CHAPTER VII THE INSTALLATION AND OPERATION OF REFRIGER- (Part 1)

CHAPTER VII THE INSTALLATION AND OPERATION OF REFRIGER- ATING SYSTEMS Installation It is scarcely necessary to state that every detail connected with the installation of the piping for ammonia, or other gases employed as a working medimn, should be executed with the greatest care. Only such materials as have been found by re- sponsible builders to be well adapted to their respective purposes should be employed. For ammonia, good full weight wrought- iron pipe is to be recommended for the expansion or low-pressure side, and extra heavy pipe and ammonia fittings of approved design for piping the compression side. In damp places, where the low-pressuro gas headers are liable to rust abnormally, these also should be of extra heavy pipe. Ammonia pipe joints are usually made up with lead or rubber gaskets in male and female flanges, sweated on the pipes. Some builders employ a litharge and glycerin cement in making up screwed joints instead of solder, and there is little difficulty in making such joints tight if scrupulous care is exercised in seeing that only true, sharp, properly formed threads are used; that they are thoroughly cleaned, for which purpose gasolene is to be recom- mended; that the litharge is freshly and thoroughly mixed into a thin paste; and that the joints are made up tight. It seems trite to suggest that gasket and flange joints should be drawn up squarely, but many a charge of ammonia has been lost through lack of attention to this detail. Rubbei* gaskets are par- ticularly likely to blow out of improperly drawn up flanges months later when the rubber has become softened by the oil. When the erection of the plant is complete, and the piping thoroughly blown out to free it from dirt, scale, metallic chips and other foreign substances, an air pressure of not less than 300 pounds should be pmnped on it to test for leaks. Leaks resulting from split pipes and improperly made up gasket joints may be readily located by the soimd. In fact, in a still cooler, sound is the most efficient means of detecting very small Digitized by V^OOQIC INSTALLATION AND OPERATION 101 leaks, especially of ammonia, when the air has become so laden with the fmnes as to make the usual methods of testing difficult. When all of the leaks have apparently been stopped, it is advisable to pmnp a pressure on the piping and let it stand for ten or twelve hours. The drop in pressure, provided there is no appreciable change in temperature, will indicate the amount of leakage. As a final precaution the air may be allowed to escape and the system again charged with air into which a sufficient amount of ammonia has been fed to make any leak easily detected either by smell or by means of sulphur sticks.* The approximate location having been f oimd by means of the sulphur fumes, the exact position of the leak may be located by oil applied to the leak by means of a long- nosed oil can or soapsuds applied with a brush. Where there is a likeUhood of existence of leaks in pipe or sub- merged condensers, or other places where escaping ammonia would not readily be detected because of its entering into solution in the cooling water or cooled brine as the case may be, it is advisable to test these liquids periodically with some reliable reagent. Where there are not too many foreign substances present, the litmus and turmeric papers are fairly reliable. A more satisfactory reagent, however, for use under the varied operating conditions of refrigerating and ice-making plants, is Nessler's solution, a few drops of which added to the suspected water or brine will show a yellow discoloration for sUght traces of ammonia, increasing with the amount of ammonia present until with large quantities a reddish-brown precipitate is formed. Repairing Leaks Many small leaks such as occur in ammonia fittings may be stopped by the judicious use of a set of small calking tools. Porous spots in iron and steel castings may sometimes be remedied by the judicious use of some rusting solution such as sal ammoniac or hydrochloric acid. Where the leaks are occasioned * Sulphur sticks used for testing for ammonia are made of pieces of white pine, or other wood which bums with little smoke, split into splinters half the size of a lead pencil and from 6 to 8 inches long. These sticks are then dipped into molten sulphur so that about 4 inches of the ends are thoroughly coated, and after being cooled are ready for use. In testing for leaks, the sticks are ignited and held close to the suspected pipe or fitting. If there is escaping ammonia, it will, on coming in contact with the burning sulphur, produce a very noticeable white cloud. Digitized by V^OOQIC 102 ELEMENTARY MECHANICAL REFRIGERATION by blow holes of considerable size occurring where the application of pressure will tend to drive the substance into the porosities of the iron, some of the patented rust-joint preparations may be effective. ^ Troublesome leaks due to imperfect welds in the seams of pipes may be effectively repaired by first cleaning the pipe with a file and some suitable soldering solution, then applying a closely laid course of bright steel wire. The layer of wire should then be saturated with the soldering solution and the whole surface thor- oughly coated with solder, special care being taken to see that it is thoroughly sweated in at the point where the leak occurs. The steel wire supplies the tensile strength, the lack of which in the solder would often allow the ammonia under pressure to lift oflf the solder coating. A hard solder should be employed and the steel wire should be thoroughly "tinned" to protect it from rust. For soldering iron and steel pipes two soldering solutions should be employed, the first being simply a cleaning solution of con- centrated hydrochloric acid, and the second a saturated solution of zinc chloride, commonly known among tinners as 'cut acid." This second solution is prepared by dissolving metallic zinc in concentrated hydrochloric acid. Some builders add to the zinc chloride thus formed an equal amount of ammonium chloride. Leaks, both in pipes and castings, may be repaired and sepa- rate pieces of pipe may be welded together to form continuous pipes, coils and headers, by means of improved processes of electric and oxyacetyline welding. Ammonia receivers, as well as larger shells such as are used for constructing absorbers, condensers and generators of absorption machines, are also made by this process. Charging a Refrigerating System After it h^s been found that the system is perfectly tight, the air and the ammonia should be allowed to escape, after which the whole system should be pumped down to a vacuimi. Even pumping a vacuum does not insure the expulsion of all the air, but it becomes greatly rarified as it expands under the reduced pressure and the remainder may be allowed to stay in the system until displaced by purging at the condensers, or a large percentage of it can be driven out of the system by the judicious manipulation of the ammonia. If, for example, the ammonia be admitted very slowly at one end of a long run of pipe, it will drive Digitized by V^OOQIC INSTALLATION AND OPERATION 103 the air before it without mixing with it to any great extent, and, if at the other end of the pipe line a valve be opened or a flange union be cracked after sufficient ammonia has been admitted to produce a pressure above that of the atmosphere, the air can be allowed to escape until it contains too large a percentage of am- monia, when the opening is closed. To initially charge or recharge the system, connect the shipping drums of anhydrous anmaonia, one at a time (or more if the plant is of large capacity or the initial charge is being put in and one wishes to save time) to the charging valve usually placed between the master expansion valve on the liquid line, where it leaves the receiver, and the expansion coils or brine cooler. This connec- tion is most easily made by a special fitting built up with two swing joints, one end threaded to fit the valves on the shipping drums and the other provided with a flanged or threaded end to connect to the charging valve. When the connection has been made the air in the pipe may be expelled by slightly opening either the charging or the shipping-drum valve and loosening the flanged swing joint nearest the opposite end. The connection having been carefully made, the main valve on the receiver is closed and the low-pressure side is "pumped down" by allowing the compressor to continue operation after the liquid has been shut off. By the "pumping down" process the ammonia in the expansion side of the system is compressed and discharged into the compression side of the system, where it is condensed and flows to the liquid receiver which it may fill as well as the lower pipes of the condenser. When the low-pressure gauge indicates that the pressure in the expansion coils has been reduced to zero, or atmospheric, pressure, the charging valve may be opened wide and then the valve on the shipping drum may be " cracked," allowing a small stream of the liquid to poiu into the system. The valve on the drum virtually becomes the expansion valve of the system and its manipulation should be governed by the same rules that govern the other expansion valves when the machine is in normal operation, except that it is better not to carry the back pressure quite as high as usual. This pressure may be anything above atmospheric, but there is an advantage in not reducing the pressure below atmospheric as the vacuum would tend to draw air into the system through the charging connection when the Digitized by V^OOQIC 104 ELEMENTARY MECHANICAL REFRIGERATION drum is disconnected if the charging vaJve is not absolutely tight, and a considerable inrush of air is obvioudy less easily detected than a very slight leak of ammonia outward. While the pro- duction of a lower pressure within the refrigerating system, than that of the atmosphere without, undoubtedly hastens the opera- tion of charging, there is always the tendency to draw air or water into the system. A vacuum should, accordingly, never be pumped until it has been demonstrated beyond all reasonable doubt that the system contains no leaks. Small leaks into the system are not readily detected, and it is evident that much more trouble can be made by what water a small leak will let into a system than by the amount of ammonia the same leak will let out. When an open connection is made between the shipping drum and the system, the hquid is forced out of the drum into the system by the pressure of the gas above the liquid just as water is forced out of the blowoflf of a boiler by the steam pressure above the water. The only difference is that it requires a higher tempera- ture than that of the atmosphere in the engine room to raise steam pressure, while any temperature above zero will give a pressure above atmospheric in the case of ammonia. This is made mechanically possible by the construction of the shipping drum valve, which, after passing through the head, turns down to within about half an inch of the side of the cylinder. When, as is advisable, the opposite end of the drmn is elevated a few inches, there remains only a very small volume of the drum below the level of the outlet, and this has its advantage in that heavier impurities tend to remain in the drum instead of passing into the system. That the liquid ammonia will pass from the shipping drum into the system without the necessity of pumping is evident. If the engine-room temperature be 80° Fahrenheit, for example, the pressure in the drimis will be 140 pounds gauge, or there will be 140 pounds difference in pressure between the anmionia and the atmosphere to cause the flow. If the drums are exposed to the sun, the temperature may rise much higher than that of the sur- rounding air and even dangerous pressures may result. It is accordingly advisable to store ammonia drums in a cool place. The shipping drimis are designed to carry any reasonable pressures, but there is a remote possibiUty that the drmn may be filled too full. Digitized by V^OOQIC INSTALLATION AND OPERATION 105 In this case, since there is not sufficient vapor space to take care of the expansion of the Uquid as the temperature increases, and since liquids are practically incompressible, there is no limit to the amount of pressure that may be produced except that of the ultimate strength of the drum. There is the same danger in tightly closing the valves on all the outlets to the Uquid receivers when it is not definitely known that they are not completely filled with liquid. Explosions due to such causes are second only to boiler explosions in their disastrous results. It is obvious that the vapor generated in the drum will drive the liquid out into the system so long as the temperature of the liquid is such as to produce a gas pressure higher than that in the system. If, for example, the system is operating under 15 pounds gauge back pressure, 16 pounds vapor pressure in the drum would suffice to expel the liquid. The temperature of the liquid corre- sponding to a pressure of 16 pounds is about 0® Fahrenheit. From this it will be seen that the only disadvantage of charging against back pressure is that the liquid will not flow so rapidly into the system because of the decreased difference in pressure. The slight- est reduction in pressure within the drum, due to a removal of part of the liquid, causes the ammonia to boil more vigorously, generating more vapor to fill the increasing space above the liquid. The temperature at which the liquid boils gradually drops, how- ever, until at a pressure of about 47 pounds, which corresponds to a temperature of a little less than 32° Fahrenheit, the pipe lead- ing from the drums will become sufficiently cold to precipitate and congeal atmospheric moisture, and is said to "frost." The melt- ing of this frost when the pressure is 47 pounds or less indicates that th^e is no more ammonia passing through the pipe. The valves can then be closed, the empty drum removed and a full drum connected. Amount op Ammonia Charge It is easier to form an opinion as to the amount of ammonia that the system needs while it is operating than it is to determine when a sufficient amount has been added. Except in the case of initial charges, it is better to add a comparatively small amount of ammonia and then to operate the system for a sufficient length of time to restore normal conditions. The height of the liquid in the gauge glass of the receiver, or the general performance of the plant when no gauge glasses arc used, will give the engineer an Digitized by V^OOQIC 106 ELEMENTARY MECHANICAL REFRIGERATION idea as to whether or not more ammonia is required. It should be remembered that refrigeration is produced by the absorption of the heat required to change the liquid ammonia to a gas, and since it takes only a very small amount of heat to raise the tem- perature of any gas that passes the expansion valve in company with the liquid, little cooling effect can be expected from the gas. So far as the production of cold is concerned, there need be only sufficient liquid refrigerant to insure a solid stream at the expansion valves, so that no gas may enter the expansion coils at any time. The passage of gas can be readily recognized by the intermittent hissing sound produced by the passage of quantities of liquid and gas. The condition in which there is just enough liquid to give a solid flow at the expansion valve is the minimum charge that can be economically employed. A lesser quantity must result in loss of both capacity and efficiency. To provide for unforseen contingencies, such as losses of ammonia through leaks, temporary trapping of liquid in low parts of the system, etc., it is always expedient to have the liquid charge somewhat in excess of this amount, a kind of credit balance in the bank to guard against the embarrassment of overdrawing one's account if collections do not come in from the expansion coils, condensers, etc., as expected. Increasing the charge of anhydrous ammonia above that actu- ally required to insure an uninterrupted flow at the expansion valve, will work no harm to the system up to the point where the liquid fills the receiver and begins to encroach upon the condensing surface. To be sure, the more liquid lying in the compression side of the system under the usual conditions of operation, the less space there will be for the storing of additional anhydrous am- monia should it become necessary to "pump out'' the low-pressure side. The additional ammonia occasions an additional investment, but in the majority of plants it is expedient to carry a small stock of liquid to provide for contingencies; and aside from the rather remote possibility of an accident that would result in the loss of the entire charge, it is better to have the ammonia in use in the system than lying idle in shipping cylinders. An overcharge of ammonia in a system can usually be detected in two different ways. Since the condensed liquid soon becomes several degrees colder than the uncondensed gas, the parts of the compression side that contain liquid ammonia, whether the liquid Digitized by V^OOQIC INSTALLATION AND