(21f) Good Distributor Design for High-Velocity Feed Debottlenecks a Crude Preflash Tower

Bazan’s Crude Unit No. 3 (designed by others) was bottlenecked by the preflash tower. The tower experienced frequent episodes, especially with light crudes (high naphtha percent), in which the naphtha stream leaving this tower turned dark, accompanied by a high tower dP. The episodes most likely originated from crude entrainment. These episodes restricted the throughput of the entire crude unit. The tower contained 12 trays above the feed and 6 baffle trays beneath the feed.

From the lowest downcomer, 4" pipe drains the liquid into the top baffle tray. At two recent outages, the 4" pipe was found broken and falling down, with a hole in the seal pan from which it originated. As the pipe was located right in front of a high-velocity feed entry, it was likely that feed impingement at high velocities generated vibration and caused the 4” pipe to collapse. It was believed that due to the pipe breakage, the high feed vapor velocities picked up the descending liquid or rose up the bottom downcomer, causing the naphtha product to become dark. The initial solution sought was reorientation of the 4” pipe away from the feed.

Suspecting that there may be other issues, Bazan requested Fluor to assist with the problem analysis. KochGlitsch were later added to the team to consult on hardware changes. Our analysis determined that the breakage of 4” pipe was insufficient to explain the entrainment and dark naphtha. We determined that the root cause was excessive feed velocities, with feed entry too close to the bottom tray.

To solve the problem, it was necessary to break the incoming momentum and to divert the feed to a lower point in the tower. This was achieved by a unique feed entry design for the very high-velocity feed, by removing one of the baffle trays to permit feed entry at a lower point, and by adding a specially-designed impingement baffle to break the incoming momentum.

When the tower returned to service, the dark naphtha episodes completely disappeared. The unit throughput could be raised by more than 30% without any adverse effects on operation. The entrainment problem has been eliminated.

Our paper details the problem, analysis, and solution.