Icing up of valve on blow down

[dbday]

Hi,
I have a 50 cubic foot air reciever which is kept at 10 barg.

Occaisionally I need to blow down the vessel quickly and have noticed that ice forms on the valve and sometimes in the pipework after the valve, when I do this. There are restrictions in the pipework after my valve which resist the flow as well as locally increase the flow velocity, and these suffer from icing.

I realise that ice formation does depend on the moisture content in the vessel, but is this the only factor, if the air was "dry" would it still happen ? also as the output of the compressor does not go through a drier is there another way to remove moisture at the output of the compressor before it gets to the vessel to reduce the incidence of ice ?

Any pointers to redude the ice would be appreciated.

 

[Latexman]

This may help you out:

http://en.wikipedia.org/wiki/Joule–Thomson_effect

Good luck,
Latexman

 

[btrueblood]

"I realise that ice formation does depend on the moisture content in the vessel, but is this the only factor, if the air was "dry" would it still happen ? also as the output of the compressor does not go through a drier is there another way to remove moisture at the output of the compressor before it gets to the vessel to reduce the incidence of ice ?"

#1 - yes: no water = no ice. More succinctly, if the dewpoint of the air is less than the temperature reached in blowdown (adiabatic expansion or JT expansion), then no ice.

#2 - removing the water from the output air constitutes drying the air, so "is there another way to remove moisture" would have to be answered "no". There are a variety of ways to dry the air, from refrigerant consensate removal systems (aka cold traps), to dessicant wheels, to simple condensate traps at points of use. The latter are the most useless. You could theoretically heat the tank (being careful not to exceed safe operating pressure/temperature limits of course), raising the total temperature of the air, so that the subsequent expansion cooling during blowdown would be above the dewpoint. In practice, doing so would vaporize condensate that is likely puddled in your receiver, and raise the dewpoint of the air as well, negating any benefit.

 

[MechEngNCPE]

You could install a refriderated dryer or blower/purge dessicant dryer before the reciever.

 

[dcasto]

When I read the term "on the valve", I assume you mean on the outside of the valve. In any case, the temperature inside the vessel must be below 40F in order for ice to form inside or outside.

 

[btrueblood]

The OP wrote "on" the valve, but also "in" the pipework, which I assumed to mean internally downstream of the valve.

Specifically:

"There are restrictions in the pipework after my valve which resist the flow as well as locally increase the flow velocity, and these suffer from icing. "

If ice on the o.d. surface of pipe or valves is a concern for some reason, the usual corrective action is insulation on the pipe o.d.

 

[dalcazar]

If I understand what you're talking about correctly, there is a very well known and very deadly condition that is similar to what you're talking about in aircraft (and cars) called carburetor icing. humid air drawn into the carb is cooled by the vaporization of the fuel and water condenses and freezes in the carb. The temperature of the outisde air isn't necesarily the deciding factor as much as the dew point and moisture content is, the effect of carb icing is that the engine quits and you crash. The solution is to heat the carburetor and melt the ice preemptively by drawing hot air into the carburetor.

I agree that trying to dry the air as its drawn in is a ridiculously expensive and/or impractical solution

My suggestion would be to either heat the piping with an external heat source, depending on the pipe material you could use IR, an electrical blanket, induction or direct flame; OR you could inject hot air into the blowoff stream with a venturi system

 

[ione]

If air drying is not affordable, go for electric tracing (I won’t go for steam tracing to fix this, not even if steam were available).

Check page 5 in the link below for a possible arrangement.

http://www.thermon.com/catalog/uk_pdf_files/PN50207U.pdf

 

[chicopee]

VW beetles used to be notorious with icing carburator throat on rainy day.
Reduce moisture inside and outside tank; raise air temperature of the room to reduce RH; insulate valve and dicharge piping to eliminate icing on the outside surfaces of valve and discharge piping.

 

[zdas04]

I have to ask "what problems are you having because of the icing?" I've seen pipework that has external ice on it for decades (usually downstream of a pressure regulator) with no ill effects. If it is just ice on the outside of the pipe that melts after the blowdown is complete, get a mop.

If ice is forming inside the pipe and restricting flow (pretty rare in air, less rare in gases that form hydrates) then you might look at getting rid of downstream restrictions that allow the ice to accumulate and moving your isolation valve to closer to the end of the pipe.

David

 

[dbday]

The restriction on the outlet pipe is a silencer, there is an inlet pipe to the silencer which seems to have a plate welded across it and there are small holes drilled in the wall of the pipe to let the air get into the silencer. The silencer manufacturer calls this part a diffuser, it is these small holes that get blocked with ice.

I have asked about drilling out the holes, but am told that this will reduce the back pressure the silencer puts on the system which in turn will increase the mass flow through the unit, which in turn will compromise the warranty on the silencer.

When we have taken it apart (while still iced up) there is ice both on the inside and outside of the pipe leading to the silencer.

 

[moon161]

This phenomenon also contributed to the loss of USS Thresher after a flooding casualty. They started to blow the ballast tanks, but they were soon blocked by ice.

 

[dalcazar]

I was assuming that internal ice was the only problem, I still say: heat the unit somehow, is the alloy of the perforated pipe a magnetic metal that you could heat with induction? that way you don't have to have to run wires in there or rely on inefficient heat transfer from the outside.

 

[AA1344]

This question could be some how related to the icing problem, discussed in this forum: Antoine equation is used to predict the vapor pressure of water at different temperatures.If the partial pressure of water in a gas stream is more than the vapor pressure of pure water at the stream temperature, water droplets will form (definition of water dew point), which can turn into ice or hydrate. To avoid this problem, the gas should be dehydrated to a degree which the partial pressure of water becomes much lower than pure water vapor pressure (at the lowest expected temperature of gas stream). My question is:
When the lowest operating temperature is below water freezing point (water tripple point), how can we predict the pure water vapor pressure? Is Antoine equation still valid?

 

[sailoday28]

dbday (Mechanical)
Is the source of the blowdown your "reciver"?
If so, during blowdown, how is the 10barg maintained and is icing noticed on it (the reciever)?

 

[dbday]

sailoday28 -
The vessel is blown down from 10 BarG to atmospheric pressure, the icing is only noticed on the outlet pipe from the valve between the silencer and the valve.

This bit of pipe is about 2 feet long, interestingly the icing is first noticeable about 8 inches after the valve outlet, there is some ice build up on the outside of the pipe, due I think to condensation which freezes.

There seems to be lots of filters on the market with auto-drains on the bottom, so I guess they cause some of the moisture in the compressor discharge to drop out, even though they don't appear to dry the air I am going to try one of these, they seem quite cheap.

 

[dcasto]

Your problem appears to be water on the diffuser. I'm wondering if moisuter in the air is settling on the diffuser and when you blast the cold air on it, that water freezes and not nessasily the water in the compressed air.

You should be draing the water off the bottom of the tank that holds the air before blowing down.

 

[sailoday28]

If the air were a perfect (dry)gas with constant specific heats AND the receiver transfered a negligible amount of heat to the air
Conditions in the receiver would be:
T/Ti= (P/Pi)^[(gamma-1)/gamma]
where T is absolute temperature
P is absolute pressure
gamma is the ratio of the specific heats, Cp/Cv
subscripts i represents initial conditions in the receiver.
Clearly the final air temperature would be quite low compared to initial conditions.
Further considering that flow velocity would lead to lower temperatures in the piping, it is reasonable to assume icing is possible within and outside the piping and fittings.

 

[willard3]

If you reduce the rate of blowdown to meet the rate of heat transfer to the pipe, the freezing will go away.
Alternately, heat the pipe/valve.

 

[moon161]

Plumb it like this:

Tank - adjustable valve- pressure gage - cyclone separator- blowdown valve- silencer
When the blowdown opens, the air takes an expansion to a dewpoint state, say 5C through the partially open valve, goes through a cyclone, and takes the final expansion dry at the blowdown valve.

Push the expansion math around:
T/Ti = (P/Pi)^(gamma-1/gamma)
P=Pi (T/Ti)(gamma/gamma-1)
Set T=278K
Gamma air = 1.4.
solve
P=8.6 barg

Check the math, it's been a while.

For the first trial, set your partially closed valve so that you have 8.6 bar at the gage after the valve when the blowdown opens.