0" , ~ -t-ll)

' ~

~ < l)

< z

Copy

" , "-'10

RESEARCH MEMORANDUM S ¥I ~ ~ I:l VJ E-1 !fl V.) CJ

AXIAL-FLOW COMPRESSOR ROTATING-STALL AND riJJT~-S §-::;, r-J a ~ . ?5<

BLADE VIBRATION SURVEY E-1 ~ [.2 r-f

~""" . §j!:J .l.tJ (\J ..

By Howard F. Calvert, Arthur A. MedeIros, a~ t~tJ r-f

Floyd B. Garrett r4 f:i3f) t:l (.) E-1>:> iff ~ ~fJ ~

Lewis Flight Propulsion Laboratory f1 ~.~ ~. Cleveland, Ohio ~ "-'::? t:-.

:::; >.; q f!t t:-. kJ IJj t;j p,. OJ fJ I'-(

if~' I

CLASSIFIED DOCUMENT

This material contains !nformatlon affecting the National Defense of the United States within the meaning of the espionage laws, TIUe 18, U.S.C., Sees. 793 and 794, the lransmission or revelaUon of which in any manner to an unauthorized person Is prohibited by law.

NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS

WASHINGTON May 23,1955

https://ntrs.nasa.gov/search.jsp?R=19930088575 2020-05-29T18:13:56+00:00Z

~ CJ1 (J)

I-'

.--l I

(J) U

- --- . ~ .~ ~---~

NACA RM E54K2 9 CONFIDENTIAL

NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS

RESEARCH MEMORANDUM

AXIAL-FLOW COMPRESSOR RGrATING-STALL AND RGrOR-

BLADE VIBRATION SURVEY

By Howard F. Calvert, Arthur A. Medeiros, and Floyd B . Garrett

SUMMARY

A compressor-rotor -blade vibration survey was conducted on a pro­duction turbo,iet engine incorporating a 13-stage axial - flow compressor with a pressure ratio of approximately 7 and having an air flow of 120 pounds per second. Resistance -wire strain gages were used to measure the vibratory stresses i n the first three stages and hot-wire anemometers were used to detect r otating stall . A vibratory stress of ±3l, 200 pounds per square inch excited by a six-zone rotating- stall pattern was measured i n the first stage at approximately 5540 rpm. The f a ilure of a f ourth-stage aluminum rotor blade concluded the investigation.

INTRODUCTION

Fatigue failures of c ompressor rotor blades have been a major prob­lem in the development of the axial - flow compressor. The source of many failures has been r otor-bla de vibration excited by rotating-stall (refs. 1 and 2 ) . The v ibration survey reported in reference 1 was conducted to determine the cause of t he first-stage rotor -blade failures occurring i n serv ice in a l3 - stage production compressor . The source of excitation for the critical vibration was found to be rotating stall. The blades in the first seven stages were fabricated from an aluminum alloy . Alu­minum alloy, with its extremely low endurance limit, is very susceptable to fatigue failure due to rotor -blade vibration . This aluminum-bladed compressor was replaced by a design in which the blades in the first three stages were fabricated from a 13-percent - chromium steel. The number of blades in the first stage was reduced by one and i n the s ec­ond stage by tWOj the airfoil area taper ratio for all three stages was increased.

As a result of these design changes, the investigation reported herein was conducted to determine the rotating- stall and blade- vibration characteristics of the new production compressor . The investigation was conducted in a NACA Lewis static sea- level test stand . The vibratory stresses were measured by the use of resistance -wire strain gages, and

CONFIDENTIAL

L

2 CONFIDENTIAL NACA RM E54K29

flow fluctuations of rotating stall wer e detected by t he use of constant­current hot -wire anemometers. Steady-state pressures and temperatures were measured at several stations throughout the compressor. The engine was equipped with a variable-area exhaust nozzle to vary the pressure ratio or air flow at a constant speed.

APPARATUS

Turbo j et engine . - The investigation was conducted on a turboj et engine incorporating a 13-stage axial-flow compressor with a pressure r atio of approximately 7, rated engine speed of 8300 rpm, and air flow of approximately 120 pounds per second . The engine, which was installed and operated in a Lewi s sea-level static test stand, was equipped with an adjustable exhaust nozzle to permit va.riation of the pressure ratio or air flow at a constant speed .

Methods of detecting and measuring vibratory stress. - Commercial resistance-wire strain gages were cemented to e i ght blades in each of the first three compressor stages at approximately midchord position as close as possible to the base of the airfoil section . The eight blades per stage wer e divided into two groups of four blades approximately 1800

apart. Lead wires were run from the strain gages to a 19-ring slip-ring assembly mounted on the front of the engine starter. The slip-ring assembly and strain - gage circuits were the same as reported in reference 3.

A 12 - channel r ecording oscillograph was used to record the strain­gage and hot-wire anemometer signals.

Hot -wire anemometer. - A c onstant-current hot-wire anemometer sys­tem was used to detect rotating stall. A O.CX)l-inch-diameter by 0.01-inch-lengt h wire kept at a c onstant average temperature was used in each anemometer . The flow fluctuations were detected from the instantaneous variations in wire temperature. A resistance-capacitor-type compensator was used to obtain the necessary speed of response. Anamometer probes were installed in radial-survey devices located in the first, second, and third stator passages.

The hot -wire anemometer signals were viewed on a dual-beam cathode­ray oscilloscope, and permanently recorded on the 12-chanel recording oscillograph. The method used to determine the number of stall zones in a rotating-stall pattern was the same as used in references 1 and 4.

Pressure and temperature instrumentation. - The engine pressure and temperature instrumentation is shown in figure 1. Radial pressure and temperature rakes were installed ahead of the compressor inlet guide

CONFIDENTIAL

..Id () en p

~ I

If.l o

NACA RM E54K29 CONFIDENTIAL 3

vanes, compressor outlet, and the exhaust tail pipe . The air-flow meas ­urements were calculated from the pressure and temperature measurements at station 1. All probes were spaced at area centers of equal annular areas.

PROCEDURE

The investigation was conducted in a sea-level static test stand with the compressor inlet and turbine exhaust open to the atmosphere. The engine wa.s operated over a range of speeds from approximately 4500 to 8300 rpm with the ad<lustable exhaust nozzle set at various positions. The minimum nozzle area used was limited by maximum allowable tail -pipe temperature (11660 F), or the minimum area obtainable with the adjust ­able nozzle. At each condition, data were obtained to determine the compressor performance, and rotating-stall and rotor-blade-vibration characteristics.

RESULTS AND DISCUSSION

The data obtained in this investigation are discussed in their relation to the compressor performance and the effect of rotating stall on the rotor-blade vibration . The compressor used in this investigation was of the same production model as reported in reference 2.

Rotating-Stall Characteristics

The performance of the 13-stage compressor used in this investi­gation is presented in figure 2 as total-pressure ratio and adiabatic temperature-rise efficiency plotted against equivalent air flow . The data points along the solid line of figure 2 (a) were taken with the exhaust nozzle set to give the approximate rated tail-pipe temperature at rated equivalent speed (8300 rpm). The data points along the lower line were taken with open exhaust nozzle or set to give maximum avail ­able area, and the data point.s taken along the upper line were taken with an exhaust area reduced from rated area.

The shaded section in figure 2(a) represents that area of opera­tion where rotating stall was encountered, and hence, the unshaded sec­tion represents that area of operation where the compressor was free of r otating stall. Examination of the r ated operational line indicates that there was no "ma.rked" change in performance as the compressor became free of r otating stall . In the performance data in figure 2 of refer­ence 2, a line was drawn indicating the operational limits of limiting turbine temperature, sta.ll or surge, and at the lower speeds, minimum available exhaust area. As the result of a failure of a fourth - stage

CONFIDENTIAL

4 CONFIDENTIAL NACA RM E54K29

rotor blade during this investigation, data were not obtained to draw this limiting type of line on the performance map of figure 2(a).

Figure 3(a) is a plot of vibratory stress against actual engine speed and percent of equivalent rated engine speed. A vibrating blade is affected by the actual speed; therefore, the vibration data will be discussed in terms of actual speed. The percent of equivalent rated speed can be used to locate the rotating-stall data on the performance curve in figure 2(a) . Complete compressor data were not obtained for these data points . The data for this figure were obtained with the com­pressor operating with reduced nozzle area . At the bottom of the curve in figure 3(a) is a bar graph presenting the detailed rotating-stall data. As the compressor speed was increased, the stall-zone sequence was 5, 4, 3, 6, 5, and 4 . The three-zone stall was the strongest and persisted over the longest speed range . At approximately 6240 rpm (74.5 percent of rated equivalent speed) the compressor was free of rotating stall.

Figure 3(b) is a plot of vibratory stress against actual speed and percent of equivalent rated engine speed. These data were obtained with the compressor operating on the open - nozzle line of figure 2(a). At the bottom of the curve is a bar graph presenting the detailed rotating­stall data. As the speed was increased, the stall zone sequence was 4, 3, and 5. As in figure 3(a), the three-zone pattern was the strongest and persisted over the longest speed range. At a.pproximately 5870 rpm (70 percent of equivalent rated speed) the compressor was free of rotating stall . The operations of the jet engine with the reduced nozzle area or the compressor closer to the stall line increased the severity of the stall, the rotating-stall speed range, and the number of patterns. The six-zone pattern was observed only while operating with the minimum exhaust area..

Rotor-Blade Vibration

Figure 4 is a plot of frequency aganist "actual" engine speed. The dashed lines represent the natural bending frequencies for the first-to fourth-rotor stages corrected for the effect of centrifugal force (ref. 5). The lower series of data points represents the relative stall frequencies for the three-, four-, five-, and six-zone rotating stall over the range of speeds observed. The frequency of excitation due to rotating stall for a rotor blade is the stall frequency relative to the rotor blades. The relative stall frequency is defined by the following equation:

where

CONFIDENTIAL

NACA RM E54K29

fl relative stall frequency s

N compressor speed

CONFIDENTIAL

A number of zones in a stall pattern

fs absolute frequency determined from hot -wire signals

The upper series of data points represents the actual measured blade vibration frequencies for the first three stages . Over most of the speed ran ge, the blade frequencies followed the dashed lines repre ­senting the static frequency corrected for the effect of centrifugal force . However, over the speed range of from approximately 4700 to 5700 rpm, the first - stage blades followed a line "B" repres enting a frequency equal t o t wice the relative stall frequency for the three-zone stall. Over a speed range of from 5300 to 6000 r pm, the second­stage blades als o f ollowed this two times three - stall line . These first- and second- stage blades were being forced to vibrate by the three - zone rotating stall . In reference 2, the first - and second-

5

stage blades were also forced to vibrate at twice the relative fre ­quency of the three - zone rotating stall and not their own natural fre ­quencies . However, an exception did occur at approximately 5540 rpm ( fig . 4) . At this point the first- and second- stage blades were forced to vibrate at the frequency of a six- zone rotating stall . Shown in figure 3 (a ) are the stress data for the first three stages as the com­pressor was operating on the reduced -nozzle line in figure 2(a) . The first stage had peak stress valUes of ±l9,OOO and ±31,200 pounds per square inch at approximately 5050 and 5540 rpm, respectively . The first peak stress was observed with a three -zone stall and t he second peak with a six- zone stall . The stresses observed in the second stage were consid ­erably lower than for the first stage . The average stress of the many peaks between 4800 and 6000 rpm was appr oximately ±12,OOO pounds per squar e inch . The three - and five - zone stalls were observed with these vibrations . No str esses over ±6000 pounds per square inch were observed in the third stage.

I n figu r e 3 (b ) , with the compressor operating on the open - nozzle l ine, the first stage again had a peak stress at approximately 5000 rpm appearing with a three - zone stall . No peak stresses wer e observed at 5500 r pm because the six - zone stall was not present .

An examination of figures 3 and 4 indicates that the peak stresses existing at approximately 5000 rpm occurred at the point of intersection of the blade frequency line, and the l i ne representing t wice the relative f r equency of the three - zone r ota t i ng stal l . The peak stress of ±3l,200 pounds per square inch (fig . 3(a)) was excited by a six- zone rotating­stall pattern . This stall pattern was observed over an extremely short speed range.

CONFIDENTIAL

6 CONFIDENTIAL NACA RM E54K29

Figure 5 is a reproduction of sections of two oscillograph records . Section A is the record of the blade vibration occurring with a three­zone pattern at 5500 rpm. Here the first-stage blades were vibrating at a low maximum stress of ±7284 pounds pel' square inch . At this compressor speed, the three-zone stall had little effect on the blades, because the reltaive stall frequency was approximately 15 cycles higher than the natural blade frequency. Section B is a reproduction of the oscillograph record showing the maximum stress recorded in the first stage in this investigation. The compressor speed was 39 rpm higher than that for sec­tion A. However, in section B the compressor was operating with six- zone rotating stall. The rotational speed of the six- zone stall was higher than the three-zone stall and, as a result, the relative stall frequency for the six- zone stall was approximately 18 cycles lower than for the two times three stall (fig . 4) . This reduction made the relati e stall frequency equal to the natura l blade frequency. With this fundamental excitation, the exciting forces were high a.nd, as a result, the maximum vibratory stresses were observed .

Fourth- Stage Rotor -Blade Failure

After approximately 30 hours of engine operation at various speeds, the compressor case was removed to install additional instrumentation . As the case was removed, it was noticed that about 60 percent of a fourth ­sta.ge aluminum r otor blade was missing . The blade section was found wedged in the eleventh-stage stators. Figure 6(a) is a photograph of the failed blade and figure 6(b), a photograph of the fracture surface . An examination of this surface indicated that a fatigue failure had occurred .

Line A representin g a frequency four times the relative stall fre ­quency of a three - zone stall (fig . 4) intersects the fourth - stage blade frequency at approximately 4950 rpm . At this intersection, resonance would occur between the blades and rotating stall . Since strain gages had not been installed on the fourth stage, no blade vibration was re ­corded; however, it can be assumed that the fourth - stage blades would respond to the three - zone rotating - stall excitation as did the first and second stages (fig. 3).

SUMMARY OF RESULTS

A rotor -blade vibration and rotating- stall survey was conducted on a production l3 - stage axial - flow compressor to determine the rotor-blade vibration and rotating-stall characteristics of a compressor with pro­duction changes . Strain gages and hot -wire anemometers were used to de ­tect the blade vibrations and rotating stall, respectively . The results may be summarized as follows :

CONFIDENTIAL

Vl CJ1 CJ)

I-'

NACA RM E54K29 CONFillENTIAL

1. A maximum vibratory stress of ±3l, 200 pounds per square inch was measured in the first stage at approximately 5540 rpm . This vibration was a fundamental excitation by a six-zone rotating stall.

2 . A three-zone stall pattern was the predominate stall pattern observed.

7

3. Stall patterns with three, four, five, and six zones were observed .

4. The investigation was ended by a fourth-stage aluminum rotor-blade failure due to blade vibration excited by a three - zone stall pattern.

Lewis Flight Propulsion Laboratory National Advisory Committee for Aeronautics

Cleveland, Ohio, December 2, 1954

REFERENCES ,

1. Huppert , Merle C., Calvert, Howard F . , and Meyer, Andre J., Jr.: Experimental Investigation of Rotating Stall and Blade Vibration in the Axial-Flow Compressor of a Turbojet Engine . NACA RM E54A08, 1954 .

2. Calvert, Howard F., Braithwaite, Willis M., and Medeiros, Arthur A.: Rotating-Stall and Rotor -Blade-Vibration Survey of a 13-Stage Axial­Flow Compressor in a Turbojet Engine. NACA RM E54J18, 1955.

, 3. Meyer, Andre J . , Jr., and Calvert, Howard F .: Vibration Survey of

Blades in 10 -Stage Axial-Flow Compressor . II - Dynamic Investigation. NACA RM E8J22a, 1949 . (Supersedes NACA RM E7D09.)

4 . Huppert, Merle C.: Preliminar y Investigation of Flow Fluctuations During Surge and Blade Row Stall in Axial -Flow Compressor. NACA RM E52E28, 1952 .

5. Timoshenko, Stephen: Vibration Problems in Engineering. Second ed . , D. Van Nostrand Co., Inc., 1937, p. 386.

CONFillENTIAL

e CONFIDENTIAL NA.CA RM E54K29

(a) Cross section of engine .

Station 1 3 5

I CD-3672 I

(b) Views of several stations ( l ooking downstream) .

Symbol Type of probe Station

1 3 5 0 Total pre s sure 20 12 15 0 Total temperature 12 6 18 6. Wall static tap 4 - -0 Stream static 4 - -

Figure 1. - Cross section of engine showing stations at which instrumentation was install ed.

CONFIDENTTAL

----~~--- -------~--------~------

r ,

NACA RM E54K29 CONFIDENTIAL 9

r-f 7 . 0 to L{) to

6 . 0

(\J I

CI.l (.)

0 5 . 0 -rl ....,

'" H

Q)

H ;:l

~ 4 . 0 Q)

H 0. , rl

'" ....,

~ 3 . 0

B perJen t ;a ted equivalent f----Rated nozzle area (pos i tion 1)

)~ speed

------ Open (position 2) !>I 00 --- Closed (position 3)

J, VJ5 ~ II / /

J) 'J10 ~/ /l 85

~ ~/

80

A ~;: Rotating- ~~ ,:0 Free of rotating stall stall region"

2 . 0 ~ ~

L ~ ~

~ ~ V t ~ :::::::; E ~ ~ ~ 1.0

(a) Total - pressure ratiO . 9

0

J-p9-9- N tq>,- (. I"

.cJ V v

~ It'--'

r

~, 'V

Y

u 6 0 20 40 60 80 100 120 Equivalent weight flow, Ib/sec

110 160

(b) Adiabatic temperature-rise efficiency .

Figure 2 . - Compressor over-all performance.

L CONFIDENTIAL ~-~------------'

..... ., 0.

+1

(") .;; .,

~ ., H ....,

CJ) ; ~

32 (

[==1--1--JL--~-+--4---~-t--+-~---t--t--1---r--t'~IIIi'--!---j--t--~~~-r--r-~ 28 I

24 I [~1--1--JL--~-+--4---~-t--+-~---t--t--1---r--tlrltltl--!---j--t--~~---r--r-~ 20

~ 16 b II 11

/IV I I \ 12

j 1- I I \1 K I In;: f\\ \ /~ ,~\ 10-1 ~\ \ L I f"'{

8 lP-, II '~~-d I t/F\f I /\ _l ~ laD '\.

'\. '\ ~tage

1

~# 4 t" ---f'r-- I""

f- - 2 ~~ I~ I LJ ~ I ~~~J -r~X~

~vr~tf:_~~n\ _Jl\~ l ' rl~ ·

~ .... 04 .0 4 . 2

IV .. [ 3

our Four zone~--~--~ ............... :.; .... ;.:.:

r-o

4 . 4 "~······"·"·"·"'·""·"·"·········"···'···""···r·"m".'.'·1···· ... ····1···· .. f·· .. ············ 51. 6

4 . 6 4 . 8 5.0 5 . 2 5 . 4 5 . 8 6 . 0 6 . 2 6:4XI03 Actual engine speed, rpm

I I I I I I I I I I I

47 . 9 50.3 52 . 7 55 . 1 57 . 5 59 . 9 62 . 3 64.7 67 .1 69 . 5 71 . 9 74 . 3 76 .7 Percent of equivalent r ated speed

(a ) Cl osed exhaust nozzle .

Figure 3 . - Vibratory stress es and rotating-s tall patterns measur ed i n first thr ee s t ages .

19S£

t-' o

(")

~ ; ~ t-l

~ ~

~ t:9

~ N to

oM CD 0-

+1

CD til

(")

~ Q)

1-< +' CJ)

i ~

CS-2 back 3561

20 XI 03

1 2

~ I I I I '\ g~_ A~q.l/~\Ljl /fh I I srBl,O- ?\ 8

Q 1

tI ~_LC'{' I -r-r _IA '~I~N1t:LJ~ r,,~ stage

'I 1 1 tttfti1\tffFr~~rrrHt~~r 1 =~ Five zone 3

I , t Three zone ~ I I I I I j..!!:.;.: .. :.:.:.:.:.:.:.: ::: .. :.:.:.:.:.::::::::1'::.:: .. :::.:::.:.:.:.:1'.:.:.:.:.:.:.:.:.:.:.: •. :.:.:.:.:.:.:.: •. : ... :.:.:.:.:.:.:.;,'.:.:.:.,.'.:.:.:.:.!

o 3 4 . 0 4 . 2 4 . 4 4 . 6 4 . 8 5.0 5 . 2 5.4 5.6 5 . 8 6 . 0 6 . 2xlO Actual speed , rpm

1

47 . 9 50.3 52 . 7 55 . 1 57 . 5 59.9 62 . 3 64 . 7 67 . 1 69 . 5 71. 9 74.3 Percent of equivalent rated speed

(b) Open exhaust nozzle .

Figure 3. - Concluded. Vibratory stresses and rotating-stall patterns measured in first three stages .

~ &;

~ l:r:I

~ t\) to

(")

~ H

i ~

t-' t-'

"0. 450 _~~~~rrrr='==. rrrrrrrrrrrr"=rrrr~==.~=~=.==· '='='='=H===_=,==_~='=>=H=='===="''''.rrH=m' .m''='"HH'= ' Four t imes relative

frequency of three-zone stal l Twice re l ative fre-quency of t hree - zone

400 .11 ••• I ... IIIII ___ la ___ I ... 1 '-. stall

Measured blade frequen cy

0 Stage

350 1~1mlIm!1llllili1lOOl!!ll~I!!Ill!m~mmll-!!l!!!m!:'!!lliillmr!!!m!:'!!llii~~I'lmillJ'!!lliimlllilJ'!lilltr!'.mfE l 1 0 2 0 3

Re l ative stall (") If frequen cy

~ ., No. of zones 0. 300 0

~ 3 2 H ;;; 4

§ 0 l:l 5 .::: \l 6 ., 1 " g' 250 11

~ s..

(k,

:.: ,Ji!! n ~m

i , 1m "" ,.

200

150 .

100

5. 0 Actual engine speed, rpm

Figure 4. - Rotor-blade and rotating-stall frequencies for l3-stage aXial-flow compressor.

199£

~ l\)

(")

~ H

§ ~ t-i

~ ~

~ t?i!

~ l\) to

" LI

'vv yl fl

" I. .' ~ ), ~'\JI , .\

(")

~ H

A' I, h \~ tN I ~ ,

~ ~V Ii Vv IN

§ ~

I' " r. VI, lip , I I IV

01' t! .. I ~ I \ '/

", ,\ . \ .,1 I, .,

" \\ ~ ~I W ,,",

r0. ~AJ\ f.\oI ~"J

ff ~W h rv

~"'(;re5S, pSl..

~~~ ,. ,I ~ ~ : I--~, t:;: f."i: :<5990

, ~ ... " .

"'" fv\ /\1' \!~ r'V' Vv Ilf 'h rlf Ih. I • • A ' l h V, II tV IVV rV\ VII f'\' 1/

+7284

N' \ N~d h ~, ~ I'IV ~V' ~~~ -V 'Vv y

"""rv'/ ~/ .. ~A ~W~~ I/~~' ~ \1\ \~ w fJ" ~ IA rJv IhA

'V IV' ' V rV . +10,104

IN1 W \~~ rY-I vh' rN ~vW tNlv\ 'N~ 1{ rvu ~ A, Ll A IAA \A W I~ W IV IV IV IV IV'

... 9~ V'I ,,, rV iN hM\.viAP ~ IV'; N (tv W W M IA f'I' VV ~V

±l-2,640

I~ ;~ ~'" \' " \1 1\ A~~ WI t~' 'v ~:I ~ I{~ ~ VV N\ \Iv 1\\ rV' 1" It I ~ rV / 1_" ±1:3,480 '1/ \, ~\ tV "V III lV' ":;V " ~V' ~V Ail 'I! I' r~~ ~ ~ W Iv "1\ /J \!V ,I \' V ~ .! I! '.' VI) ~

\ ;t3738 1 /

~)f\~ ,W II~ ~, '" ,,\1, '~'N\ltl ' ~' 1M ~~ ~~ ~ IA: ~~ 1'-1., 11# i,~ VA M " " :,: " I, I,' ~ +5543

L .... 'v'I ~ I\ • ..,.,-~\ l ,,. ~A\~ •

~~ffiw. rV

ln~ ~ ~VahY rv ~ ~ IVV\ rv (a ) Three- zone stall pattern.

' . l ~ 'I

\II( '>I" IV

.. 1M IV

W~ ~ W ~ IItv W ~' IV

,~~

~ ~ ~ W M ~ Il ~~, I t~ ~~

'Y

~ IV W In

35,61

Stress. psi %25, 704 ~ •

~~ I ~ ~A I An ~ A I An\A I A I\ AH~ A I A ~ A I A ~AI ~1 ~ A~l l A M A I AIA I A\A I A ~ A I A\A I Mrr lf~

~A~~\A\!~ ~ ~"A~ ~ ~~~: I ~~~, ~~~AAI \AIA~ ~A~A IA; ~A I AI \A IM\A I Alnn~ ~A IM~M !t IVI1V' V"I I\l/i\l\ I\I\lI\lIVVIV ~V IV V IV rVV V IV V' IV IV IV VI/V IV rv

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~IA~ I A .~~~~~~~r~~~rV N \~ W~ WIV~ ~~~~Llrv. L.~~ l~ I A~ IA ~~ rv \h I A! 'ht ~lAIMI AA I A nI ~ IAAI"N\fu ~IAA~IAH~~~ VV I" V VV ,· I V ~V rv VIV ~VI\ 1V vy!Vm""',· .. IVlrV\ rV ty

W~~~W~~ MN IN1~~AW~M~ I~~ W ~h.I.'h\ A ~.V ~M~ V lN~/ A ~J\ I~i ll1'~Y V~~ f~ ~~~VVY\INf~ ,~II/\ \t\1 W~fNMW~~ Mw,~ \4 ,~~.,' ~/llt ~ l! ,W 'IW~fJ ,'(I/I~l~AA ,II ,Vl ~If It ',Ii rV' ~Iy ~lI.,... ~

''''~3~\ r' T"'T 'I ,... 3

Six-zone _tall •• L . I.. ~J. J .~ ,III I. IA. ~ I VV,JJNWlIlltt.ll1.,""""mllW'lrV,\MAM, y,,'Nv1"V1, lfVI' rw ~!. ~ IV".'" "V 'V

5539 rpm

(b ) Six·zone .tall pattern.

Figur e 5 . - Oscillograms of vibration caused by r otating stall..

~ ~

~ I;zj

~ N to

(")

~ H

§ ~

f-' ~

14 CONFIDENTIAL NACA RM E54K29

(a) Over-all view. C-35173

Figure 6. - Failed fourth- stage blade.

CONFIDENTIAL

z > ()

> t-., " .. ii) '< , '" , .., ';" '" '" Co>

'" 0

(")

~

~ ~ t-t

3561

_. - ---_. --

!4 ~

C-35l74

(b) Fracture surface.

Figure 6. - Concluded. Failed fourth-stage blade.

~ ~ §1 t:>:j U1

~ N W

(")

~ ; ~

r' CJl