Mr. Emmet presented the following paper:

RECENT STEAM TURBINE DEVELOPMENIS

From a consideration of the general principle and from some of the literature that has been produced on the subject, it might be supposed that the design of steam turbines was a very simple matter and that many good designs simply awaited the capital for their commercial production. In reality the design of turbines involves very great practical difficulties, and is in its nature dependent upon very difficult experimenting. Steam is an elastic fluid, carrying much energy, which is convertible into heat, pressure or velocity. In passing through the small spaces of a turbine at lightning speed, it is subject to eddyings, vibrations, condensations, rarefactions and re-evaporations which may have radical effect upon the efficiency. The value of fuel and power is such that in our day inefficiency in prime movers can not be tolerated, and if we are to build engines at all, we must build good ones.

In a turbine of the Curtis type the necessity for elaborate experimental development is particularly urgent; its fundamental process being the complex action of a single steam impulse upon two or more moving blades, while in other turbines that have been produced each row of blades has a relatively independent action.

These conditions were clearly understood when production of these machines was begun, and it was not expected that the type would at once attain its highest development. At that time. other forms of steam turbine were imperfectly developed, and there was demand for such undertakings as were fully justified by our early experiments. All of our early experimenting was done with small apparatus made for the purpose, and all the Curtis turbines installed prior to the last six or eight months. were designed from data obtained from this experimental apparatus. These machines are of several successive types and have generally shown improvement in the order of their production.

The designs of the improved machines now being produced have not been simply dependent upon data obtained from small experimental apparatus, but are largely the result of analysis of performances of some of the earlier large machines. It is

only within recent months that the results of this actual experience with large commercial apparatus have taken material form.

Aims

The main objects sought in the development of the Curtis turbines have been: High steam economy; the use of a small number of rugged parts instead of a very large number of fragile ones; the elimination of all small clearances; the economic recovery of the large energy available in high vacuum, and the production of flat load curves, particularly with reference to high overload economy. The purpose of this paper is to show the progress that has been made in pursuit of these aims. The test results given in this paper do not all apply to machines of our latest types, but they show good progress in the desired directions, and a satisfactory degree of consistency of results.

Small and Large Machines

In studying the tests here given, it will be noticed that the steam consumption of smaller machines is relatively large. This is a necessary condition in all steam turbines, unless the speed can be increased in almost inverse ratio to the capacity. The natural limitations of alternating-current frequency and the desire for mechanical reliability have led us to adopt very moderate speeds for smaller machines, and thus to sacrifice degrees of steam economy which might easily be attained if we used, for the same sizes, speeds as high as those adopted by makers of turbines of other types. In spite of this disadvantage, our results in smaller sizes are relatively good, and serve to demonstrate the flexibility and excellence of the Curtis idea.

TESTS OF 9000-KW TURBINE UNIT IN CHICAGO

During the past year four 9000-kw units have been installed in the Fisk street station of the Commonwealth Electric Company of Chicago. One of these machines has been tested at different loads by engineers employed for the purpose by the owners. These tests were part of a general investigation of station conditions, which has not yet been published. They were conducted with great care and were repeatedly verified before being accepted. Their accuracy is also proved by the results of 5000-kw tests as explained below. The following figures give the results of five tests that were made:

TESTS OF 9000-KW TURBINE GENERATING UNIT AT CHICAGO

The results of these tests reduced to a common condition are shown in the accompanying curve, Figure 1. The steam economy shown by these tests we believe to be higher than any previous result within the same ranges of temperature and pressThis machine at full load with a 28-inch vacuum delivers

ure.

FIG. 1-9000-KW, 5-STAGE CURTIS TURBINE, 750 R.P.M.

LOAD-WATER RATE

CURVE-200 POUNDS GAUGE PRESSURE, 125 DEGREES SUPERHEAT, 29-INCH VACUUM

This curve is derived from the test results given above, corrected to conditions of contract under which the machine tested was sold

to the switchboard 21 per cent of the total energy in the steam and 67.3 per cent of the total available energy in the steam. It delivers to the shaft 70 per cent of the total available energy in the steam. It is interesting to consider what relation this result may bear to the ultimate possibilities of turbine improvement, and we have some information from which conceptions on this subject can be formed. The simplest application of the turbine idea is found in the delivery of steam through a straight nozzle against a single row of moving blades having a velocity best suited to the speed of issuing steam. Such simple arrangements have been tested, and efficiency curves have been traced with friction losses completely eliminated, and with all degrees of . velocity in steam and in wheels. Many such investigations have

been made in Schenectady, and similar investigations on elaborate scale have been made in Germany. The best bucket efficiency produced in such test has been 75 per cent of the available energy corresponding to the temperature range. Thus in this large commercial machine we get a shaft efficiency of 70 per cent as compared with a bucket efficiency of only 75 per cent in the most ideal turbine arrangement that we have been able to devise. This comparison leads us to expect that we may be nearing the limits of possible improvement in the simple production of power from steam.

Characteristics

An examination of these results will show that they in all respects realize the ideals that have been aimed at in the development of this kind of apparatus. The vacuum produced is almost unprecedented, and the turbine uses efficiently the large increase of available energy which this high vacuum affords. The machine gives a very flat load curve with particularly high relative economy at heavy overload. This feature, as has been explained, is characteristic of the type, and is worthy of special consideration.

Overload Economy

High economy at overload is the direct equivalent of increased peak-load capacity, since it makes it possible to obtain a larger maximum output from a given installation of boilers, plant and auxiliaries, and from a given operating force. If units in one station have a peak-load efficiency 25 per cent higher than those in another similar station, it implies that the peak capacity of one is 25 per cent greater than that of the other.

When the matter is considered in this way, it will be seen that these 9000-kw machines have very great superiority of value as compared with any other large prime movers now in use. The plant in which these units are installed was originally designed for 5000-kw machines, and for each generating unit a battery of eight 520-hp boilers was installed; these batteries of boilers being the equivalent of those which had been adopted in large engine stations built at about the same time. In the test results given in will be seen that the machine tested was operated at 13,900 kilowatts, taking its steam from these eight boilers. That the

firing of these boilers during this run was not abnormally heavy is shown by the fact that the superheat carried amounted to only 140 degrees, while the superheat was higher in other tests, and is known to be higher when boilers are forced. The superintendent of the plant in which these machines are installed has stated to the writer that they could without difficulty supply steam for 15,000 kilowatts if the generators were capable of carrying such a load.

The plant in Chicago in which these machines are installed is the handsomest and most substantial steam plant ever built, and yet if the units were assigned ratings equivalent to those of other prime movers its cost per kilowatt would unquestionably be the lowest that has been produced. Consideration of the economies from this standpoint will show any unprejudiced investigator the wastefulness of using inferior apparatus when such results are obtainable. It will also show the great advantages of power production on a large scale, and will indicate that great changes in existing methods are to be looked for at an early date.

Vacuum

Another feature of these tests which will immediately impress every one who is familiar with steam practice is the extremely high degrees of vacuum that are carried. The machine tested is equipped with a base condenser which forms an integral part of the turbine unit. This condenser contains 25,000 square feet of tube surface with 1-inch tubes. The record shows that there was no appreciable diminution of vacuum with increased load. This high vacuum is due to the free access of steam to the tube surface and to the absence of any air leakage for which this arrangement provides. The value of such high vacuum is very great when the prime mover used is capable of effectively getting the benefit of it. The available energy with a 29-inch vacuum is 10 per cent higher than that with a 28-inch vacuum, and in this machine at full load the gain of water rate accomplished by increasing the vacuum from 28 inches to 29 inches amounts to 7 per cent. The machine was designed for 28.5 inches of vacuum, and between 28 inches and 28.5 inches the gain is very nearly proportionate to theory.