GER-80

PROTECTION OF STEAM TURBINE DISK WHEELS FROM AXIAL VIBRATION

By Wilfred Campbell

TANGENTIAL VIBRATION OF STEAM TURBINE BUCKETS

By Wilfred Campbell and W. C. Heckman

Reprinted from THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS PAPER NO. 19~2O AND 19~5

GENERAL

ELECTRIC

CONTENTS
Page No.
PART I

Purpose and scope of the investigation

Exposition of the nature and theory of vibration in turbine wheels.
. .

7
17 43 71 77 85

PART II

PART III Method of design and testing for the protection of turbine bucket wheels from axial vibration DIscussIoN---~Protection of steam turbine disk wheels
TANGENTIAL VIBRATION OF STEAM TURBINE BUCKETS DISCUSSION

Tangential vibration of steam turbine buckets

WILFRED CAMPBELL

WILFRED CAMPBELL
Wilfred Campbell was born in Manchester, England, in 1884 and served an apprenticeship with the Lancashire and Yorkshire Railway for seven years. During the latter part of his apprenticeship he was selected as one of two to attend the Manchester Municipal School of Technology where he received several prizes of distinction for his ability in design. Later he was employed as a draftsman by the Northeastern Railway Company. In 1907 he came to America, and for a year was assistant to the master mechanic at the Arnold Print Works, North Adams, Mass. In 1908 he was employed by the General Electric Company as a draftsman in the Direct-current Department. During the war he was associated with Mr. C. E. Eveleth, now manager of the Schenectady Works, and Dr. Irving Langmuir of the Research Laboratory, in connection with the design of submarine detectors by the General Electric Company for the United States Government. Since 1919 Mr. Campbell has been an engineer in the Turbine Engineering Department of the General Electric Company where he has had charge of the investigation of turbine vibrations of all sorts. In recognition of his outstanding contributions to turbine design, he recently received an award from the Charles A. Coffin Foundation. The results of this exceptionally valuable research were presented by Mr. Campbell at the spring meeting of the American Society of Mechanical Engineers, May a6 to 29, 1924. The paper was reprinted in full in the GENERAL ELECTRIC REVIEW, June, July, and August, 1924. Mr. Campbells untimely death occurred at Schenectady on July 7th, after an attack of acute appendicitis. The loss to the turbine art will be appreciated by all who are familiar with his work. The breadth of his scientific knowledge is well illustrated by this paper, although it was only one of many interests. His ready ingenuity was supported by a keen intuition which was invaluable to an engineer of his type. His genial good nature and unfailing optimism endeared him to all his associates. His integrity and fearlessness won the confidence of turbine operators all over the land to an unusual degree. His intense interest in all modern inventions gave him confidence that no problem was beyond solution. Francis C. Pratt, vice-president of the General Electric Company, in charge of engineering, in commenting upon Mr. Campbells death, made the following statement: Mr. Campbells sudden death was a great shock to his many friends and associates in this Company. At the time of his death he was the engineer in charge of investigations of turbine wheel vibrations. When he undertook this important work in 1919, it was thought by some engineers that the limits in the size of individual turbine units had been reached. Mr. Campbell had a rare genius for analyzing and solving abstruse problems of a mechanical nature and did not share these adverse opinions regarding the limitations in the size of steam turbines. He discovered the cause of the trouble with large turbine wheels and buckets, devised remarkably ingenious methods of detecting and measuring the phenomena connected with them and provided means for overcoming them. He has made a fine and lasting contribution to the industry and his work will be long remembered by engineers who design and operate steam turbines.

The Protection of Steam Turbine Disk Wheels From Axial Vibration*

By
WILFRED CAMPBELL TURBINE ENGINEERING DEPARTMENT, GENERAL ELECTRIC COMPANY

Prepared with the co-operation of A. L. Kimball, Jr., Research Laboratory, and E. L. Robinson, Turbine Engieeering Department; both of General Electric Company

PART I
One of the most important features in the design and manufacture of a steam turbine is the elimination of the possibility of vibration occurring at the various natural frequencies of its disk wheels and buckets. To accomplish this purpose the General Electric Company has for several years made an extensive investigation of various forms of vibration and waves which may exist in steam turbine disk wheels. One of the results was the determination of the laws of dangerous critical speeds that must he guarded against, together with other minor resonant conditions that it is advisable to avoid. Testing machines with complete auxiliary apparatus were developed for the verification of predicted frequencies and critical speeds, and a program for the routine testing of turbine wheels was instituted. This resulted in the determination and adoption of the procedure, necessary in all cases, for the definite protection of steam turbine bucket wheels from axial vibration, as justified by several years of successful manufactureAUTHOR.

PURPOSE AND SCOPE OF THE INVESTIGATION

turbines of ratings exceeding 5000 kw. each

The total generating capacity represented by

these machines was over 2,404,000 kw.
year 1919, when this particular investigation was started, there have been 206 more turbines installed exceeding 5000

The purpose of this article is to present the main features of the work done by the General Electric Company which led to the solution of the problem of vibration of turbine disk wheels, and to describe the way wheels are designed and tested in order to insure freedom from vibration. The investigation was undertaken in order
wheel failures of a peculiar and erratic nature which could not he explained on the basis of high stress alone. The number of failures was small considering the total number of wheels in operation. These failtires were not confined to any single type of machine, hut they did show a preference in general for thin wheels of large diameter. That this difficulty has actually heen overcome with no major alteration in the turbine is emphatically brought out by results obtained by the General Electric Company in the past three years from its use of disk wheels properly designed and tested. Up to the end of 1923, this company had manufactured and installed over 9000 steam turbines aggregating in total generating capacity more than 15,000,000 kw. This investigation is chiefly concerned with large size machines, that is, of over 5000 kw. capacity. Before the year 1919, the General Electric Company had manufactured and installed in operating plants a total of 227
to account for
*This article was presented as a paper at the Spring meeting of she American 5ocicty of Mechanical Engineers, May 26-29, 1924, at Cleveland, Ohio.

Since the

kw. each, increasing the total of generating capacity of the larger size turbines to 5,864,500 kw. at the end of 1923.
TABLE I

the investigation, giving a few figures as to the number of machines and the number of turbine wheels investigated. The capacity of the testing machines now in operation is sufficient to provide for the testing of 600 wheels annually, under all conditions of speed. That the methods developed are effective is attested by the successful elimination of all serious wheel and bucket troubles due to axial vibration from the operation of recently built turbines.

FOR TURBINES OF OVER 5000 KW. STALLED BEFORE MARCH 1, 1924 Number of wheels installed Number of wheels tested (standing) Number of wheels rotated in wheel testing machine Number of tests in wheel testing machine Number of wheels tested in customers plants (standing) Number of machines investigated in customers plants Number of machines tested under load in customers plants Number of wheels replaced to avoid possible trouble Number of wheels tuned for vibration Table I shows the magnitude of

IN4399

3596 320 405 1683 291 24 497 212

PROTECTION OF STEAM TURBINE DISK WHEELS

PART I

TRAVELING WAVE IN A THIN DISK MODEL OF A TURBINE WHEEL High speed motion picture of a 4-node wave traveling backward one-half revolution in the wheel which is clamped
stationary. Alternate exposures have been omitted in this reproduction

PART I

PROTECTION OF STEAM TURBINE DISK WHEELS

Fig. 1. Broken Turbine Bucket Wheel 9th Stage of l5000-kw., 1800-r.p.m., 9 stage Turbine. This was caused by axial vibration of a type which has been eliminated by the investigation described in later parts of this article

Fig. 2.

Detail of the Fracture in Fig. 1, Showing Characteristic Fatigue Failure

Fig. 3.

Fatigue Crack which Started at a Hole

Fig. 4. Small Turbine Wheel Broken by Accidental Over speeding, 2nd Stage of 500-kw., 3600 r.p.m., 3-stage Turbine

PROTECTION OF STEAM TURBINE DISK WHEELS

HISTORICAL OUTLINE

PART I

vibrations to which disk wheels are subject, a brief narrative of the work of investigation will be presented. This seems necessary in order to give a proper perspective of the problem as a whole.
Growth of Capacity

Before discussing in detail the nature of the

The design of turbines with one row of buckets per wheel took place long prior to the entry of America into the war. These designs used higher linear bucket velocities and were produced in unprecedented quantities during the war period. The rapid

Fig. 5. Fatigue Failure which Did Not Pass Through Holes; 11th Stage of 30,000-kw., 1500-r.p.m., 12 stage Turbine

Fig 7. Turbine Wheel Failure as a Result of Fatigue Bending; 3rd Stage of 6000-h.p. Turbine

increases of turbine capacity which took place at the same time were in a large degree

accomplished by the use of larger diameters introduced to give greater bucket speeds, and by using longer buckets. The mechanical
possibilities were pushed to the limit. While some improvements in steam conditions came at the same time, the important thing to note is that the real period of increased capacity due to improved thermal processes occurred later and is still going on whereas, almost at one leap, the early designs were pushed to the limit from the point of view of structural strength. Several principles of design pointed in the direction of light wheels. The maximum wheel stress is at the bore and this could be reduced by using lighter and thinner disks having less centrifugal bursting tendency. In fact, these were cut down in thicimess as much as could be without creating a new maximum in the web due to the pull of the

Fig. 6.

Detail of Fig. 5, Shoa~ng Characteristic Fatigue Failu,e

10

PART I

PROTECTION OF STEAM TURBINE DISK WHEELS

Figs. 5 and 6 show a typical fatigue fracture which did not originate at a hole. Fig. 7 shows a break in which cracks originated at more than one hole. Figs. 8 and 9 show details of this fracture at each of the two holes. Although Fig. 7 bears a resemblance in its completeness of failure to Fig. 4, the type of fracture is entirely different and characteristic of a fatigue failure. In other cases the only damage was loss of buckets. Fig. 10 shows a wheel from which a number of buckets have been broken. In this case the breaks were due to axial vibration.

buckets. And not only did the desire for conservative stresses point toward light wheels, but also the desire for a stiff rotor. Heavy wheels are accompanied by lower critical shaft speeds. The unquestioned advantage of a stiff shaft, when possible, also dictated light wheels. These various influences were perfectly natural at the time. The subject of vibration had not been brought to prominence. It was hazy and uncertain and no difficulties had been experienced on this account. The plain path of reason seemed to be along the lines indicated.

Fig. 8.

Detail of Fig. 7, Showing How the Crack Started at One Hole

Fig. 9.

Detail of Fig. 7, Showing How the Crack Started at Another Hole

Types of Failure

In order to visualize the sort of difficulty which led to the present investigation, it will he well to examine a number of failures. Figs. 1 and 2 show a break which originated in a small tapped hole anti passed through a large steam balance hole. This was a vibration fatigue failure. An examination of the fractured surface shows the characteristic central line and progressive curves. Fig. 3 shows another crack discovered before complete rupture in the same wheel. Fig. 4 shows a wheel which completely burst. Note that the line of fracture has passed through every one of the holes in the web of the wheel. This was due to accidental overspeed

In Fig. 11 various failures occurring in the bucket dovetail al-c shown. In each case the marking indicates vibration in an axial
direction.

Fig. 12 shows another class of turbine trouble in which a diaphragm has been scored in two diametrically opposite spots by a rubhing wheel. Fig. 13 is a section of the same diaphragm at the point of greatest rubbing together with a profile of the wheel showing the shape taken by all of the buckets. These examples will selve to illustrate the various types of failure. Most of them were plainly due to vibration resulting in fatigue. A few were clearly the result of accidental overspeed such as Fig. 4. In certain cases the fractures avoided holes, but in general there

11

PROTECTION OF STEAM TURBINE DISK WHEELS

was a distinct affinity for holes owing undoubtedly to the higher localized stresses around them.
Hole Stresses

PART I

The first remedial measure used was to give immediate attention to the localized stresses in the neighborhoods of the steam balance holes and other discontinuities in the webs of the wheels.

It is not the object of this article to discuss hole stresses, but it is necessary to note that they become serious only in connection with vibration, by constituting a place for a fatigue crack to start. If a wheel is properly protected from vibration, there will be no repeated stresses at any point Rather elaborate experiments have shown that the reinforcement of holes will serve to increase the resistance of a wheel to vibration in case it should be necessary to design with the expectation of fatigue stresses. But the greater safety lies in guarding against the stresses themselves by the precautionary measures developed for use in manufacture.
Bucket Lacing

The early experiments on the vibration of buckets indicated that a lacing wire paralleling the shroud band and connecting different groups of buckets would remove various secondary vibrations. Although the fundamental period was not much influenced this expedient was actually used in a number of turbines. The lacing wire, however, failed to remove the cause of the trouble and became itself another hazard. During these experiments a taut wire attached to the end of each bucket raised the frequency thus illustrating the similar effect due to centrifugal force.
Fig. 10. Roots of Buckets Broken by Fatigue Bending Due to Vibration; 12th Stage of 10,000-kw., 1500-r.p.m., 12-stage Turbine

The Idea of Wave Motion

About the same time, certain types of vibration of standing wheels were investi-

7th stage bucket broken at neck of dovetail 12,530-ksv.. 1500-r.p.m., 8-stage

6th stage bucket broken at neck of dovetail 12,500-kw., 1800-r.p.m., 9-stage

9th stage bucket broken at neck of dovetail 18th stage buckets broken at root of blade and at neck 7500-kw.. 1800-r.p.m.. S-stage of dovetail. 20,000-kw., 1800-r.p.m.. 23-stage Fig. 11. Details of Bucket Failure Due to Axial Vibration

12

PART I
gated by means of sand form was a series of symmetrically arranged ference and extending separated by radial lines radii or nodes. These detail later.
Windage Theory

PROTECTION OF STEAM TURBINE DISK WHEELS

pictures. The usual vibrating segments about the circuminto the web but of quiet called nodal will he discussed in circumferences developed a form of wave motion as shown in Fig. 15. This was examined by means of an intermittent spark either synchronized with the speed of rotation or adjusted to occur at slightly greater or less frequencies. The shapes of the waves and their rates of progress were thus examined visually, and it was found that these waves progressed around the wheel in the direction of rotation, but at a less speed than the speed of the wheel, that is, relative to the wheel

In connection with the study of strains about holes in wheel webs, a series of India

Fig. 12. Turbine Diaphragm which has Suffered Rubbing in Two Diametrically Opposite Regions; 17th Stage of 30,000-kw., 1800-r.p.m., 17-stage Turbine

Fig. 13. Drawing Showing Depth of Rubbing at Worst Parts of Fig. 12. All ofthe buckets on the wheel were rubbed alike to the shape shown

nlbber wheels as shown in Fig. 14 had been made and photographed by means of instantaneous electric sparks while rotating at high speed. In order to produce representative stresses throughout the rubber disks, metal weights were attached about the circumference, shaped to simulate the loading due to turbine bucket blades. In running these wheels it was discovered that above certain definite speeds their

itself the wave was travelling backward, seemingly driven by the windage encountered like the fluttering of a flag. This gave rise to the so-called windage theory that the waves were developed and driven backward in the wheel by the atmosphere in which it was revolving. Investigation of this theory led to the rotation of paper disks in a vacuum. The wave motions, clearly observable in an atmosphere disappeared as

13

PROTECTION OF STEAM TURBINE DISK WHEELS

the atmosphere became rarefied. rfhese experiments were immediately extended to exceedingly thin steel disks which were found to behave in a similar manner. Since the waves could not exceed the speed of the actuating wind without encountering resistance, they could only be supported in wheels
Windage Theory Abandoned

PART I

Oscillograph coils were placed in two large turbines and during operation complete surveys of these machines showed that several stages developed wave phenomena of the same general characteristics. Turbine wheels were thus shown capable of

Fig. 14.

India-rubber Wheels Used to Examine the Strains d.oout Holes in the Web

whose natural frequencies corresponded to waves travelling at less speed than the speed of rotation. It was therefore made a condition of design of turbine wheels that the natural wave speed should exceed the speed of rotation wherever possible. This resulted in a general thickening of all wheels being designed for turbines constructed at that time to give them a greater rigidity to withstand the supposedly detrimental effect of the wind action.

supporting travelling waves. In these turbines the waves appeared only when the turbine carried more than a certain definite load and died out when the load was removed. However, waves were found to be travelling in the wheel in a direction opposite to the rotation of the wheel and at a higher speed, which could not be explained by the windage theory.

14

PART I

PROTECTION

OF STEAM TURBINE

DISK WHEELS

Wave Phenomena Records Made from a Thin Steel Disk

The first oscillograms taken in which a 1-evolving coil was used were made with a thin sheet metal disk, Fig. 16 is typical of the type of oscillograph record obtained. The upper curve A is taken by a coil stationary in space and opposite the rim of the disk; it shows that there is a wave motion in the disk. The smoother portion of the curve corresponds to the part where the wheel disk is most remote from the coil, and the more disturbed portion indicates that the wheel rim is in close proximity. These more disturbed portions, which may be called beats, occur at a much lower rate than the speed of rotation so that the corresponding wave crest producing them must have progressed around in the disk itself. Curve B is taken by a coil made to revolve with the wheel but mounted on a separate arm. When this coil passed near the supporting pillow block a voltage was induced which makes the long narrow lines in the curve, there being one of these for each revolution. The V-shaped points in this record show that a transverse motion is taking place in the wheel disk, resulting in change of clearance between the disk and the revolving coil. It can he shown that this is not in any sense synchronous with revolutions but simply records a transverse vibration of the wheel itself. Curve C is taken from the 40-cycle alternating-current line and is here used as a timing wave. The small interruptions in this wave are made by a contact on the shaft carrying the model steel disk and they correspond therefore to revolutions of the disk.

Fig. 15. Photograph by an Instantaneous Spark of an India-rubber Wheel Exhibiting Wave Phenomena at a Speed of 650 r.p.m.

The record shown in Fig. 17 was made by a thin steel disk revolving slowly. This disk was vibrated by means of an alternatingcurrent magnet carried on a rotating arm at the same speed. The upper record A was obtained by a fixed coil while the lower record B is a 40-cycle timing wave. The irregularities in the record obtained by the

Fig. 16. One of the First Oscillograms Recording Disk Vibrations. Trace A was recorded by a stationary coil. Trace B by a coil revolving with the disk, and Trace C is a 40-cycle timing wave. Model disk wheel ofthin sheet metal

Fig. 17. Record of 6-Node Vibration in a Thin Steel Disk Excited by an Alternating-current Magnet Revolving with the Wheel. 4 Trace A was recorded by a stationary coil. Trace B is a O cycle timing wave

15

PROTECTION OF STEAM TURBINE DISK WHEELS

TABLE II

PART I

EARLY SUMMARY OF WHEEL AND BUCKET TROUBLES The methods by which such troubles have been eliminated are described in Parts II and III of this article
Date Trouble Occurred
Backward Speed of Wave R,P.S. Operating Speed R.P.S.

Rating

Stage

Trouble

Nodes

350001500--20 1500018009 1500018009 500018007

3000Variable4
30000150012

19 9 9 3 3 11 10 13

1918 1918 1919

1920 1921 1921

Wheel Wheel Wheel Wheel Wheel Wheel Buckets Buckets Buckets Buckets Buckets Buckets Buckets

4 4

25.6 27.9
29.5

25
30 30

4
6 6

4
8 8

28.5 4S.1 25.2 30.2

30.1

30 48 25 30
30

20000180012 30000180017
30000180017

11917
1919 11920 1918 1918 1918

17
21 2 2 11 12

45000 120021 750015008

500036005 300001800 17 30000180017

4 4
6 4 8 6 4 8 6 8

32.6
19.8

30
20

1919
1919

25.4 57.4
29.6 28.8

25
60 30 30

1920
1920

Buckets
Dovetail Buckets Buckets Buckets

15000180023
1000018009

21
8

1920 1921
1921

29.5
30.6

30
30

35000150022 30000150020

17 14

1921

24.2 25.9

25 25

stationary coil are caused by the passage of the rotating exciting magnet of which, it will be noted, the electrical frequency is one-half the mechanical frequency. This record, briefly, shows wheel vibration of the 6-node type including both forward and backward waves which will be explained in the next part of this article. This record is reproduced in the fifth edition of Stodolas Steam Turbines.
Recognition of Wheel Critical Speeds

~alledawheelc~ In making a statement of the importance of this phenomenon, it should not be assumed that it is the only type of wheel vibration of a
serious nature. This type of vibration causes by far the largest number of failures, in fact,

so large a fraction that breaks caused by

other types of vibration may fairly be treated as exceptional. At this time the importance of obtaining test data on full-sized wheels under actual operating conditions was first fully appreciated. This resulted in the design and construction of the first wheel-testing machine.

Subsequently Table II was compiled which definitely established the importance of the wave stationary in space, th~j~~~ave pro~~jn backward in tj~j~eL~L.tj-~e ~twhi~eirotatesforwar~

16

PART II

EXPOSITION OF THE NATURE AND THEORY OF VIBRATION

IN TURBINE WHEELS

To illustrate standing vibrations in turbine

disk wheels, the following method was used. A turbine wheel was mounted in a horizontal
position on a stub shaft. An electromagnet was clamped with its poles close to the edge of the wheel. On passing an alternating current through the coils of this magnet a series of pulls was exerted on the wheel tending to deflect it in a direction transverse to the plane of the disk. The frequency of these pulls is twice the frequency of the alternating current used because every complete electric cycle corresponds to two current pulsations in the magnet, and an electromagnet exerts a pull when current flows in either direction through the coil. The alternating-current generator was driven by a variable-speed direct-current motor by means of which the frequency of the magnet pull could be varied over a wide range.
Sand Pictures

Sand was scattered over the wheel surface and the frequency of the magnetic pulls was varied until a particular frequency was reached at which the wheel responded. Fig. 18 shows a case where the wheel vibrated in four segments. In this vibration each segment springs up and down, scattering the sand over to the quiet or nodal zones where there is no up-and-down motion. If, however, the frequency of the deflecting pulls of the magnet is altered even a very small amount, the vibration immediately dies out, although the magnitude of the impulses of the magnet remains the same as before. On rais-

ingthe frequency of the magnetic pulls another point is found at which the wheel responds. It vibrates in segments, as before, but with Six nodal radii or nodes equally spaced around the wheel circumference instead of four. Figs. 19 and 20 illustrate a 6-node vibration and show that its location is not necessarily dependent on the position of a series of symmetrical discontinuities such as the steam balance holes. Not only does the disk wheel respond when the pull frequency corresponds to four or to six radial nodes, but it may respond readily to frequencies corresponding to 8, 10, 12, or even a larger number of nodes, the number of nodes always being even because for every segment which springs upward during a vibration, the segment next to it on the other side of a nodal line must Spring downward. These photographs illustrate the case of a small wheel with short buckets. Figs. 21, 22, 23 and 24 show cases of 4, 6, 8 and 10 nodal vibrations for the case of a disk wheel carrying long buckets, the total diameter of wheel and buckets being over 8 ft. This wheel was photographed with a layer of paper on the buckets to hold the sand. In the cases of four and six nodes it is seen that the regions of amplitude large enough to move the sand do not extend so deeply into the wheel as in the wheel with short buckets, while in the cases of eight and ten nodes the sand figure is confined to the bucket zone entirely.

The following general observations may be

made on this type of vibration in which segments around the edge of the wheel spring up

17

PROTECTION OF STEAM TURBINE DISK WHEELS

and down, being separated from each other by radial nodal lines: (1) Every disk wheel responds readily to vibrations of four, six, eight, etc. radial nodes, each type of vibration having its own characteristic frequency.

PART II

(3) The higher the number of nodes the more difficult it is to force the sand figures towards the center of the disk. (4) Both the disk wheel and the buckets vibrate together as a continuous disk and must be treated as a unit in this type of vibration.

Fig. 18.

4-Node Sand Picture Made by Vibration of a Wheel with Short Buckets

Fig. 19.

6-Node Sand Picture Made by Vibration of a Wheel with Short Buckets

Fig. 21. 4-Node Sand Picture Made by Vibration of a Turbine WbeeLwith Long Buckets Covered with Paper. The Active Region Extends into the Wheel Fig. 20. 6-Node Sand Picture Showing Independence Pattern from Hole Location. Compare with Fig. 19 of

(2) The higher the number of nodes the higher the frequency of the vibration and the less easily is the vibration excited. 18

Vibrations may also take place with two nodes, as will subsequently be discussed. This type exerts a couple on the shaft transverse to its length, while the types described are balanced in their reaction on the shaft.

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

speed. The frequency of a given type of vibration is determined by two factors, (a) the stiffness and (b) the mass of the vibrating body. The stiffer the body the faster it vibrates, and the more massive it is the slower will it vibrate. Now centrifugal force has no effect on the mass of the wheel, but it

Many other types of vibration exist, including concentric ring nodes and combinations of ring and ~ A hybrid form resulting from a combination of six- and twelve-node radial types is shown in Fig. 25. These types of vibration are not readily excited and do not enter into this discussion,

Fig. 22.

6-Node Sand Picture Made by Vibration of a Turbine Wheel with Long Buckets Covered with Paper

Fig. 23.

8-Node Sand Picture Made by Vibration of a Turbine Wheel with Long Buckets Covered with Paper

Fig. 24. 10-Node Sand Picture Made by Vibration of a Turbine Wheel with Long Buckets Covered with Paper. The active region is confined to the buckets

Fig. 25. Complex Sand Picture with 12 Nodes at the Edge and 6 Nodes Near the Center Made by Vibration of a Thin Steel Plate. This is a rare type of motion

because they have not been found to be the cause of serious trouble.
Effect of Centrifugal Force on Vibration Frequency

After the natural vibration frequencies of a turbine disk wheel when not rotating are determined as described, a question which arises is the effect upon these vibration frequencies of the rotation of the wheel at high

has a powerful stiffening effect. This force acting radially outward around the edge of the wheel stiffens it and raises its vibration fre~uency. This may be compared to the raising of the vibration frequency of a kettle drum by drawing the membrane outward around the edges by the tightening screws. Therefore it may be inferred that centrifugal force raises the natural vibration frequencies of a turbine disk wheel.

19

PROTECTION OF STEAM TURBINE DISK WHEELS

It is well known that a particle of mass m with an elastic support, of such stiffness that a force R5 is required to produce unit deflection, will have a natural frequency of vibration, f.~, expressed by is = ~ ~y (l) 2ir ~ m lithe same particle is supported in another manner with an elastic factor R~, its new frequency will be I R (2) fc=
Traveling Waves

PART II

Thus far disk-wheel vibrations with radial nodes and the effect of centrifugal force on these vibrations have been discussed in some
detail. The type of vibration which has been found to be responsible for serious wheel failures will now be taken up. This type of vibration results when, instead of the wheels vibrating in segments with stationary radial nodes, a wave train travels around the wheel circumference. Before considering traveling waves, a dia-

grammatic representation of radial nodal vibrations of a turbine wheel will be presented.

Fig. 26 represents diagrammatically the edge

l~ Z~~fl l 7 4 Fig, 26. 4-Node Standing Vibration. The diagram represents the developed edge of the wheel during three successive phases

Now when both stiffnesses act at once the

frequency will be 1
fr~

/R~+R~
m

(3)

Suppose R represent the stiffness 5 to supports furnished by elastic and R~the stiffness contributed by centrifugal effects. Assuming the latter proportional to the square of the speed, N , in revolutions per second, 5 this proportionality may be expressed by the use of an arbitrary coefficient B defined by the relation R, = B (4~.2 mAT52) (4) Making use of this relation and eliminating R5 by the use of equation (1) the frequency of the particle, fr, due to the combined effects of stiffness and rotation may he written 2 f = ~J2 + BNS (5) This formula, here derived for the case of a particle, has been justified many hundreds of times for use with a complete turbine bucket wheel by actual measurement as described in later sections of this article. Stodola arrived at the same conclusion on theoretical grounds. The speed coefficient, B, varies with the design of the wheel and the type of vibration. If the vibrating sectors extend a considerable distance into the wheel so that the deflection curve extends well toward the wheel center, B has a lower value than when most of the bending of the wheel is near its edge, as in the case of a larger number of nodes. The value of the speed coefficient is generally from 2 to 3, and a coefficient as small as unity is rare.
Schweitzerische Bauzeilang, May, 1914.

of a turbine disk wheel, and shows the curves assumed by it when the wheel is vibrating with four nodes. The drawing shows the edge of the wheel developed as though all points along the entire circumference could be seen at once. Evidently the two ends of each curve correspond to the same point on the wheel and are, therefore, numbered identically. Curves I, II, and III show three successive stages one-quarter of a complete period apart. The point on the wheel edge marked P is chosen half-way between nodal points and vibrates through the maximum amplitude. The points 1, 2, 3, etc., remain stationary as
they lie in the quiet nodal radii between the vibrating segments. First the wheel edge is bent as shown by the full curve I; one-quarter of a period later the edge becomes straight as shown in curve II, but the portions between nodal points have a rapid motion which carries them over to the maximum deflection in the opposite direction in curve III, onequarter of a vibration period later, or one-half a period from the initial position. At the
1~1
Fig. 27. 6~NodeStanding Vibration. The diagram represents the developed edge of the wheel during three successive phases

end of a full period, the shape evidently is again the same as it was at the start as indicated by the full line curve I. Fig. 27 shows the same sequence for a 6-node vibration. Fig. 28 shows the developed edge of a wheel in a similar manner and represents the case of 4-node traveling waves instead of standing vibrations. The difference between the case of traveling waves and standing vibrations is seen to be that the nodal points 1, 2, 3, etc.,

20

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

traveling wave from the standing vibration frequency. For instance, turning to Figs. 26 and 28, it is seen that if the vibration frequencies of each point on the rim are the same in each case, the traveling wave must move to the right one whole wavelength while the standing vibration goes through one comulete cycle. ~ travelj~veer second e uals the ~iZiT7l~r o corn lete fit rations o the corres ondin standwave er second mult , e ~toacon~te wave.

move along the edge of the wheel instead of remaining at fixed points. Curve If shows the wheel shape after the nodal points 1, 2, etc., have moved one-quarter of a wavelength to the right, and curve III shows the shape after another one-quarter wavelength motion where the nodes have moved to the right one-half a wavelength in all. At this instant the shape of the wheel is the same as for the ease of the standing vibrations previously considered. The difference lies in the motion only. In the case of the standing vibration the nodes are stationary. In the case of the traveling waves the nodes are moving to the right. Fig. 29 shows the same sequence for a six-node vibration.
Comparisons Between Standing Vibrations and Traveling Waves

The following comparisons may be made between the standing vibrations and the corresponding traveling waves for a given disk wheel: (a) In each case there must be an even number of nodes, that is, for every upward portion of the deflection curve there is a corresponding downward portion because of the continuity of the circumference. (b) In standing vibrations the nodes are stationary in the wheel; in traveling waves they move around it. In the first case we have true nodes in the sense that they represent parts of the wheel which are always quiet so they may be observed by the eye. In the second case we have traveling nodes; every part of the wheel edge vibrates and no quiet zones can be
I It
~ ~

Fig. 29.

6.Node Traveling Wave. The diagram represents the developed edge of the wheel during three successive phases

(d) For the standing vibration the amplitude of the particles varies along the edge of the disk from zero at the nodal points to the maximum vibration amplitude at points half-way between the nodes. For the traveling waves, all particles around the edge of the wheel vibrate in turn

through the same amplitude.

(e) For a standing vibration all of the particles along the edge of the wheel vibrate in the same time phase, that is, all particles vibrate together so that each comes to rest at the same instant and each has its maximum velocity of motion at the same instant

during the vibratory motion.

For

Fig. 28.

4-Node Traveling Wave. The diagram represents the developed edge of the wheel during three successive phases

traveling waves, however, the particles along the wheel edge do not vibrate in time phase but vibrate one after another in turn, successively

coming

to rest and

successively

seen. A rapidly moving traveling wave can be seen by the eye only by means of instantaneous illumination. (c) The frequency of vibra~~f~qy.~y ~l~deofa ~n dTsk~h eel ~ oJ~stai~I~vibrationor for traveling ~ is the same. This important ~5oint ~ll~j~s~tly be explained. A knowledge of it is requisite to the determination of the velocity of a

acquiring their maximum velocity of motion during vibration. Since they all vibrate one after another through the same amplitude a wave shape results of constant amplitude traveling around the wheel edge. To sum up the last two paragraphs: for a standing vibration, the particles along the edge of a wheel all vibrate in the same time phase, but their amplitudes vary successively between nodes; for a traveling wave all particles vibrate through the same amplitude, but their time phases vary successively along the wheel edge.

21

PROTECTION OF STEAM TURBINE DISK WHEELS

Relations Between Standing Vibrations and Traveling Waves

PART II

The important point to be understood is

that the natural frequency of vibration of the

A well-established principle of wave motion

wave trains travel~g~,jnopposite ctwa~ e7!27l~7~T ~ that[Tl~e ~ standi~,~fration. A

princi~T~ob-

served when two stones are thrown on the

surface of a pond giving rise to two outspreading wave trains. Midway between the two stones the two identical wave trains approaching from opposite directions are superposed upon each other. There results in this region
a series of standing vibrations of the surface of the pond with stationary nodal points between them. The particles of water vibrate up and down with the same frequency

particles is the same for a standing vibration as for a traveling wave. It has already been explained that the frequency of a particle depends only on its mass and a stiffness factor represented by the restoring force per unit of displacement which, for isochronous vibrations, is the same for each unit of mass throughout the entire structure. The principles of
elasticity show that these proportional restoring forces, acting upon the various particles of unit mass, depend on the shape of
.. ~ -.-. ,.,-*, .----~

..,.

7~S~.
~

for the standing vibrations as for each of the

wave trains of which these vibrations are composed. This illustration taken from mechanics has other parallels. The sound vibrations in an organ pipe with fixed nodal points are similarly explained by a combination of oppositely moving sound waves. In long-distance electric transmission lines standing vibrations with fixed nodes between them may also be
/
~

::1... ::.

.:~
C
...~

...

~
~

produced by the combination of two oppositely moving wave trains. A consideration of Fig. 30 will be useful as an illustration. A, B, C, and I) show successive stages of a standing vibration for each

~__~

quarter of its period, resulting from the superposition of two identical wave trains moving in opposite directions. The crosses represent the wave progressing toward the right, while the circles show a leftward-moving wave of equal amplitude. When the stage E is reached, the cycle is completed and the deflection curve is the same as at the first stage
A. In the stage A the two oppositely moving waves are exactly superposed upon each other so that they add. When each traveling wave has moved onequarter of a wavelength as shown in B they

,...

.~ ~

~
in

Fig. 30. Composition of Two Equal Waves Traveling Opposite Directions to Form a Standing Vibration

deformation only.

Since the shape is the

same for either type of motion, it is seen that

the vibration period of each particle is the

cancel, so there is zero up or down displacement at all points, resulting in the straight line. After the second quarter of a wavelength of motion both waves are again superposed so that the displacements add, but the displacements are all opposite to those shown in A. In stage B the deflections cancel again and in stage E after each wave train has
moved a complete wavelength the deflections

same in either case.

This explanation shows how the speed of a wave train in a turbine wheel with a particular number of nodes may be calculated from the standing vibration frequency of the wheel when vibratingwith the same number of nodes.
Wave Speeds

add again, giving the original curve.

22

deflection

~fhe~ntmjh~W~e train of a particular number of wavelepgt.~

at

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

26 times a second instead of 25 times due to this cause. The revolving coil will then register 2X26=52 cycles per second when the wheel is revolving 10 r.p.s.

one ~particuiar chq.w.dcrjstjc sleed oytly,,. Just as a 4-node standing vibration responds at one frequency so also the 4-node wave trains of which the standing vibration is composed must travel around the wheel at one particular speed. So also for wave trains of 6, 8, or 10 nodes, etc. Each has one particular speed with which it must travel in the disk wheel.
Detection of the Presence of Traveling Waves in a Revolving Disk

On the other hand, the stationary coil now

registers a higher frequency than it registered when the wheel was stationary. The wave train on the wheel is carried forward by the wheel motion at a speed of 10 r.p.s. besides its natural speed in the wheel of 26 r.p.s., so the wave passes the fixed coil at a speed of 10+26=36 r.p.s., and the frequency registered by this coil is 2 X36 = 72, because, the wheel carries a train of two waves in the case assumed. Therefore, the forward-traveling wave registers a frequency of 52 cycles per second on the moving coil and 72 cycles per second on the stationary coil, whereas when the wheel was stationary both coils registered 50 cycles per second. Now consider a case where the 4-node wave train is moving backward in the wheel while the wheel is revolving at the same speed of 10 r.p.s. The effect of the centrifugal force is the same as before so the wave must travel in the wheel with the same speed of 26 r.p.s. as before, but in the opposite direction. The frequency recordad by the revolving coil is 2 X26 = 52 cycles per second, the same as for the forward-traveling wave, since this coil records the same frequency for the same wave speed whether the wave moves past it forward or backward, The effect upon the frequency recorded by the stationary coil, however, is different. Since this is a backward-traveling wave, the forward motion of the wheel of 10 r.p.s. allows the wave to travel backward past the fixed coil with a speed of only 2610 = 16 r.p.s. In other words, the forward motion of the wheel subtracts from the backward motion of the wave as measured by the fixed coil. The frequency registered by the fixed coil is, therefore, 2>< 16 = 32 cycles per second for the 4-node backward wave train. Again, take the case of this disk wheel carrying.both wave trains simultaneously and also revolving at 10 r.p.s. For the forward wave train it will be recalled that the revolving coil registers 52 cycles per second and the fixed coil registers 72 cycles per second; while for the backward moving wave train the revolving coil again registers 52 cycles per second and the fixed coil registers 32 cycles per second. The revolving coil registers only one frequency of 52 for either wave train separately or for the combination, hut the fixed coil registers 72 for the forward-traveling wave and 32 for the backward-traveling wave,

Fig. 31 represents the developed edge of a turbine wheel carrying a 4-node wave train, which moves to the right with a certain particular velocity. C is a small magnetic 5 coil fixed to the stationary diaphragm so it can register the to-and-fro motions of the wheel in an oscillograph by means of the inductive effect of the wheel as it approaches

~CR
Fig. 31. Developed Edge of a Wheel Carrying a 4-Node Wave

(C~ is a stationary magnetic exploring coil and C, a similar coil revolving at the same speed as the wheel.)

and recedes from the magnetic coil during vibration. A coil C,. is attached to an arm which is carried around with the revolving disk wheel, so it also can register the vibration frequency of the revolving wheel in an oscillograph. When the wheel is stationary, both coils register the same frequency. Assume that a wave travels around the wheel 25 times a second. The frequency registered by each coil would be 2X25=50 cycles per second, because for the case of four nodes shown the wheel carries two complete waves. Assume now that the wheel is revolving at 10 r.p.s. in the direction in which the wave moves. Since the wave always has a definite speed in the wheel, the coil carried around with the wheel should register almost the same frequency as before. The frequency would be exactly the same were it not for centrifugal force, the effect of which we have already discussed. The wheel is stiffened by it so that the vibration frequencies are raised and the wave speeds are increased. The effect would not be very great at 10 r.p.s. Assume, for example, that the wave train travels around the wheel

23

PROTECTION OF STEAM TURBINE DISK WHEELS

and both 32 and 72 simultaneously for the two waves superposed, that is, for the vibration in the wheel. To sum up these last two paragraphs, it may he said that when a disk wheel carrying a standing vibration is revolved so that the radial nodal lines are carried around with the wheel, the frequency recorded by a coil carried around with the wheel slowly rises due to centrifug~dforce as the speed of the wheel increases. The frequency recorded by a coil fixed on the diaphragm, so that it does not revolve with the wheel, registers two frequencies, the higher frequency due to the forward-moving component wave train and the lower frequency of the backward-moving component wave train. These two frequencies diverge more and more as the wheel speed is increased.
Frequency-Speed Diagram

PART II

her of waves on thewheei rim, ).~n,bythenumber of revolutions nor second of the wheel, N . 5 This product )/ nN. is the frequency in excess 2 of that of the wheel as measured by the revolving coil, that is, in excess of f~. If H is the higher frequency recorded by the stationary coil and represented by the upper curve, then H fr+ ~ nNs (6) 2 In the same way the lower curve shows how the frequency of the backward-traveling wave as measured at a fixed point is decreased because in this case the wave motion is opposite in direction to the motion of the wheel.

These facts are shown graphically in Fig. 32 which gives a diagrammatic representation that has been found to be very useful. The vertical scale represents the frequency registered in an oscillograph by the magnetic coils, and the horizontal scale the rotational speed of the disk. The middle curve gives the variation of frequency with speed as recorded by the revolving coil. The upper and lower curves show the two frequencies registered by the fixed coil, the upper giving the frequency due to the forward component wave and the lower the frequency due to the backward component wave train of the 4-node vibration. When the wheel is at rest the figure shows that both coils register 50. As the speed of the wheel is raised to 10 r.p.s. the revolvingcoil freqtlency rises to 52 and the two frequencies recorded by the fixed coil diverge, the upper rising to 72 and the lower falling to 32. The gradual rise of frequency of the wheel as its speed is increased is expressed by equation (5) previously derived. Jr = \/f52 + BN52 The upper curve shows how the frequency of a forward-moving wave train, as measured at a fixed point, rises relatively to the frequency detected by the revolving coil, because this wave train is carried forward by the wheel, and is thus passing the fixed coil at a higher speed than it would were the wheel not rotating. This rise in frequency is measured by the number of wavelengths per second that the wave train is carried forward by the wheel rotation which equals the product of the num-

10 Spead
Fig, 32.

20 Rev. per 5ec.

FrequencySpeed Diagram for 4 Nodes

Thus if M is the lower frequency recorded by the stationary coil and represented by the lower curve, M = Jr (7) If the backward-moving component wave train is absent, only the upper frequency is registered by the fixed coil corresponding to a forward-moving traveling wave. If the forward-moving component wave train is absent, only the lower frequency is recorded due to the backward-traveling wave. The frequency recorded by the revolving coil, however, is always the same whether one or both of the component wave trains exist and in whatever relative amplitudes they exist. It is therefore evident that by the use of two exuloring coils as described, one revolving with the wheel and the other being fixed in space, the presence of a forward- or a backward-traveling wave train or both can be detected.

24

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

record lies behind the lower, due to the 30 deg. between the coils, the most likely explanation is that there are high spots on the wheel, 4X30 deg., or 120 deg., apart, which cause this amplitude pulsation. This can he macic clear from a consideration of Fig. 34. The wheel revolves in a counter-clockwise direction as shown by the arrow. The two coils marked 1 and 2 are 30 cing. apart, and are fixed in space. The diagrammatic oscillograph record shows the heats recorded by the two coils, No. 1 lagging behind No. 2 by onequarter of a period. Since it takes four times as much time for successive high spots to reach a given coil as for a given high spot to pass from one coil to the other, these high

The first observation of traveling waves in a turbine wheel was made by means of fixed oscillograph coils installed in the dianhragms of an operating turbine in 1919. Early in 1920 during the investigation of another turbine, several stages were equipped with two such coils 30 deg. apart. One of these stages yielded the records reproduced in Fig. 33.
Observation of Traveling Waves Fixed Coils 30 Deg. Apart

by

Means of Two

The upper curve of Fig. 33 is produced by the 25-cycle alternating-current generator being driven by this turbine. Since the generator has two poles, it revolves once for every
F

Fig. 33.

Oscillograph Records Made by 17th Stage of 20,000kw,, 1500rpm,

23Stage Turbine

(The upper curve is a 25-cycle timing wave, The other two curves were made hy stationary coils 30 deg. apart. This is a case of 6-node forward- and backward-traveling waves.)

cycle, and this alternating-current frequency curve marks off the generator revolutions on the film. The two lower oscillograph curves are the records of the two fixed coils, 30 deg. apart. Time is measured to the right. The upper of these two curves is the record of the first of the two fixed coils. The lower curve is the record of the second coil, set in the diaphragm 30 deg. beyond the former so a given point on the disk wheel reaches this coil somewhat later than the first one, The records of these two coils show a close correspondence, both having a high-frequency oscillation which goes through a low-frequency pulsation in amplitude giving the effect of beats. Furthermore, the upper curve lies behind the lower one by about one-quarter of the distance between the low-frequency amplitude pulsations or beats. Since the time recorded by the anipl itutIe pulsations is four times as great as that by which the upper

spots must be 4X30 deg., or 120 deg., apart

on the revolving disk. The inference is that

these high spots are wave crests corresponding to a train of three waves, 120 deg. apart, and that the waves cause the disk wheel to ap-

proach and recede from the recording coils

periodically. When the wheel is close to a coil, the oscillograph responds strongly, and when it recedes the oscillograph responds less strongly. If there is such a wave train the question now is to determine its d rection and speed. 5 The high spots evidently move in a clockwise direction opposite to the wheel rotation, because they reach coil No, 2 before they reach coil No. 1, as shown in Fig. 33 where the record of coil No. 2 shows a time lead over that of coil No. 1. This means that the wave which causes these low period pulsations is traveling in the wheel against the direction of wheel rotation, and that this wave train travels

25

PROTECTION OF STEAM TURBINE DISK WHEELS

backward even faster than the forward rotational speed of the disk wheel. Therefore, to
find the backward speed of this wave train in the wheel it is necessary to add the speed with which it travels backward past the fixed coils to the forward rotational speed of the wheel. From Fig. 33 it is seen that 63/b high spots pass a given coil for nine revolutions of the wheel, or since the wheel revolves 25 times per second, 6 1/2X25/9=18 1/18 high spots per second. Since there are three wavelengths on the wheel corresponding to the three high spots 120 deg. apart, 18 1/18 high spots per second corresponds to a wave speed past the coils of 18 1/183 = 6 1/54 r.p.s. The wheel revolves forward 25 r.p.s. so the wave speed in the wheel=25+6 1/54=31 1/54 r.p.s. The conclusion is thus reached that this wheel carries a train of three waves 120 deg. apart (corresponding to 6 nodes) and that this wave train moves backward in the wheel 311/54 r.p.s. which is 6 1/54 r.p.s. faster than the wheel revolves forward, so the wave passes the fixed coils 6 1/54 r.p.s. in a backward direction. Thus far there has been no attempt to explain the cause of the superposed higher frequency which shows strongly in the record of Fig. 33. This higher frequency is found to have 60 1/2 periods while the wheel revolves 9 times, or a frequency of 60 1/2X25/9= 168 1/18 cycles per second as measured on the fixed coils, since the disk wheel revolves 25 times a second. This frequency is due to a forward-moving wave train of exactly the same type as the backward-moving wave train, that is, a train of three waves 120 deg.

PART II

that this wave speed is the same whether the wave train travels forward or backward in the wheel. This coincidence is therefore striking evidence of the truth of the statement that the high frequency registered was caused by a forward-moving wave train of

six nodes, that is, of the same type as the

backward-moving wave train. There is further evidence that the wave train recording the higher frequency is moving forward and the wave crests are 120 deg. apart as in the case of the backward waves.

COIl

Record of No.1

Record of No.2--~fi~k~

4
(The polarity

Fig. 34. Diagram Showing 6-Node Forward and Backward Waves Such as are Recorded in Fig, 33

apart, or a 6-node wave train. If such a wave

train registers a frequency of 168 1/18 cycles per second on the fixed coils, its speed past these coils must be 168 1/183=56 1/54 From a close examination of the film record of Fig. 33 it will be seen that the higher frequency of coil No. 1 leads that of the lower

r.p.s., because there are three wavelengths on the wheel rim. Since this wave train is assumed to move forward and the wheel is also moving forward, its speed in the wheel must be less than that registered by the fixed coil by an amount equal to the wheel speed, that is, 25 r.p.s., because the wave train is
carried forward by the wheel rotation. This gives a wave speed in the wheel of 311/54

record by a fraction of a period.

of the two coils happens to be opposite in this record.) This means that the disturbance producing this harmonic moves forward because it reaches coil No. 1 before it reaches coil No. 2. Furthermore this lead is as before about one-quarter of a complete period. This again corresponds to waves which are

r.p.s. But this is exactly the characteristic speed of a train of three waves as it checks
with the speed of the backward-traveling

4X30 deg., or 120 deg., apart. As to the relative amplitudes of these two
wave trains, a casual inspection of the film shown in Fig. 33 might lead one to believe that the amplitude of the forward wave train

wave train of this type already found. It will be recalled (from the first paragraph under the heading Wave Speeds) that a given type of wave has a definite speed in a disk
wheel which revolves at a given speed, and

producing the higher frequency was as great as that of the backward wave train. It is necessary to keep in mind, however, that in an oscillograph record the amplitude is depend-

26

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

lated by formulas derived from equations (6) and (7). ~=H ~/~nN5 (8)
f,.=M+Y nN 2 5 (9)

ent on the induced voltage which in turn depends on both the amplitude and the frequency of the vibration so that higher frequencies have amplitudes recorded which are magnified in proportion to the increase of frequency. For instance, since the higher frequency is about nine times as great as the lower in Fig. 33, the higher frequency would be expected to be amplified about nine times as much as it should be compared with the lower frequency recorded.

Thus

f,.=1681/1875=931/18 f,.= 181/18+75=931/18

Suppose the 6-node wave train should still persist with the speed of the disk wheel whose

In

all

probability the amplitude of the

6-node characteristics are shown in Fig. 35 reduced from 25 r.p.s. to 15 r.p.s. Then the fixed coil would record the frequencies A and B instead of A and B. If there were a revolving coil, it would record the frequency C instead of C. If the wheel were brought to rest with the wave still persisting, both coils would record the same frequency or the standing frequency E for six nodes. It can therefore be seen that if the standing frequency B is measured for a given number of nodes such a diagram can be constructed to a fair approximation, because an approximate value of the speed coefficient can be assumed. For dependable results, however, rotational
tests are necessary with the wheel-testing

30

5pe~d Rev, per So~c,

Fig. 35. FrequencySpeed Diagram for 6 Nodes, Points A, B and C were determined from the record in Fig. 33

backward-traveling wave is greater than that of the forward wave. There are other reasons for believing this, to be considered later. The film record in Fig. 33 which has just been analyzed is the one which is reproduced on page 916 of the fifth edition of Stodolas

book on Steam and Gas Turbines. Fig. 35 shows the frequency speed diagram for this wheel for six nodes. A and B correspond to the high and low frequencies recorded by the film shown in Fig. 33, that is,

machine which will be described in Part III of this series. The frequency-speed diagram may have on it curves for all of the usual types of radial nodal vibrations as, for instance, 4, 6, 8, and 10 nodes. Fig. 33, the record just discussed, is the record of the 17th stage of a 23-stage, 20,000kw, turbine. This record was taken while the machine was carrying 13,000-kw. load. Fig. 36 shows a record of the same wheel, but with only 5000-kw. load. No vibration phenomena developed at this load. The jagged and irregular record repeats exactly for each revolution. It may be regarded as the wheel autograph, and due to slight irregularities in the rim opposite which the coils are placed. These are magnified because of the high speed. Fig. 37 shows where the load has been raised to 9000 kw. Not until the load reaches 11,000 kw. as shown in Fig. 38 do the vibration
phenomena distinctly develop. Fig. 39 shows

the record at a 20,000-kw. load or full load. The wave phenomena when once developed
appear to remain about the same up to

168 1/18 and 18 1/18 cycles per second or to

the wave speeds 56 1/54 and 6 1/54 because in the case of six nodes the carries three complete waves. C the value of ,fr, the frequency of the r.p.s., wheel equals wheel

20,000-kw. load.
Observation of Traveling Waves by Means of Two Fixed Coils 90 Deg. Apart

Fig. 40 gives the frequency-speed diagram for the 17th stage of a 23-stage, 1800-r.p.m.,

itself rotating at the normal running speed of 25 r.p.s. This may be calcu-

15,000-kw. turbine, where wave motion in a wheel was detected by means of two fixed
27

PART II

PROTECTION OF STEAM TURBINE DISK WHEELS

From the frequency-speed diagram shown
in Fig. 35 corresponding to the film in Fig. 33 previously discussed, H = 168 1/18 cycles per see,, M = 18 1/18 cycles per see,, and N =25 r.p.s. 5

coils on the diaphragm. In this case the coils were 90 deg. apart. Fig. 41 shows the record from which the diagram was made. This

case differs from the one previously described

in that the higher frequency recorded due to the forward-moving wave develops only

after the load is removed. The removal of load is shown by the upper curve which registers the electrical frequency of the
generator becoming a straight line, remembering that time is measured to the right. The

higher frequency develops in about one second

Thus from equation (10) HM 1681/18181/18 N 25 n=-=-- =6nodes 5 For the case shown in Figs. 40 and 41 an exact analysis is difficult because where the higher frequency comes out clearly the machine has doubtless increased slightly in speed, due to the sudden dropping of the 6500-kw. load, There can be no doubt, however, that the following interpretation is
very close to the truth.

H = 130~/~ cycles per sec., M = 103z~ cycles per sec., N5 =30 r.p.s.
U, (U U
,~2

HM130)4103/ 2 N 30 5
Traveling Waves

4 node

Feathering Action of Buckets a Possible Cause of

Co

so

LI

ZO

Spe.eUl P~v.ptrSec.
Fig. 40. Frequency-Speed Diagram for 4 Nodes. 17th Stage of 15,000-kw., l800-r.p.m., 23Stage Turbine (Points A and B were determined from the record in Fig. 41.)

of time, corresponding to 60 alternatingcurrent cycles of the upper curve. As this is a 4-pole machine two alternating-current cycles correspond to one revolution, and 30 r.p.s. is the running speed of the machine.
This is known to be a case of four nodes or

of two waves 180 deg. apart, because from the record it appears that it takes half as long for a high spot to go from one coil to another as for two successive high spots to pass one coil. The high spots are twice as far apart as the coils, that is, 2X90 deg. = 180 deg. Another good check is obtained by the use of equations (6) and (7) derived in connection
with the frequency-speed diagram.

H =f,.+3/2nN. Mfr3~flNs
Subtracting,

nN , or n = 5

H-M N

(10)

Thus far nothing has been said about the cause of vibration in the two cases just described. Further study brought out the fact that waves of this sort rarely occur in turbine disk wheels and the phenomenon is confined to unusually thin types of wheels in which waves are easily built up. After these films were analyzed, a satisfactory explanation of the cause was sought. Referring to Fig. 42 in which the circumference of the wheel disk is formed into a wave shape, the relative angular twisting of the two buckets A and B will be somewhat as shown. It will be seen that the axial component from the energy left in the steam at the point of leaving the buckets as shown at C and D will be greater at C than at B, both reactions being toward the left in the diagram, Assuming the wheel to be stationary in space and the wave form to move upward in the wheel to a new position as indicated by the dotted line, it will be seen that the buckets A and B are moved to the positions shown dotted, This means that the force C, which is larger than B, is operating on the bucket A in the direction that A is moving due to the wave transition. At the same instant the force D is opposing the motion of bucket B moving toward the right due to the wave transition, However, since the forces acting in the direction of bucket motion, as at C, are greater than the forces opposing the bucket motion, as at D,

29

PROTECTION OF STEAM TURBINE DISK WHEELS

PART II

Fig. 41.

Oscinograph Records Made by 17th Stage of 15,000 kw., 1800-rpm.,

23-Stage Turbine

(The upper curve is the 60-cycle line current. The other two curves were made by stationary coils 90 deg. apart. 4-node forward- and 1 ackward-traveliug waves are indicated.)

30