Revamps 2026 Issue

loss. Third, the bottoms liquid level control (flow) becomes erratic. None of these was occurring, and that theory was ruled out as the leading problem. After eliminating the possibility of the lowest tray flood - ing, the author concentrated on the next most likely zone to flood: the feed zone. The feed zone design was reviewed many times and showed no significant issues. Tower feed quality (such as percentage of vaporisation) was dou - ble-checked and did not reveal any problems. Operator rounds Having operators involved during a troubleshooting case is always beneficial and necessary. During one of the trouble - shooting visits, the author noticed a distinct popping sound near the base of the tower. This distinct sound was simi - lar to that produced when wet steam is injected into a hot product side stripper. It was a subtle pop. Then, eight min - utes later, another pop. Then, eight minutes later, another pop. The operator was asked whether he had heard this sound before the revamp, and he did not remember hear - ing it. Searching for the source of the sound led the author to the reboiler. After a while, the author concluded that the sound was not coming from the reboiler. However, what was it? The only consumer of steam was the debutaniser reboiler within 100ft. The sound was not due to a steam issue; it was tracked to the feed line (around the point where the combined feed line turned vertical), which was very close to the reboiler. Figure 3 shows a simple PFD of the debat - aniser feed system. One of the author’s mentors always stressed the need to touch items in the unit. He was always gathering infor - mation to see if they were hot or cold, whether they were pulsing, or whether they were vibrating. In this case, the feed temperature was <250°F and the insulation could be touched without any safety issues. The team could physically feel the line jumping/moving around; the movement aligned with the ‘popping’ sounds they heard. The next step was to find out what was causing the sounds and oscillations in the line. A low-point bleeder was located about 5ft from the origin of the vertical run for the feed pipe, alongside the debutaniser tower. It provided a critical piece of information. A local pressure gauge was installed on the bleeder at grade. During the survey, there was ongoing consultation with the board to make sure that the tower top pressure was constant. During the evaluation, the tower top pres - sure did not move more than 0.50 psi high-to-low. The pressure gauge on the feed showed a rhythmic cycling of 72 psig to 81 psig. After recording a few points, the author recognised that the cycles were within 15 seconds of the peak-to-peak events on the stripping section pressure drop excursions. There was now concrete evidence that the feed circuit pressure was causing a stripping section pressure drop excursion and flooding. Figure 4 s hows the pressure read - ings at the beginning and mid-cycle. Immediately, the team focused on the feed enthalpy and

Feed pressure after slug relieved and tray flooded

Feed pressure when slug forming and tray unflooding

Figure 4 Feed piping pressure at grade during a cycle

vaporisation. Was the feed temperature cycling resulting in problems? In Figure 3, it can be seen that the feed comes from two separate deisobutanisers (DIB#1 and DIB#2). DIB#1 has a feed preheater, while DIB #2 does not. Effluent from the DIB#1 feed preheater did not indicate changes in heat balance that were sufficient to cause the downstream issues. Since DIB#1 was direct from the tower, it was also eliminated since the bottoms temperature (i.e., composi - tion) was very steady. The author knew that since the temperatures of the upstream towers were not changing and the composition of the feed was not changing, the only variable left to set the enthalpy was the feed pressure. What would cause the pressure to change radically from the feed nozzle to grade? The answer was identified after mapping the flow regime. Vertical flow regime maps are not readily understood (upflow versus downflow) and are highly variable due to the effects of static head that accompany vertical flow Analysing feed line flow regimes Flow regimes are readily used when designing highly flashing streams, including reboiler returns. Experience from recent jobs indicates that this evaluation is not fre - quently done on feed streams. Based on the results of this case, it is recommended that all distillation tower feed pip - ing be reviewed for flow regimes. Horizontal flow regimes are well understood, and the flow regime maps are well defined.2 , 3 , 4 Vertical flow regime maps5 , 6 are not readily understood (upflow versus downflow) and are highly variable due to the effects of static head that accompany vertical flow. This is especially true for services that experience large changes in vaporisation, where a discrete element analysis is required. There are also proprietary methods that are used to analyse flow regimes.

Revamps 2026