and a MBA in corporate management from Getúlio Vargas Foundation. antonio.cortines@sulzer.com Revamps