Revamps 2024 Issue

Tray 15

Flooding

Draw-o

46’

31 ± 3% 28 ± 3%

Tray 14

FloodingFlooding

44’

38 ± 4%

36 ± 4%

Tray 13

FloodingFlooding

70 ± 4%

42’

Feed inlet

41 ± 4%

Tray 12

FloodingFlooding

40’

57 ± 4%

34 ± 4%

Ring

Tray 11

FloodingFlooding FloodingFlooding

38’

Tray 10

36’

Inlets

53 ± 6%

42 ± 4% 39 ± 6%

36 ± 4%

Draw-o

Tray 9

34’

40 ± 4% 39 ± 4%

37 ± 4%

26 ± 4%

Tray 8

32’

54 ± 4%

35 ± 4% 29 ± 4%

16 ± 4%

Tray 7

Platform

30’

39 ± 17%

31 ± 4%

25 ± 4%

22 ± 4%

Tray 6

28’

Overash draw-o

Blue pen: flooded scan, northeast (NE) chord Red pen: flooded scan, southwest (SW) chord

Gold pen: ‘ normal operation’, NE chord Green pen: ‘normal operation’, SW chord

Figure 2 Active area gamma scans under flooded and ‘normal operation’ conditions

Initial theory: Blockage of tray 10 downcomer The initial theory was that a side panel in the downcomer from tray 10 was dislodged, partially blocking the over- flow of diesel out of the diesel draw box, possibly allowing vapour entry from the active area into the draw box. Vapour entering the downcomer would impede liquid flow into the draw box from the tray above. Tray layout (see Figure 4 ) shows the plan and elevation of trays 6-9. The trays were 10.5ft ID, two-pass sieve trays with ½in holes, a hole area of 12.4% of the active area, with multi-chordal downcomers. The centre and side down - comers were straight, approximately 8.5in and 7in wide, respectively. The 2.5in outlet weirs were picketed, with the top of the pickets 14.5in above the tray floor. The pickets occupied about 86% of the weir lengths. Downcomer clearances were 2.5in. There were 3in inlet weirs located 2.88in away from the downcomer walls. The floor of the centre sump below tray 9 was sloped,

with a depth of 8in at the end opposite the draw nozzle, gradually increasing to 16.5in at the nozzle end, with the 6in draw nozzle flush with the bottom of the sump. Beneath the deep end of the sump, the picket-fence weirs were shortened to permit a 2-4in gap between the sump and the top of the picket fence weirs. Evaluating downcomer blockage theory The downcomer blockage theory readily explained the initiation of the flood on tray 10, as seen by the gamma scans and the erratic behaviour of the flood. Either a liquid build-up step followed by a dump step, or vapour entering the downcomer, will explain the sudden and unpredictable loss of liquid flow at the diesel draw and below. Arguing against this theory is the low velocity, about 0.1 ft/s, in the tray 10 downcomer (including product, reflux, and pumparound), compared to typical maximum design velocities of 0.4-0.5 ft/s1 for non-foaming systems. For the velocity to exceed this maximum, the panel blocking

Revamps 2024