TURBOMACHINERY ANALYSIS AND PROTECTION by

John S. Sohre Turbomachinery Consultant

Vernon, Connecticut

Born and educated in Germany. Graduated State Institute of Tech­nology Chemnitz, Germany 1951. Professional E n g i n e e r ( Pennsyl­vania). Member, American Society of Mechanical Engineers.

Mr. Sohre has had twenty-eight years experience in shop, operating departments, and design work. Eight years with Elliott Company (Section Head, Mechanical Design and Engi­

neering Mechanics Section, Turbines and Compressors). Seven years with The Terry Steam Turbine Company as Chief Development Engineer. Had responsibility for development of Terry's type IS line of turbines for high­speed, high horsepower service. Since July 1970 in prac­tice as an independent consultant for Turbomachinery Design, installation and problem correction.

ABSTRACT a . Predic t ion and Analysis o f the R eliabil i t y o f

Turbomachinery Insta l la t ions-Influence of var ious com­ponen ts on over -a l l re l iab i l i t y , such as o f turb ines, com­presso rs, gears. motors, p ip ing, foundat ions. Cr i ti ca l i tems i n reliab i l i t y evaluat ion .

b . Iden t i f ica t i on o f Problem A reas i n Existi n g In­s ta l la t ions (" trouble-shooting" 1-How to pred i c t and iden t i fy problem a reas and probab i l i t y of shu tdown o r componen t fa i lure. Applicat i on of protect ive dev ices and supervisory instrumen ta t ion t o preven t or m in im ize losses.

c. Analysis of Turbine Opera t ion to Pred i c t Future Problems-Cri t ica l areas i n turb ine operat ion . Common fa i lur es caused by mal-operat i on . Instrumenta t ion and protect ive devices t o min imize risk.

PREDICTION AND ANALYSIS OF THE R ELIABILITY OF TL'RBOMACHINERY INSTALLATIONS

During the l as t few yea rs the size of petrochemical plan ts and chemical plants h as i ncreased considerably . This means that the turbomachi nery which is used in t hese plants h as a lso become l arger. The machi n es operate a t h i gher speeds, and they are n o w far m ore complex than before . In add i t i on , t h e t ime between turn-arounds h as now increased t o two, t hree, and even f ive years. As a resul t of the i n creased output o f p rod­u c t of these large p lants, the l osses which are experi ­enced as a r esu l t o f an acciden ta l shutdown are very h i gh , and n umbers of 100,000 do ll a rs per day and even more are n o t u nusua l .

As a resul t o f this si tua t ion i t became evident tha t there is a n eed for more systematic means t o a i d i n

1

t rouble-correc t i on and to reduce shu tdown t ime . Also, means are requ ired t o predict rel i ab i l ity in advance, and t o eva lua te re l iab i l i t y versus i n i ti a l cos t , opera t i ng cost, and shutdown costs.

To indicate the rel i ab i l i t y of a t u rbomachi n ery i nstal la t ion a ser i es of cu rves can be developed, which are keyed to a common denomina tor , for compara t i ve pu rposes. Fig. l shows a set o f such c urves. This approach h as t he advan tage of giv i ng a genera l idea of t h e i n terac t ion o f the var ious fact ors. For example, i t can b e demonstrated why a perfec t ly good mach i ne may no t r un wel l i n a poor insta l la t ion. Vice versa, i t can be demonstrated why a poorly applied m ach i n e o r poor ly designed mach i ne may run qui te we l l i f i t i s wel l i n ­stal led, carefu l ly opera ted, and capably ma in ta ined. The i nd i vidua l re l iabil i t y factors are mu l t ip l ied to f ind over­a l l factors for m ajor componen ts or for the en t i re i nsta l ­l a t i on . For a complete explanat ion o f t h is method see References ( l I and ( 2 I .

O f course there are many shades o f gray between the two extreme examples q uoted above. We need a method to determine wh ich shade o f gray is invo lved in a par t icu lar trouble job. The approach presen ted can also show which type of equi pmen t i s best for cer­t a in ranges o f operat i ng cond i t ions, such as speed, pres­sure, temperature, and o the r operat i ng parameters. In Reference papers ( l I and ( 2 J w e used t h i s approach t o a id t roub le-shoot ing . The procedure wo rks as fol lows: P lo t ted versus speed is i n a l l t hese c urves a so ca l led "Rel iabi l i ty Facto r." This facto r var i es between zero and 2 .0 i n most o f the p lo ts . We assumed a rel i ab i l i t y factor o f ONE to represe n t a n average rel i ab i l ity for a given appl ica t i on . In o ther w ords, i n a petrochemical p lan t , a given mach ine is expected to run con t inuously between turn-arounds and to requ i r e a cer ta in amount o f ma in tenance in between. A lso, average figures can be compiled for the so-cal led "norma l" break-down r ates for th is type of p lan t. We h ave arb i t rar i ly chosen t o use t h e facto r o f ONE f o r an i ns ta l la tion w h i c h h as th is break-down rate and rel iab i l i t y . A greater factor shows that the equipmen t is l ike ly t o run l onge r be tween break­downs. For example, a facto r of TWO would ind icate that the uni t wou ld run twice as long as a u n i t w i th a so-ca l led "normal" break-down ra te. The rel i ab i l i ty factors for var ious i tems can be p lot ted versus speed, load, pressure-or any o ther cr i t i ca l parameter.

The first plot shows the type of equ ipmen t: A compressor w i l l have a re l i abi l i ty ra te somewha t d iffer­en t from a turbine . The motor and gear combinat ion wi l l aga in have d ifferent re l iab i l i t i es. This si tua t ion wi l l b e i n fluenced by speed-a turb ine w i l l usua l ly be more re l iable at h i gh speeds, whil e a motor-and-gear combinat ion can have very good rel i ab i l i t y at low speeds, b u t i t can have fai r ly poor rel i ab i l i t y at the h i gher speeds. In the nex t p lo t the equipment s ize is shown .

2 PROCEEDINGS OF THE FIRST TURBOMACHINERY SYMPOSIUM

PLANT AND EQU/Pt1ENT lJES/GN TYPE OF &OUIPI"'ENT

0 10 XJ S I' E E ZJ - /IJ()() IIIW

START liP TIME

EQlJIPHENT SIZE

() /0 10 Jl'eE]) -l()f)I)IIPH

PAESSIJRE

2

I

NO. OF BEARINGS IN rRA/It

-

-�

__..,

:-::: � "'

3

so 6 8 10 20

0 10 20 .SP�EJJ -I(J()()IIPH

TE/'1PERA TlJRE 2 �----�------� 2 \

I

2

� I

0 so /()() COl.� .START TIHE,I'IINVTES

TO I'UI.I. .SPI6])

COUPLING TYP£

0 /0 20 .JI'Ee])- /()()()ltP/'1

\

� .....

I

0 /()()() 1/)()() .lfJIM 11M)

2

� I

HAX. I'RESSI.IIl�, P.Siti

TYPE OF CASING .SUPPORT

0 10 2(J SPEED - !()()() RPH

Figure 1.

! • � ·I I

z

� I

0 S(J() /IJ()() NAX. TEHPeRA TUilE �

STARTING FR£/JtiEICY

0 10 2() JPEID -ltJtJ()Ifi'H

TURBOMACHINERY ANALYSIS AND PlWTECTION 3

The span between hearings was selected as the significant parameter. I t can be seen, t ha t the l a rger m achines wouid n o t be very reli able a t h igh speeds, because the faster a machine runs, the shorter and smal ler i t mus t be, m os t ly for reasons o f cr i t ical speed and rotor dy· namics . The next curve shows the n u mber o f hearings in the ent i re un i t-or in the train as i t is o ften cal led_:_­and the curves are p lo t ted h ere for three, four , s ix, eight , t en , and twenty hearings . The n umber o f bearings is a good ind ica ti on o f the complexity of a u n i t. The more bodies and the more a u n i t the more com­p lex i t mus t be and the more opport un i t y w il l i t h ave for get t ing i n to t roub le . These curves a l l represent da ta perta in ing t o the equipment i tself .

On the same chart , Fig. I, are curves showing operat ing condi t ions , such as pressures and tempera tures . Also shown i s start -up t ime and n umber of s tart-ups per year. Qu i te evident ly , a un i t which i s s tarted only on ce a y ear and then runs cont inuously wil l be far less l i ke ly t o s uffer a break-down than a un i t which i s s tarted and

da ily , Thi s i s reflected i n th i s curve . The faster !:he u n i t rur;s , the more pronounced wil l th is effect he .

One th ing i s o f i nterest t o n o te here: Once the operat ing condi t ions are given hy the p rocess de:>igner, the designer of the rotat ing equipment has very l i t tl e choice concerning the span he would l ike t o use between bearings, o r the number o f heari ngs. This i s essent ia l l y fixed, especia l ly where the equipment approaches the l imi t o f capac i t y o r speed obta inable wit h rota t ing equ ipmen t . Such i s the case , for example , w i th syngas compresso rs for ammon ia syn thesis , w here turb ines m a y he rated 35,000 H P a t ] 1,000 RPM, and where extrac­t i on m ay also be requ ired . In such cases, there are n o t many sol u t ions, and a designer has t o compromise and l ook for t he be s t po::�s ib le so lu t i on concern ing the n um·

her o f bodies i n the t ra in , the hearing spans, and the number o f s tage�:' for each machine .

Thi s br ings u s t o a very impor tan t poin t , n amely , tha t a design can often he grea t l y sirnpli fied ---and tbe rel iab i l i ty greatly improved----by bet ter cooperat ion be­tween process designer and equ ipmen t designer . One example i s the vac·uum on a condensing tu rb ine . For example, a smal l ehange o f d esign vacuum on a con ­densing mach i n e can sometimes re.su l t i n a s ingle-flow turb ine in s t ead o f a double-f low machine , w i th a cor­responding improvement o f the rel i ab i l i t y factor . The hearing span would get shorter, the i n i t i a l cost w ou ld be lower, and of l en the i s bet ter under n o rma l operat ing cond i tions . Qui t e o ft en t h e reduced vacu u m would o n l y he n eeded d ur ing over load cond i t ions such as dur ing s tar t-up o r emergt�ncy , o r dur ing o ff-design condi t ions . Dur ing normal operat ing condi t ions the machine would operate with the des ign vacuum. B u t the way t h e speeif ica ! ions a r e of ten written, a designer has n o choice h u t t o quo!<' his mach i n e for the operat ing condi t ions as specified. The resu l t may he considerahle economic disadvantage and redu ced reliab i l i ty . This may serve as a n example h ow c urves as shown may he used t o op t imi ze an in s ta l la t i on .

Fig. 2 shows the effect of pip ing on rel iab i l i ty . The p ip ing af feds the equipmen t lly the forces and moment.s i t c reates a t the connec t ions t o t h e equipment . The p ip ing may he pushing t h e en t i re machine o u t o f alignment and eau;;c v ibra t ions o r rubs . Esp1:cia l ly on h o t maehines---sueh as h igh tempera ture turbines---the p ip ing can warp the This may resul t i n leaking flanges o r i n terna l rubs, caus ing ser ious v ibrat ion prob­lems . Also , the p ip ing may c ause t he cas ing guides t o h ind , o r the casing may actua l l y l i f t off i ts suppo r t feet . This , then, ean create a l l s orts o f vibrat ion problems

PIPING

PIPE SUPPORTS EXPANSION :JOINTS

��--4-�nW���-4 � ��--������

0 II) 20 .SPEED -10(}() RPH

Figure 2.

4 PROCEEDINGS OF THE FIRST TURBOMACHINERY SYMPOSIUM

and interna l rubs. To correct problems of thi s type i n t h e f ie ld c a n be very costly . Curves as shown may help to f ind the most economica l solution .

F ig . 3 shows the i n fl u ence of the foundatio n upon re l iab i l ity o f r otatin g equ ipment. Very b r ie fly , the fun ction of the foundation i s to ho ld a un i t i n a lign­men t , n o t only dur i ng n o rma l conditions , b u t also du r­i ng abnormal condi t i ons . The forces the foundat ion has to absorb are the p ip ing forces, and such forces as may be developed b y expansion joints in the p ip ing system o r on the condenser o f a turb ine . Other factor s whi ch can cause misa l ignment a re sett l i ng, shr inkage, o r un­equa l heatin g o f the foundati on .

I t i s v e ry important t o preven t build-up o r t rans­m i ss ion of v i b rations . The curves shown indicate the e ffect o f foundat ion r ig id i ty upon rel iab i l i t y . Evidently a h igh speed machine must be he ld in closer a l ignmen t than a l ow speed machine , and the effects o f m isa l ign­men t are more severe . Therefore the rel i ab i l i ty fac tor drops off with speed for a founda t ion wi th poor r ig id i ty . The v ibrat ion characteristics are expressed i n t e rms of foundat ion weight . The n ormal re lationshi p be tween foundat ion mass and un i t m ass is 1 .5 t imes and this i s used f o r a rel i ab i l ity factor o f on e . If t h e mass i s less than that , the r el i ab i l ity wil l drop off, because the ma­ch ine becomes more sensiti ve to smal l upsets and smal l operat ing problems, or , in case o f emergency , the risk of ser ious damage is greater with a l i gh t foundat io n than w it h a massive one . The effect o f major s tructura l resonance i s a l so shown . This , o f course, wi l l great ly reduce the rel i ab il i ty of the un i t , especia l ly dur ing start· u p conditions , u psets and emergencies . The faster a machi ne runs , the more pronounced wi l l thi s effect be .

The th i rd graph shows the i so lation from the sur­round ings. Transmiss ion of v ib ration from and t o the uni t h as a s ign i ficant effect upon the r el i abi l ity o f ro ta t­ing equ ipment .

T h e effect o f operat iona l a n d ma in tenance p ract ices i s shown i n Fig. 4. This effect a l so becomes more pro­nounced with increas ing speed. A machi n e runn i ng a t , say , 15,000 RPM i s a very delic a te piece o f m achin ery, and it can easily be damaged by care les s ma intenance o r opera t ion .

Some t ime ago the m agazi n e "THE LOCOMO­TIVE," i ssued by the Hartford Steam Boi ler Inspect i on Insurance Company , had a ve ry i nterest ing a rticle , whi ch t ies i n w i th th i s type o f rel iab i l i ty eva lua t i on . The a rticle was headed "YOU CAN NOT ALWAYS BLAME THE O PERATOR." It po in t s ou t tha t opera tor f ai lur e i s i n rea l i ty o ften management fai lure . When an opera­tor fa i l s , thi s may mean tha t he has n o t h ad su ff ic ie n t t ra ining, o r tha t h e wa s n o t fami l i a r enough with safety prac tices and, perhaps, that he was n o t proper ly selected t o begin with . This a r tic le a lso poi nts ou t t ha t opera tor educa t ion and safety practic e i s no t a n o ut -of-pocket expense, b u t rather an i nvestmen t which wi l l y ie ld re· turns . Thi s same ph i losophy i s wha t we wa n t t o demon­s t ra te wi th the rel i ab i l i ty curves : We want t o show r el i ab i lity as a funct ion o f i nvestment , b e the i nvestment i n the form of a t r a in ing program, safety prac t ices, o r b e i t ma intenance shops o r , i n an ear l ier s tage o f the game , i nvestment i n equ ipment , foundat ions , p ip ing, and so on . I do n o t mean t o s ay tha t ju s t a n y i nvestmen t i s desirable , bu t r ather t ha t the i nvestment shou ld be made consc ious ly , w i th a c l e a r i dea o f the returns t o b e expected.

IDENTIFICATION OF PROBLEM AREAS IN EXISTING INSTALLATIONS

A fter we have seen how to predict and eva lua te r el i ab i l ity , we mus t n ow provide the man i n the f ie ld w i th too ls t o enable h im t o deal wi th a problem qu i ck ly and effic ient ly . Turbomachinery has a way o f tel l i ng u s when something goes wrong : I t jus t s t a r t s t o v i brate .

FOUNDATION CHARACTERISTICS RIGIDITY- TO HAINTAIN ALIGNI1ENT

0 10 20 S PE EJ) -10()() IIPH

2.

I

VIBRATION CHARACTERISTICS

0

F()U. V.D. wr., J•UNJWEIJ ,. /.SIC

/.(}JJt H. :JORJ" ,.§, -

RES'c 'NifNc� �

10 20 SPEEJJ -tOO()ItPH

Figure 3.

ISOLATION FROH SUAROUNIJINGS

2 r-----..--.----r---,

I

0

No NIJT/Cf SLE �

/(} 20 .SPEE.D -1000 IIP/'1

CHART"/

-=-=�� --�

�� ==

� ..... ---..... .-Soo1 .... ...... ""'llldol

:::u---= �l\J' .....,.-

��-� ......

--�� ��-

;::::"'...,

=-· --�

·.=, �·.:e· u..1

:::!:f

= �.,,

leu4ot 1-1-

==-===· ·· --

_'_ol_!O.sl/r· .. ...:..... --=· �� Orlu.ol-

··-·-· OUtildll -·OJIIUl

�llldrl ==.!:. -==

.. _ .... _ .....

� �-��i � ��;· :.. � ���� !:"uploo �-�=--

!.

-

,

, ., 1 ... I4W

1111

N

,

,. II" r lt1 lr r

-- --,II lr r

-"' lr r , I I

I'll 1 ..... Ill

]i-;, A6 1.-.t .. /#. .. ,. Nl -- •• I• ,. ,. "'

• ,lfl Nl

- IN Nl

"' IU *' IM i-

,10 Nl I• I• M

--·u.-· --�- -

.. I• Ill ..._ Ill 1-- "

• M

I" I• M l.tt� Ill

-

/1111 I AIM /tAifl -� _,., Rr J!I/8

IIJ Ill ll

IINl

-

7tl INI 'llf Nl •• M

Jl . SRTal .uV 1$ 11D1117

.i-'- -, /Ill

,., IH

'lli!J Mlllll'l'llt ..

16 Ill H I• .,,

.,.,

'"' ilfl ,

,., 11:111 ri'Y Mil

.r . .;� -.��---�

·� •

i

,. "'

"'

II

.,

T!

61)

lfi.J

ltNII

...... ���· Slip ....... ......

, ,, .. , Ill Ill ,, ... .. , .. ,

----t- ... .

_ ..... ,.,._, ____

IH ..

-

l'tlllfATIOitl ANALYSIS ... �"· .... ot ... _,_ ...... f•rt. • Hol'ls. Adal ... ,, ...... I C:.•""l:�� Plp1nc'"""""

... , 1"11 14 � I,

"" 417 Kl Ill "'

411 N 111 ,, 111 Ill Ill ¥-11 .Ill 7' Ill .Ill

.Ill lu .{'11 I (I "' /(I

.Ill _,, S"ll "' 117 If?

.. , S'tl 117 ,() Ill

311 N 317 , �(I 1/1 ,, �'' Ill S'll Zll .�, ./P

Ill? S"ll Ill fll Ill Jll

,, 311 S"-11 I' "' ""

,., S'tl ,, Ill lu .�, 'Ill ,., ,, ''· 1117 117 ., Sll ,, Til lu ,

J, ,r, -.Ill ,, Ill ,,

�II ��� 411 111 ,, Ill

" " Ill 711 311

;Ill I•' N , Ill ,II ,, S'll Ill -- " .,,

, 1111' Ill , ,., Ill I, f'll S'll Ill I' r11 ,, Ill M ., 311 , fp_ lfl'_ ,,

311 IH Ill ....... ,., IQ Ill 1/-11

ll I•, .Ill .. , ¥11 ,,

311 .II .1/) IHI � ,Ill Ill ., �_,_ ,, ,, Ill

II# �- !_II. .111 ltl 41 . _ _. ... ,. .�, .Ztl

, , .117 .Ztl II II ltl Ill JQ lfO !1() /() ,, I .Ill

_lid 4"0 Ill #11 #11 _/It? WI .IQ Zll IQ .zo 917 .1(1 ., G'll ,, IIJ 211 "" ro IQ .111 I .til z, .Ill Jl 311 'IIJ Jtl "' lt.� Zll Ill Ill

'HI S'17 ,, Til Ill 111 IQ TIM.f(ll'!,'rt -· I 1,;. r•" ··-

Ill

� F"i«i fi(J .. , ,

SYf1PTOt1S AHO ])/STRESS 11ANIFESTATJO_H.£ "" ...... ....... .......

, .,.. .... =• lncl'ftHa Dlcrt�•"• lh&k• � =t

IJ=- b:::::_ � � bC B:.. ""' I � -

.311 �p r S' .111 S'l1 r .r 1/J ., ill!? ,t(l I (I I! 1/J 711 1111 ,

Ill fl(l /() Zll ,, "' .111 J() /1(1 Ztl "" .-!". "" Zl1 'Ill rl7 /()

IIJ S'll Ill .Ill Ill

/Q 9tl

., S'll ,

"" S"ll , , Ill

,, 111 911 "' '" Ill

;1,11 ,, ,JII /Ill _Ill '"

Ill Jill lz'' � /Il-l'

"" 80 Zll IQ . .JII 111

"' 80 ., " ,, 10

, .II. rill, Mr.-.. ,

10 IQ Ill Ill

.Ill ,, J(l Stl 2() PI ,117

� /()

.,, ! Ill "" ill

, 1(1 Zl 8tl Jill

8() JQ .ZII IJo J(J ,zq

Stl �q Zll

n•r•AY"UJ, ,,., - �-·

. ..... EfC!ct of Qp!r!.tiM CondH.Jsu

l01owm:- C..a !a or GeNis out at: �� 1 lnerooooa ... ,....., C.Oo .,..po 1 .. ottoct !Mllood !No loodll'or>....., SLut • ..,. �' ! =:"

or 1__. -.,. .,..,..,. I"' ''" I"' !ou' I"' '"'' .i.,;r 1;,;;;,;1.. 1'-"=- I � I Y II!:- I I , : ... :::� .. )t' I

- -

IN IN tl .1/J s- S"ll S' ,, 1.-. ,IQ 111 110 .II? u

.111 r S"ll tr "' I• i*' "" I() ,, ,,

"' ,, ,

ttl "' .,, I• I-" '" ,, It? Ill Ill' 711 If! , //} 1, II S' 1/} H I-'' S' Ill S"l1 ,, Ill S' IJ11 111 r ,. , IJII

.Ill 4'11 ill? Ill f(l J(J I" ,r r r s-. UJ "

;Ill ,., �� .Ill Ill Ill I"' l.rl7 Ill Ill S'll IN

, II S'' Ill "' ,

"' .Ill ,r r Is S" 1211 "'· ....m.y- Ill /{} , ,..,, �,..

., " ,

I .til II (II , , 1111 r r _,, .II

Ill ,17 4fl I .ill Is- lr 1- 411 "' ZO N7 ,, "' Ill

,. ttl Ill .r r r l.r IJt1 .Ill Ill 1111 Ill iJI " Ill .I(J I' II IN Ill I,, ,,

11l1lZ1!I. 1/1 , t;J � � ., r:wr "!/ ' ,, ,.,_Ill 11 �� IV "'

.Ill 9(1

Ill 1�14 , .:. � ,,_. � .IP It'

811 fa ,, , - f(l

,() � H r lr r lr i""' 1-'' u

*' .((1 , "' ,(1 J"() S'll 10 Ill Ill Ill IN Ill •Ill 1M 1, Ill Ill

.Ill _,, JP 'Ill lu Z4 .Ill

l2f? .l(l /() ti? /(I

Ill 9Q 'II? �� rr- i;' rM •- .,._ .... �,.. --���

.211 I ill? ��� li?P

IIIII 6(/ 20 lzo ll7 Ill S"l1 I s-o. $'11

6() 217 '" ID "

1217 .ZIJ Ill ., 2111 ��

lP N .SIJ .J(J so "' "

.117 MJ lfl(l

w .. , .r ou... . ... r.-""'"-• ........ p lwroooo t

I botto, .. m bottorioor .. lbottorlooroo

J

-··· . --.0.CI'U.Mt. bottOI .....

Ill A1lMIPI7 Mit# .fll 16· !!.'t:.:."":r! •. i"!.�t. ":.;..

-1--

"" /Ill ltl, "" '

IWP I,'FI'I ·rr

TRY16DI' .• 11" .... ��

'';;"· ·J' 1/J ,, u , ,. Ill ,

/() str• f' IQ • kJ" "7D

, , , � , I' " Ill' ,, Ill , r11 ,, , " 1,, , , 1St�

ISHM . .1M ...,... ri'FI"4 'r.s 111 f'I.Y - lwr I l\\l··' t '

Figure 5. Example showing "Trouble-Shooting Chart"- Vibration Analysis.

CHART #fJ.2 .

=:u;. �.

_ ..... lelt. ................ ,_., �:':

Caq ""--;;;;;;;;�

-loll loll .... - ... Oidol lfl.oo1Spo -Plpiltc -· -· ::::;. .... --. -..

...... .._.. -sw..unt.s. (oUwlllrlo, ... =: .WIMH -· ..u.ol - .... -.. �-��.r ill =-,_......,

·-

'� - -

:=:.7,;Jo *'JI!· Uool

:::!:1

= -Wnl -

, -.. I�"' I�

·-..-- --trioolll' _..,_, ............ -Ya1ft-.

� ----·

�sr CrlUool opood _,_, CIU-l -·-Jlr710b11-1 Cl.ooliuoo Ia--....s.•u. -,.._.

l'::r!:!

................ ,_,... YIMAr .. ll "Jr Ill�) "'hoqv. ...,. ..

,. ,

/fl Zfl Ill

ltl

"'

II

N

Ill ,

" ""

�-,•·11

II II

,

"" ,

,

Ill

fl lfl Ill

/(J "

liNd ._.. C..1 ucl I Hicb Very hlP ...... .... pi�btd. loud � '\lb1M111Hre.• ("1Met.•J •

Ill Ill Ill 211 7tl , , 111 , ,,

Fll Ill Ill N Ill M H ., H N M

ftl Ill Ill ""

40 Ill. " "

" Ill

Ill Iii .Ill Ill

, ,

:Ill "' 111 .Ill

u .Ill Iii •Ill Zl! IN/ Ill

N /() H MM/,.111

n .11 ""

, .. "" ll " ,., ,

,, II .lt1 111

, II , �� *"

, '" ,,

.II II , .,

.Ill

1111 II "' ,, 1.11 Ttl Ill .II

,, '" , I'

lfJ 20

, Zfl Ill Ill Jl "' '"

'ltlttiWH) Ztl

'•1'7 h1&b Ultra..S.. "•Hl" -·· --11

, II) 1(1

Ill

"' rtAt.tl�" $ �t.WD

If I

"" II

Z.il

t!l.11i11X •' IIUbW Ini.U&lllot 1-lD lt.u'\up ,..,. ,_

., ...

v ...

,, ....

"

I' .... "' , ., ,

, .... .... "' "'

VI" .... "'

;v

, "' ,

I"Y y

"'"'

"'"' , I"

, ,,..,. V' ,, .. , ,,., ., ., I" II'

Yl" .,

, .... ,

""' ,

""' V'

"""' V' ., .. , .,

., .... ,. , .. , ""

""' ,/ .... .... , v

.,., , y

vv ""'

I'Y w .,

,.., Y#" .,. ,, , , ,.., I' , ., ., , �""' "' I" I" #" I" II' y

-lD ,. ...

,/

y

,,. .,

.,

"""'

.,

....

1/

.,

.,

,

-

. '

OPEKATIONAL £1/IJ)ENt:E CONT'/; I -

�- II: lllk'M•��J.&Dil loolo 1•n ........ bent

I 111

,., + "' /(1 /(7 I() , lfl

-f- IS' , Ill

S'l Ill S', "

9111 '"

'II "

/(/

WI .u

"' Ill "' , u

Ill

J'l Ill

:lt1 "

.1'# Zil

Zl 211 ����

iltl

., ,

Ill "'

'"' 1/J � Ill z• u • Ill IS"

Ill Ill '" "' r .r

Boorlop toil .. ..... d1o- Qat ot \ ...... .. al""'""'l ..... ··-!-"' .......

_ ..

r· i '

"" ,,

Ill Ill "' I I ,, .Ill !

I'll "" " + AI

S'll .111 S'll IT " " " H S'l? '

Ill .ttl .ttl +� " "" , H

.II "

I 1(1 ¥111

IO 40 10 '

"' "' :

"" M

��� ,, 40 ' 111 �- ,., j 1#4 II , u "" , u J

.R S'll r .1'1 S'l i II �� ,

•'

111 z, Zll ill 'II al1 . 3� ( .II .Ill 2;0'

I 2D ';Ill 20 2t# 3"

'

l :ZII Ill i 1 u 24 ! \ II " u ; .I'll Stl I 1 ' Zl Z.D I# '0 .Ill , ID (ltl ' : 1 ,, 26' 4-1 l :

Ill .Ill ZO ' : 1 r Zil I

.f' r

..._, I c ...... � lo.or ... � t...r• I

c.w.c ••

I lNrDed or broken 01' ...rit.ed.;Ji · cneked treUiD& pl.t... pl.t ... ,-4'1 .. .... . ..... bNk.n di.tc.or bubo

,, "" "" .r "'

Jfl Ill

:Ill II ffl

/Ill

211 u Ill Ill INI

Zl 2D 14 " '"

Ill

ltl ,, I'll 211 .I'll eo zo 20 Ill S' 0 20 � ltl

"" - , .,

1t1 Ztl "' S' 111 ,, .r

20 �0 311

:;" '" '" .r 30

"' .kJ lfl r "" 31 " r

.to IHJ ¥11 lt1 S'll 111 ., /(J .r Zl .Ill Zll II Ill ,, ¥11 30 20 2.0 20 " r .r » S'

IO zo 20 II .r #II 1'0 ' " (Ill .r » f'o t:o Stl Ill

_, ........ '-ltA ,

.. tt.l.t - .. , .....

..

.r

+ "

14 Ill HJ II

Zll z.o

M

Ill I#

Jll Ill

h? .F

.r '"

$' "

,,

:ftl

Ill 111 711 Ill

fll .riJ ntiJIIS IIIJICATI TB LII&WIOOD II J1: THAT THE sJio.r� (TOP. HOIU.ZOJri'AL) WILL OCCUR M11R1 CAUS1S (LIP'T, YIR!'lCA.L) _.. IlnOLYitl.

I Figure 5. - Continued.

SoleplatA• -looM or ... r ... ....... '""'"""

Jo 3# �ll , Ill

u S'(/ " " .II

tll1 Zil

•• ID Ill "' :Z.D

,

'"

ltl II

-1 ...... ion ..........

. .Ill

Ill " "" S'l n

Iii

Jill

1.11

nu• r-Ml'k1 c.,oraents .. ...... .. ...... b

llld RlfU .IU ,PU �"'

"

II Ill

"' Ill

H ,,

,., 10

oolldo ........... .. ...... • rotor

./I? rr IUAIUJ '-I� I'AI

"

1111

4'11

........ i nMco• looldoa Wt I� ,�. I em :a..us.nc 1Dt.emal•·

Zl? ,�lfAO. J),h ,-.r,rr�. '"tf,l Ill

STAESS """''s ,,

ill. ,,

u " N 1'1

,., "

S'l "

Ill "" Ill

Iii

�(I ,.,

Ill u Ill

Ill

21

:zr

" ,, ,

II'

,

116

_ .... "f"T ...

·-·

..

"

-Jll

''1''1.,

� &;; · � ?'

� l .,.. � ;;· 0'!

� &. � en � 0

�· [

� i:: .... s· t � 0"'

r

�11 1! I' ii 1i ! tIP'! ff�!�fi 'i!HI �II� J: Ill". u , I !1.tiJ1!1 'J Iiili� !JW• 111� lf ilfll ���� ����; qJ f �! ��� "I' ... '

§., � '

'\I\ '

'- I \ j ' Jl I I I -14 I I\ I I I\ ' \q \ 1'-'\ �J'\1 '\ ,, i ' \ '\

���

I \I� I I 1'1\l\l l\1�1' 1'1 1'1 � I . . , --�� ' < ' ' � \ \ ' � � \ � \ ' i�,.. . i .. ' -ll' f i i ' . ' " ' . � � ' . ' ' , ' \ ' ' ' -�J I l i � �· " , � I � , � ' ' � � � �" ! ' _i l !

� 11111111111111 U !IIIII 1 11111, Ill I Ill U I, II Jl ftl ��:! i !Ill I II Ill I I I I II I I I II I I I I I 111·11� I I I I I I I I ll\l I' I� ll' II! iiJi.

� 111111 I I� 11111111111111 I � 1�1111 �II ; I I I � IIIII �11 , If I i -- . �

il'- q ,1,

\I'-

\I\

'\I � � '\ \

\I I� 'I ' 'I' ' \

.. '\ 'l'l'ltl'f I' I " I' '''" \\ '

i' l �

"';.

�I� 1�1 \I �I� I �

' " '

�

\

'

� I �

' '

"

'\ \

·\

'

V� ' ' ' I �

�I\" \I�

'\ I I I I 1\

'\

�

' '

"' '

'

\

'\

'

� \

� l\hll\1\ �

\

� ���ll\1\ 1 � I�HII'I'

' "

� � n

�Ill

,.

'

'\

' � � ' \

\

\

' '\

�

'\

\ 1\ ' '\ I\

\ \ \

� I\ \

\

�

' I' ' \

l I �

' '

,\ \ \

'

� I I\

'\ I� � 1\ I� I�

I I I

"' ' "' ' '

\

'

\ '

\I� \I�

'

'\l\l\l\1\

\ · n-u l ·t i '\ I'- I \ U"ii' �b

�I \ "' '

�I \

�I '\

����[ .a i U8s!il' . . '

3 " '= ! g(j: ·i� f HHf ' i 8i

Jlf i • � h l :l

\

lj,l� �' l .. 1!� I'\ .. !II -� '"

"t((•d � \I \' 1\'

'

Jr i iUi \! �1

�� '"tR =t?

I i �t . ... . .. .;:

1'"( tit 1='i ���

\ ,., !JaH _,._,.__ , \I '1'·1 . Hf i \I '\ I\ ·n �:1 ,, '\ ,, ;,

I

t .;

i! 'I �� I l£

' �I\ I\ n\1 \ !�'l !h'•

:1 �uti ' I' ' \ �I� I\ II\ I '

" l� • u 11 •1r1': ... .. � .

I

I I

. I

----�-- --------

-·/N.STALL.A rtt?AI-�::�. -JffARI l1l!laL PII'IN6

...!!PJ:. Soleplatllt Ri.Wt>' -... - ... ......... I ... loo .. tnMttflct.nt. tore• • 6 JoloU ,.. _...... .._,.,.. -· .... ...... - · .. ,.. .........

1not.o1lod I =:. --.

- - ... loR rot.or puol.t ......... hu .. )

-r..;.o.:or7· --..... '!"" , � """" ,, ,. ,., Ole- ""

--- - ·· - · · -� VY v"' .......... , , Sool '- � ,. ,. ,,.. "'"' ,., yv - 11111ulel , , , v ,.,;' -� II' J/Y ,, "'" ,..,,.... Pip1ec ,, + I'll' "'"" --· --riel'¥ -... Y' ,, """' - "' II' ,

-. . . _ -lt..l 'fllllstoS.. I'll' ,, "' II' o/ (oU 111alrle, .w.

= -

, .uttoooo , -- � = ""'· """ -tie-� le

ru:; ·- "'"' , _ _ ,. ,; o/ .....,. _ 1 -::::... ,, ,.,. ,.II'

- ... -CooopUoo � .. ,. "'"" _ .. .

=:.•Je· Y' II' II" u..J,, :!:f . ·

== ��.,

,-.. y ,.,. , ......... ,, - , ,,.._. ,, "' , 1"1' ,,

- ,, Jll1a---.ur - -· -· - ,. I" I" , -Yalw 1'iltrr.

-..:u. -- ,,.., , "' , -......w. , , v -- · v I"� - -1 Crlu..:l ..... "' v , ,.

; ...... , 'dllr. ,, fi'V ,., ,, QIJ. wldQ - - ""' Yll' ,., ,., .,. _1 v ·

ca- ID- ,.., """' ..... � ..... ,, """ _, -�

POSSI/Jl £ CAVS£5 OF JJIF r 1&/ilXY, CONTINUED

·-· �·· WM� � ·- ��--�� � �� � -� � �_u� off at top eco.W..t.llt run to � not pot.t linn .-..tricW b7 or .. ..._ onr waJ,.,.. � or worldzw; ...,_.. under- contact with

.

-

hNder I aiat tise tound&Uon or J or :lato water otua p1pN , I

.,, .,, "" , v "" "" ,., ,, 11',. ,., ,, ,, "" , , , ,.,. , """" ,,.

II' "'" ,,. ,,. ,., """ ,., '"" I' I" ,, 11' ., ,., ,, ,;

. .. ,., """

II' , ,. ,,.... "' "' ,; "' ""

, "'""

,, ,..,, II' ,,.., """' "'"'

Ill! II' II' ""' w II' II' ,..,.... "' "'"'

,, ,, ,, II'Y ,, , , , ,;./ y 11'11" ,,

, v

, II'

,, I'Y "'"" ,, ,,. ,,

,.,. YY

""" ,. ,.

II' ,

�

- �--· llut not. nlltri.ctins drainod -

""' , , ,,

,.,. ,;

, ,, "'""

, ..V ,, ,,

, "'

,

.,

,., .,

,

I'll' .,, ,

v

v ,

,

'

I

:

I

�

I

I

l

i

l

I

':.

�--�--r-� CIU'ftllt. -otor qele nl.Q' at.art.tnc �r)¥ t&llecl palutsah t:laecl

_ ..... tor ..,,.._

at�� Gs":,.-"' v "" ""'

,. ., ., ,. , .,. ,

"" "" ,

, , v II' ,. ,.

, v

, .,

o/ v

,., Yl/

Figure 6. Continued.

, .

- ----·- - -- - - - - - - - - - - - - - - - - - - - - -- --- - - - - - - - - - - - - - - - - ·

J'rostll 0."- m o r &putacturiJy: - - ------ · - - -- -

' I - - -- ' Jlot co......t 1n detall 1A thb paper • ' tlwn bf&h probablllQ' of .....,. 1a 1ndicat.ed. COCJt.Rt •nutactunr.

v v' , y

""' , y , ,

,.,

v

,..

II'V

y

,..,. ,, ,. v · ,...,

�,

II' #I'

YY

,,

, ,,;' ,.,

VY

, ,.,

, v

..V �v ,. "' ,. "" ,. ,v ,,

' C011!1EtiTS

i � hlch-<epnd l"flt.on ori.n -nquire tltld���: .'::.' :f':!,j� -�m.• or rat.or - • .-..uon _

a...t nton cua NMt.S.• b1 .trala:;trt..Md b7 "hA ... pot" �. but t.hlt ��aou£G_,�,.• ��-·· • • 1 oolutl•o bo,.... "*.r.w1U c- bock In ·� ��uroo :"aV.JJiiJ'J7.oi"��lco� =�·�'.:, .:::r.�:w� , Mit , • "''' /Ill �-- - ---·------ · - -----St.ralcfat.en � �1ow17, �llw en tumlltc ,..,. o:._:� l�ar .,... • . 1:.._�� occura, trl�l}.";-,.:rr.:,.."/.

!:;:.�tn�:.c;-::::.,:t :�-s.��':!t •;-.:::.;:::-;,:�O:C!!-:!\i'..��-::= .. ca::-c:!:CW ... :!,.::*" foundation, or Ullt..-1 .t.,... (INat. •pot.• 1 or UMCNU ..... ___. . v .. .u,r ,...,.....

� I'IIN M:t clW, but. trlp un t '-dS.t.et, u • __._,._ ..- .-• ..... . l 'hra bJ' 1IMd .rt.ll clear. •

... ... _;;:;;;.! -- - ·- .. � • -·- -- R -

" �� ca;::. ':r -::-:!-:.:!':..:'::W.-:!: ........- .....e.�a� ...e r.-.u .. , ._ e..tt.n a1ao .., 1au1 oA troUble -i.a· c ..... bi"PaOr -pr,; aupport.s-�.....W ... ....... ......... 1 lllproperJ., ...,. _,..ton ·JDla\s _. poor PiP' u.,...., at cutnc c-ntan.. POIIIIS&tlon Mt.US.C abo aa UY• ...... .t.Nla. 1

: ... rlnp _, pt dlllt.cried tra. heat . .... bMheck, 1t poulltle, clwelrllll: «at.aot . ..

, W.\.oh for .._ dlecoloNtl.• lllhlch ott.e ..,._.., � tal.l.taNo. '11111 Sadlcat.eo ...., Mch 1eca1 oU-tU. ! �·· Cllock rotor for ribrat.lon. Cbtok t1ear1ac ,_.. llld Ita cl ......... s.

;IC� �� end ,....._,, ot JOI.U'Ml, •• •U • .-..t. _. t lcht � tU. Ill eue . 1111t.et1 011t trw � ==-�-::,n;:;,�:,rr,.,.�-s---;;.;r �� =�r•��.���=--1 C... Ullllt.e ,.._.,, ud crltlc&l• ud .,_.t.-t.t ... u.r..t at 2 a ,._Inc r.......,.. ��. �1.-.-1�'--�-I!!':C:" horl-:=_ ...... a.o:.�..;:�..:.�!-'..':s :"': .:,.:. -::-:.::.:lc• .,.... n .. .. �

! u..u., tho' naUt of olutlllw ,._ -.ht.. trl\h tlu.td, ..u.. ; : .. ,..iallr' -.tJw. llldl.t liP • .._, w oft4Nip .,....,.,

1:: :: �-"'::.-' .... !! ... 1e --le. �=-.::....� .... - l.eR ..... _ .. _

·Ofta -�� oU ...... 1 -- � -::"..:-:::.."-:=..::: -· .. :-·- -- .... �� ........

.. ......

.,

........ r- -- .... "'" -..... ... ....... .... - .. ·�-.- ·-· ,.._ -·

�---- ------� .... .... 11A "'"'-c:-: � "=.--..:-·� : .. ;... -..:.. ""' - ;��;� -·- - - -.. 1/J11!51 -· � ............ .-.;.... ��.-· �·� · �'"':�: ..., �· or oo11l c.,U.C. Dl.tfled\ to Cfii'NO\ Ia tl.old. a-t.S.. fiiMlllc --

at. .... 1111 ... Jtelpo ·-tdorUl.J',. �,. .. ,, ., � ... ,,.,.. ,.., .,.

'·;��.;.r,:!:-=:.:o;:.-�·-�- "'"' ll&oll -. - � t ldt ol .•

1.,..... u ltpt. _. nut .. ,..ow. ( � .. �-==�·�t&t :=-t.=a...Jf.=t.��lJ:'.:=L� .._. ... .... loll . .. ... .. - - - .... ....u .... .. .,., ,...,..�,-

I �--�

I ..

; 16) ' (15)

lll) (19) (20)

(2)' (17)

,., ( 17) (1•) 19) (20)

(lll) �1!)_J�<) !

- · - ·

' 0(10) ( U )

i ,(16)

I

tm tu

·- ..

m m m

:1 m (t)

-

-

::.r:n .... ... -�:u:- ::...�·�- �;:;..-;:;--.w,. �- 1e _. .. ;: - ! ...... l&l.bo -=-• ot ... lble nbo Oftd ....... r.u-o oo ..U-u -IIIAI --f.oo ot - ri .. �..eo::.=. "':0'� ::'� 1.=1::·.:=::·�-... • &loo.:r m.� .. �'U�d:i ;;:!'O:: :r.� = .. �-r. �=--'! =:--��':!Lc· �":'!:"!u...-::.=..!�Jr::-���':::'.;· .. -.=.�·

l � -=-� na-t.t.. .. _ .. llo.uiltl fau.r.o. J_ pl _ _ _ .. _ _ _ _ _ _ '

��"!. � ,.::,:-:;->;::;:: ..... .......... . , .... .... -.-. .._ ..... _ .. _ _

·-lal 1111 �· I !2K10) IUl � .. ���� (10) (U)

! �"':.':":....-:!',��·: �� � ..=:.:.:.1-;: :r 'i::i:" ... �w:;::--u..- i cMck OMl �. ,....._ �. COiqlljllp, _. � � ottaot.e. . � -=."'C::::.. "1.:::..::·=-.!:":!i. , ... - - -· - 1e ... - .. , ... ...... ! eu. M ...,. tlllf'lod. If tnt.ndu.om. loclc tMo ............ Wl'ial.t.o. a...u, ntw .n " ......Ut, �-olteiiJA

'i ;:;' •• ..,";:' """!:. _..,. �· w.w _. <•m -..!:. ...__. -.. - ... ��r-. ... ..,.

....... .. .;.=.:; :;:;<, � -.u -=·"!.:"""..:.;:-:."'f.;:.:-...--· .... .. ...... -=--= :.. a:--.u�=--... "=u.r...:.u:.. = � :-== .... ..... ... """=-��. . :�il!£'1,f�-.:r- . ....... crt\icd ..,..., .. ... t.tlt. ....... ....... .. -· � .... ......... _ _ .. ........ -..:.!::::"'�· -· .. � nHMDt. ...._.. Md HIII'COe or •dtat.t•. ft1t llhoo .,. '-'· CHD

�"'·= . ., - -

·= = �'

"'- -· �� le �· � _.. ��� ' .. �� - ....... '-'"'" �":! =;;;. -;.:-v::::.. �� "=·:"

'!��::" ==-��--=..- -= ��':.��.=}":"'

trn.n" '" ll! l;! 151

... ... ltftl "" llrll fnl

I ,, _J.f , r-�:�·"' . ·�.r;.,

TURBOMACHINER Y ANALY SIS AND PROTECTION 5

OPERA TION AND HAINTENANCE

OPERA TING PERSONNEL

11A INTENANCE P ER SONNEL 2.

HAINTENANCE FA CIL I TIES

0 10 2() 0 10

I

20 0

EXCI. II LENT .�

AYER �GE

/'()()�

/0 20 .S PEE]) - /OIJ() NPH .SP E E .D - /IJ()I) IIPH SPEEIJ - ltJI)O RPH

Figure 4.

Bu t v ibrat ion can mean many th ings, and if we know how to i nterpret v ibrat ion phenomena, we can get a very good idea of wha t i s w rong .

O f course, much h as been wri t ten and sa id about v ibrat ion, bu t much o f th is i s h i gh ly technica l informa­t ion-wri t ten b y v ibra t ion experts for o ther v ibrat ion experts-and l i t t l e i n format ion gets t o the man on the job. The main reason for th is i s the fact tha t v ibrat ion phenomena are h i gh l y complex and requ i re a great deal o f ma thematics . Unfor tunate ly, one rare ly h as the t ime or opportuni ty t o read a mathematical d i sser tat ion when h e i s faced with a ser ious f ie ld problem. In other words, there i s s t ill a w ide gap between theory and pract ica l appl icat ion, a l though much progress has been made in the theoret ica l i nterpretat ion o f vibrat ion phenomena dur ing the las t few years . There i s a great n eed to transmit the essence of th is theoret ica l i n format ion to the man on the operat ing f loor . Natura l ly, generaliza­t ions mus t be m ade, and many explana t ions may have t o be oversimpli f ied t o accomplish this . But wha t we n eed i s a cookbook-to s impl i fy the f i rs t s teps when trouble occurs-and such a too l h as t o be s t r ipped of a l l excess baggage which i s not abso lu te ly required t o do the job. To do t h i s, we can develop a se t o f charts, showing the corre la t ion between causes and effects, so as to show wha t type of v ibrat ion to expect from a cer­tain cause--or the other way around-wha t type of problem may arise for cer ta in faults i n i n s ta l la t ion or operation ( Fig . 5 and Fig. 6 ) ( Ref. 3 ) . Numbers shown ind ica te the percentage o f un i ts wh i ch wil l de­ve lop the symptoms shown at the top o f the chart . These i l lus tra t ions sho w a par t o f such a tabula t ion . The ver t ica l column l i s t s var ious types o f v ib rat ions, or cau ses of v ibra t ion, wh i l e the hor i zonta l head ing shows the symptoms or evidence which the trouble shooter sees .

For example, i t can be seen tha t a ro tor u nbalance has a t yp i ca l frequency a t l x R PM, misa l ignmen t frequen­cies are l or 2 x RPM, l oose ro tor parts cause frequen ­c ies a t the r esonan t freq uencies o f the ro tor-wh ich may b e any odd frequency i n the lower-frequency group. Thi s char t then shows which components a re suspect, and which components cou ld n o t cause a specif ic prob­lem. The char ts are extended horizonta l ly t o inc lude ins ta l la t ion problems and opera t ion problems.

Fig . 7 shows the frequency characteris t ics o f some major problems. This should help to ident i fy the basic mechan i sm i nvolved. The vert ica l ax is ind ica tes fre­quency of v ibra t i on, the h o rizontal ax is shows speed, both expressed as mul t iples of cr i t ica l speed. Plotted versus speed i s a l ine showing 1 x R PM, wh ich i s cha r ­acter is t ic o f u nbalance. A lso shown are o i l whi r l fre ­quenc ies, cha racter is t ic frequencies o f i n ternal f r i c t ion w h ir l s, harmonic and subharmonic resonances, and m i s ­a l ignment frequencies . For example, i f a machine w i t h a h ysteres is whi r l problem we r e to be run up t o speed, the frequency would come i n somewhere a long the resonan t frequency l i n e and wou ld rema in constant as speed i n creases. The frequency of a pure o i l whi r l wou l d go u p a t abou t %RPM wi th speed . A resonan t wh i r l wou ld suddenly appear e i ther a t lf:! frequency o r a t resona n t frequency, a n d s o o n . With th is d a t a one can then en ter the char t s and make a l i s t o f the thi ngs t o l o ok for, s tar t ing with the i tem hav ing the greatest probabi l i ty, as indicated by the percentage f igures s hown in Fig. 5 and 6.

This way, the choice can be nar rowed d own to some half d ozen poss ib i l i t i es, and these must then be explored i n deta i l , which may involve gett i ng more da ta and checkin g i n to references .

1 tl ?:; � �

,Z (J �----

� � \:!.. • ; ., ,,

f, o I , \ • . 8J.l'"

f, IJ .t o

Figure 7.

t�i)i . . I . � fr .

, c,��/ FREQUENCY YS. .SPEFD w�� � FOR C0/1110N 1'/BRATitiN

i PROBJ.£HS

( 1/3 pf&J � -

� I

'HI6'HL Y UNSTABL E �EGlON, AYOIO PIPIIIG I FfJVNIJATION RE'.SONA/IeF.S AT I• RPI"/ I 3S % TO S'S% RPH, ANJ) /(OTtJN C�IT!,AL FRFQVE/ie Y, US£ TIL T - SIII'JF B£ARIII6S.

I

�}

){ \ ���t\'�

J, (J ()PENA TtN6 J?PH ----... � /ST CHIT/CA L 1(1'/'f

a>

""' c ::0 ttl 0 a:: > 0 :I: z tz::j ::0 ....:: > z > � lf.l ...... lf.l > z tl "tt ::0 0 ""' tz::j 0 ""' ...... 0 z

Tl'RBOMACHJ NERY ANALYSIS AND PROTECTION

SOME PROBLEi\TS C\CSED BY IMPROPER OPERATION

Man y v ibrat ion problems and shaft fai l u res are caused by l oose fi t o f wheels and sleeve� on the shaft . Thi s may he a resul t o f design 11 eakne,o;s, poor qua l i ty contro l , o r faul tv ma in tenance. b u t o n e should remern­her that q u i te f;.equentl v a l iH;se s leeye o r even a loose d isk can b e the resul t o f c areles� opera t ion , main ly ch t l l ­i ng o f a hot turbine ro tor \1 i t h co ld s team o r water . A qu i ck drop o f in le t tempera ture eau.ses the sleeve� and wheds to cnniract . S ince !he shaf t does no t cool as qu ickl v as the sleeves o r t h•; coo l .sl�ceves squeeze the �haft : and a very h i g.h tens i le s t ress a t tl1e bore o f the >;!eeve i "' the re.sul t . H the temperature d i fferenee between the sleeves and the �haft i� enoul!h , the sleeve wi l l y ie ld i n teu� ion .

c

\Vhen the tempi"'alu re of sleeve and shaft equalizes again under no rmal opera t ing cmHli t ion� , the s leeve wi l l be permanen t ly loo;;e on the shuft and th i s may l end to a fri c t ion - induced whi r l o r i t m a y i n i ti a te n shaft fai lure . The temperature d i fference f)et 11 •�e11 shaf t and ;;:]eeve t o get y idding: i s around 2:)0 ' F,. and w i th 700"F i n le t temperature o r more, a s lutr o f wa ter can eas i lv prod uce th i� ternperalure d i ff,>.rence. So tan a sudden drop of � t eam temperature, especia l ly \1 h en the t u rb ine earr ies .;;ub;.;tan t ia l load .

The same prohlem can resu l t f rom improper opera· t. ion o f the sea l ing steam svstem. If w ater can col lect i n improper ly i n�tal led lea

.k -o ff o r dra in l ines----or i f

these l i n es ;:;o i n to a commnn tank whieh may f i l l wi th w�;� te r---eold water may be sucked up or flashed up i n to t he h o t turb ine dur ing load changes o r a fter t r ip . After a C:fmdensin p: t u rbine tr ips , the en t i re turb ine i s exposed to the condenser vacuum., and th is nm suck the wa ter np i nto the turbine , o r f lash the water to 1 •roduce coo l s team which then enters the turb ine . Thi s h as o flf�n caused severe damage such as loose sleeves and bent rotors .

Espec ia l ly t o be deplored is the pract ice o f comb in ­i ng dra in l i neo and leak-off l ines in to one l ong co 1nmon h eader, o r combin ing any leak--off o r d ra i n l ines . The sea l ing sys tem on a turbine o ffers many possib i l i t i es t o damage the machine , and one should always be extremely careful with i ts instal la t ion and opera t ion . Steam and water in th is svstem can flo�,. back under cer ta in comli ­t i ons . I t can ven wel l f low back up into the shaft seal i f no provis ions <�re made to prevent i t from doing so . A lso , seal ing steam l ines which are s hu t o f f du r i nf' no r ­rna[ operat ion may f i l l 1 1 i th water , and when the t u rb ine is un loaded o r t r ipped, the seal i ng steam repu la tor opc>rls to supply "team to the glands. Thi s 1n a y sho o t cf•ld water in to the >;ea ! and against the hot shaft.. where i t can loosen s leencs o r cause the ro tor t o bene! . These p roblems are u�ual ly not under the contro l o f t h e manu ­fac turer, t h ey a r e often caused by oversigh t in i n -;tatla ­t ion o r operat ion .

An o the r i t em deserv ing close at ten t ion i s t he prob­lem o f c orros ion fat igue. Th is i s caused by corrosives in the a i r , or gas, o r steam .. as the case may he . Even very smal l amounts o f cer ta in corros ives c a n have a ser ious e ffect o n mater ia l properties , especia l ly on fatigue strength . This has caused many fai lu res on compressor

blading, wi th both a i r and p rocess ga:"es. personne l can do a lot abou t t h i s. for examp le ln obsen·­ing the air goin;r into a compressor . An (�b�en·ant pe.n;on can o ften see a c ompressor take a p:ulp o f u i t rou� ac id o r ammonia, o r hyd rog.en su lf ide , or other process )lases wh ich mar be b lo 11 n off or leak o u t o f the �ys!em, espec ia l ly dur i;1g a plan t � tar t -up . Wi th the wind blowing i n the righ t d i rect i on , the can eas i ly get in to the i n lets . :- tartin ::;; fa i lu re�. blades o

.r

impel ler-vanes may. fa i l at �nee, or ther mav fai l years

l a te r - -and i n th i s case i t can be e xtretTteh· · d iffif·u l t t o establish t b e cause of the fa i lure and to

" prov i:le the

p roper corrective acl ion . This ::; i tua t ion i s even more d i ff icu l t t o detect where gas compre:>�ors are i nvo lved and process gnse� are contaminated with corros ives, or where co rros ives ):!el i n to the steam of a l u rh ine . ]Han \' bucht fai l u res have been e<m'ied by this. but those faiL u res have of ten been a t t r ibu ted l ;l o ther causes. wi th the resu l t tha t the fai lure� recur red .

,

ll shows th<; fat igue :-; trength d i ap,ram for a typiea l 1:-l ';i ero.mium stain less i-ited, a;o; U;<ed for turbine buckets . Shown i s the stren gth versus the n umber o f ncles for the u�ua!, pol ished

'sample, in a i ; , Nnte how

t h e curve fl a t tens ou t bet 11·een one mi l l ion and t en mi l ­l ion C) c les . Thi s i s c al led the endu rance s tren!!tll uf the ma ter ia l . The rnakria l has an in f in ite l ife for s lre;;;:;;es belo w thi s l ine , i n o ther w o rds , i ncreasin ::; the number of cycles \1 i l l rHil. produce a fa i lure .

'

The 11ext curve show� the ,.arne fati ;.we strem:.th da t a i n pure condensate . A f iean l

'reduct ion'

in fat igue s t rength i s (;v ident .

The !'arne pol i-;hed exarnple in a sal t solut ion i s a l so shown .

. The fa t igue � l rength i ;;; only a fraction of i ts

or igina l va lue , bu t even more importan t i s the fact tha t t h e curn� n o lonrxer l<"veh out , b u t con ! i nue5 to d rop. This means. the par t no longer has an endurance l imi t··-· on ly a l imi ted l i fe. The par t w i ll fa i l u lt ima te ly , n o mat ter how l ow the applied s t rc·�s may be. pro

.v ided

enough c:ydes are applied . . -- and th i s i s usua l ly the case i n an indus t r ia l appl icati on .

-

A t u rb ine b lade i s n o t a pol ished specimen. It has many geomet r ic s tress concentra t ions , for example where the blade joins the root o r i n the roo t i tself . The l ow•�r cu rve i s for a sample w ith a s imi lar ;.; tress concen t ra t ion fac tor I, about 2 .2 ) . I t can be seen tha t the mater ia l has los t f)WX o f i ts fati;,::1w " lrength at T OO mi ll ion cycles, and tha t the curve is s t i l l going d own. In o ther wonk the pa r t mus t fa i l , no mat ter how low the ac tua l applied s tress m igh t he . This i s t he s i t ua t i on we face i n a t u r · h ine i f w e h ave chemicals i n the s team. How severe the s i tua t ion actual ly i s depends on the type and con­centra t ion o f the corros ive, a s well as upon static and v ibra torv s t ress level. Fai lu r es can occur w i th i n m in ­u tes, be::�ause the blades are exposed t o very h igh fre­q uencies o f exc i ta t ion as they pass the s ta t ionary r;ozzles. To get a mi l l i on cycles may t ake from f ive m inu tes to several h ours�-depending on the n umber o f n ozzles and on the speed -�-hu t sometimes the fai l u re occurs after years , because of cumulat ive dfeets . A blade may get ;;ome chemicals todav and some more next w eek and f inal ly , after enough �C) des h ave accumulated whi le un­der the i n fl uence o f chemicais-the b lade fai l s .

8 PROCEEDINGS OF THE FIRST TURBOMACHINERY SYMPOSIUM

EFFECT OF CORROSIYES ON FATIGUE STReNGTH 13 CR STAINlESS STEEL

\ -;.:-HIJ I

/N AIR --� IN CONJJ:NJA 11�___:� �

� '!: 30

� � !.20 t; � 10 j::: � '!:10

t s

lfJTUISJSI'lOifEN� o<.lc • 2.2

- - - - - --------Ill

NV/t8ER OF CYCI.ES TO FAII.URE

Figure. 8.

Such fai l ures somet imes have a pecu l iar appear­ance, which c lear ly ident i f ies them as corrosion fat igue fail ures .

Fig. 9 * from Ref. ( 4 I shows the fracture sur face o f a s imple pul l t es t i n a i r , oxygen, hydrogen su l fide , and hydrochlor ic ac id vapor . Below th i s are fatigue tes t specimen for the same gases, hydrochlor i c ac id vapors , and hydrogen sulf ide . At the t op i s a fractured turb ine b lade showing the same type o f fracture d isco lora t ion as the samples .

Fig . 10* shows a fractured b lade from a turb ine which operated on l y 70 hours dur ing plant s tar t -u p : I t shows the type o f co lors o f the hydrogen sulf ide con­tamina t ion . There were about 16 cracked b lades i n th i s wheel, two b lades were ac tua l ly los t . The sur face o f th i s fracture was exposed by breaking a c r acked blade open . The whi te frac ture area where the b lade was cracked open i n the vise can be c lear ly seen . The golden area i s the corrosion fatigue fai lure which progressed dur ing operat ion .

Fig. l l * shows a corros ion -fat igue fai lure caused by ch lorides . Aga in , i t can be seen where the blades were b roken in the v i se, as con t ras ted t o the corros ion fa t igue fa i lure area . Abou t 60 blades were l o s t o r cracked i n f ive o u t o f ten stages o f th is machine before the s i t ua ti on was c orrected. The mach ine ran for seven years before the f irs t fa i lure occu r red . Apparen t ly , bo i l e r feed water , con ta in ing m ake-up, was gett ing in to the desuperheater th rough a l eak ing va lve .

I f this type o f prob lem occurs , no amount o f re · design or bucket s t rengthening wil l prevent fu ture fa i l ­u res wit h any degree o f cer tain t y . A t be s t t he mach ine may r un a l i t t l e l onger between fa i lures, bu t s ince t he ma t e r i a l actua l l y ha s n o endurance l im i t, on ly a l im i ted

*Color slides. Not reproduced in the proceedings .

l i fe , the fa i lures wil l re -c u r sooner o r l a ter , u nless the corros ives are e l iminated .

Fina l ly I would l ike t o point out another source of turbine b lade fa i lures, espec ia l ly i n the las t s tage. If a turb ine i s operated a t fu l l throt t le wi th excess vacuum, the pressure drop across the las t row o f blades becomes very great and the s team loading on the blade wil l be­come excess ive . This causes cycl ing s tresses in the blade which may exceed the fat igue l imi t , causing fa t igue fai l ­u res . I f the coo l ing water gets cold in the winter and vacuum inc reases above the design l im i t, something should be done about i t e i ther by throt t l ing the c i rcu ­la t ing water supp ly , o r b y bleeding a i r i n t o the con ­denser . Man y people be l ieve tha t they ge t more power o u t o f a turb ir..e when they have a vacuum exceeding the des ign po in t . Thi s i s n o t n ecessar i ly true, because the steam velocit i es in the last row of blading may be­come supersonic , and no more power i s real ly ex trac ted from the s team. In fact, efficiency i s often reduced. The resu l t o f excess-vacu u m operat ion i s an ex t remely heavy b uffet ing o f the b lades which in many cases has led t o blade fa i lures .

This s i tua t ion actual ly presents an excel l en t oppor ­tun i ty t o improve turb ine water-rate under n ormal and part - load cond i t ions . If a vacuum-contro l ler i s used in such a way as t o in crease vacuum a t par t - l oad and to reduce vacuum a t over load , the b lade ex i t ve loc i ty ( Mach. No. ) can be he ld constant . The resu l t ing im­provement i n water rate , overload capaci ty , and r el i a ­b i l i t y can b e ve ry sign i fi can t .

I t goes withou t say ing tha t , i f chemica ls are ad­m i t ted whi l e opera t ing a t excess vacuum, b lade fa i lures are very l ike ly to occur .

O f course, we rea lize that there are a lso o ther fac ­tors which can cause b lade fai l ures, such a s excessive design stress at a b lade resonan t frequency, poor sur face

TURBOMACHINE H Y ANALYSIS AND PIWTECTION

f inish, i n accurac i es i n the blade roo ts , and so on . Bu t these a r e des ign problems, w h i le we are concerned w i th opera t ing prob lems and \1 i th methods to preven t fai lures and breakdowns hy proper i n s ta l la t ion and operat ion of the equipment .

REFERENCES

l . J . S . Sohre : "Rel i ab i l i tv- -Evalua! ion for Trouble­shoot in g o f H igh -speed turbomachinerv ." Pre�en ted at the .! 970 ASJ\lE Petroleum Mechanica l ing Conference, Denver , Co lorado . Cop ies avai lable from the Author , J . S . Sohre, ()3 Gr it•r Hoad. Vernon, Connec t i cu t 06066.

2 . J . S. Sohre : ' 'You Can Pred i c t Rel iabi l i ty o f Turbo-machinery , ' ' Hydrocarbon .J aimary 1 'J/0, page 1 00 - 1 06.

:t J . S . Sohre : ' 'Operat ing Problems w i th Turhomaddnery, Causes and Correct ion ."

a t the 23rd Annual Petroleum Mechan ica l Eng ineer­ing Conference, September l ()68, Dal las , Texas . Revised vers i on , November 1 5 , I '>6H. Copies ava i l · able from the Au thor , J , S . Sohre. 1)3 Gr ie r Road, Vemon, Connect icut 06066.

l. E. J. Poh l : " Betriebserfahrungen m i t grossen D ampf­turbinen . " Der Masch inensdwden , October l ')6-J., Heft 9 / 1 0, page 1 H.5 · } <)6. Mun ick Germany,

28. 5 . R . M. Curran, J .

Corros ion Paper Electr ic Exchange, l %').

F. Qu in lan , S . Toney : ' 'S tress­o f S team Turbine l\'lateri a l s ,"

Prcf'ell ted a t the Sou theastern Clearwater , F lor ida, Apr i l 2 1

6 . H. D . Barer and B . F. Peters . "Why Meta ls Gordon & B reach, London , l (f/0, page J 1 " 1 6.

7. E. J . Gun ter, J r . : Bear ing Sys l em;;." Print ing O ffice,

Stabi l i t y o f Holnr ­S P- l L ) . L5. Government

D. C .