PTQ Q1 2023 Issue

reactor using a conventional basket filter/cartridge filter was first tested using a dirty (low-quality) kerosene feed stream. To minimise plugging problems and pressure drop in the reactor, the unit was run at only one-fifth of its design capacity for fresh (but dirty) feed with a four-fifths clean recycle stream. Even so, the feed stream was so dirty that front-end cartridge filters needed to be replaced in just a few days of operation. However, when the filtration system was enhanced with the Universal Filter/conventional cartridge filter, it demon - strated an even higher ability for removing smaller solid particles (especially nm-size particles) than the larger ones, as shown in Tables 2 and 3 and the following discussion on basic principles. Therefore, the combined Universal Filter (advantageously to remove smaller particles in nanometer sizes) and con - ventional cartridge filter (for removing larger particles in micron sizes) are complementary to each other in filtration. Their combination may create a much more enhanced per- formance to protect downstream reactors and conceivably provides near-total solid particles removal. A schematic diagram with the combined units is shown in Figure 4 . To solve plugging problems and improve unit operations, the basket filter in front of the cartridge filter was replaced by a Universal Filter, significantly reducing clean recycle feed requirements. However, total throughput was at 50% reactor capacity for a few months due to reduced recycle and a lack of fresh feed supply. Stable operation was achieved after three months at around 75% of total design capacity, with significantly reduced clean recycle in the feed stream. With more fresh feed supply available after eight months of operation, the unit could reach full capacity (around 30,000 b/d) with min - imal or no clean recycle stream. Monthly operating profit was approximately US$7MM (or annual profit of more than US$80MM) with full capacity and no clean recycle. Universal Filter also keeps conventional filter car - tridges cleaner, avoiding frequent replacements. Since the Universal Filter was installed in front of the existing car- tridge filter, filter cartridge replacement was reduced from every few days to every few months, with annual cost sav - ings of approximately US$225,000 just for the cartridge costs. This validates the enhanced performance of combin - ing the Universal Filter and conventional filter cartridge in commercial HDS unit operations. Treating straight-run naphtha to CCR reformer’s HDS unit To reduce plugging problems at a 30,000 b/d straight-run naphtha HDS reactor (for a CCR reformer), a Universal Filter was installed in the feed stream to the HDS reactor. The HDS reactor run time increased from 3-6 months to two years for uninterrupted continuous operation. This result generated a significant profit for the CCR opera - tion because each HDS reactor turnaround took 10 days, causing a US$2,860,000 loss in CCR production. The HDS reactor turnaround has been reduced from four (every six months) to one in two years, and the total annual cost savings were estimated to be US$4,290,000 using the Universal Filter.

Comparing PSD between filter slurry and inlet liquid 30 minutes after shaking

Solid particles in collected

Solid particles in inlet liquid

filter slurry

(shaking/settling)

(Curve plotted in Figure 3) Particle size (nm) (%)

(Curve plotted in Figure 2)

Particle size (nm)

(%)

6.6 7.6 8.7 10

0.308 0.448 0.920 1.705 3.183 5.668 9.204

6.6 7.6 8.7 10

0.385 1.205 2.858 5.233 8.184

11.4 13.1

11.4 13.1

11.619 16.497 21.690 18.983

15

15

17.1 19.6 22.5 25.7 29.5 33.8 38.7 6000

13.926 18.538 19.496 14.989

17.1 19.6 22.5 25.7 29.5

9.201 2.495 0.370 0.066 0.150 1.067

8.114 2.790 0.517 0.195

4472.1 5122.3

6000

* 7-30 nm particles in the inlet liquid matched well with those removed with filter slurry. ** There were no detectable “nm” size particles in the outlet liquid from the filter.

Table 4

liquid matched well with 7-30 nm particles in the sample of solid slurry removed by the filter. Both results imply that 7-30 nm solid particles in the inlet stream were effectively removed by the Universal Filter. Treating a kerosene HDS reactor feed The Universal Filter was very effective in solving extremely difficult filtration when processing a 30,000 b/d dirty kero - sene feed through an HDS unit. To demonstrate its effec - tiveness, filtration of a kerosene feed stream to an HDS

Liquid feed distributors

Universal lter

Catalyst bed

Cartridge lter

Fixed bed reactor

Furnace or heater

Figure 4 Combination of Universal Filter and conventional cartridge filter for treating fixed-bed reactor feed

90

PTQ Q1 2023

www.digitalrefining.com

Powered by