• The diesel draw on tray 9 was clear from mechanical obstruction. • There was relatively less fouling on trays 8 and 7 com- pared to trays 9 and 10. • Tray 6 had a light layer of debris/foulant on the panels and was much cleaner than trays 10 and 9. • No apparent physical or mechanical damage was noted for trays 6-10. • Trays 11 and up were clean and no sign of mechanical damage or fouling. Nature of the foulants Lab analysis showed ~8 wt% phosphorous, as well as excessive amount of FeS (~70-80%) in the foulant deposit. It is important to note that chemical cleaning (RTI) was used before opening the tower, which may have washed off some of the phosphorous. No analysis test was conducted for drilling mud because the X-ray deflection analysis showed small amounts of some parameters common for drilling mud. However, there was no bentonite, barite or hematite, so this was not indic- ative of the presence of a water-based drilling mud. Most experiences with phosphorous fouling have been in the jet pumparound or near the jet draw, but there have also been some phosphorous-plugging experiences in the jet-die- sel section. The phosphorous compounds usually deposit at colder temperatures than those at the diesel draw, typically those near the jet draw. Also, deposits from sections fouled by phosphorous usually contain 10-50% phosphorous. The 8% phosphorous in the deposits of trays 9-10, as well as the observation that trays 11 and up (including the trays in the jet section, even trays 11 and 12 in the diesel pumparound) were clean, argue against phosphorous fouling. The formation of FeS in a hot flowing oil is believed to involve two successive processes: the corrosion of iron by sulphur compounds and the decomposition of iron salts, for example, of naphthenic acids.6 The corrosion of iron by organic acids occurs on the surfaces of equipment. High temperatures favour FeS formation reaction, with temper- atures at the diesel draw (around 580ºF) well within the range. Most importantly, in the previous turnaround, trays 8-10 in the tower were replaced by in-kind, but the tray material was changed from 410 SS to carbon steel, which is far more prone to corrosion and FeS formation. So, while the source of the plugging could be either FeS or phosphorous, or both, it is likely that FeS was a major actor. Solution Trays 8-10 were replaced by three new fixed fouling-re- sistant valve trays with 0.375in openings. Upon return to service, the two temperature indicators in the vapour space between trays 8 and 9 read the same temperatures again, indicating that the maldistribution issue has been resolved. Stable operation at maximum crude and diesel pumparound rates with low dP and no flooding has been restored. Takeaway There are several invaluable lessons taught by the trouble- shooting investigation:
needs to flow into panel 9A to satisfy the downcomer back-up equation in the centre downcomer from tray 10. This is analogous to the previously described valve float pop-out (based on reference 5). Due to the popping of the valve floats (the A side), the B side had a larger pressure drop, so the downcomer back-up equation directed much more liquid, 67.3%, to the A side to balance the pressure drop difference. A similar mechanism can be postulated for plugging or damage to one or more of trays 6-10. The only premise is that the plugging is more severe on one side of that tray than the other. Due to the low liquid loads, the downcomer back-up equation will direct large changes in the liquid split to counter even a relatively small vapour imbalance. What will vary, depending on the plugged or damaged tray(s), is the direction and magnitude to which the liquid will be preferentially directed. The scenario may be further complicated by possible downcomer seal losses on the trays that approach dryness. The area under the centre downcomers is 57% of the hole area. If it can find the path, the vapour has an incentive to attempt rising up the downcomers. The inlet weir’s ability to protect against seal loss depends on how much liquid arrives at the inlet weir sumps compared to how much leaks out. For a tray nearing dryness, little liquid arrives into the inlet weir sumps, and the downcomer seal is likely to be lost. This shifts vapour distribution and, in turn, the liquid distribution. Further, our sealing calculation shows that once the seal is broken, the velocity of vapour in the downcomer is likely to exceed the system limit, preventing liquid descent. The liquid will pile up above the downcomer, and when enough builds up, it will eventually dump. This explains the erratic nature of the flood experienced in the tower. Plugging of the active area parabolically raises the dry pressure drop, almost always resulting in vapour-sensitive downcomer back-up flood. The liquid sensitivity came from tray 10. As this tray handles the pumparound liquid, as well as the diesel product flow, it has several-fold higher liquid loads than the low liquid load trays below. The liquid load affects the pressure drop of tray 10 and, therefore, the flood. In summary, the plugging theory was in line with the maldistribution observed in the gamma scans, as well as the erratic nature of the flood and the vapour and liquid sensitivity of the flood. Its ability to explain all observations put this theory in the lead. Tower inspection Based on the plugging theory, the tower was shut down for cleaning when the opportunity came. Figure 5 shows the findings: • The active panels of trays 9 and 10 were heavily fouled, with the holes barely visible. There was approximately 1/8in of tightly adhered sandy scale on all active surfaces of tray 10 that could easily be removed with a metal scraper. On the underside of Tray 10, the scale was 3/16in or ¼in thick. The top of Tray 9 was absolutely caked with foulant approximately ½in thick. It got worse near the tower wall. The tower wall had no fouling and was clean. The down- comer box surfaces had some foulant, but the downcomers were fairly clean
20
Revamps 2024
www.digitalrefining.com
Powered by FlippingBook