PTQ Q1 2024 Issue

In one case, a new preflash drum was sized per Figure 7, but upon start-up the drum experienced severe entrain- ment. Following the Fluor FEED design, the detailed design was turned over to a local contractor who installed a baffle in front of the crude inlet nozzle that was open at the top and bottom and closed on the sides, causing an impinge- ment jet onto the liquid level and shooting crude upwards. In another case (again, not a Fluor design), a preflash tower12 experienced frequent episodes – especially with light crudes – in which the naphtha stream leaving the tower turned dark, accompanied by a high tower dP. These episodes originated from crude entrainment and restricted the throughput of the entire crude unit. The root cause was diagnosed to be excessive feed velocities with poor feed entry design. The entrainment problem was eliminated, and the naphtha make could be raised by more than 20% by a unique feed entry design by Fluor for the very high-velocity feed. This was achieved by removing two segments of the top shed deck and by a specially designed impingement baffle to break the incoming momentum. Key takeaway: It is imperative to correctly specify the feed inlet design and/or thoroughly check inlet designs by others. Trays or packing in preflash towers Most preflash towers recover naphtha as the only prod- uct. Some also draw an additional kero stream as a side draw (usually a small flow) a few trays above the feed. When drawing kero, a pumparound that preheats crude is sometimes used right above the kero draw. The kero may be sent to the stripper or the kero-naphtha section of the atmospheric tower. Preflash towers have trays and reflux to fractionate overhead product from the kero side draw or bottom product streams. Random or large structured pack- ings have also been used with success, although trays are preferred. It is important that the trays, at least those right above the feed, are fouling-resistant, as crude carryover, entrainment, and foamovers (especially during upsets and power cuts) tend to entrain foulants into the trayed sections. Entrained waxes and resins can bond together and crystallise out on the lower trays. Fouling of the lower trays has been experi- enced even when VTC has been installed in the flash zone (see Figure 9 ). Cases have been observed where moving valve trays and fixed mini-valve trays right above the flash zone plugged. Salting out and corrosion near the top have also been experienced and, again, favour using fouling-re- sistant trays such as Sulzer’s SVG fixed valves with 0.5in opening, LVG fixed valve trays, or Koch-Glitsch’s ProValves. The trays near the top should also have corrosion-resistant metallurgy. The economics dictates using a reflux ratio just enough to achieve the naphtha product specification (typically the 95% ASTM D86 temperature). Adding reflux beyond this recycles some of the naphtha back into the crude, which in turn counters the benefits of the tower. There are typically 15-20 trays used above the flash zone, giving about 8-10 theoretical stages. Minimising the reflux often generates spray regime con- ditions (low liquid rates and high vapour rates), with tray

Figure 9 Plugged mini-fixed valves in wash section of preflash tower

drying and/or break of downcomer seal, accompanied by fouling, plugging, and instability. This is especially severe when drawing kero, as the ‘wash’ section below the draw is particularly liquid-lean with weir loads often below 1 gpm/in. Due to the temperature gradient, the driest spot is just above the flash zone, and here the potential for plugging and seal loss is maximised. A good spray regime design (avoiding excessive open area, keeping down- comer clearances low (1-1¼in), and judiciously using picket-fence weirs) is essential. Liquid reflux and the wash below the kero draw (if it exists) need to be on flow-con- trolled pumpback to avoid fluctuations that can dry out the trays and break the downcomer seals. Adding vortex tubes at the flash zone (Figures 5 and 6), or at least ensur- ing a good feed distributor, will improve reliability, alle- viate foamovers and entrainment, and may be relatively inexpensive. In many preflash towers, especially those recovering kero as a side draw, designers add trays or packings below the flash zone and use stripping steam to recover more naph- tha or kero from the crude. We prefer not to do this due to foaming concerns (see next paragraph), but some clients insist. If stripping steam is used, its ideal rate is 10 lb/bbl, but many units use less than that. There are also reboiled preflash towers using fired heaters. A study by Sloley² found that using a stripping section saves energy. Regular trays (fouling-resistant) or packings in the strip- ping sections work fine as long as severe foaming is not experienced, which is probably 50% of the time. However, trays or packings below the feed can be disastrous if foam- ing is experienced, which is the other 50%. There was one case in which it was necessary to remove the stripping trays to stop foaming in the stripping section trays from severely bottlenecking the entire crude unit. Trays or packings are, therefore, not recommended for the stripping section. Shed decks and grid packings have shown good performance in the stripping section of preflash towers. With their large open area, they can be designed to handle foam. Correct design of the shed decks and grid packing is essential in this service.

PTQ Q1 2024