Effects of Water Contamination in Ammonia Refrigeration Systems

By: Per Skaerbaek Nielsen, Mech. EngineerDanfoss A/S, Denmark

Copy right Danfoss - November 2000 1 RZ.0X.A1.02

Preface Water contamination is often an underestimated problem in industrial ammonia refrigeration systems. The aim of this paper is to point out which problems, often seen in industrial ammonia refrigeration systems, can be related to water contamination. Often these water contamination related problems are the reason for a severe increase in service, maintenance and running cost without the cause of the problems being realized and cured. A. Water contamination sources 1. In industrial ammonia refrigeration systems, the water contamination sources can be divided into two groups. The contamination sources in the construction and initial start up phase, and the contamination sources after the system has been put into normal operation. B. During construction and at initial start up 1. Causes: = Water remaining in new vessels, which are not properly drained after pressure test. = During construction, water may enter through open piping or weld joints, which are only tacked in place when either are exposed to the elements. Condensation, which may occur in the piping during construction. Condensation, which may occur when air has been used as the medium for the final system pressure testing. Water, which remains in the system as a result of inadequate evacuation procedures on start up. = The use of non-anhydrous ammonia when charging the system. 2. Solutions: Water contamination during construction and at initial start up, can and should be avoided by: = Using pure anhydrous ammonia = Ensuring that all vessels, evaporators, condensers, etc. have been properly drained before installation = Applying proper procedures during construction, pressure testing = Flushing, using hot dry nitrogen. = Sufficient vacuum pulling at sufficiently high ambient temperatures, before charging the ammonia. = Using dry clean tubes when charging, etc. C. After the system has been put into normal operation

Copy right Danfoss - November 2000

RZ.0X.A1.02

= Rupture of tubes on the low-pressure side of the system, especially in chillers. Improper procedures, when draining oil or refrigerant from vessels or pipes in vacuum range into water filled containers. = In systems, which are operating below atmospheric pressure or which are making pump down so the pressure goes into a vacuum range: Leaks in valve stem packing, flexible hoses, screwed and flanged piping joints, threaded and cutting ring connections, pump and compressor seals, and leaks in the coils of evaporator units. Improper procedures when evacuating the system or parts of the system, while service and maintenance work is carried out. Complex chemical reactions in the system between the ammonia, oxygen, water, oils and sludges can create more free water in the system.

Water contamination, after the system has been put into normal operation, can be very difficult to avoid. Very often it will happen without being noticed by the staff operating the system or the service and maintenance engineers. It is important to remember that unless steps are taken to drain the water, all of the water that gets into the system will stay there, and the concentration will rise over time.

Let us have a closer look at how the water gets into a system, and especially how this can happen without being noticed. 1. A bottle test is a very useful example to show the strong affinity between ammonia and water. If a bottle is turned upside down, filled with ammonia vapour from the system and placed upside down in a bucket of water, the water will be sucked slowly into the bottle until an equilibrium level is reached. This occurs because the ammonia vapour mixes with the water, decreasing the pressure within the bottle and drawing water inside. This little test shows why it is necessary always to use a check valve in the drain line, when evacuating system, or parts of the system into water filled containers, while service and maintenance work is carried out. Very often a valve connected to a tube leading into water filled containers, is left open during this kind of work. This is to make sure the pressure in the system, or parts of the system, will not rise during effectuation of the work. If a clear tube is used to make the water visible, it may be noticed the water sometime is being sucked into the system, if a check valve is not used. 2. When draining oil or refrigerant from vessels or pipes with the working pressure in the vacuum range, into water filled containers, a check valve should also always be fitted in the drain line to avoid possible back flow of water. It should always be ensured that the pressure is above the vacuum area while draining, but leaks in valves or mistakes in the procedures, could allow the pressure to drop into the vacuum range

Copy right Danfoss - November 2000

RZ.0X.A1.02

and create back flow of water. 3. On processing equipment in areas with the presence of water, especially ice machines, the pressure is normally not in a vacuum range. However, when this equipment performs a pump down sequence before shut down, the pump down may be adjusted to take the pressure into the vacuum range. This is done in order to save time and/or prevent compressors from starting and stopping many times during this process. If this equipment has seals, which are only meant to seal against, pressure from inside the system and out (for example U-lips sealing rings) and not from outside and into the system, the pump down into the vacuum pressure range can suck water into the system. Even if the sealing arrangement is meant to be tight in both directions, small axial movements of the shaft or the sealing tips (due to the change in direction of forces coming with the change in direction of pressure) can easily make the seals leak water into the system. On this type of equipment, it is recommended to use a pressure regulator in the suction line, to ensure the pressure in the evaporator never reaches the vacuum range. 4. Leaks on the low-pressure side of a system, where the pressure is in the vacuum range, will allow air with humidity to be drawn into the system. The moisture in the air will immediately react with the ammonia and stay there, while the air will pass through the compressor and be trapped in the condenser and/or receiver. When too much air has accumulated on the high-pressure side of the system, the condensing pressure will raise, because air will take up the space in the condenser, leading to an increase in power consumption on the compressors. NOTE: If air is caught in the condenser/receiver, and the high pressure receiver runs out of liquid, or if a high pressure float valve system with a small high pressure gas by pass is used, the accumulated air with its content of especially oxygen (02) will pass continually through the compressors. This will allow chemical reactions with the oil, and speed up the oxidation of the oil and other chemical reactions in the system considerably. 5. Air, which is accumulated on the high-pressure side, can be purged from the system manually or by means of an automatic air purger, but the moisture from the air stays in the system, and accumulates over time. The amount of water that enters into a system this way will depend on the amount of air drawn into the system, and the water content of the air. If the air is purged manually when needed, the operating personnel should get an idea about the amount of air drawn into the system, and should be alerted by increases in the need for purging. However, if an automatic air purger is used, leaks in areas,

Copy right Danfoss - November 2000

RZ.0X.A1.02

which operate in the vacuum range, can go unnoticed for years. If no high-pressure problem on the system is apparent, the capacity of the air purger may handle the load. For this reason it is strongly recommended always to use a counter device on an automatic air purger - if possible with alarms and / or warnings for increasing purging activity. A logbook for the air purger where running conditions, system and purging activity is noted, is also recommended. This could be an important help in tracking down where leaks in vacuum pressure range of the system are located. To try to get an idea about how much water could be accumulated unnoticed in a system because of air drawn into the system, a worst case situation with a very leaky system can be considered: = Information given by an air purger manufacturer indicates a maximum air purging capacity of approx. 10 l air/min (at atmospheric pressure). = It is also recommended by air purger manufacturers to mount two air purgers on large systems to handle peak load situations. Based on this information, the following theoretical worst case assumption is made: An automatic air purger purges continually 5 l air/min. The ambient temperature where the leaks are located is 20C, and the humidity of the air is 80% in average. The air entering the system will then contain 13.84 grams of water each 1000 l air (table values).

This means: 5 l x 1/1000 l x 13.84 g x 60 min = 4.15 g H2O each hour or: 4.15 g / hour x 1/1000 l x 24 hours x 365 days = 36.35 l H2O each year After 10 years we have: 363.5 l H2O in the system If the small amount of water, which will be on the high-pressure side, is not taken into consideration, it means: A system has on the low pressure side approx. 363.5 l / 10 x100 = 3635 kg refrigerant This is equal to approx.
1 0 x 3635 = 6058 l refrigerant, 0.6

Copy right Danfoss - November 2000

RZ.0X.A1.02

will have approx. 10 % water in the ammonia on the low-pressure side. With approx. 1817.5 kg equal to approx. 3029 l ammonia on the low pressure side we will have approx. 20% water in the ammonia on the low pressure side. 6. A major contamination can occur within a very short time because of rupture of a tube or tubes in the system. Incidents of this type usually alert operating personnel because of the upset to normal operating conditions, and/or the resulting release of ammonia. After such an incident it is very important the amount of water contamination is detected, and the necessary steps are taken to drain the water from the system. D. How to Detect the Water Contamination 1. How the amount of water contamination is detected, and what can be done to drain the water from the system, are very well described in the IIAR Bulletin No.108. Please refer to this IIAR Bulletin for further information on these subjects. 2. It is recommended to make water contamination measurements on ammonia samples as a normal procedure during the periodic service and maintenance work on the system. 3. When a sample of ammonia is taken, and the water content is detected, it is very important to understand what is being measured. It will be the percentage of water contamination at the given place in the system, at the given time. It is necessary to consider the system design and the exact running condition at the time the sample was taken when evaluating the results. An ammonia system with water contaminated ammonia can be considered as a big distillery, where the water will concentrate wherever there is evaporation, because of the large difference in vapour pressure between water and ammonia. It is therefore important to know how much of the systems total charge is actually on the lowpressure side when the sample for water detection was taken. In many pump circulation systems more than one third of the total charge can be stored in the highpressure receiver during some running conditions. The same system could experience situations where the high-pressure receiver runs nearly empty. As almost the entire amount of water will stay on the low-pressure side, this will have a big influence on the water percentage measured in the low-pressure receiver and evaporators.

E. System Design Influence on Water Contamination Effects 1. Some systems have evaporators with hot gas defrost, where the liquid during defrosts is drained to the high-pressure side of the system. These systems will drain the water

Copy right Danfoss - November 2000

RZ.0X.A1.02

from the evaporators to the high-pressure side of the system every time they make a hot gas defrosting. This also means quite a high concentration of water can be found in the high pressure receiver on these systems. Due to this, these types of systems are especially sensitive to water contamination, when compressors with high-pressure liquid injection systems for oil cooling or intermediate stage cooling are used. The injection systems will lead the contaminated ammonia directly into the compressors. Once in contact with the compressor oil, it will quickly break down the oil and create high amounts of sludge. Incidents of this kind have been seen, where an extension to an old very water polluted system was made. Two stage reciprocating compressors with liquid injection for intermediate cooling were used on the extension, and more ammonia was added to the old polluted ammonia. After a short while the reciprocating compressors could only run three days at a time on new mineral oil, before the oil got so soapy that it clogged the oil filters, and the compressors lost oil pressure. 2. Systems with water contamination and inefficient liquid separators can have water flowing back to the compressors, leading to oil and corrosion problems. High amounts of free water have often been found trapped in valves or low points in the dry suction lines on such systems. 3. In flooded systems with liquid separators at the evaporators and pump down in connection with water or electric defrost, the water will always stay in the evaporator. In such a system the water content in these evaporators can reach a very high level, while the water content in the rest of the system is much lower. One advantage of this type of system is the possibility of making pump down on these evaporators, and then drain them for water. In this way the evaporators can be used as water rectifiers and the system can be kept drained for water. 4. On dry expansion systems, the water will normally be dropped by the evaporating ammonia and driven as free water with the high velocity superheated suction gas back towards the compressors. If a suction accumulator is mounted as normally recommended, the water will settle there. If no suction accumulator is used, the water will go to the compressors, where some of it will mix with the oil. Part of the water will pass on to the condenser, to the high pressure receiver, on to the expansion valve and back to the evaporator. Water contaminated ammonia liquid will increase the wear in all kinds of expansion devices due to increased cavitation, erosion and corrosion in the orifice area, as described later in this paper. 5. If water gets the possibility to accumulate in evaporators, working with thermostatic expansion valves, and regulating on superheat, we are facing another problem. It could be because of low suction gas velocities, part load operation, or a flooded evaporator with an expansion valve working as a liquid level regulator. The problem will be due to the change in saturated temperature at the same pressure for the

Copy right Danfoss - November 2000

RZ.0X.A1.02

ammonia water mixture (this change is described later in this paper). The thermostatic expansion valve will see this change as superheat, and will not be able to control the evaporator correctly. This will be the same problem for both mechanical and electronic thermostatic expansion valve systems. In some situations, this could lead to overfeeding of the evaporator and possible liquid carry over to the suction line, even with a perfect working thermostatic expansion valve system. If these relations are not known, valve systems often get the blame for these waterrelated system problems.

F. How much water is actually found in systems 1. A large investigation of the water content in more than a hundred ammonia systems was carried out in Denmark, Norway and Sweden. The method used to determine the water content in the ammonia was a simplified version. This method shows a slightly exaggerated water content, when the sample is taken at evaporating temperatures higher than -33C, because the flash gas created when taken the sample is not taken into consideration. The method used to determine the water content in the investigation is enclosed in appendix 1. Diagram showing the determined water content in the investigated 136 systems is shown in appendix 2. As can be seen from the diagram, water contents between 2% and 6% is seen very often. Water content of 10.8%, 18.5%, 24% and 26% were measured, but these measurements could be showing slightly exaggerated values, as they were taken at temperatures over -33C (-27,4F) and no correction for flash gas was made. As described earlier, the measured water content will depend on where the sample was taken, the design of the system, and the running condition when the sample was taken. Other test samples taken on the same systems in different running conditions could show different results. 2. On 64 of the investigated systems with water contamination, water rectifiers were installed. The diagram Quantity of drained water in appendix 3. shows how much water these rectifiers drained from the systems. As can be seen from the diagram, up to 250 l of water was drained from some systems. 3. Based on the diagrams, it can be concluded: = Many systems have a much higher water content than the normally recommended

Copy right Danfoss - November 2000

RZ.0X.A1.02

max. of 0.3%. = Some of these systems have such a water contamination, that it must be expected to cause serious problems with capacity, COP value, and chemical reactions in the systems. = None of the system operators involved in the investigation had any knowledge of any water contamination in the system, or which problems could be related to water contamination. = Water contamination checks should be carried out regularly in connection with service and maintenance work.

Effects of water contamination on system capacity and power consumption 1. When ammonia is contaminated by water, the thermodynamic and physical properties of the solution change. A lower pressure in the evaporator is needed to maintain the same temperature (see enclosed tables in appendix 4.) This will be a penalty for the system in terms of reduced capacity and higher power consumption on the compressors. Example: If a 500 kW (142 TR) chiller with plate heat exchanger (as evaporator) is designed to work at 3 bar abs. (43.5 psi) the evaporating temperature with 100% pure ammonia will be -9.23C (15.39F). Chiller units with plate heat exchangers often have a high capacity with a very little ammonia charge, and a temperature difference between the evaporating temperature and the chilled liquid of 3 to 4C (5.40 to 7.20F). If we are talking about a plate heat exchanger working as a flooded evaporator, the charge in the evaporator and liquid separator could be approx. 22 kg on a 500 kW (142 TR.) unit. This means 2.2 kg water gives 10% water contamination, and 4.4 kg water gives 20% water contamination. A unit with dry expansion could have a charge of 5 l in the evaporator on a 500 kW (142 TR.) unit. In this case 0.5 kg water means 10% water contamination and 1 kg water means 20% water contamination. With 10% water in the evaporator, the evaporating temperature at the same suction pressure 3 bar abs ( 435 psi) will be -6.69C(19.96F). With 20% water in the evaporator, the evaporating temperature at 3 bar abs will be -3.16C (26.31F). Such a chiller unit will no longer be able to keep the capacity or the temperatures it was designed for, and very often a service engineer is called to solve the problem. If the engineer is not familiar with this change in saturated temperature because of the

Copy right Danfoss - November 2000

RZ.0X.A1.02

water ammonia solution, he will easily draw the conclusion that the unit is operating with too high superheat on the evaporator. In the search for reasons for what seems to be superheat, many things can be anticipated as the cause of the problem: Valves, filters, heat transfer in the evaporator, temperature, and pressure sensors etc. In these situations it is very important to understand, what can be read on a pressure gauge with a temperature scale on it, is not necessarily true for the refrigerant any more. Note! On systems controlled by thermostatic expansion valves, water contamination could be the cause for liquid flood back to the suction line as the expansion valve will try to compensate for what it will see as an increase in superheat, by injecting more and more liquid. 2. The penalty in terms of power consumption and capacity will be worst at low evaporating temperatures, as the COP value of the compressors will be affected more at low suction pressures. In order to get an idea about how bad the situation can get, a worst case calculating example is made: If a system is running at an evaporating temperature of -42C (-43.6F) and a condensing temperature of 30C (86F) with screw compressors. What will happen to power consumption and capacity, if we assume a water contamination in the evaporators of 10% and a worst case of 20%? Computer calculations on a screw compressor give the following results: (See appendix 5.) Note: The lower pressures, the compressor has to maintain in order to keep the evaporating temperature constant, have been put into the computer calculations on the screw compressor performance as a suction line pressure drop. The thermodynamic and physical properties are taken from the enclosed tables. (Appendix 4.) 100% NH3 and 0% H2O in the evaporators: ET = -42C (-43.6F) EP = 0.64 bar (abs) (9.28 psia) CT= 30 C (86F) Cooling capacity: 361.7 kW (102.7 TR) Power consumption: 252.6 kW COP = 1.43 90% NH3 and 10% H2O in the evaporators: ET = -42 C (-43.6F) EP = 0.59 bar (abs) (8.56 psia) CT= 30C (86F)) Cooling capacity: 327.1 kW (92.9 TR) Power consumption: 248.6 kW COP = 1.32

Copy right Danfoss - November 2000

10

RZ.0X.A1.02

80% NH3 and 20% H2O in the evaporators: ET = -42 C (-43.6F) EP = 0.51 bar (abs) (7.40 psia) CT=30C (86F) Cooling capacity: 278.8 kW (79.2 TR) Power consumption: 241.1 kW COP = 1.16 As it appears the compressor looses both capacity and COP value. If it is assumed that a system is running with these screw compressors 10 hours a day, 300 days a year under these running conditions, how many kWh do we have to use extra to produce the same refrigeration effect? By using the definition of COP: COP = refrigeration.effect power.consumption

At 100% NH3 and 0% H2O in the evaporators: For every 1000kW (284 TR) cooling capacity power consumption is: 1000/1.43 = 699.3 kW 10x300x699.3 = 2097900 kWh each year. At 90% NH3 and 10% H2O in the evaporators: For every 1000 kW (28.4 TR) cooling capacity power consumption is: 1000/1.32 = 757.6 kW 10x300x757.6 = 2272800 kWh. each year Additional power consumption each year: 2272800-2097900 = 174900 kWh each 1000 kW (284 TR). At 80% NH3 and 20% H2O in the evaporators: For every 1000 kW (284 TR) cooling capacity power consumption is: 1000/1.16 = 862.1 kW 10 x 300 x 862.1 = 2586300 kWh. each year. Additional power consumption each year: 2586300-2097900 = 488400 kWh each 1000 kW (284 TR). Apart from the additional cost and lost capacity, this is also an environmental issue,

Copy right Danfoss - November 2000

11

RZ.0X.A1.02

as the increase in power consumption means more pollution. G. Water related chemical reactions Pure anhydrous ammonia is not very chemical reactive in a refrigeration system, but water contaminated ammonia is very chemical reactive. Actually pure ammonia without any water at all will not attack copper. As soon as a little bit of water is dissolved in the ammonia, it will form the very chemical reactive ammonium hydroxide. NH4 is an ammonium ion and OH is a hydroxyl ion. A solution, which contains ions, is carrying electrical current ( an electrolyte) and can create a galvanic cell with metals, which have different electric potentials. This gives the possibility of galvanic corrosion in valves, pipes, etc. especially in areas of the system where oil is not present. Ions are chemically reactive and can lead to chemical reactions in the system. Some of the chemical reactions will create more free water in the system. Reactions with acid from oxidation of oil: AMMONIA+ ACID AMMONIUM CARBOXYLATE AMIDE + WATER NH3+ RCOOH RCOONH4 RCONH2+H2O RCOOH is an acid created by oxidation of oil. RCONH2 is an AMID which can be a solid sludge and can settle anywhere in the system. Atmospheric air always contains some CO2. Reactions with air containing carbon dioxide: AMMONIA + CARBON DIOXYDE AMMONIUMCARBAMATE CARBAMIDE+WATER 2NH3 + CO2 H2NCOONH4 H2NCONH2 + H2O HNCOONH4 is ammoniumcarbamate, a substance that is very corrosive to steel, and H2NCONH2 is carbamide (a sludge). H. What happens to the oil in the system? The oil is broken down in the system under the influence of water and oxygen in three ways: Oxidation, Nitration and the resulting formation of nitro compounds. OXYDATION is:
+ -

Copy right Danfoss - November 2000

12

RZ.0X.A1.02

WATER + OXYGEN + OIL ACIDS).

PRECURSERS + ORGANICAL ACIDS (WEAK

Precursors are very complex compounds, which colour the oil (cognac colour). These processes are amplified by heat and catalysers (metals). NITRIDING is: WATER + NITROGEN + OIL ACIDS). PRECURSERS + ORGANICAL ACIDS ( WEAK

These processes are also amplified by heat and catalysers (metals). NITRO COMPOUNDS are: ORGANICAL ACIDS + AMMONIA SALTS, SOAP PRODUCTS) NITRO COMPOUNDS (SLUDGES,

NITRO COMPOUNDS will only be formed if oxidation or nitriding have taken place and only with the presence of water. The created nitro compounds have a catalytic effect, speeding up the process of creating more nitro compounds. The nitro compounds are not soluble in oil, but partly soluble in ammonia and can escape with the ammonia vapour out through the oil separators into the system. Due to this fact, the sludge from the nitro compounds will be found both in the compressors, valves, pipes, vessels, filters and anywhere they can settle all over the system and cause operating problems. The dissolved nitro compounds gives the ammonia a yellow/brown colour.

I. Water related problems on valves and controls 1. Rust and sludge The aggressive environment created by the water-polluted ammonia can lead to the formation of rust in valves, pipes and compressors. The rust together with the sludge can cause problems with valves, controls and regulation devices. The filters get

Copy right Danfoss - November 2000

13

RZ.0X.A1.02

clogged, and valves get stuck in the dirt and rust, unable to operate correctly. The higher amount of hard particles in the refrigerant from the rust and sludge will increase wearing problems due to erosion in orifices and expansion valves. 2. Fretting corrosion A normally not so well known problem is fretting corrosion. Definition: Fretting is a wear phenomena occurring between two surfaces having oscillatory relative motion of small amplitude. Fretting corrosion is a type of fretting in which chemical reaction predominates. Fretting corrosion is often characterised by the removal of particles and subsequent formation of oxides, which are often abrasive and so increase the wear. Fretting corrosion can involve other chemical reaction products, which may not be abrasive. In many places where valves and controls are located, they will be exposed to vibrations and/or pulsations. Valves and controls mounted on compressors or in connection with other machines will be exposed to vibrations. Valves and controls in many lines in a system will be exposed to pressure pulsations, especially in the pipelines close to the compressors. These vibrations or pulsations can give reason to many oscillatory movements of small amplitude between parts in valves and controls. These small movements can lead to fatigue of the small asperities on the surface of the metal, and as a result they brake off. When this happens in a very corrosive environment, the small particles are attacked by corrosion, and subsequently grow in size and get hard. This will increase the friction and create more wear. Due to this, valves and controls can get stuck and be unable to control the system correctly. The involved chemical reactions are the following:

2Fe+O2 2FeO 3FeO + O2 Fe3O4 2FeO4 + O2 3Fe2O3 Fe2O3 + H2O FeOOH

Black Blue / black Red / brown, abrasive leads to wear Red / brown - Rust

If fretting corrosion takes place, increased friction and severe wear will take place

Copy right Danfoss - November 2000

14

RZ.0X.A1.02

rapidly. To avoid fretting corrosion, it is necessary to eliminate the vibrations / pulsations or control the corrosive environment. As the pulsations / vibrations always will be present in a system, the best protection is to try to avoid a corrosive environment in the system. 3. Cavitation, erosion and corrosion On systems with a high degree of water contamination, an increase in wear problems in all kinds of expansion devices will be noticed. The exact reason for this is not known or proven yet, but some suggestions are made: = More erosion in the orifice due to more small hard particles from chemical reactions coming with the liquid. = The introduction of corrosion together with the erosion and eventually cavitation in the orifice area. = The collapse of water vapour from the flash gas into very small water drops with a kind of cavitation effect on the orifice area. As mentioned none of these suggestions have been proven, but the explanation is probably a combination of these suggestions. Problems of this kind will especially be seen on systems where the water gets to the high pressure receiver during hot gas defrost, or because of liquid carry over from the low pressure receivers, as previous mentioned. J. How can the water be taken out of the system, and is it an economical advantage? 1. The water can be removed from the system by emptying the entire system, drying it with nitrogen, pulling vacuum and recharging with pure anhydrous ammonia. This procedure can be recommended in small system with a limited charge. For example liquid chiller units with plate heat exchangers. The cost of recharging these systems, and the time involved to do the job is limited. 2. An evaporator can be dedicated to act as a water rectifier and drained regularly. This can only be done if there is an evaporator suitable for this function on the system. Suitable evaporators could be: = Evaporators with gravity circulation and liquid separators in connection with electric defrost on water defrost. = A flooded chiller with liquid separator. = Intermediate coolers with possibility to make a pump down sequence. = Applying a water rectifier to the system. Is often the only way to clean a pump circulation system. With a water rectifier, a batch of liquid is taken from the

Copy right Danfoss - November 2000

15

RZ.0X.A1.02

low-pressure receivers and/or the intermediate coolers and carefully evaporated, making sure no water drops return with the vapour. This is done many times until the water concentration in the rectifier is very high. A pump down is performed, and the rectifier vessel is emptied. Experience shows it is very important the rectifier is controlled, so heavy boiling is avoided to make the rectifier work satisfactorily. How this can be done, is described in the IIAR technical bulletin No. 108, so the issue will not be treated further in this paper. 3. When evaluating if it is a good idea to get the water out of a system, the following should be taken into consideration: = How much does the water contamination cost in lost capacity and extra power consumption? = How much of the service and maintenance cost could be related to water contamination? = Are the risks of system break down increased because of the water problem, and what will be the consequences? = What is the cost of changing the charge? = Can an existing evaporator be used as a water rectifier? = What is the cost of having a water rectifier built into the system? = Can a water rectifier be rented for cleaning the system with an economical advantage? 4. For very small ammonia systems Danfoss has investigated if it is possible to make ammonia drying filters. Until now these investigations have not lead to the development of any efficient working filter cores. The filter cores on the market at the moment have also been tested, and proven not to work. Further research and development on this have been stopped at the moment for the following reasons: Very small ammonia systems (0 to 10 kg charge) can be manufactured and charged so that they are very dry. If such a system should have a drying filter, a trained service man would be needed for changing the filter core in connection with normal maintenance. The price for filter cores - if they can be developed - will probably be somewhat more than the cost of changing the charge on the system with pure dry ammonia. Another problem is to get an indication of a wet system, and due to this, when to change an eventually filter core. CONCLUSIONS = Ammonia systems very often contain much more water than recommended. = The relations between water contamination and problems seen on systems are normally not very well known among service and maintenance staff.

Copy right Danfoss - November 2000

16

RZ.0X.A1.02

= System design plays an important role in which effect water related problems will have on a system. = Improper procedures during service and maintenance work are an important source for water contamination. = Air purgers can solve the high-pressure problems due to incoming air from leaks, but they let leak problems go unnoticed for years, and so hide an increasing water pollution problem. = Water contamination detection should be carried out as a part of normal service and maintenance, taken the actual system design and running condition into consideration. = Installation of water rectifiers or other methods of water draining from a contaminated system can prove to be a very attractive investment.

Copy right Danfoss - November 2000

17

RZ.0X.A1.02

References:
= Trelle Pedersen, J.: The refrigerant ammonia, Internal compendium, LA 23181, Danfoss = Finn Broesby-Olsen, Laboratory of Physical Chemistry, Automatic Controls Division, Danfoss AS: Chemical Reactions in Ammonia, Carbon Dioxide and Hydrocarbon Systems Paper presented at Conference Applications for Natural Refrigerants 96 3-6 September 1996 = Michael Norden Hydro Texaco: Re use of Refrigeration Oils: Lecture presented for Selskabet for Kleteknik 18. and 21. November 1996. = Svenn Hansen Danish technology institute and Michael Norden Hydro Texaco A/S: Investigations on water contend and oil in industrial ammonia refrigeration systems, Internal papers. = Organisation for economic corporation and development, Paris 1969: Friction wear and lubrication, Glossary terms and definitions. = Sabroe Match Master computation program = IIAR Bulletin 108: Water Contamination in Ammonia Refrigeration Systems = International Institute of Refrigeration, 177 Boulevard Malesherbes F75017 Paris France: Tables and diagrams for the refrigeration industry: NH3-H2O Thermodynamic and physical properties.

Copy right Danfoss - November 2000

18

RZ.0X.A1.02

Curriculum Vitae

Name: Born: 1987 1987 1988 1992 1994 1995 1997

Per Skrbk Nielsen, Mech. Engineer, Danfoss A/S, Denmark 1959 Graduated from Aarhus Teknikum Project Engineer - Stal Kulde, Denmark Product Engineer - Sabroe Product Division, Denmark Service Manager - Sabroe Industrial Division. Denmark Technical Manager - Sabroe Canada Marketing Engineer - Danfoss A/S, Denmark Product Manager, Danvalve stop valves and Danfoss mechanical industrial products, Denmark

Copy right Danfoss - November 2000

19

RZ.0X.A1.02

WATER CONTENT IN RELATION TO EVAPORATING TEMPERTURE

30

INTERMEDIATE COOLER EP. TEMP [-50;-31] C

SEPARATOR OR EVAPORATOR EP. TEMP [-50;-31] C

26

25

SEPARATOR OR EVAPORATOR EP. TEMP ]-31;5] C

175 PLANTS IN TOTAL

24

20

MEASURED VALUES 77 : = 1% WATER 37 := 2% WATER = 25 : = 3% WATER 18.5

% WATER

15

10,8

10

6 4 4 4 4,6

6 4.6

4,5

4,6

Appendix 2

0 0 -37 -35 -35 -35 -35 -30 -27 -20 -18 -17 -15 -15 -12 -10 -10 -10 -10 -10 -10 -10 -10 -10 -10 -8 -5 -40 -40 -38 -38 -38 -3 5

-45

-45

-45

-45

-40

-40

-40

-40

-40

-40

-40

LOWEST EVAPORATING TEMPERATURE [C] ON THE PLANT

DRAUHT WATER [L] 100 150 200 250 50 0

DK01 DK03 DK04 DK05 DK06 DK07 DK08 DK09 DK12 DK14 Dk15 DK16 DK18 DK19 DK21 DK23 DK24 DK25 DK27 DK28 DK29 DK30 DK31 DK32 DK33 DK34 DK35 DK37 DK39 DK40 DK42 DK43 DK47G DK50 DK53 DK55 DK56G DK58G N05 N26 N27

57,09 EP temp. ]-31; 5] C EP temp. [-50;-31] C

65 PLANTS IN TOTAL

158,8

100

50

152

199,04

102,9 108,4

Appendix 3 Chart 2 QUANTITY OF DRAINED WATER FOR THE CURRENT PLANT [L]

Appendix 3

Page 1 No. PLANT

50

N37 S 38 system 2 S01 S02A

100

82

S02B S03 S05 S12 S14

98

150

70

S22-1 S22-2

78

S22-3 S23-1 S23-2 S26 S27 S38 system 1 S38 system 3 S42 Fryslager

105 100 120 250

80

S42 Glasstillverkning S43 S45 S47

150 250

65

S52