Modern dynamo electric machinery

MODERN DYNAMO ELECTRIC MACHINERY.* BY

ALEXANDER GRAY, M.Sc., Professor of Electrical Engineering, Cornell University, Ithaca, N. Y.

ALT~nATI~G-CURR~nT OEN~.~ATORS. We have already seen that direct-current machinery had become somewhat standardized by I89o, and that the Hobart machine, described on page 22, has design constants and a type of mechanical construction which do not differ much from those of F r o . 32.

Belt-driven single-phase alternator, 90 k.w., I6,OOO alternations.

similar machines of recent date, but at that time, 27 years ago, there was no one type of alternator that was considered superior to all others. The first Niagara machines, shown in Fig. 3, had an external rotating field system, this type of construction being considered necessary to give the large flywheel effect required to maintain * Continued from page 48, July issue. 205

206

A L E X A N D E R GRAY.

[J. F. I.

the speed :approximately constant with changing load. Rotating armature machines, such as that shown in Fig. 32 , were still listed in 19o3, and inductor alternators of the type shown in Fig. 33 were on the market as late as 19o 7, but the advantages of the present type of internal rotating field construction were such that by 19oo most of the other types had been discarded, so that Guilbert, in describing the alternators that had been on exhibit at the Paris Exposition in 19oo. states as follows : " One of the most striking features was the triumph of the three-phase system even for lighting; of equal interest was the Fw,. 33.

Inductor alternator with vertically split armature.

fact, as had been foreseen before the opening of the Exposition, that the inductor type of alternator was being abandoned." Many data have been published on these Paris machines. 1~ F r o m these data, and from some additional inforlnation published by Rothert, 1~ the seven machines in Table III have been selected as typical alternators of 19oo, a;nd these we shall use as a starting point in our study of modern machines. Let its l*L'12clcdrge Electrique, vol. 29, p. 276, November 2,3, 19Ol; Electrical World, vol. 37, PP. 113, 154, 194, 231, 274, 302, 352, 398, January to March, 19Ol.

"~L'l~clairage I~lectrique, vol. 29, p. 307, November 30, 19Ol. See also " Engineering Evolution of Electrical Apparatus," Lamme. Electric Journal, vol. II, p. 73.

Aug., IOl7.] ~'~ODERN I)VNAMO t{LECTRIC 5IACHINERY.

207

first, however, take up the various limitations in the design of polyphase alternators. TABLE Kv.a .................... Volts ................... Amp4res ................. Phases ................... Frequency ............... R.p m ................... Poles .................... I n t e r n a l d i a m e t e r of s t a t o r in inches ............... Frame length in inches .... m. Pole pitch in inches .... n. A i r - g a p 21earance in inches ............... Ratio m/n ............... Slots per pole ............. S i z e of s l o t in i n c h e s . . . . . . Conductors per slot ....... Connection ............... Tooth/slot ............... Maximum tooth density at n o l o a d in l i n e s p e r s q u a r e inch ................... (2ore density at no load .... Pole density at no load .... a. A m p 6 r e c o n d u c t o r s p e r i n c h of p e r i p h e r y . . . . . b. C i r c u l a r m i l s p e r a m p 6 r e , Ratio a/b .................. Peripheral velocity of rotor in feet per minute ...... c. A m p h r e t u r n s p e r p o l e a t no load .............. d. A r m a t u r e a m p 6 r e t u r n s per pole ............. Ratio old ................ Pole enclosure ............. Output factor X ioo .......

215 220 565

800 2200 210

III. 800 2200 210

lOOO 5000 II5

133o 5500 140

14o0 3000 270

3

3

3

3

3

3

3

32 12o 32

42.5 80 64

5° 79 76

50 94 64

50 94 64

25 75 40

5o 72 8,1

lO6 io lO.4

236 IO 11.6

236 lO.7 9.8

197 ii 9-7

230 11.8 11..35

214 17 16.8

252 IO 9.4

.312 37 6 .87 X I . 4 6 delta 1.24

.275 35.5 6 .6 X I . I 5 5 YY 1.72

.312 36 6 I.I dia. 6 Y --

.43 39 6 I X2.7 9 Y 1.8

.235 40 6 .83 X 2 . 2 6 delta I.O

II5,OOO 30,000 118,ooo

lO4,OOO 18,ooo lO9,OOO

77,ooo 84,000 90,000

IOO,OOO 26,ooo 114,ooo

450 620 .72

370 725 .51

450 lO6O .42

570 480 1.18

197 52.5 3 1.85 d i a . 9 YYYY --

lO4,OOO 29,000

118,ooo 365 74 ° 49

11o,ooo 30,000 12o,ooo 375 430 .87

65,000 21,ooo III,OOO 320 450 .71

,275 35 3 1.7 X 2 . 5 7 Y .9

3320

4950

4900

4850

5650

4200

4750

5385

7600

4600

6950

5175

8050

6675

1600 2.87 .8

2200 3 15 .6t 2.15

2100 2.46 .66 1.72

3800 2.12 .65 2. 3

2750 2.42 .64 3 I

1920 2.8 57 1.64

2200

3.45 .5 1.82

2.i2

Volts per phase =2.22 Xconductors Tooth

860 2400 207

p e r p h . )4 e# X f X I O -s

pole×

area

per pole =minimum tooth X slots per net iron in frarae l e n g t h . X qJ Pole density =gap fluxXleakage factor/pole area A m p S r e c o n d u c t o r s p e r i n c h of p e r i p h e r y = c u r r e n t p e r c o n d u c t o r X c o n d u c t o r s p e r p h a s e X p h a s % rr X s t a t o r i n t e r n a l d i a m e t e r A m p e r e t u r n s p e r p o l e a t n o l o a d = o p . P i g . 38. Armature ampi~re turns per pole =current per conductor Xconductors per pole 2 Output factor =volt amp&res rpm XD~L

PERFOIIMANCE CHARACTERISTICs.--Fig. 34 is a section of a three-phase alternator showing the conductors of only one phase. \Vhen this machine is supplying current, these conductors are encircled by the lines of force of the alternating magnetic fluxes VOL.

I84,

No.

11oo--I6

208

ALEXANDER

[J. F. I.

GRAY.

q,x and ¢r. The former, often called the armature leakage flux, generates a voltage of self-induction in the conductors which is proportional to the current and is equal to IX where X, called the leakage reactance, is constant, since the flux g,~ is proportional to. the current which produces it, the reluctance of the FIG. 34.

A r m a t u r e fluxes produced by one phase of a three-phase alternator.

Pla. 35-

N ~

MOTION

A r m a t u r e f i e l d o f a t h r e e - p h a s e a l t e r n a t o r w h e n t h e c u r r e n t lags 90 degrees. P I G . 35 A.

Ptt.l

PH.2 PH.3

Currents in three phases.

leakage path being practically all in the air part of the path. The magnetic flux ~r enters the poles and so modifies the main field. Fig. 35 shows a section through the same three-phase machine and shows the current distribution in the conductors when the power factor of the load is zero and the current lags 9o degrees

Aug., 1917.] MODERN DYNAMO ELECTRIC MACHINERY.

209

behind the terminal voltage. Under these conditions the current is a maximum in the conductors that are between the poles, because in these conductors no electromotive force is being generated. The three fluxes q,,., produced by the three phases, combine to give a gliding field exactly as in an induction motor, which field is proportional to the current flowing and moves at synchronous speed in the same direction as the rotating poles. The phase relation between this armature field and that produced by the poles depends on the power factor of the load and, as shown in Fig. 35, directly opposes the main field at zero power factor. Fi6. 36.

Eo

/

_

~~x F / I

Vector d i a g r a m per phase of a polyphase alternator.

These effects of armature reaction can best be indicated by a vector diagram, as shown in Fig. 36 . The no-load m.m.f. Fo produces an alternating flux in the winding of each phase, and the voltage Eo generated in each phase lags Fo by 9 ° degrees. The m.m.f, of the armature produces a synchronous gliding field of constant magnitude and causes an alternating flux to thread the windings of each phase. It may be seen from Fig. 35 that the armature flux threading phase I is a maximum when the current in that phase has its maximum value; thus in Fig. 36 we have :

2 IO

A L E X A N D E R GRAY.

[J. F. I.

F o is the no load m.m.f. E o is the voltage per phase

due to F o. I is the current per phase. F a is the m.m.f, of armature reaction which, as pointed out in the last paragraph, is in phase with 1. Fg is the resultant m.m.f, due to field and armature. ~, is the re'suiting synchronous flux. Eg is the voltage per phase due to q6m 9!,, is the armature leakage flux per phase. .. I X is the leakage reactance drop. Et, the terminal voltage per phase, ~ E , j - I X - I R taken as vectors. T h e voltage d r o p ce due to the combined effect of leakage reactance d r o p and d r o p due to a r m a t u r e reaction is generally called the s y n c h r o n o u s reactance drop. ]PIG. 37,

FIG. 38.

1.4 L9 CD _J

1.4

]-2

/

~ >"Y.je[.--u

/~.~

> 7i

0.8 o

12

~" -

? 06~ / / -';~'~ 1-0

I

Ea

10 08

?

~i"_:

0.6 F--- 0.4-

F-- 0-4

L)

z.0j~

l 02

',2~.-'" ~:..--T

_i

i

"'o

i

02

1,0~N

I c AMP.TURNS P. POLE

Characteristic curves of an alternator.

"

F0

c

P

q AMETURN5 P. POLE

T e s t c u r v e s of a I s o - k w . ,

600 r.p.m.,

three-phase alternator.

I n the particular case w h e n the p o w e r f a c t o r is zero and the c u r r e n t lags the t e r m i n a l voltage by 9 ° degrees, the leakage reactance d r o p is subtracted directly f r o m the g e n e r a t e d voltage, and the m a g n e t o m o t i v e force of a r m a t u r e reaction is subtracted directly f r o m the exciting m . m . f . T h e v o l t a g e d r o p due to a r m a t u r e reaction, h o w e v e r , decreases as the poles become saturated, because the s a m e m.m.f, produces a g r a d u a l l y decreasing reduction in the flux. T h e s e two effects are well s h o w n in Fig. 3 7 , w h e r e curve I is the no-load saturation, a b is the m.m.f, to o v e r c o m e the d e m a g n e t i z i n g effect of a r m a t u r e reaction, and b e

Aug., I917,] MODERN DYNAMO F.LECTRIC ~[ACHINER¥.

2II

is the voltage drop due to this effect, bc is the leakage reactance drop, and c is a point on the full-load saturation curve at zero power factor. It is found in practice, however, that the two curves are not parallel to one another as shown, but gradually separate as in Fig. 38, which gives test data on an actual machine. The additional voltage drop is due to the pole leakage. The m.ln.f, required to generate a voltage Eo on no load is op, Fig. 38, but, when the alternator is loaded and the power factor is zero, the m.m.f, for the same generated voltage is oq: the flux crossing the air-gap is unchanged, but the leakage flux from pole to pole is increased in the ratio oq/'op, and, under these conditions, an increased excitation is required because of the higher densities in the pole core. The relative amounts of these three effects can be shown as follows: The distance ab, Fig. 38, is determined from the formula : Demagnetizing amp6re turns per pole = o.35 Z lc/p. where Z / p is the total conductors per pole, L. is the current per conductor, and the distance bc must be the leakage reactance drop; this can be checked approximately by calculation. If the triangle abc be now placed so that the full-load saturation curve found by actual test is the locus of the point c, then the locus of point a is what designers call the no-load saturation curve figured with the fullload leakage factor, and the excitation ad is the additional excitation l,~r the poles over that required at no-load for the same air-gap flux. LINITATIONS IN DESmN.--If it is desired to have good regulation, pa.rticularly with low power factor loads, it is necessary to keep the synchronous reactance drop fc, Fig. 38, small compared with the normal voltage. To accomplish this result: I. Keep the leakage factor small and the pole density below saturation. 2. Keep the leakage reactance small : this, as shown in Fig. 39, means limiting the number of amp6re conductors in the phase belt .r or, what is equivalent, limiting the number of amp6re conductors per inch of armature periphery. 3- Keep the demagnetizing amp6re turns per pole a small

ALEXANDER GRAY.

212

[J. F. I.

fraction of the exciting ampgre turns per pole and sa.turate the magnetic circuit so as to make the voltage drop be small. One may well ask why the output of this particular machine is limited to 15o kilowatts. W h y would it not be possible to increase the main excitation and thereby allow a larger current to be carried by the armature without any sacrifice of regulation?

i"x-'i F I G . 39.

1

I

Winding of one phase of a three-phase alternator, showing the leakage flux.

Fm. 40.

¢9

P'C p

ll!r //

\\\

¢.

The main field and the pole leakage field of an alternator.

Fig. 40 shows part of a machine to scale. Of the flux q~p which enters the pole, the leakage flux q~ at no load is 2o per cent., and the useful gap flux Cg is 8o per cent., under which conditions the pole density at the root is ioo,ooo lines per square inch, and the tooth density 95,oo0, while the length of the pole is just sufficient to give the necessary radiating surface to the field coils.

Aug., I9x7.] MODERN DYNAMO ELECTRIC ]~ACHINERY.

2I 3

If, then, the field excitation is increased 5° per cent., the pole length will increase in the same ratio to dissipate the additional heat, the air-gap will be increased to use up the increased m.m.f., since ~g, limited by the tooth density, remains constant, but the area as well as the m.m.f, of the leakage path will be increased, so that the leakage will become (1.5) ~, or 2.2 5 times the original value, and the flux density at the root of the pole will be excessive. FIG. 41 . 120C 110C 100C 90C .wz

600 7OO

cO b-

600

2 500 7" 0

400 500

"< 200 100

l

2

5

~- 5 6 7 KV-A.OUTPUT

8

9 lOqO00

Values of ampere conductors per. inch of periphery for slow-speed alternators,

For a satisfactory machine, therefore, all of the dimensions in Fig. 4o must increase as the ampfire turns per pole are increased, so that one constant in design is the ratio amp6re turns per pole/pole pitch, the value of which ratio can be increased either by raising the permissible temperature rise of the field coils or by improving the ventilation. It is of interest to note, however, that, in order to obtain a certain desired voltage regulation, the armature ampere turns per pole are made a definite fraction of the main excitation, so that the constant used by the designer is rather the ratio of armature amp6re turns per pole/

214

ALEXANDER

[J. F. I.

GRAY.

pole pitch, of which the ampere conductors per inch of armature periphery are a definite measure. This constant, therefore, which we found to be of great importance in direct-current design, is of equal importance in the design of alternators, but, since its value depends on so many factors, the writer who gives a table of values is subject to criticism. Let us, however, examine a few machines and see what values are found in practice. TABLE I V . Kv.a ................................. Phases ............................... Frequency ............................

R.p.m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Poles ................................. Date I n t e r n a l d i a m e t e r of s t a t o r i n i n c h e s . . . . . . . . . . . . . F r a m e l e n g t h in i n c h e s . . . . . . . . . . . . . . . . . . . . . . . . . Centre vent ducts ............................. End vent duets ................................ m. Pole pitch in inches ......................... n. Air-gap clearance in inches ...................

Ratio

I ooo 3 5° 94 64 19oo 230 11.8 None None I 1.35 5/16

m/n .................................... 36 Slots per pole ................................. 6 Size of slot in inches ........................... I.I dia. Tooth slot .................................... 0.8 M a x i m u m t o o t h d e n s i t y a t n o l o a d i n ! i n e s p e r sq. in. lO4,OOO Core density .................................... I8,OOO Pole density ................................... lO9,OOO a. AmpSre conductors per inch ................... 37o b. C i r c u l a r m i l s p e r a m p e r e . . . . . . . . . . . . . . . . . . . . . 725 Ratio a/b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Peripheral velocity of rotor in feet per minute .... 5700 c. A m p e r e t u r n s p e r p o l e a t n o l o a d . . . . . . . . . . . . . 45oo by test d. A r m a t u r e a m p S r e t u r n s p e r p o l e . . . . . . . . . . . . . . 21oo Ratio c/d ...................................... 2.15 Pole enclosure ................................. o.66 Output factor ................................. o.oi72 Guaranteed temperature rise .................... 4 °° C.

1917 130 13.5 3 of ~ inch 2 of I i n c h 6-4 3/I6 34 6 .5X2.75 1.12 IOO,OOO 5o,ooo 95,ooo 9oo 75 ° 1.2 3200 37oo 288o 1.28 0. 7 o.o47 5 °0 C .

Table III gives data on several of the machines that were exhibited at the Paris Exposition; the i4oo-kv.a, machine is of special interest because Rothert claims that it is so much superior to other machines of that time. The values of amp,~re conductors per inch found in these machines are plotted in Fig. 41, and on the same diagram are plotted figures given by S. P.

Aug., ~9~7-]

MODERN ])YXAMO

]TLECTRIC M A C l l [ N E R Y .

2I 5

Thompson in I9o5; these indicate the improvement in ventilation in five years. The figures of I9Io are higher because operating engineers were willing to accept poorer regulation in order to get a cheaper machine, while the enormous increase in the last seven years is due partly to the fact that the larger machines are now rated at 5°0 C. rise with no overload guarantee, but more particularly to the fact that regulation has been thrown to the winds and the voltage maintained at all loads and power factors by voltage regulators such as these first invented hv Tirrell. FI(;. 4 2.

/

'rlS.5~

<

b

<

1000 KV-AI ,v 94- R.RM.[/

1000 ffV-A. 94 R.P.M.

JDAT~.-1900[/

DATE-1917

164 P0LESl/

POLES

oo

7"

i,,"

/L Comparison between a machine built in I9oo and one built in I917 for the same output.

Perhaps the best conception of the progress made in design, and of the entire change in performance characteristics, may be obtained from Table IV, which gives comparative data on the ~ooo-kilovolt-amp&re machine exhibited at the Paris Exposition and on a unit as designed to-dav for the same output. The two machines are drawn to scale in Fig. 42, and the performance characteristics are given in Fig. 43. With an excitation neces-

216

ALEXANDER

GRAY.

[ J . F. I.

sary to give normal voltage on no-load, the current on shortcircuit is 2.75 times full-load current in the early machine and is only i. 5 times full-load current in the more recent design. In the case of alternators of large output, when the air-gap clearance is not limited partly by mechanical clearance, the short-circuit current with normal no-load excitation does not greatly exceed the full-load value. The regulation guarantees expected at various times have been : Date

Regulation at IOO per cent. Dower factor Per cent.

I900 1910

Regulation at',8o per cent. power factor Per cent. :~

15

.................... ....................

5 8

20

1917 . . . . . . . . . . . . . . . . . . . .

20

4O

The method of obtaining good regulation in the early machines was, as shown in Fig. 43a, to saturate the magnetic circuit and to use a small number of amp6re conductors per inch of periphery. Since it was not possible to use high tooth densities as in direct-current machines because of the high frequency and the stationary armature, it was usual to run the pole density above saturation, and this often caused an excessive drop in voltage, due to pole leakage. The machines, therefore, would not give an overvoltage if this was desired: the iron going into the machine had to be carefully tested for permeaMlity and the leakage fluxes, and saturation curve closely calculated; the machines also were large for their output. The tendency in America until almost I9oo was to build a cheaper machine, of poorer inherent regulation, and to compound or compensate the machine for voltage drop ; series excitation was obtained by SOlne kind of rectifying device such as that shown on the end of the shaft in Fig. 32 . This scheme worked well on the small single-phase units used for lighting service, because the power factor was approximately constant, but it gradually lost favor as the polyphase system came into general use. The extent to which alternators have been gradually rated up is well indicated by the curves of output constant given in Fig. 44It is of interest to note from these curves that, contrary to general opinion, the core dimensions of an alternator are greater than those of a direct-current machine of the same output. This

A u g . , 1917.]

~'IODERN DYNAMO

ELECTRIC

217

~'IACHINERY.

is due principally to the fact that, whereas in direct-current machines tooth densities of 15o,ooo lines per square inch are not exceptional, values of IOO,OOO lines per square inch cannot safely be exceeded in 6o-cvcle alternators, because the loss at that freFiG. 43 a

1'2

~

LO < 10

/ ~ ' ~

I '

\

Z 0

•

06

z

~: 04 7--

/

0~-/

5'0

S

/

U

2.0,.'o j ~o-r

/

I

/

L

2 5 4 5 AMP.TURNS PER POLE

)°~_ ~-

6× 1000

Performance characteristics of the machine built in t9oo. Fie;. 43B.

1-2

/ f

1.0 o 08

<-J z::

z

I

/

//

N 0"6

F"'~ c¢: (..)

'"

0-2

'=~

1O_z

"

I--

/ 1

2

3

4-

5

6~'1000

AMP. TURN5 PER POLE Performance characteristics of the machine built in I917.

quency is large, while the ventilation of a stationary core is not nearly so effective as is that of a rotating armature. VENTILAmmx.--Manv of the earl}, alternators were so large for their output that no special provision was necessary in order

218

A L E X A N D E R GRAY.

[J. F. I

to keep the temperature rise down to a safe value. The machine shown in Fig. 42 , for example, does not even have vent ducts in the stator. Even in the case of modern machines, the slowspeed alternator does not require that any special precautions be Fro. 44-

O08 & 0-07. ~ ;

--

, ~,~,~~

006~/@

I.~\~,~>~t

==__0'05

~

t

~

I

-'~R~

/

~004 0.03 / ""

_ I. . . . I Mp ,, 40N. ]90'5.P.THO

o ~ 002 / f

K:

0.01 ,

VOLT-AMPERES

D~ix'RF'M

, o KV-A. OUTPUT

7

o

,ox,ooo

Output constants of slow-speed alternators.

taken to keep it cool, and it is only as the speed for a given output increases, and therefore the dimensions decrease, that the problem becomes difficult. watts loss

The

temperature

r i s e is p r o p o r t i o n a l

to

~DL(a + b v)

where

¢r D L is the armature surface V is the peripheral velocity of the armature a and b are radiation and convection constants respectively. . . . . .

The output m ~v. a .

D2L X rpm . . .

(DL V) X a constant, K • watts loss V therefore temperature n s e = - ( a @~V) × outPut X a constant [ V \ = (per cent. l o s s ) X ta + b V )

so that, for a given output and a given efficiency, the higher the peripheral velocity the more difficult it is to keep the machine cool. But note further that V =~D X rpm I2

=

=

7rD X I2o f T2 p I0 X pole pitch X frequency

A u g . , 1917.]

~,IODERN ])YNAMO

ELECTRIC

~,'L,\CHINERY.

2I 9

so that, for a given output and given frequency, the greater the pole pitch the m o r e difficult it becomes to keep the machine cool. This is shown in rather a striking way if we compare two waterFro. 45.

/ /

I

I.

I.

rm

r'~

Comparison between two IO,Ooo-kv.a. water-wheel generators, one of IOO r.p.m, and the other of 6oo r.p.rn.

wheel generators, one to operate oll a high head and the other on a low head: compare, for example, Kv.a .................... Rpm .................... Poles ...................

IO,ooo 6oo 12

IO,Ooo I0O 72

Probable dimensions are : ~ n t e r n a l d i a m e t e r o f s t a t o r in i n c h e s . . . 80 Frame length in inches ............... 37 Pole pitch in inches .................. 21 Peripheral velocity of rotor in feet per minute .......................... i2,5oo P e r i p h e r a l v e l o c i t y a t r u n - a w a y s p e e d . . 22,5oo

250 26 lO. 9 65oo 11,7oo

220

ALEXANDER GRAY.

[J. F. I.

These machines are drawn to scale in Fig. 45, and it can now be readily seen why the high-speed machine is so difficult to ventilate: it has not the radiating surface of the slow-speed machine. In slow-speed machines the air is stirred up around the stator by the rotor itself, no fan blades being necessary. The whole construction is open, as shown in Fig. 47, and also in diagram A, Fig. 46. With this type of ventilation the air is Fro, 46.

A

Different methods of ventilating alternators.

not always directed where desired, and the air streams are rather unstaNe, so that one side of the machine is often cooled better than is the other. With moderate-speed machines it becomes necessary to properly direct the air, which result is generally accomplished by the addition of fans to the rotor, while the housings are often made solid, as shown in Fig. 48, and also in diagram t3, Fig. 46, so as to deflect the air over the back of the stator windings and core. In the case of high-speed water-wheel units, under which

Aug., ~9~7.]

~/[0DERN D Y N A M O ELECTRIC ~{ACHINERY. FIG. 47"

Generator with

open end

bells.

Fm. 48.

Generator with solid end bells.

22I

222

ALEXANDER

GRAY.

[J. F. I.

class would come the 6oo-r.p.m. machine in Fig. 45, the frame is generally long in order to keep the peripheral velocity down to a safe value, and in such cases it is necessary to create an airpressure in the end bells to force the air between the poles and out through the vent ducts in the centre of the stator core, as shown in diagram C, Fig. 46, and also in Fig. 49; comparatively small openings are left at D, Fig. 46, to ventilate the end bells. MECHANICAL CONSTRUCTION.--0ne feature of interest is the growing popularity of the vertical shaft or umbrella type of lqc,. 49.

Generator with enclosing end bells and forced ventilation.

water-wheel generator. This has been due partly to improvements in the design of this type of water-wheel, but also in large measure to the development of a reliable thrust bearing. The Kingsbury bearing makes use of that feature of the horizontal shaft bearing on which its ability to carry large pressures depends; namely, that, as shown in Fig. ~o , the oil is drawn into the high-pressure space as a wedge, by the rotation of the shaft itself. ~Phe stationary thrust plate of this type of bearing is made of a number of segments, as shown in Fig. 52, these seg-

:\ug., 19~7.] -~'IODERN DYNAMO ELECTRIC ~,{ACHINERY. FIG. 5 ° .

°1 < 0

~ O I L

FILN

Oil film in the bearing of a horizontal shaft.

FIG. 51.

°1 < 0

_1

MOTION ~.s s , a ~ . ~ . ~ . s ~ s J

~ j

I

j ff~.~tJfJ

fJ

OIL FILM Oil film in the Kingsbury thrust bearing.

FIG. 52.

i VOL. 184, N o .

Parts of a Kingsbury thrust bearing. 11oo--17

223

224

ALEXANDER

GRAY.

[ j . F. t.

ments being free to tilt, so that when the upper plate rotates in the direction shown in Fig. 5I, the lower plates tilt and the oil wedge is formed as indicated. Such a bearing will carry an average load of 35o pounds per square inch, which is about seven times the safe value for an ordinary thrust bearing. Fig. 53 shows the arrangement of the guide and thrust bearings of such a vertical generator, Fig. 54 shows the very substantial support used to carry the weight, and Fig. 55 shows FIG. 53.

Section of Westinghouse generator with two guide bearings.

one method adopted to force the air into the centre of the long core of these machines. The core length of water-wheel generators of large output is generally longer tl~an it would be for engine-driven units, because the diameter has to be kept small so that the stresses due to centrifugal force will not be excessive should the generator run away, due to faulty operation of the governing mechanism of the water-wheel. It is generally assumed that the cost and weight of a machine for a given kilowatt output go down as the speed is increased.

Aug., 1917.]

MODERN DYNAMO ELECTRIC ~IACHINERY. FiG. 54.

Three 78oo-kv.a., I44-r.p.m. vertical generators with Kingsbury thrust bearings. FIG. 55.

Rotor for a io,ooo-kv.a., i44-r.p.m, vertical generator.

225

226

I\LEXANDER GRAY.

[J. F. 1.

It is true that the value of D'aL decreases, but when the speed reaches such a value that a radical change in the type of construction becomes necessary, as for example when one has to change f r o m the simple cast-steel spider with bolted-on poles to the expensive construction shown in Fig. 45, where the r o t o r is built up of steel plates, then the labor co,st becomes a larger proportion of the total cost, and the f a c t o r y cost of the machine will not always be reduced. 1G

(To be coutinued.) America's Longest Railway Tunnel. A>~ox. (The Contract Record, vol. 31, No. 21, p. 449, May 23, I 9 I T . ) - - T h e Rogers Pass Tunnel, officially known as the Connaught Tunnel, on the line of the Canadian Pacific Railway through Selkirk Range of the Rocky Mountains in British Columbia, was placed in service on December 9, I916. The completion of the Connaught Tunnel is without doubt the most notable achievement in the art of tunnelling ever accompl'ished on the American continent. T o complete a rock tunnel, five miles long, in practically three years, under conditions by no means easy, is an undertaking that might well be termed a brilliant success. The growing traffic congestion of this division of the Canadian Pacific Railway lines made the greatest speed construction essential. To meet the urgency demanded by the railway, methods somewhat removed from the conventional were obviously a necessity. The rapidity of headway was a result of the combination of two elements--a new mode of heading attack and a revised plan of enlargement. In both of these distinct progress is evident upon the practices heretofore accepted. First, a pioneer bore, inaugurated as an experiment, was looked upon with distrust by many authoritative tunnelling engineers, because it did not follow the precedent set by prior rock tunnelling. The enlargement operations further d e m o n s t r a t e d the success of progressive ideas. T h e adoption of a central heading of such size as to permit radial drilling for the final enlargement proved its value, as the speed of driving, averaging 16 to 20 feet daily, constituted a record. Blasting in rings on planes perpendicular to the tunnel axis, although an innovation, contributed to phenomenal progress. Engineers throughout the world have expectantly watched the progress of this experiment. The experiment "has been successful, stereotyped processes are at least doomed to revision, and the achievement at Rogers Pass will be an incentive for still further progress in this important branch of engineering. l"Behrend, Eieclrical Re~ezv, New York, vol. 45, September io, 19o4 (P. 375).