Ammonia condenser debottleneck 1

[tu1]

We are looking at adding a second ammonia refrig condenser to provide additional surface area as the existing unit could not keep up with the cooling required. My question is: should the new condenser be tied in parallel or in series with the existing unit? I've heard they should be connected in parallel because of 2-phase service. If they are in parallel, how can the flow be controlled to prevent each from overperforming or underperforming? The units are not identical in sizes.

Thank you for your help.
Tu1

 

[Montemayor]

TU1:

You don't state how the two condensers are proposed to be connected or what cooling fluid you are using. I have to assume this is a conventional NH3 condenser system that uses a TEMA type AEM shell & tube unit with NH3 in the shell side and cooling water (CWS) in the multi-pass tubes.

The parallel shell connection is what should be used in order to have simplicity of operation and control. A series connection makes no sense for this service. I've never heard or seen one - and I've done a variety of NH3 condensers - horizontal and vertical. Unless you have a very special application or expectation of the refrigeration results, I cannot justify your concern for controlling the flow to prevent each unit from "overperforming or underperforming". Seldom, if ever, is there any concern for controlling the sub-cooling of NH3 condensate that is subequently used for direct expansion into a refrigeration evaporator. Conventionally, everyone I know tries to get as much subcooling as possible in order to reduce the flash vapors and increase the refrigeration effect and efficiency. I would not (& have not had need to in the past) try to control the subcooling of the NH3 condensate. Normally, users set the CWS to a set flowrate, based on the maximum condenser duty (& this is normally a constant). If the CWS is colder than usual, you get more subcooling of the condensate and your NH3 refrigeration compressors work less with the same evaporator load because of reduced flash vapors.

Connecting the hot NH3 discharge vapor header to both condenser inlets and allowing each unit to do its work as it receives vapors is the easiest way to operate the system. The vapor will distribute itself according to its availability and the capacity of each unit. The excess load of one unit will be automatically handled by the other unit. Don't forget: the condenser units are normally overdesigned anyway to allow for fouling, capacity changes, upsets, CWS temperature variances, etc. This is the normal way of designing this type of equipment. The resultant liquid condensate should be routed to the normal liquid surge/storage tank receiver that is usually directly under the condensers. The condensate line is usually extended into the surge receiver in the style of a dip pipe. I would use independent dip pipes, one for each condenser.

From the Surge receiver forward, the process behaves and operates the same as it is now and no further changes are required. Don't forget to install an equalizer line between the receiver and each condenser - or design self-venting condensate drain lines.

I hope this information helps you out.

Art Montemayor

 

[tu1]

Art, thanks for the tips. The condenser, actually, is a horizontal air-cooled heat exchanger and the NH3 is a R-717 (ASHRAE)refrigerant. So, if the 2 units are going to be in parallel, do you foresee one will have more flow going through than the other? I guess my concern is one unit will be overworked and the other will be starved unless there is some sort of pressure/flow control at the inlets. I am considering putting in a buterfly valve, one on each inlet piping, to redirect the ammonia where it is needed. Any comments?

Thanks.

TU1

 

[Montemayor]

tu1:

It doesn't make any difference if you have an air-cooled condenser rather than a water-cooled unit. The work of the condenser is to effect a phase change - to convert the gaseous stream into a liquid stream - and, as such, the units create their own available capacity as they continue to condense NH3 vapor, with the resultant condensate dropping out by gravity. This characteristic of a condenser allows you to just keep the vapor feed always available to it and the maximum capacity of the condenser will be fulfilled on its own. You do not need to install throttling valves or other flow control devices to try to keep both condensers fed with the calculated design vapor load - both units will continue to condense as long as they have the available clean internal surface and the design coolant flow rate and temperature - in this case, the ambient air. I have designed and operated steam condensers in this manner and they have worked perfectly. The application in NH3 service is no different, except in the pressure being higher.

Air-cooled condensers present a different problem in northern climates: the potential for over-cooling the process fluid by subcooling the liquid such that you might reach the solid phase. In your case, I don't believe you face the danger of freezing the liquid NH3 into the solid state, but what will happen is that the equilibrium condensing pressure will be lower than the saturated pressure. Depending on the resulting temperature, this might be below 100 psig - a pressure I prefer to work with as entering the refrigerant expansion valve. I don't know if you've allowed for this, but it might be of minor concern. If accounted and designed for, this effect doesn't pose a problem for your refrigeration load; it actually makes your refrigeration unit work more efficiently because of the lower heat sink given to you freely by Nature during the Winter months.

In summary, I don't see any reason to complicate matters and create more instrumentation costs in this application. Both units will work very well as I have described previously; one will not bias the other, nor will you have to favor one or the other. It is a simple and direct process design that you don't get very often in this business. I would take this simple design and run all the way to the bank with it, never asking any questions.

Art Montemayor

 

[tu1]

Thanks, Art. I appreciate your advice on this. I will let you know in August on how the units are performing.

TU1

 

[matin]

Dear tu1
If you want to add an identical surface condensor to the original one and if it would be installed at same level as the first one. Simply connect both inlet line via a symmetricaly designed manifold to the compressor outlet and do same for the coneecting phase to the receiver. now the condensors are "connected vessels" the role of the connected vessels says they are in the same level. As long as you have same level, you would have same surface area and same condensation load. (once more remember you should have same lenght of piping for both condensors)
About current configuration, I am interesting to know why the first condensors is not carrying the load of the system.
please write more about:
1- The routed to the receiver.
2- Is there any installed equalizer line between the receiver and condenser - or design self-venting condensate drain lines
3- Is your system a big industrial refrigeration system. Do you have monitoring system on this process and finally is your process smooth controled?

 

[25362]

Little can be added to what matin and art montemayor have said. I'd only add that deareators connected to the cool side of the condensers may help in the plant's overall performance.

 

[tu1]

MATIN:

Thank you for your inputs. I will give you more information about the existing condenser (C-1)and the new condenser (C-2):

1. C-1 and C-2 are not identical in sizes. C-2 duty is half of C-1. The units are separate.

2. The length of piping from the compressor discharge to each condenser is not the same. The distance from the compressor to C-1 inlet nozzle is 6 m, while the distance to C-2 is about 20 m. Piping to the condensers are coming off from a common compressor discharge header.

3. There is an equalizing line connecting from C-1 to the receiver. C-2 will also have the equalizing line to the receiver. There is only one receiver with two separate inlet connections.

Background:
The NH3 refrig unit is currently used to chill natural gas to -10 C to meet pipeline dewpoint specification. We are processing about 340E3 m3/d raw gas. The existing C-1 has given us some grief last summer, especially when we hit 30 C ambient temp. The C-1 had trouble condensing NH3 eventually lead to compressor shutdown on high discharge. The intention of adding C-2 was to use it only on a hot day. Otherwise the system is running quite smoothly for most of the year.

Though I agree with Montemayor's comments, I still think there should be valving installed on the condenser inlets so the flow can be directed to the units if needed.

Any comments/advice is much appreciated.

TU1





4. You mentioned the condensers be kept at same levels. I assume you mean

 

[25362]

Gate valving the condensers is normal for shutting then down and for maintenance purposes. The system pressure is determined by the pressure in the common receiver balanced out with both condensers, itself fixed by the saturation temperature of the condensed ammonia. Why are you considering the need of regulating vapour's flow rates ?

Art Montemayor likes 100 psig for ammonia entering the expansion valves. Such a pressure can generally be achieved after an economizing flashing stage, not straight from the condensers. Air as a coolant at 20-30^oC couldn't render that condensing pressure level. I'd assume 15-16 bara compressor's head pressure to be more suitable for the condensers' operation. Chilling the gas to -10^oC would mean expanding anhydrous ammonia to about 1.5-1.7 bara. Whatever the compressor's limitations, it would normally be working with a total pressure ratio close to 10. This is a pressure ratio I've seen -for ammonia refrigeration- in one-stage oil-flooded screw compressors. Reciprocating units would usually use two stages.

Don't forget the deaerators. Good luck !

 

[Montemayor]

tu1:

Although I hate to keep stirring a thread that I consider simple to resolve and straight-forward, I feel I should clear up some points that I've stated (or failed to state):

1. Matin is incorrect in stating that the vapor lines should be symmetrical. Symmetry is not a requirement in this application. What is a requirement (& what I consider routine and detail engineering) is that the vapor lines feeding both condensers should have an equally minimal pressure drop. This, I have presumed you are very aware of and will design appropriately.

2. Matin is also incorrect is stating that both condensers have to be at the same level. This is NOT a requirement for both units to condense vapor and drop out the resultant condensate directly by gravity into the NH3 receiver. All that is required is that vapor flow easily into either condenser and that both condensers be above the NH3 receiver, just as you have described. The liquid condensate will drop out by gravity and settle into the receiver regardless of whether one condenser is higher than the other. Additionally, the piping does not have to be of equal length. Pressure drop is the criteria that should govern, not symmetry.

3. 25362 is absolutely correct; any valving or attempts to try to distribute vapor will wind up being a control nightmare and ineffective, to boot. The vapor will seek the available heat transfer area of either condenser as this area is made available to it. It is as simple as that.
25362 also states that I like 100 psig for ammonia entering the expansion valves. I don't think I said that; what I said was that in the Winter time an air-cooled condenser will furnish a lot of cooling of the condensate, which will in turn yield in a lower (less than 100 psig) receiver pressure. I further added that I actually prefer the lower pressure (100 psig is the saturated pressure for approximately 56 oF - which is very achievable in areas such as Iowa and Minnesota where I've run these units). So, contrary to what 25362 says, you don't need an economizer or interstage flash drum to achieve 100 psig prior to expansion into the evaporator - it is very possible to obtain less than 100 psig as High Pressure liquid NH3 if you condense it at 50 oF.

Just wanted to make sure I haven't confused or shaken up anyone. It's an easy subject to mix up if one is not careful or fails to communicate well.

Art Montemayor

 

[25362]

Art Montemayor was indeed clear enough and I must admit I didn't correctly interpret his statements.

If weather temperatures are quite appart between morning and night, and between summer and winter, the compressor's design -and probably also the (induced-draft?) air cooler- should include provisions to adjust for this variability.

By having 50% more air-condensing surface, the temperature approach would be a little better, however, in summertime, with high air temperatures the compressor's head pressure may result in prohibitively large compression ratios and the unit would still have to be shut down.

 

[matin]

Dear Art Montemayor and tu1
1-You are right, I tink I did not explain may points, as clear as you discribed it. But when we are ,as chemical engineers, talking about symetrical arangement in piping, there should be a main aim that is same pressure drop which, we can easely get in symetrical piping.
2- For the elevation of the condensor, I did not see, in tu1 explanation about your point which is, ""both units condense vapor and drop out the resultant condensate directly by gravity into the NH3 receiver"". As long as we do not know the configuration, there is posibility to have reciver installed above condensor, as I have seen in many different plants. But you are right if we have condensors installed above the reciver,we do not need symetrical piping for condensor outlet.
3- Assuming condensors above reciver, with longer lenght piping from compressor to the second condensor, I think the load ratio (actual condensation capacity/design capacity) for the second condensor will be lower than the first one, but I foresee overally the problem of the system will be solved, Although the system arrengement is not an efficient one and condensors ar not in same load ratio.
thank you