Lube vacuum column revamp
Primary goals of a lube vacuum column revamp were to increase lube product fractionation and minimise HVGO product contaminants
Kevin Basham, Marathon Petroleum Company
Edward Hartman, Process Consulting Services
Viewed : 6156
The lube column at Marathon Petroleum Company’s (MPC) Cattlesburg, Kentucky, USA, refinery was revamped in 2006 (Figure 1) to improve lube product fractionation, reduce the heavy vacuum gas oil (HVGO) product’s micro-carbon residue (MCR) and metals, and improve the vacuum bottoms specifications for the production of asphalt. Prior to the revamp, the HVGO product was black and cylinder stock yield was excessive. Cylinder stock was fed to another vacuum unit so that the lube column bottom product could meet the asphalt product specifications, as well as recover a portion of the HVGO boiling range material in the cylinder stock. However, reprocessing this stream consumed some of the other unit’s capacity and increased its heater firing. All the project’s justification benefits were met following startup. Moreover, a lower column operating pressure and improved stripping efficiency led to a 4.0 Mbpd higher crude charge rate due to the lower heater’s cracked gas production freeing up some compressor capacity to process more crude.
Process flow scheme
The previously mentioned Figure 1 shows the simplified process flow scheme for the atmospheric and vacuum column prior to the revamp. Off-gas from the atmospheric crude and lube vacuum columns was handled with a common compressor. Atmospheric column heavy gas oil (HGO) product was routed to the vacuum column to recover some of the lube-quality material. The vacuum column produced light vacuum gas oil (LVGO), side stream (SS) #1, side stream (SS) #2, HVGO, cylinder stock and asphalt products. It was necessary to yield cylinder stock to meet the asphalt specifications on the bottom product. LVGO and HVGO were routed to cat feed hydrotreating. SS#1 and SS#2 were lube-quality base stocks targeted for further processing.
The vacuum unit is a wet design with coil and residue stripping steam, plus a precondenser prior to the first-stage ejector. The top column operating pressure on a wet column is set by the ejector system load, which changes with the seasonal cooling water temperature. During winter, the first-stage ejector load is lower due to the reduced cooling water temperature; hence, the column’s top pressure is as low as 35 mmHg. This pressure increases to 55 mmHg in the summer when the cooling water temperature is higher. The vacuum column internals consisted of three packed beds and 12 trays to remove heat and fractionate the feed into five side-cut products (Figure 2). The tray pressure drop was approximately 3.5 mmHg per tray, with the packing contributing only a small amount, which resulted in a flash zone pressure of 95–105 mmHg, depending on the ambient temperature. The HVGO product was black from entrained vacuum residue. Furthermore, the trays were prone to damage and leaks, resulting in poor fractionation, excessive cylinder stock yield and difficulty in meeting asphalt specifications.
Typical unit charge was a blend of Middle Eastern crudes and was limited by off-gas compressor capacity. During the summer, the atmospheric column’s overhead receiver temperature increased, producing more off-gas. At the same time, the vacuum ejector’s off-gas rate was higher because the heater’s outlet temperature was at maximum. The crude rate had to be reduced once the off-gas compressor reached its maximum capacity.
In 2005, a study was completed to determine the cost/benefit of revamping the unit. Major economic incentives were improving the lube products fractionation, minimising the HVGO product MCR and metals contaminants, producing on-specification asphalt and improving internals reliability.
Lube products fractionation
Fractionation between lube products depends on reflux and the number of theoretical stages. Increasing either one improves fractionation. This requires high-efficiency mass-transfer internals and minimum column flash zone pressure. Prior to the revamp, the column pressure drop was high, stripping efficiency was low, and fractionation between lube cuts and HVGO was poor. Before the mid-1980s, most lube vacuum columns were designed with bubblecap trays and occasionally valve trays. Yet, trays have the disadvantages of high pressure drop and inherently poor efficiency in lube column service. In 1984, the first large-diameter lube vacuum column was revamped from bubblecap trays to structured packing, with several others modified since then. Some were successful, while others were not due to poor-quality liquid distributors.
Trays have inherently low efficiency in lube columns because the liquid rate is low and the tray weir length is large (because of large diameters). Conversely, structured packing’s inherent efficiency is good at a low liquid rate, assuming a high-quality liquid distributor is used. A structured packing bed’s efficiency is largely controlled by the liquid distributor’s performance. Since column diameters are large and liquid rates are generally low (~1 gpm/ft2 of column area), distributing the liquid uniformly is a challenge.
MPC’s lube vacuum column flash zone pressure was high because it had 12 trays above the flash zone. Trays produce a high pressure drop per theoretical stage because each tray generates about 3.5 mmHg pressure drop and three trays are needed to achieve a theoretical stage. Thus, each theoretical stage creates approximately 10 mmHg pressure drop. Conversely, structured packing produces only 1.5 mmHg per theoretical stage.
Minimising HVGO product contaminants
HVGO product contaminants consist of volatile MCR and metals in the product boiling range as well as entrained vacuum residue containing high amounts of MCR and metals. There will always be some volatile contaminants present, irrespective of the column design, but the amount of volatile contaminants depends on the efficiency of the wash zone and residue stripping. Surprisingly, many vacuum columns produce black HVGO product because of poorly designed column flash zone, wash zone and stripping section internals. In this case study, because the unit processed low metal crudes and operated at low cutpoints, eliminating entrainment would reduce the HVGO product’s metals and MCR to low levels.
Add your rating:
Current Rating: 2