The details of the work involved in the substitution by the ship's force of the Arkansas, of an electric motor for the wrecked starboard main circulating pump, have been here assembled for the reason that, while they demonstrate nothing new, at the same time they involve certain principles of electrical and mechanical engineering in a most instructive manner. In addition this work has emphasized one fundamental principle, namely, that all apparatus delivered to the service should be as near the commercial standard as possible, and that service operation should conform to the commercial standard, so that in emergency the resources of the whole country, rather than a limited reserve, are available for use. This point was demonstrated in the attempt to obtain a suitable motor from shore. Whereas most commercial motors are built for a voltage of 220-230, the navy power circuits carry 125 volts, so that it was impossible to obtain on short notice a motor fulfilling the requirements. Instead it became necessary to adapt one of the ship's motors to the service, and take extraordinary precautions to maintain it in operation.
The main circulating pumps of the Arkansas are of the centrifugal type, and are driven by impulse turbines manufactured by the Terry Steam Turbine Co. of Hartford, Conn. The turbines developed about 75 I. H. P. at 600 R. P. M. when driving the pumps. Steam is taken from the auxiliary steam line at boiler pressure and expanded in the nozzles down to the back pressure of the auxiliary exhaust, about 15 lbs. gage. Consequently the casing is not ordinarily subject to greater pressure than this back pressure and is designed on this basis. A relief valve, 3½ inches in diameter, is fitted to the exhaust side of the casing and set to lift at 20 lbs. gage. At the time of the accident the pumps were being secured. The machinist had closed the throttle by hand and, upon seeing the pump stop, had closed the exhaust. Seeing steam leaking from around the glands, he started to close the throttle with a wrench, but before he could do so, the turbine exploded, blowing out the exhaust side of the casing as shown in the photograph. The relief valve had been frequently lifted by steam before the accident and afterwards relieved at a pressure of 20 lbs. gage.
Examination of the turbine showed it to be damaged beyond repair by the ship's force, though the casing itself was the only part rendered unfit for use. Steps were immediately taken to rig some sort of substitute, and, in the meantime, the two condensers were cross-connected by way of the main drain so that the one circulating pump remaining could serve both condensers. This method limited the ship's speed to about 10 knots, with circulating water at the temperature of that locality, about 74° F., though this speed could have been increased a knot or two with the colder water of higher latitudes.
Two alternatives presented themselves, (1) to drive the pump by chain or gearing from the main turbine shaft adjacent, and (2) to remove the old turbine and install a motor in its place. Of these (2) seemed the simplest and most reliable, provided a suitable motor could be had. Inquiry at the navy yard at Norfolk developed the fact that while no motor at 125 volts was available, there were several at 220 volts which might have been used. In order to use these however, it would have been necessary to alter the armature circuits of one main generator, connecting the two windings in series rather than in parallel as they ordinarily are, so that the terminal voltage would be 250 instead of 125 volts. This method of operation was not attractive one for the reasons that one generator would be required for the service of the pump alone, and that a separate circuit would be necessary from the dynamo rooms to the motor.
Since no suitable motor was available from an outside source, it was decided to take one from the ship itself. Investigation developed that the largest motor available was rated at 35 H. P., or about half that of the turbine. There were three types of motors of this capacity: (1) a shunt motor built for continuous service at 900 R. P. M., (2) a series motor built for intermittent service at 350 R. P. M. and a compound motor for intermittent service at 350 R. P. M. The first belonged to a forced-draft blower, and was eliminated for the reason that 350 revolutions of the pump required the full power of the motor, so that it would draw an excessive armature current in its attempt to maintain rated speed. The series motor might have served, but it belonged on a boat crane and was very necessary to the ship for this service. The compound motor belonged to a deck winch aft, where it was speed variation were necessarily restricted. It was also easily accessible and was therefore chosen for the duty.
The electrical characteristics that entered into the choice are shown in Fig. 3. Here are plotted (not to scale) the curves of speed versus armature current of the different types of motor. It was estimated that about 350 R. P. M. of the pump would require 35 H. P. output. At 125 volts this would call for an input of 245 amperes with an assumed motor efficiency of about 85 per cent.
From the curves it is seen that the shunt motor was eliminated right here because at 245 amperes it would ordinarily run at 900 R. P. M. However, when the pump was driven at 350 R. P. M. it would require the full motor horsepower. As a result of the low speed the counter E. M. F. would be so small as to permit an excessive armature current to flow, burning out the motor, or blowing the safety devices. In addition to this it is evident that even if the motor had been able to carry this current, the shafting would have been over stressed, because of the greater torque required to transmit the same power at the lower speed. As for the other two it is seen from the curves that both the series and compound motors were adapted, as their rated speeds and armature currents corresponded to the speed and H. P. of the pump. The disadvantage of these two motors, however, lay in the fact that they were designed to carry full-load current for intermittent service only, while in the new duty they must operate continuously. However, since they were fully enclosed on deck it seemed reasonable to suppose that if opened up and properly ventilated below so as to get the heat away, they would be able to stand the load without burning. On this assumption, the deck-winch motor was removed with its controller, thoroughly overhauled and started below.
In the meantime the turbine half of the flexible coupling for the pump was bushed and fitted to the armature shaft so as to render the coupling up easy. At the same time the carpenter's gang had begun the installation of a wooden foundation of 6 x 6 timbers on the old foundation of the circulating pump which had been removed. This foundation was set up as nearly in place as could be from measurements of the motor, and certain holes in the planking were slotted so as to permit of adjustment fore and aft and athwartships after the motor was located in place. Adjustment vertically was had by leaving the top timber a little low and shimming up with strips of sheet iron. This foundation seemed amply solid as the vibration from both motor and pump was negligible. Since the pump itself had two rotors on the same shaft taking water from both ends, the end thrust was balanced and there was no need of a thrust bearing or of any stiffness to take up thrust. The long bolts were made in the shop while the motor was being sent below, so that by the time the motor was ready its foundation was in place.
The repairs to the motor consisted of the renewal of a shunt field coil, both bronze bearings and the armature, together with necessary varnishing and cleaning. The object of this work was to give the motor every chance to stand up under the work it was to do, and the necessity for these precautions became very evident when the motor was put in operation. At the same time it was much handier to send it below in sections as the distance from its old location to the new one was quite long. While the motor was being assembled in place the electricians were installing the starting rheostat and connecting the leads to the circuit for number six turret which passed through the wiring passage close by. A voltmeter and ammeter were installed in a convenient place, so that by the time the machinists had made the alignment everything was ready for operation.
In aligning the motor and pump use was made of the large flexible coupling. By means of the slotted bolt holes the motor was moved about until the flanges on the two shafts were parallel and their edges in line. After turning the motor over free to check the alignment, the coupling was bolted together and the motor started up. Predictions as to speed and load were immediately exceeded, as the pump was driven at 400 R. P. M. with 250 amperes at 120 volts, but the motor promptly began to heat. Inasmuch as the armature current was quite large it was assumed that the heating was due to this, but it soon became apparent that the trouble was in the bearings. Investigation showed that the new armature shaft had been made with the collars, which limited the end play, in such a position that there was no end play of the armature permitted. This was overcome by putting a gasket between the shields which held the bronze bearings and the casing so that the end play became about of an inch. This relieved the difficulty when running free, but as soon as the pump was coupled up the bearing nearest the coupling ran warm again and the whole motor became quite hot. It was then observed that in drawing up the coupling the shoulder on the shaft was drawn hard against the bearing and that this was the source of heat. The coupling was then broken and the motor shifted toward the pump until, when the coupling was drawn tight, the armature end play was halved between bearings. As a result of this, the armature, while running with full load, worked fore and aft under the flexible coupling, so that it was not only free in its bearings but permitted a certain variation in the brush position on the commutator which was of value in keeping the latter cool. This eliminated all bearing trouble.
In the meantime every effort was being made to keep the motor cooled off. The commutator end of the armature was cooled by a blast of air from the ventilating system, which was led in by a canvas chute. The forward bearing was under a blast of the same sort. At both ends the blast as first installed blew directly across; but when this failed, holes were drilled in the forward-bearing housing, and the air which went in at the commutator end passed the length of the armature and out at the forward bearing where it met the second blast and was carried away. This of course made the air about the bearing very warm at the same time it was rubbing on the shoulder of the shaft. When, therefore, the adjustment was made for halving the end play, the process was reversed. The chute on the forward end was secured to the shield so that all its air was forced through the holes toward the commutator, where the blast from the second chute, sweeping across, carried it away. This was a fairly satisfactory arrangement, but even with this, though the bearings remained cool, the heat was not getting away from the field coils and armature. A third air chute was then made to deliver a blast directly across the casing at its center, and while this helped on one side the other was still warm.
Finally it was decided to resort to water cooling and this was accomplished in a very simple way. First all possible leaks to the casing were plugged, then a ridge of putty was placed around the casing at each end. Next the intervening space was wrapped with a few turns of burlap and a small pipe was led from a handy salt-water pipe with a valve directly over the center of the motor. From this source the burlap bagging was saturated, and as the air blast swept across the motor, the water was quickly evaporated. The heat necessary to cause the evaporation was taken from the casing itself, so that a rapid flow of heat from the coils and armature was established. Of course the latent heat of evaporation being large, a much smaller amount of water was required than if the water simply ran over the casing and increased its sensible heat. It was possible, if desired, to so regulate the rate of flow to the casing that it was equal to the rate of evaporation, and thus prevent water from dripping to the deck. At any rate, the motor promptly cooled down and remained cool thereafter. Thermometers had been installed, one in each bearing as shown in the pictures, and one in the path of the air discharged across the commutator. While these were not located to give accurate absolute results, still they were a good indication of the conditions at each place and thus were guides for operation.
When the installation was completed and found running well, an electrician's watch was established on the motor and a test begun. From the curves, Fig. 6, it is seen that the motor gradually increased its temperature along a typical warming curve; then settled down and maintained this temperature. The final temperature was that at which the difference between the motor and the outside air was such that the heat was radiated as fast as it was generated and equilibrium was maintained. The actual temperatures then were dependent upon those of the air supply, and it was found that the curve of temperatures followed that of the outside air while cruising. When the ship moved to northern waters, then, motor temperatures went down promptly, and no more trouble was had from this source.
The change of latitude, of course, had the same effect on the vacuum. The motor, operating at 400 R. P. M., was only 100 turns slower than the turbine on the other side. For a given speed the vacuum depends upon the temperature of injection, since the same quantity of steam condensed causes the same temperature rise and the higher the injection the higher the final temperature and the lower the vacuum according to the steam tables. It was expected that the temperature of the circulating water discharge would limit the speed of the ship, since the temperature of injection was high and the quantity of water circulated by the motor-driven pump was less than that of the turbine-driven unit. Were the temperatures allowed to become too high; damage might result to the condenser. At just what speed the temperatures would become too great was hard to estimate, but a glance at the trial data showed a 40° temperature rise at 20.5 knots with 600 R. P. M. of the circulator. Assuming the rise to be directly proportional to the R. P. M., it appears that 60° might be expected. However, it is well known that the frictional resistance in pipes varies as the square of the velocity of the fluid, so that at a rotor speed of two-thirds the trial speed, something more than two-thirds the quantity of water might be expected, resulting in a smaller rise than 60°. Again, since the power required per knot at high speeds may go up as the fifth power of the speed, it seemed likely that reducing the speed to 20 knots would reduce the rise very appreciably. Finally, if necessary, it was possible to make an average of 20 knots by making turns for 19 knots on the motor side and 21 knots on the other. Assuming a temperature rise of 50°, which was hardly to be expected, and placing the limiting temperature at 120°, which seemed safe, it was predicted that the ship could maintain 20 knots with an injection temperature of 70°, and her maximum speed of 21.50 with an injection of 55°.
Shortly after the motor was installed the ship began operations in connection with the strategical problem, but was not called on to make more than 17 knots before its completion. While proceeding to port, however, the commanding officer, desiring to determine exactly what could be done, ordered a two-hour, full-power trial. By this time the motor had settled down to a steady temperature difference above the outside air, and was running well at about 395 R. P. M. Watching the discharge temperatures carefully, the speed was increased to 20 knots, and maintained for two hours, with a temperature rise of about 44° on each side, the port shafts making about 5 R. P. M. more than the starboard. At the end of the two hours the throttles were opened wide for 10 minutes, during which turns for over 20½ knots were maintained on both engines, with a temperature rise of 50° in the starboard condenser. This confirmed the estimate based on the law for friction in pipes, and since by this time the ship was in the latitude of New York and the injection was 54°, the final temperature was but 104, leaving a wide margin had it been necessary to work up to the maximum speed. Upon arrival in port the motor was shut down after over five days' continuous operation, being in condition to continue indefinitely.
The motor having proved successful, there were consequently many people who advocated the installation of motors for this service in place of the old turbines. Against this proposition there were two important arguments. In the first place the engines became dependent upon the operation of a generator at a distance from them, over which the engine-room force had no control. The blowing of a circuit breaker or an accident to the generator would stop the flow of water through the condenser, and such an accident while operating at high powers might result in burning up the condenser before the engines could be stopped. To avoid this, two generators were kept in operation and a careful watch was stood at all times. A signal was arranged between the pump and throttle, so that if the motor stopped for any reason, the officer of the watch would be warned in time to shut down the engines and save the condenser. If through design or accident the power were shut off the motor circuit and the motor thus stopped, then before the operator could throw the rheostat to "off," the power were cut in again, there would be a dead short across the armature which would draw a large current and at best blow all the safety devices in the circuit. Against this the operator was carefully instructed and drilled. Unreliability, then, militates against the motor.
Even if the motor were practicable it would not be desirable. It is easily estimated that the motor would require about half as much steam as the turbine, and thus have the advantage. However, the steam leaving the pump turbine is not wasted. Part of it goes to heat the feed water, and the rest does work in the lower stages of the main turbines. It is likely that a pound of steam that enters the circulator turbine and expands down to the back pressure of the auxiliary exhaust, then passing to the main turbines expands again to a very small back pressure in a turbine capable of utilizing a big vacuum, does quite as much work as another pound that goes through the turbine of a 300 K. W. generator operating off its best load. Each of these pounds gives up its latent heat of condensation to the circulating water where it is lost. That portion of the steam that goes to heat the feed water, however, does a certain portion of its work in the turbine; then passing to the feed heater, gives up its latent heat of condensation to the feed, whence it is returned to the boilers and saved. These facts account for the general principle often enunciated in power-plant practice that economical auxiliaries which rob the feed water of heat it might use, are not economical at all.
SUMMARY
The above description of the installation shows the following general principles to have been utilized: (1) The electrical characteristics of the motors available were utilized in the choice; (2) the type of service (motor drive without vibration or end thrust) called for a simple wooden foundation, and this was made adjustable; (3) a differentiation of the sources of heat made possible the elimination of one, the mechanical; (4) the motor was operated at a large overload by getting away the heat, (a) by air cooling and (b) by water cooling, and in the latter the large latent heat of evaporation was utilized in place of the small sensible heat; (5) the whole plant was set up as if for a laboratory test, a log was kept, and the data used to produce the best operation; (6) using all the information available it became possible to predict accurately the possibilities of the plant, then to confirm the prediction by actual results.
CONCLUSIONS
In general two conclusions can be drawn from the case: (1) an impulse turbine, since it is not designed to withstand all the stresses to which it may be subjected either by design or accident, must be well protected by safety devices, and the same care must be given these devices as is bestowed upon boiler safety valves, including routine lifting by hand and steam; (2) were the wiring of ships installed similar to that ashore, where by means of a three-wire system 110 volts can be had across two wires for lighting and 220 volts across two others for power, and were the appliances made to conform closely to a commercial standard, then the resources of the whole country would form a reserve from which to draw equipment in times of stress.