The acid absorber made weak acid; our scrubbers were acting as absorbers. For the next few exhausting hours we stumbled through an ad hoc checklist. Eventually, we determined a ½-in. patch of filter was plugging the absorber's spray nozzle. It probably escaped our inspection of the vessel bottoms. Packed towers demand a feedstock free of debris!
First, consider the basics. Velocity controls gas absorption but is hard to manipulate. Blower capabilities limit the influence of gas velocity, leaving only liquid velocity to adjust. Another factor is liquid distribution. The selection of the nozzle(s) and pump(s) affect this distribution. The effectiveness of the spray nozzle(s) and liquid re-distributor(s) in spreading the liquid over the packing impact the distribution. Problems will become apparent during commissioning.
Flush the tower with the spray nozzle(s) removed. Before starting the pump(s), inspect the sump, and disassemble and inspect the spray nozzle(s). Install a coarse basket strainer in the suction to avoid pump impeller damage from broken packing; keep the strainer in for several weeks. Size the mesh to less than half the spray nozzle orifice. Typically, the flow goes first through a coarse mesh then a fine mesh; 2 mesh for coarse and 12 mesh for fine is common. Filter area should be at least 2.5 times greater than the pipe flow area. I recommend a differential pressure gauge and duplex filters. Also, strain feedstock to prevent spray nozzle fouling.
After flushing for half a day, re-inspect the sump, and install and set the spray nozzle(s). The nozzle-packing distance is crucial. You likely must refill the packing level because it shrinks, typically 1–3 feet, after use. Use the time for flushing to set the levels, test the tank overflow capacity, and adjust the feed flow loop. You'll observe a sudden drop in the sump level as the packing fills; 8–15% is typical. Take care in setting the level trips. On one job the level was at 50% and tripped the entire plant off at 45% when the pumps started. It's important to determine the cavitation point for the pump during this period.
The next step is crucial: test the feed flow to ensure the scrubber or absorber can operate over its full range. In one startup we were forced to order a replacement impeller because the pump couldn't supply enough flow.
Cooling often is used to improve the capture efficiency of a scrubber where heat of absorption is exothermic. If the feedstock is cooled, carefully check the temperature loop. It's best to test the loop under high stress conditions; the heat of summer works best.
Properly setting the spray nozzle(s) is vital to maximizing the effectiveness of the liquid transfer coefficient. A viewport helps during startup but an open manway also works although the nozzle(s) will behave slightly differently under vacuum. Adjust the spray nozzle(s) manually so spray touches the wall without formation of either large droplets or atomized spray. Record the pressure versus feedstock flow. Separate gauges and valves for multiple nozzles are best.
A common problem with multiple spray nozzles is interaction. Collision of sprays from different nozzles can create large droplets. This can cause serious distribution problems and efficiency losses because of dry spots, especially in the center of the tower where the gas velocity is strongest — that's where you want the water to go.
In crucial applications, consider alternate spray nozzles to avoid poor distribution at the low and high range of equipment operation. It may be desirable to adjust the spray pressure according to the tower throughput. Also, be aware of variations in liquid density, viscosity and surface tension; the process liquid may differ in these properties from the water used during the test. Your vendor can run trials to evaluate droplet formation. Another concern is pump duty; some pumps serve many duties and this could affect spray nozzle pressure.
And, finally, there's gas distribution. Ideally, gas should enter the bottom of the scrubber and be evenly distributed — although usually it isn't. At best a distributor can spread out the gas. So, perform a visual test by running the tower for several minutes, so the spray nozzle(s) can wet the packing. Then, allowing the gas to flow, turn off the water flow for a time and touch the packing to see if it's still wet in the middle. If the packing is moist in the center and wet at the walls, the gas distributor is working well.
DIRK WILLARD is a Chemical Processing contributing editor. You can e-mail him at firstname.lastname@example.org.