Injection Quill Design

[kujah]

hello

what will happen if i use a chemical injection quill for capacities that are significantly lower than its design, say 20% of rated capacity? the system shall involve injecting liquid water at around 3000 kPag in a hydrocarbon vapor stream of around 2800 kPag. distribution of the water is important as excessive "pooling" could result to corrosion of downstream equipment.

what is needed for me to size a quill of smaller capacity?

thanks and best regards,

 

[sshep]

In my experience every injection quill is pretty much a custom design. The objective is usually to design the simpliest equipment to get the job done. Experience and good practices are probably more important than any strict quantitative method of design. The most important consideration is the objective which you are attempting to achieve downstream. I would be interested in what you case looks like.

From a mechanical standpoint the quill must be sufficiently strong, removable (at least during shutdown), of chemically resistant material, and design checked for vibration (as per thermowell). The tip design objective is often just to get the fluid into the centerline where it can disperse by turbulance. The use of spray nozzles and the like is avoided where pluggage is possble due to difficulty of cleaning online. I often like to put the injected fluid momentum inline with the main flow stream (i.e. tip bent in direction of flow) when such a design is practical for insertion. The installation can be by flanged spool piece, or through a nozzle on the pipe (like a thermowell, or removable via valved packing gland). If you have one of these simple open tip designs and it has worked well then stick with that design. When sizing the quill inside diameter split your 200kPad between the quill and the flow control valve to get a reasonable velocity at the tip. This will set your quill diameter, but don't go smaller than 1/4" dia.

The location (vertical up, down, horz) is an important consideration but presummably you will put your new quill into the same location. Also you probably have no control over the vapor velocity at the point of injection. The process considerations for your case are potentially numberous- i.e. T, P, densities, velocity, the phases involved, flashing,... For the vapor service you describe it is not clear if you are trying to disperse free water drops into a vapor stream or if the water is expected to vaporize. In the latter case if you are fortunate with a good vapor velocity and location you can probably feel comfortable even using the existing quill at 20% of rated capacity. If you are trying to disperse a mist of free water for some downstream purpose then your design is a more interesting problem. Please let us know some more detail.

This is my experience anyway, sshep

 

[kujah]

thanks sshep!

for further info, the purpose of the water injection is to clean a downstream equipment on-line. particularly, this involves the washing of air-fin coolers that exhibit ammonium bisulfide fouling.

this is the result of increasing the operating severity of a diesel hydrotreater for LSD production.

yes, there will be some flashing of the water at the injection point. but should still be enough to wash the cooler off the salt. excessive water is not advisable as a downstream water boot is not designed to accomodate this increased amount.

can you elaborate when you say "to split 200kPad between the quill and flow control valve"?

thanks and best regards,

 

[sshep]

Hey Kujah,

My comment about the pressure drop was only because you specified the water pressure to be 200kpad above the pipe pressure. I envisioned this supply to be from a header and since you had a flowrate in mind, then it seemed logical you would have some sort of control- a typical on the cheap installation might be nonreturn valve, a hand control valve, rotometer, and injection device. If you have an injection pump to control capacity then my vision may not apply.

You are dispersing water droplets into vapor for cleaning purposes. The water you put in will have to come out somewhere (i.e. drawn off in your boot) and so the goal seems to be to get the best wash you can with a small amount of water. This means we want to get small enough water droplets to be entrained to all the fouled surfaces.

The energy for dispersing the water can come from the 200kpad available (i.e. a spray device at tip), or from the vapor stream. Although my experience is that many injection points are not commercial designs, the applications are too numberous for such equipment not to exist. A quick internet check for chemical injection quills will turn-up many sites. I found this site for a device which allows connection of quills, open tip, nozzles, and strainer/heads. Their quill design terminates with a 45deg cut, and a slot:

http://www.alspi.com/injecthp.htm

Other designs are flat tips or rely on only on the normal 45deg bevel:
(

http://www.koflo.com/injection.html)

In all these designs the dispersion from a quill tip uses energy from the vapor stream via turbulance at the tip. 15ft/sec is described as excellent. No quantitative guidence is given for quill inside diameter, but as the liquid is dispersed by flowing into turbulance at the tip, we expect a low exit velocity to be fine.

If you have a quill design as described above with a high velocity vapor, then I continue to think you can probably use your existing system as is- i.e. replacing will likely result in little improvement. Have you calculated the velocity at the point of injection yet? If your velocity is low, some sort of spray tip may be required.

Best of luck, sshep

also search quill keword in threads:
thread301-75398, thread135-90634; thread378-9202

 

[kujah]

i've estimated that the vapor flow at the injection point os highly turbulent. you are right, there is probably little or no effect by resizing the existing quill with a smaller one since it is being injected in turbulence.

i'll give some feedback after the implementation during after the forthcoming unit turnaround.

thanks!

 

[unclesyd]

Two caveats based on failure investigations caused from liquids injection into hot gas lines. Both failures were thermal fatigue resulting from liquid contacting the pipe walls.

The cause of one failure was obvious, the injection quill was orientated where the liquid stream was directed at the pipe wall instead of being aligned axially. The cracking had a spiral pattern even though the flow was highly turbulent at this point and 40 or so pipe diameters from a bend.
The second failure was caused by injecting a partial flow stream from a quill with much higher capacity. The quill was pointed in the direction of the flow (downstream) and the liquid stream formed a "rope" that contacted the wall at about a 1 1/2 pipe diameters. The cracking occurred in a spiral band for several pipe diameters before the liquid was vaporizer.

I would be a little leery if I couldn’t atomize the water.

 

[sshep]

Unclesyd, what sort of redesign was used to address these failures? Your story does bring up the possibility of looking for tell-tail water marks during the outage as an indication of how good the liquid may be getting to the problem areas.

I still feel that if the system will have at least no more problems at the turned-down flow than it had at the design flow. If a spray head is felt to be required then a strainer in the injection line can mitigate clogging problems.

I don't think there is any common quill design (i.e. channel diameter and tip detail) that can do the job reliably if the bulk fluid energy for dispersion is inadequate- this is just the way quills are designed to work. I do think there may be a niche for clog free injection designs that would use the energy of the injected material to atomize. To do this I envision some sort of impingement surface integral to the tip which would brake-up the water by impact rather than by forcing it through small jets as in a spray head. This is the tact taken in cooling tower top basin distributors, deluge systems, ect, where distribution and clog free designs are important.

cheers, sshep

 

[unclesyd]

The first case mentioned was a flange mounted injector with a flow control (pneumatic actuated) also mounted on the flange. The flange was altered to where it couldn't be mounted except in one direction with the addition of a locating pin. This orifice is pointed downstream. This process still has a problem in that the gas flow will vary and the temperature will climb several hundred degrees C when there is an emergency shutdown, so this line is checked at each shutdown.

The second incidence with the rope effect was corrected by going to a higher pressure water system and a redesigned quill with the orifice pointed upstream. The orifice was made from Hastalloy C due some erosion problems.

Take a look at this website. We were doing something very similar in another process prior the introduction of these mixers in relatively low flow very hot gas stream to insure good mixing and keeping the liquid off the walls.

http://www.komax.com/Gen-GasliquidContacting.htm

When you get a look at the piping look for a very pronounced grainy appearance and if there is cracking it should be normal to the flow patterns if any. The surface patterns are quite easy to discern but the cracking can be quite tight and hard to see. It has been my experience that you don’t want to see flow patterns downstream of a contact point.

 

[roker]

we are using in a hi pressure gasoil hds water injection
with a 5%w/w on the feed and full cone spray nozzle to each air-fin cooler.

regards,
roker