The case about to be described concerns a steam plant in which there were seven cross-compound condensing Corliss engines, and two Curtis steam turbines. The latter were each of 1500-kilowatt capacity, and were connected to surface condensers, dry-vacuum pumps, centrifugal, hot-well and circulating pumps, respectively. In the illustration (Fig. 76), the original lay-out of piping is shown in full lines. Being originally a reciprocating plant it was difficult to make the allotted space for the turbines suitable for their proper installation. The trouble which followed was a perfectly natural result of the failure to meet the requirements of a turbine plant, and the description herein given is but one example of a great many where the executive head of a concern insists upon controlling the situation without regard to engineering advice or common sense.

Circulating Pump Fails to Meet Guarantee

Observing the plan view, it will be seen that the condensers for both turbines receive their supply of cooling water from the same supply pipe; that is, the pipes, both suction and discharge, leading to No. 1 condenser are simply branches from No. 2, which was installed first without consideration for a second unit. When No. 1 was installed there was a row of columns from the basement floor to the main floor extending in a plane which came directly in front of the condenser. The column P shown in the plan was so located as to prevent a direct connection between the centrifugal circulating pump and the condenser inlet. The centrifugal pump was direct-connected to a vertical high-speed engine, and the coupling is shown at E in the elevation.

Every possible plan was contemplated to accommodate the engine and pump without removing any of the columns, and the arrangement shown was finally adopted, leaving the column P in its former place by employing an S-connection from the pump to the condenser. It should be stated that the pump was purchased under a guarantee to deliver 6000 gallons per minute under a head of 50 feet, with an impeller velocity of 285 revolutions per minute. The vertical engine to which the pump was connected proved to be utterly unfit for running at a speed beyond 225 to 230 revolutions per minute, and in addition the S-bend would obviously reduce the capacity, even at the proper speed of the impeller.

Besides these factors there was another feature even more serious. It was found that when No. 2 unit was operating No. 1 could not get as great a quantity of circulating water as when No. 2 was shut down. This was because No. 2 was drawing most of the water, and No. 1 received only that which No. 2 could not pull from the suction pipe A. This will be clear from the fact that the suction and discharge pipes for No. 1 were only 16 inches, while those of No. 2 were 20 inches and 16 inches, respectively. The condenser for No. 2 had 1000 square feet less cooling surface than No. 1, which had 6000 square feet and was supplied with cooling water by means of two centrifugal pumps of smaller capacity than for No. 1 and arranged in parallel. These were each driven by an electric motor, and were termed "The Siamese Twins," due to the way in which they were connected.

The load factor of the plant ranged from 0.22 to 0.30, the load being almost entirely lighting, so that for the winter season the load factor reached the latter figure. The day load was, therefore, light and not sufficient to give one turbine more than from one-fourth to one-third its rated capacity. Under these conditions No. 1 unit was able to operate much more satisfactorily than when fully loaded, because of the fact that the cooling water was more effective. This was, of course, all used by No. 1 unit when No. 2 was not operating. At best, however, it was found that the vacuum could not be made to exceed 24 inches, and during the peak, with the two turbines running, the vacuum would often drop to 12 inches. A vacuum of 16 inches or 18 inches on the peak was considered good.

An Investigation

Severe criticism "rained" heavily upon the engineer in charge, and complaints were made in reference to the high oil consumption. An investigation on the company's part followed, and the firm which furnished the centrifugal pump and engine was next in order to receive complaints. Repeated efforts were made to increase the speed of the vertical engine to 285 revolutions per minute, but such a speed proved detrimental to the engine, and a lower speed of about 225 revolutions per minute had to be adopted.

A thorough test on the pump to ascertain its delivery at various speeds was the next move, and a notched weir, such as is shown in the elevation, was employed. The test was made on No. 2 cooling tower, not shown in the sketch, and showed that barely 3000 gallons per minute were being delivered to the cooling tower. While the firm furnishing the pump was willing to concede that the pump might not be doing all it should, attention was called to the fact that there might be some other conditions in connection with the system which were responsible for the losses. Notable among these was the hydraulic friction, and when this feature of the case was presented, the company did not seem at all anxious to investigate the matter further; obviously on account of facing a possible necessity for new piping or other apparatus which might cost something.

Approximately 34 feet was the static head of water to be pumped over No. 2 cooling tower. Pressure gages were connected to the suction, discharge, and condenser inlet, as shown at G, G' and G'' respectively. When No. 1 unit was operating alone the gage G showed practically zero, indicating no vacuum in the suction pipe. Observing the same gage when No. 2 unit was running, a vacuum as high as 2 pounds was indicated, showing that No. 2 was drawing more than its share of cooling water from the main A and hence the circulating pump for No. 1 was fighting for all it received. Gage G' indicated a pressure of 21 pounds, while G'' indicated 18.5 pounds, showing a difference of 2.5 pounds pressure lost in the S-bend. This is equivalent to a loss of head of nearly 6 feet, 0.43 pound per foot head being the constant employed. The total head against which the pump worked was therefore

feet approximately. Since the static head was 34 feet, the head lost in friction was evidently

per cent., approximately.

Supply of Cooling Water Limited

In addition to this the supply of cooling water was limited, the vacuum being extremely low at just the time when efficient operation should be had. The natural result occurred, which was this: As the load on the turbine increased, the amount of steam issuing into the condenser increased, beating the circulating water to a temperature which the cooling tower (not in the best condition) was unable to decrease to any great extent. The vacuum gradually dropped off, which indicated that the condenser was being filled with vapor, and in a short time the small centrifugal tail-pump lost its prime, becoming "vapor bound," and the vacuum further decreased. The steam which had condensed would not go into the tail-pump because of the tendency of the dry-pump to maintain a vacuum. When a certain point was reached the dry-vacuum pump started to draw water in its cylinder, and the unit had to be shut down immediately.

Vapor-bound Pumps

As the circulating water gradually rose in temperature the circulating pump also became "vapor bound," so that the unit would be tied up for the rest of the night, as this pump could not be made to draw hot water. The reason for this condition may be explained in the following way. When the circulating pump was operating and there was a suction of 2 pounds indicated at G, the water was not flowing to the pump of its own accord, but was being pulled through by force. This water would flow through the pump until a point was reached when the water became hot enough to be converted into vapor, this occurring at a point where the pressure was sufficiently reduced to cause the water to boil. Naturally this point was in the suction pipe and vapor was thus maintained behind the pump as long as it was operating. In this case the pump was merely maintaining a partial vacuum, but not drawing water. After the vacuum was once lost, by reason of the facts given, it could not be regained, as the circulating water, piping and condenser required a considerable period of time in which to cool.Before any radical changes were made it was decided that a man should crawl in the suction pipe A, and remove such sand, dirt, or any other obstacles as were believed to cause the friction. After this had been done and considerable sand had been removed, tests were resumed with practically the same results as before. The investigation was continued and the dry-vacuum pumps were overhauled, as they had been damaged by water in the cylinders, and furthermore needed re-boring. In short, the auxiliaries were restored to the best condition that could be brought about by the individual improvement of each piece of apparatus. As this was not the seat of the trouble, however, the remedy failed to effect a "cure." It was demonstrated that the steam consumption of the turbines was greatly increased due to priming of the boilers, as well as condensation in the turbine casing; hence, the ills above mentioned were aggravated.

Changes in Piping

After a great deal of argument from the chief engineer, and the firm which furnished the pump, both making a strong plea for a change in the piping, the company accepted the inevitable, and the dotted portion shows the present layout. The elbow M was removed, and a tee put in its place to which the piping D was connected. The circulating pump was removed to the position shown, and a direct connection substituted for the S-bend. The discharge pipe C was carried from No. 1 unit separately, as shown in the elevation, and terminated at No. 1 cooling tower instead of No. 2, which shortened the distance about 60 feet, the total length of pipe (one way) from No. 1 unit being originally 250 feet. In this way the condensing equipment was made practically separate for each turbine, as it should have been in the first place.

With the new piping a vacuum of 24 inches on the peak could be reached. While this is far from an efficient value, yet it is better than the former figure. The failure to reach a vacuum of 28 inches or better is due primarily to a lack of cooling water, but an improvement in this regard could be made by reconstructing the cooling towers, which at present do not offer the proper amount of cooling surface. The screens used were heavy galvanized wire of about 3/16-inch mesh, which became coated in a short time, and must be thoroughly cleaned to permit the water to drop through them. The supply of cooling water was taken from a 30-inch pipe line several miles long and fed from a spring. The amount of water varied considerably and was at times quite insufficient for the load on the plant. Instead of meeting this condition with the best apparatus possible, a chain of difficulties were added to it, with the results given.