ML20100B682

Intervenor Exhibit I-MOSBA-225,consisting of Correspondence Re Jul 1990 Starting Air Valve Problem

ML20100B682
Person / Time
Site:	Vogtle
Issue date:	10/06/1995
From:	AFFILIATION NOT ASSIGNED
To:
References
OLA-3-I-MOS-225, NUDOCS 9601190218
Download: ML20100B682 (32)

Text

e o

Cooper Cameron Corporation Cooper 4essemer T- /QOSf3A - a E5 Reciprocating Products Division 1351 Harbor Bay Parkway, Suite 1000 00CKETED Alameda, CA 94502 4 541 USHRC

'95 0CT 20 PS :15 0FFICE OF SECRETARY DOCKETIH3 & SERVICE C ooper Energy Services BRANCH CORRESPONDENCE RELATING TO JULY 1990 STARTING AIR VALVE PROBLEM COP i

a 1

1 NUCLEAR REGULATORY !.OMMISSION Docket No. 50-424/425-OLA 3 EXHIBIT NO._T ' 7. M In the matter of Go r la Power Co. et af , Vootle Units 1 & 2 9601190218 951006 O stan O Applicant Mtervonor O Other PDR ADOCK 05000424 fp G PDR O identified elved O Rojected  : Reporter Deesto/b [ Witness

1 IA g ic-05 OD 3L'A Cooper EnerOy Services. 510 748-7409 p,03 l ENERGY SE VICES CROUP l ENTERPRISE ENGtNE SERVICES r su*ic ai.aa sma g

?O se esp su i c. m-484:Al J.*4H To. "

TJehe5

~ MEMO Ftom:- ll* AP*r l A43k* s

Subject:

Smo'n Se 3*'-

\les o N/e flre b/c S&W/* es lln S l0 ~ f f J J

Reference:

5 S N / W l jew l

/

^ > ^ -ci eS lN NN /

r C"/

l T == b A'"k*.rb ll& st y>) t~ou 3 c' es 3* l#n' S / # cr ca s*r 3 is= l v en. /* e 3 e.J e an e, le{ /s )o /o bave $ees'-4s hers 's er

• $s&/m 3'e #re ,erp.w ssnwe l $cs /r, $p ~ M,nerk/x Q

/Ar,* vs. /ve e er M As <w ses, , ire *arer Z 'c- f /, A a. /o be<--

-AG. /~n n '/o~e e/$an,e h e .<rrectsurec/ oa. J 3D e/Wp

.. A ! A m~/ Ada es-eumEe ae e a ,/ /wc,' '

% s,r /9 . n/en er /dp en. '>c /.5 . Jo v 'l3 o*s c / E/] ol

//e dos o- psy J$ _ /Ds a,,, o & TborJ/e/ Aw sr7m SO/ ed sm // & k De sn/ ja")/c~r nsto N/- J'o /kes- /ko n ~

_b,s e wuerso erS >b de s / , fe /prn~/s* e /4 e= e-lP otee

/, / Wue$ /- M n /Ae,- /.a, m n 4 / -n n ,,,s N ! ~ /-

/s A> 3 es>7 )oan Webs,*/** e >J p / syn $1 o Ade v e. /s e re, A -

rr ese p /.3 Wa/n eo s // Ae el oJ $xs-l, AD A en >6 Ave An - .n A,- n./

c~es s <*e ve n e lsca _

h/en5M s e // bt>nn s.r $r //w0' k Atosr e~ fen -f syfe t/*/s es c l~

su 5~- C/'s '7'roo Y cve / $$.~ ora //o.! , /Ae !ssretx/m c A L o ,,. / 4 ,ner Asn /% o,m / S /le eno /s e de d, ,w c, smm J J/,- Aae jar c4 4,4,/ ou/ / A, < ,

s-/m nsr/ .5-b uA l le m. llP vm/nA /$r stere 53d / 4

. son 4 /.s *

. . J A -,/S v i

/4*W M a Dr>tnbution: .. _ . _ . . . File: Engine S/N: Zgde'7/

. . . _ _ . . . _ . . . . _ Pmject Admin: ._

Salew _ _ . . . .

. . . . . Other: .

L

a m

- _ J 4.-- -

, S, '

INTEROFFICE MEMORANDUM i

TO: DPQ GROUP FROM: ROBERT JOHNSTON ,

DATE: JULY 27,1990

SUBJECT:

10CFR21 REPORT NO.154

REFERENCE:

STARTING AIR VALVES The subject report was discussed between Allen Gillette, Lanny McHugh, Maurice Lowrey and myself on July 26 so as to develop recommendations for examination, corrective action and further testing. This summary is being provided to DPQ for consideration and, as appropnate, inclusion in the follow up to report No.154.

Recommended Customer Resoonse To 10CFR21:

1. Examine site operating history for occurance of unexplained failures l to start of diesel generators.  ;

l Yes No Failures _ Defer further action until next scheduled outage.

l

2. Test function of air start valves by app 1ying a 100 PSI pneumatic signal to the subcover tubing connection and listen for audible actuation, both opening and closing, of the air start valve at keepwarm or operating temperature.

No Actuation Yes, Valve Works _ Defer further action until next scheduled outage.

3. Confirm failure of air start valve actuation by boroscope examination thru injector bore while repeating test outh'ned by Item 2. Note:

Customer may elect to proceed directly to Item 4; that depends on site specific documentation and procedure requirements.

No Visual Actuation Yes, Valve Works _ Defer further action until next scheduled outage.

ta , t t

MEMORANDUM: 10CFR21 REPORT NO,154 j JULY 27.1990 PAGE 2

< t

4. Remove air start valve cap and piston only. Do not disturb valve body or intake rocker assembly. This way it will not be necessary to 4

reset valve lash or. perform hot retorque of air start capscrews.

Measure cap bore and piston diameter with components at room temperature and at least at 30' intervals around the circumference of  :

the parts. Nominal clearance required is .002" to .003". l l

Measured Clearance Clearance Is Correct _ Fault lies within the valve assembly; i i replace air i

! Is Insufficient start valve assy. ,

Mount piston in a suitable lathe and sand down the outside diameter 5.

to obtain a nominal clearance of .002" to .003" at room temperature.  !

Use emery paper only. Do not attempt to turn down O.D. with lathe tooling. Retest valve at keepwarm temperature after rework. i a  !

, Test Fails Test OK _ Work Concluded.  !

j _ i l 6. Replace valve.

Recommended Enternrise Resoonse To 10CFR21:

3

1. R&D test to determine if cap distortion results from manufacturing i operations. '

2. R&D test to determine if flange face distortion can cause valve failure t

when torqued to valve body.

+

3. R&D test to determine allowable range of nominal diametrical

clearance which provides satisfactory actuation.

i-i Based On Test Results:

1  ;

~

4 1. Revise manufacturing drawings to specify control of flange face flatness. ,

i 1 2. Revise bore and/or piston diameter to obtain proper nominal diametrical clearance. '

3. Revise piston material specification to minimiw, or eliminate, differentialin thermal expansion coefEcients. -

4. Revise manufacturing process to eliminate set up induced distortion.

a ,. .

MEMORANDUM: 10CFR21 REPORT NO.154

' JULY 27,1990 PAGE 3

5. Revise published table of clearances to reDect allowable piston to cap clearances.

RJ:djl t

I f

F l

i I

? *

- .- ,, , i

j;p o..

s.

Ccop:r Enzrgy S rvic, l 14490 Catalina Street  !

p' San Leandro, CA 9457~

l f

COOPER l

[

inter-Office Correspondence ,

i Date: November 15,1991 -

f i

To: Greg Desin  :

From: John Gildea  !

a

Subject:

Air Start Valve investigation Progress 1 i

To date, all planned testing of the subject Georgia Power Air Start Valve caps has been  :

completed (reference attached outline). I anticipate all data reduction and report writing 1

. to be completed by November 22. However, at this time I can comment on some of my  !

findings.  !

1. A review of the manufacturing history shows we have had a problem maintaining the close tolerance of the cap bore diameter in the past. Up until about 8 years ago the C and C lathe was employed to bore the parts. This technique often produced out of round bores because the v-chucks that held the caps at the bolt hole flanges were hydraulically operated and tended to exert too much force to the structure. That machining method was subsequently changed to utilize an end mill and 4 point chucks to bore caps. The chucks gripped the parts about the smaller axis (opposite the flanges) and were manually adjusted to better control clamping force and out of roundness attributed to it.

Currently, our Grove City Manufacturing site uses 3 or 4 point chucks to hold the cap cylinder (away from the flanges) during milling. The bores are then honed ,

with virtually no clamping forces to ensure bore integrity. As such, parts are now '

being manufactured without out of round characteristics.

2. The subject cap bores were measured at subscribed depths and intervals (see l attached outline) and plotted for analysis. The measurements were compared to those of new air start cap (l.D. Number tJ2735). In general, the Georgia Power ,

caps were found to have out of round bores elongated in the axis coincident with l the bolt hole flanges. The new cap bore was found to better maintain its  !

roundness. '

The caps were then chucked in a lathe to simulate the hydraulically controlled v-chuck of the C and C lathe and the bores were again measured. Although it was not possible to rechuck the caps as to produce perfectly round bores, the resulting diameters showed the material did deform and cause the direction of the bore elongation to change _to the axis perpendicular to the bolt hole flanges.

This suggests the current bore out of roundness measured in the caps is most likely the result of machining technique.

i

MEMORANDUM: Air Stan Valve Investigation Progress  ;

November 15.1991 Page 2

3. Flatness of the cap seating surface was measured for cotn the Georgia Power  ;

caps and new cap taken from Enterprise stock. All the caps exhibited a dished shape about the flange axis with a center high point of 1-2 mils. The absence of any flatness disparities tends to discount the possibility of creep pneumonia or plastic deformation of the cast iron caps.

4. Each cap was assembled with an air start valve and torqued to a cylinder head at 150 ft lbs. Bore diameters were again measured. All caps deformed in the same manner, pinching the bore about the flange axis. In that direction it was  :

found that the diameter was reduced by as much 1-3 mils. ,

Conclusions Preliminary analysis to the test results yields the following conclusions:

1. Manufacturing techniques employed in certain vintage caps induced out of l roundness of the bores. The subject Georgia Power caps have bore dimensions representative of those manufactured caps .

2. Material creep or yielding does not appear to have occurred in the bolt hole flanges of the caps. Any permanent material deformation would most likely be apparent as an impro as was the new cap. per flatness. All the caps were found to be flat within 2 mil

3. Torquing the caps to 150 ft-lbs causes the bores to pinch along the bolt hole flange axis and reduce clearance by as much as 1.3 mils. Existing air start valves with 1-3 mils clearance are subject to piston to cap interference.

/

W JRG:sr:dg Attachment 91CHRoN\NOVWE t S276 i

l l

COOPER

. e . i. j i ,

+

,i DETAILED OUTLINE  !!  !

. i OF l

e  ;

i PROPOSED AIR START VALVE TEST PROCEDURE i

i j .!

i f f

?

A. OBTAIN TEST COMPONENTS  :

i -

r

1. Request subject air start caps from Georgia Power Engine 2B (S/N 76021.  ;

4 Samples should include but not be limited to caps taken from th)e malfunctioning valves.

? 2. Select one random air start valve cap from Enterprise stock.  !

3. Select one random air start valve piston from Enterprise stock. l 4

l l' B. ~ ESTABLISH DIMENSIONAL BASELINE FOR SUBJECT COMPONENTS l 1. Define a reference point and mark air start valve cap bores at 30* intervals  ;

j about circumference. '

! 2. Using 2-pt instruments readable and accurate to .0001", take diametrical l measurements at marked intervals at bore lip and at planes approximately l

1/3 and 2/3 of the distance along the axis of the bore depth.  !

~ i

3. Mark air start valve cap flanges at 10 equally spaced points about l circumference.  !

~

4. Using " feeler" type gap gages accurate to .0005", place ca i and determine flatness at marked points. Record gap sizes.ps on flat surfac l

e i 5. Define a start point and mark piston at 30' intervals about circumference. l 1

6. Using 2-pt instruments readable and accurate to .0001", record diametrical

, measurements at reference points about planes approximately 1/3 and 2/3  ;

the distance along the axis of piston height. j i

. C. FABRICATE TESTING FIXTURE

. 1. Machine a steel plate 1-3/4" to 2" thick, approximately 6" x 4". j l

PAGE 1 OF 3

)

2. Bore a 1-15/16" hole through center of plate and (2) 13/16" holes 2-1/4* from center of 1-15/16" hole along major axis of plate. l

3. Indicate the test fixture for flatness and grind as necessary to cotain flatness within .001" and surface finish to 125 MU in RMS.or better.

(

l D. ASSEMBLY AIR STARTCAPS AND TESTFIXTURE

1. Bolt caps to test fixture with (2) 3/4" capscrews coated with 50/50 lube oil and l 4

powder graphite compound.

2. Torque capscrews to 150 ft-lbs + 25 ft-lbs/-0 ft-lbs. ,

E. MEASURE CAP BORE DIMENSIONS WHILE BOLTED TO TEST FIXTURE.

REPEAT PROCEDURE DESCRIBED IN STEP B.2 PASSING GAGE THROUGH THE TESTFIXTUREINTO THE CAP BORE.

F. PERFORM STEPS D AND E FOR EACH AIR START VALVE CAP OBTAINED.

G. ANALYZE RESULTS

1. Construct diagrams of each cap and the piston to illustrate bore distortions and flatness deviations. .

l 2. Compare " torqued" cap profile to "as machined" cap profile for each subject cap.

3. Test results should indicate the degree with which, if any, machining, torquing, and creep deformation affect cap bore distortions.

H. PLAN ADDITIONAL TESTING BASED ON ABOVE RESULTS l i

1. Review current and past machining processes.

I

2. Calculate clearance reduction due to thermal growth of assembly to establish design limits.

1 3. Calculate maximum allowable cap to piston clearance that would satisfy the five-start requirement.

4. Establish a course for corrective action based on test results.

Considerations:

PAGE 2 OF 3

'4.1 - Revise manufacturing processes to eliminate machining induced distortions.

4.2 Revise piston material to minimize or eliminate differential in thermal expansion coefficients.

4.3 Revise ca clearance.p bore and/or piston diameter to obtain proper running 4.4 Alter cap material to prevent bore distortion due to creep deformation.

91 REPORT 4mSCVIPo7coi l

l l

6 l

l i

I

.l I

.1 1

PAGE 3 OF 3

, , 4 e A

e 4

.- )

ENGINEERING REPORT NO. HE-05-1991 STARTING AIR VALVE INVESTIGATION l

NOVEMBER 25,1991 4

t 1

^

PREPARED BY: DATE:

1 REVIEWED BY: DATE:

i APPROVED BY: DATE DISTRIBUTION:

L CASTERUNE

0. DESIN B. GUNTRUM .

R. JOHNSTON J. MANNO L MCHUGH R. NBfMO ,

y- _._ _ ____ _.- _.___ __.._ _.__.______._ ---- --

j, ^. - e jt  : 4

'f; l'.

I t 1 4  ?

i 1, -

i S. 0% Tot *NG j- t l' .

r 4

j $

1-  !

i-i.

k_. .

1 1

i s .

i 1

s  !

4 .

1 i

-1 i

i  !

I i ,

I i a

)

a I.

4 e

• .I E

F i  ;

i 6

I I

t '

l

q. .

4 l i

f't )

E- l e

I i i r i i

'. l

! =

4 t-i  !

i .,

I I

1 i

I l

t e .

.4 t

! 8

3. .

I 1

I i

h 1'

h ]

5

+ r i I

• ,t i

t l '

i-  !

(

t  !

4' i

(* i i: -

l t- i i .

i  !

g i-e

d. '. -

, s I

l' l c-n,-.,~,~ ..,~,-.-.--.----,,----~,+.,-,4-, - - - - - - - - - . . . - - - - - , . - - -, - - - - - - --*

, .. , l ENGINEERING REPORT HE-05-1991 PAGE 3 i

l . l. OBJECTIVE Failure to start of a diesel generator in nuclear standby service was attributed l l to a starting air valve piston seizing in its valve cap. This report documents the  !

subsequent investigative actions performed on the suspect failure components. l

11. FAILURE BACKGROUND t in July of 1990, Engineer Bob Johnston investigated a failure to start of the l

. unit 28 diesel generator at Georgia Power's Plant Vogtle (S/N 76021). At that time the engine had accrued four separate failures to start, three of which while  :

the engine was attempting to start on the left bank air system and the fourth  !

with both systems active. In troubleshooting the system Bob " pop" tested each  !

starting air valve, pressurizing its pilot air inlet and confirming valve actuation. ,

He determined piston seizure restricted valve actuation in several of the j i assemblies. Subsequently, six of the eight left bank starting air caps were '

returned to Enterprise for further analysis.

$ 111. DESIGN HISTORY i

. f l The starting air system employs a poppet valve assembly housed in each  !

! cylinder head. Upon valve actuation, starting air from a manifold at rou y 4

250 psi is admitted to the cylinder, translates the power piston, and rotates t e  !

! crank. The valve is normally closed by a compressive spring. Actuation is

! controlled by a timed pilot air signal that is admitted in the air start cap. The i

pressure of that pilot air overcomes the spring force and depresses the piston  !

to the air cap to actuate the valve. Upon termination of the pilot signal the air l l above the piston is vented and the spring force retracts the valve and piston. l 1

i The Georgia Power air start valve assembly (FA 02-359-03-04) was released  !

2/18/78 and is a revision of the original air start valve assembly for nuclear i

! application (PA 02-359-03-01) released 12/03/74 (see Figure 1: Air Start Valve \

Assembly). The valve housing (PM 02-359-03-AK) and cap (PM 02-350 AL) are cast iron, ASTM A48 Class 40, and are bolted together in the cylinder .

head. The starting air piston (PM 02-359-03-AH) is stainless steel and is i i contained in the air start cap. Unlike previous air start piston designs, this ,

i assembly does not include piston compression seals. Instead, the piston to i 1 cap clearance is closely maintained and the piston wall is grooved to provide a i labyrinth seal.  !

4  !

j On 11/18/75 the piston material was changed from 310 to 316 stainless steel.  !

j Cost reduction was cited as the change reason. On 2/15/77 the piston O.D.

was revised from 2.249"/2.248" to 2.2485"/2.2475" which in turn cunged the piston to cap diametral clearance from 1-4 mils to 1.5/4.5 mils. The clearance  :

was again revised on 6/19/78 when the cap bore diameter was changed from 2.252"/2.250" to 2.2505"/2.2495". That revision was prompted by the desire to limit pilot air blow-by between the cap and piston which was believed to affect 4 i.. starting air time. *he diametral clearance since that revision has been 1- j i 3 mils. However, as a result of the reported failure to start at Georgia Power, i

, t

, < a ,

i ENGINEERING REPORT HE-05-1991 PAGE4 new pistons and caps are matched as a 1 A-7818 assembly to limit clearance to 2-3 mils.

Seizure of the piston could result in two sequences of events leading to a i failure to start. A piston stuck in the closed position could result in a " dead" l cylinder. This type of seizure would most likely have only a slight impact on starting time since engine momentum would roll the engine past the " dead" ,

cylinder. If, however, the engine is at rest or does not have sufficient momentum when a pilot signal is given to a valve stuck closed it is possible that a failure to start would result. The second mode of failure to start would occur when a piston seized in the open valve position. If the valve remained stuck open during the compression stroke of the cycle the entrapped starting air would oppose engine rotation and result in a very slow start or failure to start.

IV. TEST PROCEDURE Six of the eight left bank starting air caps were returned to Enterprise by Georgia Power for analysis. The following outline details the investigative work performed:

A. Desian Review

1. Review air start cap manufacturing techniques.

2. Calculate clearance reduction as a function of thermal growth in the )

assembly.  ;

3. Calculate maximum allowable j ensure proper valve actuation. piston to cap diametral clearance as to I

B. Obtain Test Components I

1. Request subject air start caps from Georgia Power. Samples should i include but not be limited to caps taken from malfunctioning valves.

2. Select one air start cap from Enterprise stock (S/N 1J2735).

3. Select one air start piston from Enterprise stock (S/N 1Kf 513).

C. Establish Dimensional Baseline For Subject Components

1. Define a reference point and mark air start cap bores at 30' intervals about circumference.

2. Using Bore gage accurate to .0001", measure bore diameters at marked intervals at four equally spaced planes along the axis of the bore depth.

3. Map measured results.

i

b ENGINEERING REPORT HE-05-1991 PAGE5 D. Measure Machinina Dimensions Of Cao Bores

1. Chuck each cap at bolt hole flanges.

2. Remeasure bores at defined intervals and depths.

3. Map measured results.

E. Measure Assembled Dimensions Of Cao Bores

1. Assemble bore gage in valve housing in cylinder head.

2. Calibrate gage.

l

3. Torque each cap to head to 150 ft lbs.

4. Remeasure bores at defined intervals and depths.

5. Map measured results.

F. Measure Cap Flatness

1. Mark cap flanges at eight equally spaced points about their circumference.

2. Place caps on flat surface and determine cap sizes using feeler gages at marked points.

6

3. Record gap sizes.

4. Place inverted caps on 3-inch joe blocks.

5. Sweep cap flange surfaces with dial indicator.

6. Record greatest height variations.

G. Determine Maximum Allowable Diametral Cicarance

1. Measure cap l.D.

2. Measure piston O.D.

3. Assemble one air start valve (FA. 02-359-03-04).

l

4. Pop test valve at 100 psi. 1

5. Reduce pilot pressure by 10 psi. .

1

6. Repeat steps 4 and 5 until valve fails to fully open.  ;

i l

4 , - * .

! i

. ENGINEERING REPORT HE-05-1991  !

PAGE 6 ,

7. Record final pressure. l S. Using lathe and fine grade sand paper turn down piston O.D.

approximately .001"  !

9. Repeat steps 4 thru 8 until final pressure s; 60 psi. j V. DISCUSSION i The following items are pertinent results to the above detailed investigation. l 1

l 1. Manufacturina History l' A review of the cap manufacturing history shows that Enterprise has had  !

difficulties maintaining the close bore diameter tolerance in the past. Until  !

! 1983 a lathe was used to bore the caps. That technique often produced  ;

out of round bores because the v chucks that gripped the caps at the bolt ,

i hole flanges were hydraulically operated and tended to exert too much l

force in that direction of the struc:ure. The machining method was j subsequently changed to utilize an end mill and four point chucks. The  !'

j caps were gripped on the flange along the minor axis (perpendicular to the

bolt hole) and clamping force was manually adjusted to better control bore  ;

out of roundness. .

=

. Currently our Grove City manufacturers use three or four point chucks to ,

j hold cap body (away from the //anges) during milling. The bores are then  !

! honed with virtually no clamping force to best control bore integrity. i i i

2. Differential Thermal Growth  !

! The gray iron air start cap (ASTM A48 Class 40A) and stainless steel piston (316 SS) are known to have different coefficients of thermal expansion. Thermal growth of the dissimilar metals was calculated to  ;

i determine its effect on the cap to piston clearance.  !

Using a coefficient of thermal expansion, a, for ASTM A48 of 6x104

in/in *F, an a = 8.9 x 104 in/in 'F for 316 stainless steel, and assuming a )

maximum temperature gradient of 100*F , the equation i i

S=aATL ,

shows the differential growth to reduce the clearance by .00065 inches. ,

This reduction of clearance alone should not jeopardize the function of the i valve at standby temperature since the expected clearance is 1-3 mils. -l However, the calculated differential growth is substantial if other clearance  !

j reducing factors are present.

- 1 l

4 l

i i

i

- _______________________-_____-______-_-____-________-____a

l T,. , j

( i

NGINEERING REPORT HE-05-1991 PAGE 7 ,.

' i i

n 3. Bore Measurements +

The bores of the six subject air start caps (L1, L3, L5, L6, L7, L8  !

designating engine bank and cylinder), and the new cap (S/NIJ2735),

. l' were measured per the above test procedure and those results are shown ,

in Attachment 1. Note the alpha variables represent the 30' intervals from  :

bolt flange to bolt flange, the numeric variables show the depth of the  :

measurements from the bottom of the bore to the lip. The measurements i are in 1/10 mils.  !

! Several comments can be made regarding the results. Firstly, all of the

cap bores conformed to the design specification of 2.2495"/2.2505" with the exception of cap L5. Measurements of that bore were found to be as

', much as 1.9 mils over nominal diameter along the bottom depth.

Diametral deviation was no greater than the other caps at that location and the remaining depths were measured within saecification. No reasons for that nonconformance were visibly apparent out it should be noted that those measurements were taken as close as possible to the 3/32" re-entrant fillet at the bottom of the bore. A slightly oversized fillet, chamfer, or taper could have attributed to the unexpected bore measurements.

Additionally, the oversized bore at that location would not induce piston seizure.

Graphical representation of the bore measurements is shown in Attachment ll. The linear graphs trace the deviation from the mean diameter at each 30* interval from bolt hole flange to bolt hole flange. The four measurement depths are shown equally spaced from the bottom of the bore to the bore lip. Measurements are in 1/10 mils. In general the bores appear to be slightly elongated along the axis coincident to the bolt hole flanges. Some exceptions to that are the measurements taken at the lip of cap bores L1, L3 and L8. Those measurements show the bores at that location to be slightly longer in the direction perpendicular to the bolt hole flanges.

The cap bores were remeasured while chucked in a lathe at the bolt hole flanges to simulate the forces imparted by the hydraulically operated v-chucks that were at one time used during cap manufactun,ng. These measurements and the corresponding graphs are shown in Attachments til and /V. Although I was not able to chuck the caps to produce round bores, 4 the plots show that material did deform and cause the bores to elongate in  !

the direction perpendicular to the bolt hole flanges. A comparison of the i chucked and unchucked measurements shows the Georgia Power caps to i be out of round in the direction one would expect if excessive clamping force was applied to the bolt hole flanges during machining.

4. Cao Flatness

. 1 4 Each of the caps were examined to determine flatness integrity. Both the  !

six Georgia Power caps and the new Enterprise cap were found to be  !

d

-,+------,------a---w

_ . _ _ _ _ ~ _ _ ___. _ _ _ . . _ . _ . . . _ _ _ . . _ _ _ _ _ _ . _ _ _ _ _ _ . _ _ .

r ENGINEERING REPORT HE-05-1991 i

PAGE8  :

.i >

slightly dished along the bolt hole flange axis with a hign point cetween 1-2 mils. The absence of any flatness disparities tends to discount the i

possibility of creep or plastic deformation of the cast iron caos. .

5. Torauina contributions

l To determine if the clamping force of the bolts on the air start cap flanges t affects bore roundness, the caps were remeasured after being torqued in a i

valve housing in a cylinder head per the above outline. A special bore -

gage with a neck length of 20 inches was obtained for this purpose and i 4

was dismantled, reassembled, and calibrated in the cylinder head. Each '

1 cap was then torqued to the head with (2) 3/4"-10 UNF-3A capscrews (P/N GB-032-113) lubricated with 50/50 oil and graphite. The required >

l 150 ft lbs of torque was consistently reached in three steps. Both new and old caps deformed in the same manner, pinching at the bore lip in the ,

direction of the bolt hole flanges. Analysis of the test results suggest the i

cap flanges behave as cantilever beams, bending under the transverse loading of the capscrews. The cap walls bow in the unconstrained region i between the flange and cap top along that axis. The cap bores were found -

j to pinch an average .71 mils with cap L7 displaying the most reduction of clearance at 1.1 mils. (See Attachment V.) Note, the valve housing flange 4

was measured with respect to flatness and determined to be within

.001 inch.

Additional testing with the new air start cap was performed. Firstly, the l

cap mating surface was coated with a moly grease prior to assembly to reduce relative friction between the cap and housing and accent any bore l 4 distortions caused by torquing. No dimensional changes were apparent. l L Measurements of the bore were also made after torquing the cap to 150 ft lbs in one step. Again, no significant change in bore distortion was l j observed. Finally, the cap was torqued to 175 ft lbs in four steps to predict  !

i' bore deformation at the high limit (150 + 2540 ft1b). That test showed the i i

bore to pinch in the bolt hole flange direction by 1.3 mils or an additional j l 44% for 16.7% added torque. )

1

- 6. Uooer Limit of Clearance f

Per design, the cap to piston diametral clearance can be between 1-3 mils.

Prompted by the Georgia Power failure to start and controlled by the .1A- l l

7818 matched cap & piston assembly, the clearance is currently -

maintained at 2-3 mils. An upper limit of 9 mils has traditionally been accepted to allow wear of the mating parts but no physical tests of record li confirm that such a large diametral clearance would allow proper valve

)

actuation. As such, the " blow by" test described in the outline was performed to determined at what maximum clearance pilot air pressure be sufficient to overcome the valve spring force and actuate the valve.

Because proper starting air admission into a cylinder is dependent upon valve actuation duration and valve travel distance, and those parameters i are govemed by pilot air pressure and engine speed, certain assumptions were made for the test. Firstly, although air start tank pressure is

( maintained at 250 psi, head loss due to friction in the tubing and venting at 2 -.. . . . . . , . , . - . . ...e r m

ENGINEERING REPORT HE-05-1991 PAGE 9 the air distributor significantly reduce that pressure to some unknown value at the air start valve pilot inlet. A value of 60 psi was assumed to be the lowest pilot air pressure to the air start cap in this test. Secondly, in performing the test, pilot air signal duration was assumed to be infinite and proper starting air flow was determined to be a function of valve travel distance. Minimal pilot air pressure for a particular cap to piston diametral clearance was considered to be the lowest pressure at which the valve could fully open.

At 5.5 mils diametral clearance the subject starting air valve could be fully actuated with a pilot air signal at 60 psi. At 6.2 mils diametral clearance the valve could not fully open with the subscribed pilot air pressure. At the minimal clearance tested,1.5 mils, the valve performed similarly at 35 psi.

Note, the nominal pressure required to overcome the spring force is calculated to be 33 psi.

VI. CONCLUSIONS

1. The stainless steel air start piston and cast iron cap have dissimilar coefficients of thermal expansion. Calculations show that initial or ambient temperature diametral clearance can be reduced by as much as

.00065 inch at operating temperature.

2. Manufacturing techniques employed prior to 1983 are known to have induced out of round cap bores elongated along the bolt hole flange axis. l That out of round condition was attributed to the high clamping force imparted on the caps by the hydraulic v-chucks that were used during I machining. The physical tests performed and measurements recorded show the subject Georgia Power caps to be slightly out of round in a manner representative of that manufacturing vintage.

3. Material creep or yielding does not appear to have occurred in the bolt hole flariges of the air caps. Any permanent material deformation would most likely be apparent as an improper flatness. All of the subject caps were found to be flat within 1-2 mils as was a new cap.

4. Torcuing both old and new caps to 150 ft ibs caused the cap bores to pinci along the bolt hole flanges and reduce diametral clearance between

.5 and 1.1 mils. The unsupported cap flanges behave as transversely loaded cantilevers deflecting under the bolt loads. At 175 ft Ibs of torque, the new cap was found to p,nch i 1.3 mils along the bolt hole flanges.

5. Pilot air flow between the air start piston and cap limits pilot air pressure and valve actuation. A maximum diametral clearance of 5.5 mils was determined to allow the test valve to fully actuate with a pilot pressure of 60 psi.

Vll. RECOMMENDATIONS ,

__m ._ . .. . ._____ _ _ _ _ _ . . . _ _ . _ _ - . _ . _

j . , . * - .

1 ENGINEERING REPORT HE 05-1991

PAGE 10

.1. Change piston material from 316 to 416 stainless steel. The 416 type has a coefficient of thermal expansion very similar to the gray iron air start cap

(5.5 x 104 in/in 'F vs. 6.0 x 104 inhn *F . That material is resistant to corrosion caused by water in the start)ing air system.Its annealed hardness at 155 Brinnel is very similar to 316 type at 149 Brinnel and should be sufficient to guard against galling. However, the material can be j

heat treated to obtain a hardness up to 390 Brinnel.

2. Revise piston dimensions to limit cap to piston diametral clearance to 2-4 mils. The tests performed determined the caps could deform under '

175 ft Ibs of torque by as much as .0013 inch. A revision to the piston O.D. would eliminate the need to supply matched piston and cap sets as

1 A-7818 assemblies.

i- 3. Change the current published acceptable wear limits for the assembly from

.009 inch to .005 inch. Although only an actual engine air start test can accurately determine an upper limit to cap to piston diametral clearance, the test performed and parameters assumed indicate the existing wear j i,

limit to be too large.

i JRG:dg 4

d I

l i

l l

! l l  !

4 I

l s .

J

s# t ,

ATTACHMENT I 4

9 e

~ ~

8 e

,4 i

i l

i i

I

\

{

I l ATTACHMENT II  :

t I

, l l I 1

1 1

I 1

I l

O 1

1 ATTACHMENT lli l

l I

l

.s

+p ,e g i

l ATTACHMENT IV j

4 e

1 l

l 1

l 1

)

7-----~___

-e .? e ATTACHMENT V I

i i

I m__ ___ _,

e ** e i

1 1

l l

i 4

)

i ATTACHMENT VI A

4 1

l 1

l 1

1 i

a 1

i i

)

i l

1 e

"'

• Coop;r Encrgy Scrvit ,

14490 Catalina Stract San Leandro, CA 94577 Y

COOPER Inter-Office Correspondence Date: December 3,1991 To: Bruce Guntrum From: John Gildea

Subject:

Starting Air Valve Cap Design Actions At this time I am sure you are cognizant of all the conclusions and recommendations reached in the starting air valve cap investigation (Ref. R&D Report HE-05-1997).

Engineering is presently undertaking the design revisions therein addressed:

1. The starting air piston (F/N 02-359-03-AH) material is being changed to Type 416 stainless steel. The corresponding E-315 form is also being revised to incorporate this change. As this materialis ferroma practical means with which to inspect future stock. gnetic you will have a

2. The new piston O.D. will be 2.2475"/2.2465" and allow an assembly design

' clearance of 2 4 mils. This revision will preclude the need to match caps and pistons as 1 A 7818 assemblies.

Note, the accompanying change notice will be marked to show a " major optional" type ,

change and the engineering disposition will be to use all existing 316 type stainless steel caps in 1 A-7818 assemblies.

V JRG:dg Copy: R. Johnston DPO Group 91CHRON\DECWE03302

~

i l

~~ ~~~ ~ ~ ~ ~ - - - " ~ ^ -

C4okr^ inruiteds~

• 8

• e Cooper Energy 52rvices 14490 Citihna Strwt S:n Leinors. CA 945P !5 4 415 614-7400 Fax 4 8 5 6 84-7409 inter-office COOPER correspondence Date: December 16,1991 i To: Greg Desin From: John Gildea

Subject:

Starting Air Cap Design Actions With regard to your questions regarding the disposition of the air start cap

assembly design

1. The piston design changes are incorporated as a revision to the ,

a existing piston (P/N 02-359-03-AU).

2. The air start cap design (F/N 02-359-03-AL)is not altered.

3. Ideally, we can eliminate the need to match pistons and caps as 1 A-7818 assemblies since the designed clearance stack up will ensure proper fit and function. However, some sort of safe guard should be

implemented to ensure that an old piston from a 1 A-7818 assembly is not later used with a replacement air start cap. That combination could result in insufficient diametral clearance in the assembly.

I f k JRG:dg Copy: R. Johnston 91CHROMDECwE18315 i

e i

0

INTEROFFICE MEMORANDUM TO: ALLEN GrT I Ei 1e FROM: JOHN GILDEA '

i DATE: JANUARY 9,1991 .

SUBJECT:

PROPOSED AIR START VALVE TEST PROCEDURE '

Attached is 'a proposed step by step procedure to perform a root analysis of air start valve ,

cap bore distortions.

Note this outline is primarily concerned with recording and analyzing bore distortions due '

to machining and torquing of the subject caps. Additional testmg as required will be detailed later.

i To date, I have spoken with Paul Hudson at Georgia Power with regard to sendmg us a .

sample of the air start caps removed from the site engine 2B which failed to start  !

t July 16,1990. He has agreed to send the six or seven caps available. I have also spoken with Larry Bernstein witi regard to fabricating a test fixture in our machine shop.  ;

Barring any disagreement with the test plan, I will proceed to work with Larry in machining the fixture while we wait for the test components being sent from the Georgia Power site.

. JG: dig 91CHRON\JAN\ME09009 I

h S

.[

i WeEROY SesMMS GACMP

DETAILED OUTLINE OF l l

PROPOSED AIR START VALVE TEST PROCEDURE ,

1 4

A. OBTAIN TEST COMPONENTS

1. Request subject air start caps from Georgia Power Engine 2B (S Samples should include but not be limited to caps taken/N from76021).

the malfunctioning valves.

2. Select one random air start valve cap from Enterprise stock.

3. Select one random air start valve piston from Enterprise stock.

2 B. ESTABLISH DIMENSIONAL BASELINE FOR SUBJECT COMPONENTS

1. Define a reference point and mark air start valve cap bores at 30 intervals about circumference.

2. Using 2-pt instruments readable and accurate to .0001", take diametrical  :

measurements at marked intervals at bore lip and at planes approximately 1/3 and 2/3 of the distance along the axis of the bore depth.

. 3. Mark air start valve cap flanges at 10 equally spaced points about circumference.

4. Using "leeler" type gap gages accurate to .0005", place caps on flat surface l and determine flatness at marked points. Record gap sizes.

5. Define a start point and mark piston at 30 intervals about circumference.

6. Using 2-pt instruments readable and accurate to .0001", record diametrical measurements at reference points about planes approximately 1/3 and 2/3 the distance along the axis of piston height.

O. FAB'llCATE TESTING FIXTURE

1. Machine a steel plate 1-3/4" to 2" thick, approximately 6" x 4".

a PAGE 1 OF 3

1. ( o

2. Bore a 1-15/16' hole through center of plate and (2)13/16" holes 2-1/4" from center of 1-15/16' hole along major axis of plate.

3. Indicate the test fixture for flatness and grind as necessary to cotain flatness within .001" and surface finish to 125 MU in RMS or better.

D. ASSEMBLY AIR START CAPS AND TEST FIXTURE

1. Bolt caps to test fixture with (2) 3/4" capscrews coated with 50/50 lube oil and powder graphite compound

2. Torque capscrews to 150 ft-lbs + 25 ft-lbs/-O ft-lbs.

1 E. MEASURE CAP BORE DIMENSIONS WHILE BOLTED TO TEST FIXTURE. ,

REPEAT PROCEDURE DESCRIBED IN STEP B.2 PASSING GAGE THROUGH THE TEST FIXTUREINTO THE CAP BORE.

F. PERFORM STEPS D AND E FOR EACH AIR START VALVE CAP OBTAINED.

G. ANALYZE RESULTS

1. Construct diagrams of each cap and the piston to illustrate bore distortions and flatness deviations.

2. Compare " torqued" cap profile to "as machined" cap profile for each subject cap.

3. Test results should indicate the degree with which, if any, machining, torquing, and creep deformation affect cap bore distortions.

H. PLAN ADDITIONAL TESTING BASED ON ABOVE RESULTS

1. Review current and past machining processes.

2. Calculate clearance reduction due to thermal growth of assembly to establish design limits.

I

3. Calculate maximum allowable cap to piston clearance that would satisfy the five-start requirement. ,

l Establish a course for corrective action based on test results.

4.

Considerations:

l PAGE 2 OF 3 l

- . . ._- _ _ _ . _ _ _ _ _ _ - ~ _ . _ _ . _ . . _ _ . _ _ _ . _ ,

. A .. c,  !

i >- r

^

4.1 Revise manufacturing prccesses to eliminate machining induced  ;

( distortions.

) 4.2 Revise piston material to minimize or eliminate differential in  !

thermal expansion coefficients.

4.3 Revise ca clearance.p bore and/or piston diameter to obtain proper running j 4.4 Alter cap material to prevent bore distortion due to creep  :

4 deformation.  !

i 1

F 1  ;

. I 1 l i  !

j . t 1

?

t 4

i

= j k

[

i

?

4 l'

i t

i 4 1 l

i PAGE 3 OF 3

,