ML20029A293
| ML20029A293 | |
| Person / Time | |
|---|---|
| Site: | Palo Verde  |
| Issue date: | 01/30/1987 |
| From: | Sorenson R Battelle Memorial Institute, PACIFIC NORTHWEST NATION |
| To: | Ruth L NRC |
| Shared Package | |
| ML20029A292 | List: |
| References | |
| FOIA-90-442 NUDOCS 9102110207 | |
| Download: ML20029A293 (33) | |
Text
7L,a nr4 cw en.
J na ry w. 198' i
0Battelle
- )
Pacific %onhwest Labratories Mr. Lawrence C. Ruth P0 80
- Technical Assistance Project Manager Technical Assistance Management Branch
$'1"$/f.M3),"
Planning and Program Analysis M' "4 Of fice of Nuclear Reactor Regulation 1
U.S. Nuclear Regulatory Commission Wasaington, DC 20555 4
Dear Larry:
TRIP REPORT: PALO VERDE, UNIT 3 EMERGENCY DIESEL GENERATOR (38) FAILURE 1
5.D12SL$C1) l l
On anuary 20, 1987, a team from PNL attendec a briefing to the NRC staf f by:
I the Arizona Public Services Company (APS) anc their vendors.
PNL was asked to attend the briefing and provide a third party review. The Te'am was comprised l
of Cr. David A. Dingoe, Dr. A. Durtron Johnson, Jr., and Robert J. Sorenson of FNL, and Paul J. Louzecky, a diesel engine consultant to PNL.
l The following is a brief report of the meeting followed by the team's 'recrvnmendations.
The Cooper Energy Services' engines at Palo Verde Unit 3 were built in 1978 or 1979. The engines were delivered to the Palo Verde plant in -1980. Since that tive the engines were evnnected and integrated into the Diesel Engine Emergency Generator system.
Pale Verde and vendor staf f provided an excellent prer.entation o'f the engine f ailure, investigation, failure analysis, recovery plan, repair method and -
startup schedule.
They showed pictures of the damaged engine, provided a_
detailed description of the damage and discussed their findings. Thy engine problem and corrective procedures were very detailed and professionally present ed.
The schedules for repairing the engine showed the influence of this engine damage on the overall schedule for qetting the Unit 3 diesel engine ~
back into the startup schedule. They provided a plausible root cause for the-failure, explained their repair procedure and plans'for prevention of its' recurrence. ' The PNL team generally found the proposed root cause plausible, but has several recommendations to provide better assurance'that this and other engines will operate safely in the future.
BJLCl".224 On December 23, 1986, the Unit 3 emergency diesel generator _ "B" at Palo Verde f ailed during-preoperational testing._ The Cooper Energy Services' Unit, which is a 20 cylinder engine Mode:1 KSV, threw a piston pair And connecting rod pair which penetrated the cylinder walls during the_ second hour. of a 24-hour test The diesel engine. contained a connecting rod that was electroplated run.
with iron in a-high stress area because it had been over-bored.
Two connecting rods in Palo Verde Unit 2i which had. been similarly electroplated.:have subsequently been replaced..Also, a connecting rod at the Byron Plant has ~ been replaced because it was also electroplated with iron.-
9102110207 901239
~
i
' ~ ~ ^ ^ ^^^^
~
Oncrittio n of t he F A.il.urs The Cooper Energy Services' engine has an articulated connecting rod joined through a connecting rod pin to the master connecting rod. This pair of connecting rods at cylinder No. 9 failed and went through the lef t $1de of the crankcase. The piston pair was also thrown out <'f the engine durir.g the incident. Air continued to enter the engine thcough the two cylinder openings, and oil vapors and spray from the engine oil continued to prwide fuel. The engine continued to run at roughly half speed for about 50 minutes after the f ailure. Fire extinguisher foam was sprayed on the oil in the crankcase to stop the engine.
Due to the f ailure, engine cooling water from the brcken cylinder liners probably entered the engine oil and circulated through the engine system, along with engine fragments, during the approximate 50 minutes that the engine continued to cperate.
ErEJJ1h110A3Lf_ttLCML0tctirgRod 14 eta 11urgical examination of the broken sections showec that e connecting reo f ailure started at the fillet on the oil hole that feeds oil to the articulated rod. Typical f atigue f ailure marks started at both sides of this oil hole.
The connecting rod pin bore had teen iron plated as a corrective masure and the plating extended into the oil hole.
The fillet on the oil hole of the connecting rod is highly stressed.
The stress concentration around the oil hole is probably 2-3 timos the normal strest It was hypothesized that the iron plating cracked under high stress and I,ress concentration.
Iron plating has very low rechanical properties compared to the parent metal.
Also, the thickness of the tron plating repair rnetal was about.050.060 inches.
This combination of low strength metal over stronger mecal subjected the thick t ron plating to stresses above its physical strength which resulted in cracks.
The cracks propagated into the parent rnetal, which then apparently failed frce fatigue.
Cooper Energy Services stated that their Quality Control procedures specified that the plating be removed from g oil hole area. This was apparently not done. There should be no plating in any high stress areas or areas where the stress could exceed the yield strength properties of the plating material.
The thicker the plating the more sericus the problem.
Scme overbored connecting rods in Unit 2 were refurbished by spraying on a nickel coating.
The load transfer capability of the sprayed nickel is questionable. The sprayed nickel may have better mechanical properties than f ron plating.
(a) Iron plating is a comron repair procedure but usually it is used to build-up worn, low stress parts, such as worn shaft surf aces.
However, using thick tron plating on a high stress part requires testing to show that the iron plating will not yield, and also properly transfers its stress to a ductile metal part.
A maximum plating thickness abould also be included in the Quality Control specifications.
_ _7_
Desides the crack in Unit 3 that resulted in failure, another crack was found in _a Unit 2 diesel showing typical fatigue u. arks. This crack had not progressed to fliilure, 4
It does not appear that hydrogen embrittlecont contributed to the failure.
Connett.ing Rot.Str.csses and Torsional vibrations Cooper Energy Services is perfonning a finite clonent stress analysis of the critical connecting g areas. This study should include the influence of torsional vibrations in these high stress areas. Cooper Energy Services should also study the physical properties of iron plating and nickel spraying vs. thickness to demonstrate the capability of these materials to transfer 1cacs in high stress areas.
Ecil.lailutc.02fn+.iQ.0_.QL the En;jAc Besices the adverse conditions caused by operating the engine without cylinder No. 9 connecting rods and counterweights, cooling water and engine debris circulated in the lube oil. The sudden failure may have damaged' the crankshaf t, it is possible that cracks may have been initiated in the oil holes anc fillets at cylinders 8, 9 and 10.
Cons.lutions It was concluded that the iron plating in the oil hole supplying the articulated rod was the root cause of the failure.
However, we have sono concern that the radius around the oil hole outlet may have also contributed 'to the initiation of the f ailure fatigue cracks.
Perhaps the articulated connecting rods need to be redesigned to lower the stress levels.
EcfmTAdi' b.,
P re-Ope rational :
1.
Cooper Energy Services should provide a vibration spectral analysis curve showing all the critical frequencies and their corresponding amplitudes f rom engine startup to engine overspeed.
It would be desirable to repeat the torsiograph of shaf t rotation f requency test and compare the results with the original data. A copy of the torsional vibration analysis should also be included in their report.
2.
Cooper Energy Services should provide the results of their finite element analysis of the articulated and master rod joint stresses and deflections.
(a) In order to evaluate tu nresses imposed by the torsional vibrations, the vibratory amplitudo curve vs. engine speed from startup to overspeed is required.
A test showing the vibratory amplitudes during a fast start is also required because this type of operation imposes full load on the engine crank system. The normal clastic curves are required and the torsional vibration analysis would be helpful in analysing the ef fects of torsional vibrations.
4 This analysis should include the torsional vibration irrpact loads and stresses at the critical connecting rod arvas.
v/
3.
If not already completed, APS should inspect critical engine partse such as the main and crankpin bearings, cam shaf t bearings and surfaces, the valve mechanism bearings, and the cylinder liners, to ensure that the additional 50 minute running time on the engine did not cause additional
- damage, d.
APS should provide the results of the crankshaf t evaluation. (hot and cold crankshaf t deflection test) showing that the connecting rod and counterweight damage did not bond the crankshaf t.
3.
Cooper Energy Services should provide a written analysis indicating that post-failure engine operation did not cause any torsional or vibration damage.
(,
APS shoulc inspect the crankshaf t to insure that there are no cracks in the Ho, 9 cylinder area, including the crankpin or journal fillets and oil holes at cylinders No. 8, 9 and 10 journals. APS should also consider the possibility that crankshaft cracks may have developed due to the rapidity of the failure.
7 APS should inspect the torsional vibration damper to insure proper function prior to the failure and continued function in future operation.
6.
APS should inspect the oil holes and fillets in the two adjacent cylinders for evidence of cracking.
Po st-Ope rational :
9.
APS should inspect the engine for water leaks and analyse the oil for mater after the repair is completed to show that the damaged cylinders do not leak water into the oil.
Also, the oli should be analysed for rmtal content to look for residual debris or post failure wear.
10.
APS should inspect the master rod crankpin bearingI8I shell af ter oporating the engine to be sure it is operating satisfectoril).
From the photographs which were shown, it appeared that there was a section of the crankpin journal that was not cleaned up. This damaged surface shovic be evalvated as a potential cause of future crankpin bearing trouble.
(a) Of ten when an engine throws a connecting rod, the crankshaf t is bent.
Therefore, to determine if this crankshaft is straight, a bot and cold crank-shaf t deflection test showing the deflection at all the cylinders should be required. Then, during the early running of the engine, a hot " feel-over" test should be required to demonstrate that all the bearings are operating ncerally.
.S.
The No. 9 crankptn journal wgreduced in dianeter by about.175 inches Slight surface damage is gill visi,ble.
"3 to remove post of the damage It is not known if the crankpin bearing,will operate satisfa'etorily on this slightly damaged journal surface.
In order to show satisf actory operation, the crankpin bearing should be inspected after sono period of full load operation.
General:
11.
No tron plated connecting rods should be used in any of the Cooper Energy Services' engines in any nuclear power plants.
12.
No nickel sprayed rods should be used in the Cooper Energy Services' engines at nuclear power plants, except in low stress areas, until it can be shown that the physical properties of the nickel-sprayed metal are capatie of transferring the stresses without developing any cracks.
13.
Cooper Energy Services should identify the high stress areas in the connecting rocs. This would assist in indentifying the Areas that need to be critically esolvated prior to being placed tu an engine.
la.
Loading the generator to shut down the diesel engine should be considered in this type of an emergency.
This would be an alternative to shutting the engine down with chunicals.
i 15.
Any used oil from the damaged engine should be analysed to show if there were any abnormalities in the oil, such as corrosive constituents,-that could have contributed to the f ailure.
(This may not be possible. )
- 16. This f ailure of a Cooper Energy Services' diesel generator has generic implications. Other manuf acturers of diesel generators tnay use iron electroplating and/or nickel spraying to repair high stress parts. The stress and f atigue f ailure implications of such plating should be investigated.
17.
Use of fire extinguisher chemicals on fires in turb as been known s
to initiate intergranular stress corrosion cracking o.;ertain materials.
The of feet of the fire extinguisher foam residue on engine materials should be carefully evaluated.
MditienLComnt s 1.
EDSinclenterf.raig._AnL0ther Repa i rs Reynolds French made a general presentation on how the broken center f rare of the diesel engine would be repaired. Because the presentation was quite general, it is dif ficult to comment on the ef fectiveness of (a) The damaged crankpin journal was ground to remove the nicked and scuf fed a re a s.
This journal at No. 9 cylinder was reduced in diaceter about.175 inches. This reduction in crankpin diarneter appeared to leave a small damaged area on the journal surf ace which needs to be repaired.
l
-6o the repair except to say that the nothod propond appeared logical and 3 is the way that many large, damaged engines and other machinep are repaired.
The netal locking prccedure discussed in the presentation is the way many cast engine frames and parts are successfully repaired. Also, the bolting on of stool reinforcing plates to strengthen weakened areas is a logical repair procedure.
2.
Other_utflitics_vith Comee.r_Entrm_.Sen_inctLattici Osines Diesel Engines manufactured by Cooper Energy Services have been installed in at least the fcilowing nuclear plants to provide emergency onsite power:
UTlLLTY F ACL1TY Pennsylvania Power a Light Co.
Susquehanna 1 and 2 Commonwealth Edison Co.
Zicn 1 and 2 Hebraska Public Power District Cooper Louisiana Power and Ligh:
Waterford 3 Commonwealth Edison Co.
Byron 1 ano 2 Comnonwealth Edison Cc.
Braidwood 1 and 2 Arizona Public Service
,Palo Ve rde 1, 2, and 3 Niaga ra Mohawk Power Nine Mile Point No. 2 Houston Lighting and Power South Texas Project 1 PNL would be pleased to previce the following additional technical assistance Third party review of the APS metallurgical analysis and Cooper Ene'gy o
r Services' stress analysis (finite element analysis),
o Review the APS or Cooper Energy Servicest analysis of the ef fccts of operating the engine at 295 rpm for approxirr.ately 50 minutes and other accident recovery of fects.
Prepare an assessment of the generic inp11 cations of iron electroplating o
and nickel spraying on high stress arces in all diesel generators.
Review the APS startup plan for monitoring the operation of the diesel o
- engine, o
Review of a torsional vitration analysis and torsiograph tests on the engine to show that torsienal vibration did not contribute to the failure.
Sincerely.
Robert J. Sorenson. Manager NRO Programs RJ S itj s
NE3 NEl Peebles - Electric Products, Inc.
I Interof fice Memorar.dum O,'
E l-3313 To:
R.B. Politi Date: February 10, 1987 from:
J.V. Pospisil Subject :
10CFR21 - Conclusion of Investigation of Probable Causes of a Defect in Class 1E Equipment Reported by El-3306. Palo Verde, Unit 3, D-G Serial 17609966 6 969.
Reference:
a.
Telephone Notes dated Nov. 25 6 26, 1986, J. A. MacKinnon of Bechtel and J.V. Pospisil of NEl Peebles, TN-E-5168 and TN E-5166.
b.
Receiving inspection Report dated Dec. 8, 1986, 2 pages with attachments, NEl Peebles, c.
Letter of Certification, Resin D-111, dated May 9,1980, Sterling Div. of Reichhold Chem. to Parsons Peebles, with attachments.
d.
Letter EF 3308 dated Dec. 9, 1986, NEl Peebles to Reichhold Chemicals, Inc., Sterling Group, c.
Letter dated Jan.13, 1987, Reichhold Chemicals, Inc. (D.E. Campbell) to NEl Peebles (R. A. Rosstnan) with attachments.
f.
Sketch of Failed Rotor Pole, Diesel Cencrator 3B, Attachment to Service Report F-1063, NEl Peebles.
g.
Low Magnification Views of Failed Coll Area, Photographic Lab Records, Jan. 26-30, 1987. _ NEl. Peebles.
h, Overspeed Test Record, Standard Test Form 2002 for Serial No.
17609964 200 dated Dec. 7, 19 77,- Por tec, inc., E. P. Div.
Our internal correspondence El 3306, a copy of which was attached to your letter of Nov. 24, 1986, addressed to Mr. James Taylor, Director of Office of Inspection and Enforcement. U.S. Nuclear Regulatory Commission, reflected the status of our investi-gations as of that date. As stated in El-33*J6, three potential causes were identified, however, one of them (a, Mechanici damage by impact or excessive overspeed) was immediately eliminated from any further consideration due to a total absence of any corroboratory evidence. Two remaining potential causes to be investigated further were identified as; a.
Improper formulation, mix or cure of the resin.
b.
Improper application of resin or contamination of the wire surfaces.
As apparent from Ref.
a., arrangements were made by telephone with Bechtel to ship the pole with the damaged coil promptly to Cleveland where it could be inspected in
.g detail and examined further by our Engineering Lab. Ref b, contains the results of NjF the inspection performed. As there were no dominant signs symptomatic of improper I
1 A wholly owned subscary of Nof thern [9ncehng indv5hics L td, of England YbY O
?
4
1
)
application of the resin and/or surface contamination, it was decided to proceed with Q
the verification of resin characteristics by physical or chemical testing of resin samp*
les obtained from the f ailed rotor pole.
If the results of these tests would confirm a very high probability that the resin used is indeed factory-mixed Sterling formulation of D-111A with no apparent anomaly due to its cure, the potential cause 6. could be eliminated from any further consideration. This then would confirm the potential cause c. as the probable cause of failure by a process of climination, if, however, significant anomalies would be detected by these tests, the potential cause b. must be retained as a probable cause (either alone or combined with c.) and its generic impll-cations must be also considered.
Relevant methodology of testing was discussed by our Engineering Specialist, Mr. C.J.
Moosbrugger, with the Reichhold Chemicals, Sterling Group's expert, Mr. Ray Cushner and it was agreed that a positive resin identification from small samples can be made by infrared (IR) Spectroscopy but the Thermal Gravimetric Analysis (TCA) could pro-vide also some information on the cure of specific samples. Since our Lab lacks equip-in order to ment required by either method, testing was subcontracted to Sterling, eliminate any doubt about the proper resin formulation and mix, the Sterling Lab agreed to prepare a sample batch of D-111A Polyester Compound, such as certified by Ref. c.
Further, our Lab v as to provide a thermally aged specimen of D-111A in addition to the samples obtained from the failed rotor pole winding in order to eval-unte thermal effects of curing and aging.
Ref. d. identifies the samples according to their origin.
Sample B was thermally aged for 2352 hours0.0272 days <br />0.653 hours <br />0.00389 weeks <br />8.94936e-4 months <br /> at 140*C af ter the cure.
On the other hand, samples A1, A2, A3 and C had no significant history of thermal aging following ne standard cure because of a very few operating hours accumulated by the Train B Generator from which the failed pole was removed. Of course, the g
freshly mixed compound iy Sterling Lab) called " standard" was initially uncured.
As reported in Ref.
e., Sterling Laboratory performed the Thermal Gravimetric Analysis (TCA), determined the filter content of samples provided and compat ed the results with those of a freshly made batch of D-111A Polyester Compound. Upon completion of their analytical study, they concluded that:
" Comparison of the TCA indexes, shape of curve and filler content indicates a very high probability that Samples A1, O t, C are cured specimens of D-111A."
These conclusions can be verified by copies of TCA Charts attached to Ref, c. and en-able us to eliminate the potential cause b. (Improper formulation, mix or cure of the resin) from any further consideration.
Thus an improper application of resin or con-tamination of the wire surfaces remain to be examined as probable causes of the failure, meager corroborating evidence premitting.
As apparent from Ref. f.. the outer winding layer, one wire thick, separated on one side of the pole from the rest of the coil "as a sheet" deflecting outward due to centri-k fugal forces acting on individual wires (turns) of the layer, in spite of the complete f
resin bond failure between this "shectilke" layer and the rest of the coll.. resin bonds between adjacent wires of this " sheet" remained mostly intact.
Although two outer l
layers were cut and " peeled of f" from the coil before it reached our Lab, aome wires of the outer layer (with black varnish coat) remained bonded together in pairs, or groups of three and four, giving an indication that the bond between individual wires I
of the outer layer was considerably _ stronger than that between the outer layer and the h
rest of the coil. This f act can be considered as conclusive evidence that the wire umnammmaxmemummmmmmmm
m s
h l, '
I sur faces were not contaminated..Were it otherwise, equally weak bonds would have been found on all three sides of the (square profile) wire. That is, between adja-cent turns (wires) of the outer layer as well as between the outer layer and the l
c oil.
Furthermore, even if it could be accepted that only one side of the square wire were contaminated, this would have to happen in regular 81 foot intervals 25 a
times over, so that the contaminated surface could position itself in the failure area.
Of course, there is no credibility to this scenario and the hypothesis of wire surface contamination should be discarded.
.a'l Relatively Inw magnification (10X) microscopic examination of the outside-facing sur-
)
faces of a portion of second layer (received still intact) revealed relatively large areas in the failure rone that appeared significantly " smoother" than the remaining surfaces.
Few small irregular " rough-appearing" areas randomly distributed over
.p the large " smooth" areas contrasted quite sharply with their general background.
Our first hand notion that the " smooth" appearance is a result of mechanical polish-q ing due to a relative motion between the first (failed) and the second layer had to be dismissed as there were no pervasive signs of abrasion, except for the " rough"
(
spots. Similar microscopic examination of the inside facing surfaces of wire groups p
(received separated) of the corresponding pnrtion of the first layer revealed the i
existence of like contrasts also there and again, no signs of abrasion in the " smooth" i
g a
aren. This implied that a contact between the first and the second layer in the l
failure :one was limited to the " rough-appearing" arcos only and the " smooth" areas I
never touched each other.
Ref. 9. illustrates the details, y
increased magnification (25X) confirmed that the " rough-appearing" areas were a i
result of a " brittle fracture" of cured Polyester resin bonds while the " smooth" arect d
were undisturbed surfaces of Polyester compound-impregnated fibrous wire covering O
of either iaver. senarnice by a void.
^ithouch this interpreietion was consistent with the postulated failure mechanism, there was one unexplainable discrepancy originating in the method of pole sinding and curing.
With the Polyester compound applied by brush over the entire surface of each completed layer before thc next y
layer is wound, the inside-facing wire surfaces of the next layer can be impregnated j
only by bringing them into contact with the coating of the preceeding layer and f
keeping them in place during the cure. Since the coil sides are held compressed 9
aDainst the pole body between two full-length clamp plates to which the pressure is applied by a series of C-clamps distributed over the entire length of the pole,
(
only small (one wire wide by a few inches long) voids con be sustained during the cure in either coil side. Major voids can exist only in the free coil ends where
]
they are benign due to their orientation relative to the forces act;ng on the coil.
y in order to create larder voids in the coli sides, it is necc$5ary to apply the Polyester
'A compound so sparingly over relatively large areas of the prcceeding layer, so that Q
the inside facing wire surfaces of the following layer do not et,teblish a normal con-tact with the coating when the cult is compressed by C-clamps for baking.
In such
,j a case, however, the fibrous covering of the loside-facing wire surfaces will not J
be uniformly impregnated.
That, of course, is contrary to the conditions actually observed. On the other hand, a relatively uniform impregnation of fibrous wire O
covering on the layer surfaces facing each other with a gap between them could g
be achieved by using Polyester compound diluted by a votatile solvent. This would j
allow a relatively uroform wetting of the fibrous wire coverings which would readily 1(
wick the more fluid compound durint the winding and clamping process. Thermal T
cure would then drive out the volat les leaving voids in their stead because the i
d@
" scabbed-over" surfaces of the crmpound impregnating the fibrous wire coverings g
of both layers would no longer bcnd with each other.
Of course. this hypothesis does not hold for several reasons.
First of all, the Polyester compound used and h
?
m@
Ud I$
Y SsY$
l
}l# "7
.g.
positively identified as Sterling D-111A is solventless and factory-premixed-from Q
1001 non volatile components. Secondly, the voidy area is highly locallred, being confined to the central section of the last layer on one coil side only, even though the Polyester compound must be applied to the entire preceeding layer before the next layer can be wound over it.
l'urthermore, all the poles were wound on the
" day shif t" by the same person, who had more than 10 years experience winding rotor poles and was known to be a conscientious worker, who would recognlic a dif-ferent " feel" of the diluted compound being brushed on the windings.
All the above reasons notwithstanding, the hypothesis of " locally-diluted" compound -
seems to fit all the symptoms found. A probable scenario can be constructed under wh!ch a coating of the properly premixed compound can be applied over the entire layer of the pole winding in 'a manner that permits its localized dilution in one con-fined area, such as observed.
it is only necessary to assume that the pole winding operation was not completed in one shif t and the last layer was lef t to be wound during the following shif t.
In accordance.with the Shop Practices.then in effect, the partially wound rotor pole would have been lef t in the ' pole-winding lathe and covered with a polyethylene sheet, the unused part of the $ gallon can of. the com-pound mix would have been put in the "in process" cold storage and the brush would have been cleaned and stored in a con of Xylene (MS 10,1 - a solvent). -Before resuming winding operations on the next shif t, the winder would remove the can with partly used compound from the cold storage, allowing it to warm up and, mean-while, would remove the brush from the Xylene can and squeeze it dry into a clean rag or paper towel. When ready to resume pole winding process, the first operation to be performed by the winder would consist of coating the entire surface of the last layer already wound on the pole.
This requires starting on one side and pro-O gressing towards the trailing end, then advancing the pole winding lathe by 1/4 e evolution to coat the end surface and 1/4 revolution again to coat the other side surface. Two more t /4 revolutions of the lathe enable the application of Po yester i
compound to the entire layer surface so that the next layer of wire can be wound over it.
Now, let us suppose that the winder was distracted by somebody. or some-r thing and did not properly dry the brush af ter removing it from the Xylene can but just laid it aside.
Resuming his work, he would stir the compound in its con-tainer fir st, then pick up the brush (no longer in the Xylene t.an), dip it in the container and start brushing the compound on the coll. side facing up without being aware that the coating is diluted by Xylene released from the brush.
He might not even feel less drag on the brush because the compound is still cooler than normal and its higher viscosity tends to compensate for the effect of solvent.
By the time -
the central area of the coil side is coated, nearly all solvent v.ould have been purged from the brush so.that the coil ends and the other side are coated with practically undiluted compound.- While there might be s.ome slight difference in the visual appear-ance of one coil side and the other, they can never be viewed simultaneously., thes making a detection of this condition by the winder or the inspector quite unlikely.
Although no claim can be made that the above-scenario represents a factual account of events resulting in an improper application of the resin being investigated, it is justified to accept the local dilution of resin as the most probable cause of failure based on the existing symptoms as.well as the credibility of the events described above. Manufacturing defects such as this are detectible af ter curing by tapping the coils for hollow sounds, which is a part of routine O.C'. procedures. Why such detection was not successful defies deterministic explanations and can be considerated only iri the realm of statistical probabilities.
Viewed in this perspective, the failure O
con be considered e Premeiore feiivre ("infeni merieiity") due io e rendem. reeeiiv I
f
$hYNWS4 N1%48 f,
t
)
.W@*
i detectible but undetected, manufacturi0g defect having a low probability of recurrence..
O ii c>n. iaerefore. ue treaico as an iselaiee incieent wiin ao sic"ificani ceneric i nii-cations.
As readily apparent from Ref, h., the inherent capability of the generator to withstand without injury the postulated overspeed of 25% above its synchronous speed was con-firmed by testing the first generator of this design built for Polo Verde Unit 1 at 750 4
RPM.
This generator, tdentified by the Serial No. 176099681-200 was overspeed tested ener0 red at its rated voltage in accordance with IEEE Std 11$, Section 2.45 while l
asembled on the factory test bed with an auxiliary Jack shaf t and inboard bearing i
for other running tests.
Driven by a direct-coupled D.C. Dynamometer and excited from an auxiliary D.C. Source, the generator ro'.or was kept at its synchronous speed It,00 RPM) long enough to obtain stable electrical readings and a set of mechanical vibration measurements at both its bearing housings.
It was _ then accelerated to 1251 rated speed (750 RPM) at a rate of 25 RPM /sec (while its excitation was reduced to maintain the rated voltage at the generator terminals) and held at these conditions with no further chanDe for 300 seconds.
Upon repeating the vibration measurements 4
and obtaining a new set of electrical readings during the clapsed time, the rotor speed was rtduced back to synchronous (600 RPM) at a similar rate and its excitation increased aDaln as required.
Upon obtaining another set of electrical readings and performing all vibration measurements at the synchronous speed again, the rig was shut down for a thorough visual inspection which revealed no thanges in any part, nor any evidence of injury. Cursory examination of the vibration measurements re-corded in Ref, h. conllrms a high degree of dynamic stability of all rotating parts and the low vibration levels at 750 RPM together with the results of visual inspection provide conclusive evidence of the generator design integrity.
O Since aii Diesci cenerators are run at 10s overspeed af ter they are instaiied ai the site in order to set or to verify the setting of their overspeed trip devices, it is advisable to verify the Integrity of their generator rotors by a thorough visual inspec-tion following the overspeed trip setting run for each.
This can be facilitated by a borobcope inserted into interpolar spaces upon removal of the outboard air intake covers. Thus the effort required to perform such inspection is minimal. While it is recognized that a limited overspeed exposure, such as this, falls short of confirming the specified overspeed capability of the generator, it certainly exposes the Diesel Generator system to the highest speed encountered in its normal operation.
It simply does not verify the margin available in the generator rotor. within the speed range normally considered unsafe for the diesel engine (
1101
'1251). On the other hand, centrifugal forces acting on the rotating parts at 110% of the rated speed reach magnitudes 211 higher than those at the rated speed-and exaggerate the critical stresses and deformations of defective parts sufficiently to permit visual identification of significant manufacturing defects, if any.
Duration of the overspeed exposv.re is of a secondary importance as long as the time at the overspeed required to set the overspeed trip devices and confirm their function is equal to or exceeds the time at the overspeed experienced by the system when the trip devices actually operate.
in any case, the magnitude as well as the duration of the overspeed condition must not-be injurious to any part of the Diesel Generator but, _ at the same time, must adequately simulate the most severe condition the system is required to-tolerate during its opera-t ion. Arbitrar!!y set overspeed duration. such as 5 minutes suggested initially, 'is not really applicable here even though it has been traditionally used by the electrical industry as a typical duration of proof _ tests.
Summary of conclusions:
1.
The wire-wound rotor polo failure is a direct result of a readily detectible but undetected manufacturing defect.
- mww wwemmmmmwraemm:Muscrimumswnswggg,my,gn,
~~' [s.e.45b
,qm
- ?y Is
- fjf, EGF 2.
The manufacturing defect resulted from imprcper application of the specified W.;
g Polyester bonding compound, most probably caused by local dilution of the re 19 7 in the area of failure.
k [!p//
3.
The design integrity and the identity of critical materials were confirmed by
}
/
Ing.
No evidence of additional debilitating ef fects was found, i
4 4.
The manufacturing defect causing the premature failure of the rotor pole appi.
4 to be a result of an isolated incident with no significant generic implications.
si) 5.
The corrective action suggested in El 3300 appears to be adequate, however,
]jh ~}
- ~ g.4 arbitrary 5 minute overspeed condition de-energired shall not be considered r '
j' gf tory.
]W p.,
It is recommended that the failed rotor pole be rewound with no change in the ors
- t4 atl tl specified materials or manufacturing technology. When accepted upon th *pecifit ;
"p f D routine inspection and tests, it shall be returned to Palo Verde renewal parts sto ;
hh.M for future use, jerg r
o h/
Q?f J J.V. Po/)i s il, P. E.
%s sp
$l- @
Manager of Engineering State of Ohio License E-039040 Y
"k n (Y
+f j
cc: NDM. CJM y
Attach Ref. a. through h.
(;
3 u
k W
M h
1 y
i b
(i
?
t, a
Y
_t
bN
. TELEPHONE NOTES F000S27Dn-t-5168 oistniavvei>H D. R. Anderson t
g tr-flevkinton-
[
g C. A. Hierzer W
-h-Or-liechs r.
J. A. MacKinnon Bechtel (602) 932-5300 T. L. Mack or-
- PHoNC
__. _ _ ugg SY3 Joe Pospisil NEI-Peebles (216) 481-1500 R. R stiens to or-PHON ggggg L
November 2$ 26, 1986 10407 M.11.04 T. C. Dallweg oATC JOB No.
FlLe gggggg Damer?d Rotor Coil in Unit 3 Diesel Generator Ray Miklos l
sunstcT d proposed disposition i
g,g,,,fkedwithJoebothdaysre6ardingthecolstatusan 7
for releasing the machine to APS. The followin6 points were discussed:
Joe wants to pursue the chemical analysis of the coil resin to rule 1.
out the problem with the cocpound. The coil is to be forwarded to the followings o.
NEI-Peebles Electric Products Inc.
170% tuc1Td7ver.ue Cleveland, Ohio 414112 ALTei)T1on s r'f. ' ricMTU 703!MMThaniger of MaTertuis 9
7.
Joe' coes not7We'3 TrnTEtrv1T7rrepnrttg the tutt mrster un n.e
'4 spare coil anc retTirrTirFTt~to = spares 77rttic vsre7scu3e.
~3 The spare coITs'enTite UnTt 3 ufTrn7TWUrtcIn vert Trer The y
samc nWTi'6T~FHIh.
y
~
~
A f'
4 Joe WB 1 Inspect the cell when rlceiveo, c6nsult witn the resin
~
k[i
[
manuf acturer to determine The couTse os action, perform Ln
~
X anoTysis, review the resiiTts. This errort as rot ecaTt s or or,e to sfe:
two weeks.
k T indicated to Joe that next week _ Gecember ist) Norwalk Home Orrice p
Engineering will contact him retar31ng support for the DDi resolution.
Y Recorded by:
~J.
Mnon k
Reviewed by:
f/pg
- j.
.K it
' W
. h i
CMH DRA: JAM:ce
- i lff x.-
v.
r ATT AC - V EN";A TELEPHONE NOTES F000S277 78-t-5166 Dis 7Rl9ufloN D. R. Anderson O
~. Hierzer
- c. A
-J-Dr-timmtwu J. A. HacKinnon/
Bechtel
"" "" (602) 932-5300
- 7. L. Hack txtr-73to- -JT-RT-ttehtu O. e. Aris nIUn Joe tospisil NEI-Peebles (216) 481-1500 R. P Stiens to:
or ruowc gg,g, November 25, 1986 10407 H.11.04
- 7. O. Ballweg oAvr
- oeno
-ritt g
g Release "B" Train Diesel Generator to APS Ray Miklos suo/cer Joe concurred with releasing the Unit 3 train "B" diesel generator to APS to Norts coeplete start-up testing based on the following reasons:
1.
Train "A" has been run on site for over 140 hours0.00162 days <br />0.0389 hours <br />2.314815e-4 weeks <br />5.327e-5 months <br />, energized and unenergized, and has perforoed the overspeed trip verification.
2.
Train *B" has been run on site foi three hours unenergized.
(It was not discussed during the telecon, but was verified that "B" diesel has verified the overspeed trip setting.)
g 3
"A" and "B" train generators' coils and spare coils were wound using the same resin batch.
_ Based on the above and the concurrence of Fred Lamcho of NEI-Peebles that the generator is in satisf actory condition, the diesel generator train "B" can be released to APS.
r
'[J. A. fab'Innon/DD.. $$fers h1T4
.h Recorded by:
NA Reviewed py:
j CA e
/;-C. H. Herbst k
CMH t DR A :.1 AMi cb O
uo ee. im
,-<y,
w, e.., : v.n -.
~~
k/,
RECEIVING INSPECTION REPORT ["{
{ "]
CUSTOMER: Cooper Energy Services SIT E : Arliona Public Service's Palc ferde Nuclear Station H
l RECEIVED FROM: Cencral Electric Co.
I Apparatus Servico Shop 6 Anderson Drive Albany, New York 12205 Il PART: One Wound Pole Assembly A-64665-50 From Synchronous Generator 17609966
,i
REFERENCE:
Customer Order CES 36:106493 Portec, Inc./ Parsons Peebles-Electric Products No S.O. 17609964-69 DATE:
12/4/86 1
PROBLEM: Outer Two layers of Rotor Coll Reported to be loose.
l
- 1) Shipping container.
Received in skidmounted wood crate, condition - falr.
7-acking of parts Orc Wound Pole Assembly with two outer coil layers and bottom
- eceived.
pole washer removed.
Wound Pole Assembly was loosely braced in center of crate. No additional damage.
The two outer coil layers, cut from top to bottom turn, were packed on each side of assembly, j
- 3) Condition of parts.
a) Pole stack was tight.
b) One end ring tab was bent.
c) Insulation washer (bottom) was in good condition - no cracks or brakes, d) Coll winding, which remained on pole assembly, appears to b-bonded and tight. There are volds in between layers on one side of pole only. These were.010 to.015 wide and one wir-deep.
Location of volds per sketch below:
M N "-
5 M
'dy h Barwrm M LSirsJ gg(y3ac wpu, b 4 *1. B ( 9 3
- 4.,
,5f.> rc, 1
- 4. a
- b +
" !p
- g a
=m m
p
~ ~,
~
- P ww--
{
)Q g
- G d-
- f -
4-O-G$
]--
g 1
5, sunmem wues
-)
L
'l Aub C i u-% T LAY E R.
fk WOOND PCLU. NIP-.W.GQ EROld 30TTOM hh
i>
- k. ;..,....
i t
L e.
- 1. I_.
j
. l
. A... - E., 0. '
i --
H i
_;. j..:..
...-....]..-........
~
' 1 l
1
. s J...
..i
[
4
.j.
I s.--._.
-...j.. _.
_4..
l
.j.
l..
.g. 6.N D..I M+',~
T~
.3.
j..
l
. 7..
g.
s i
.b
..j.
.L.
{. -.. }.
.q..
i t
C...._~~
n.
7---
J..
...... ~
-. -...._....L_
.. f. 7* U..
)4 I
l l _._.
..... l (ye/).
tg..
.I j,
~ --
.3.
s l_
j, 4
1
,q t
J l
f(
.. fu t 1).. D..r,..
_. -.g'
\\.
i
_..._dG 1 f.).... {...._..
. _ l.~._.
l.. _.
j 4.
a j
d I.
{
j
.i i
j...... ~
4
.. c y
I i
1 e
r n...
g d
1 3
4 l
j.
b)
J'
}.
l,l.
.i 1
I _ _..
I l
l i.
i I
4 l
y L..._-
-_d...
. 94.L.
1,i i _...
p+r0
.. }' _.
J...
_. _. _. _ j.g..__. J, _/1_ __
i M,.
. 1.
(
)
. U 2 0 4)..
l..
)...
/'n o r o
.d i
k
,U I
j--_.
i.
t.
g I.
1
, '; )
6
,y l
.T-jf
.__._.... p.._ _-._. y _._.... _.,
]
q N) g e
v i
p j
I y
- 3.. '
,h
@y lWI
)
k V
n j
i,
.,a i
Q ip I
a d1 y
@y
~
?jk
...p sutot e t R O N.
i' o W 4 5 5 C M O t.s' j A - 6 'L& _0l 5.'. O souo n I. o '!_9 6 '-l q
u
.a. tf_t., 2.,..
riA
- a,3_p. e
,3 yo.
9l r
}
l e
n.
w "Y S PARSONS PEEBLES-ELECTRIC PRODUCTS,INC.
- I l
d*Y l
J i I ntvisios WE egua 1725 CLARKSTONE ROAD
- CLEVE LAND. OHIO 44112 - ' -
~~
~
' o~~r._.
f
~~~
s
~-
SHE ti No -
m li' l
}
%q
........ 1....
l
- i L...h...
ms ww s,.-
- k h
(( N %. %!]1,If;$.$. [Q.h%, s,@, ff%f'T, "Q'"T.,*W,T;**r p r7,,
n T
m r ;.,
,.s, 1
t.
...,,1.r a ;,m.u.a w s,
o
. m,
, y,g, @.
~
s
,a
- pcq
.=: n 2.p>
A.f.ACHW NT C
[ rey _,
i
%k.
~
~
o g
e A;
,f 0,visicm or sitics.o.0 cn.wic Att. e4 Ult) 746 7600
- 7413 832
' i SJ TWX 710 666 3877 May 9, 1980 Parson Peebles Electric Producto, Inc.
1725 Clarkstono Ed.
Clevelard, Ohio 44112 Attention:
Purchasing A ent C
Centlement This is to certify that the 10 - $ gal, pails of D-111 furnished against yov.r order $029-3 vere e.anufactured in accordance vith and meet the requirements of our standerd canufacturing specification for this item.
Tours yeay truly, C31 Fsidg C. G. Prost V. P.
Manuta eturin,g CGF 0B F ACTORIES H A Y5 VILL E. P A
- MANCHESTER. ENGL AND
s
(
d M MS
[bSINN b N b 1
MATERIAL SPEC 1 g N t d
a4inc con W1 5"c *
<<.no CldTO -
. ny.
,11,,,79.
..,0.,......
se nsEns.ev...... n mm....to..t.....
m o
s u n a. 1..... o r......i........
so n a n o e f 0LYESTER RESIN 1.0 Descri io_n:
t Tolyester resin Varrdsh for une in " Wet Winding" of eles per 1r.suladon t
S;ec El-1.5.0. Tw part compound mixed by vendor prior to shipment DIST' lays high strength and cohersion. Taints on easily.
J. O I hysical Treperde_s:
- 2. 2 Color Grey
- 2. 3 Composition: 03'" Styrene t'onomer, 73?- curing agent
- 2. 4 Evaporadon rate: Essendally 10070 r.on-Volatile 2.5 Plash Tolrt: 93 I'setadash method
- 2. 6 tbn - e::p osive 2.7 TLV: 203 ppm 0
2.8 Shelf Life: 3 Vonths at 35 F.
- 2. 9 Gel Tirne: 10 l'inutes at 150 C.
(O.
2.10 Welgh; pec gallon: 9. 6 Tounds
^
0
- 3. 0 CuringRc'e,: 2 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> at 290 325 F, deperding on weight ard shape of work ple:e.
- 4. O Cured Prongde_s_
4.1 }!ard ess: 65 Shore D
- 4. 2 Insulator Class: F ( 155 C) 4.3 l' eat Cenection Temperature: 140 C
- 4. 4 Water.'.' sorpuor.: 0. 41% (7 Days at R. T. )
- 4. 5 Weight Loss: 1. 035 (7 Days at 180 C)
- 5. 0, Approved Sources Ma nufactur_er Catalog No.
Sterling Varnish D 111A Thermoset :hsdes L c.
DC-9 or DC 529 e'McGrak Edisbn ' Service 2A353 VarrJsh 9
- Rev. 7/a/80-FMs PORTEC INC., ELECTRIC PRODUCTS DIVISION,1725 CLARKSiDhE ROAD, CLEVELAND 44112
~
l PREPARED BY:
DATE:
APPROVED BY:
DATE:
APPROVED BY:
D AT E:
..e.>.u nvcruerg..>un n..
9 v.. %. %.et.n....'... A % 4..e< m l C"Z5sEE2WZUEFEHiMT5?WF N!7EEEER2mEWmier
w
$0d55E NEl Peebles - Electric Products, Inc,
^
ATTACHMINT D nV, 1704$ Euclid Avenue Clevetand, Ohio 44112 Tetephone:(210) 481.1500 Teler:: 241504
" I I
I O
December 9, 1986 EF-3308 i
Reichhold Chemical Company l
Starling Group Routo GS
(
Sewickley, PA 15143 i
Attention: Ray Cushner
Dear Mr. Cushner:
As we discfssed in our telephone conversation of December 9,1986, we have enciosed resin samples, labeled A1, A2, A3, B, and C.
Please determine by IR or TCA analysis, whether one of the A sampics, the D, and the C are D111A.
j A1, A2, and A3 are all from the same side of one pole. A2 may have polyester Q
mat imbedded in it. One sample from this group is sufficient. Sample B is an aged sample.
If the sampics are not identical, we would he Interested in I
learning what the difference in them is.
Your reply is urgently needed, if there is a cost for expedit!ng the answers, please contact our Purchasing Manager, Mr. Richard Rossman.
Thank you for your assistance.
Yours truly, NEl PEEBLES - ELECTRIC PRODUCTS, INC.
~$
f p a-7'Y Charles Moosbrugger Engineering Specialist CM/jmr Enclosures Po v
g hmti wned tuk Mig of N0'thern Ence,eetin indutt'ies Ltd, of En'tand.
~i k*[
h.
M s
~
. g. H 2-R 9 - '- + -
- +
+ i - - o ~ o ;-e u.,. i = 4%
A
..... t..i.
A.4.....
...~
~
- t g
,e
_.1r,_.
q
.+ g...
g
.. 3
.q...
..7..
.a
- J.k
....l r
+
....e
...n..
. -...o
.,....n
-4... g..
L...
i
.,8.-..g
.?^..
....-.,..-.i.
..JA J.
..._...4
....,,,,,,,,,.y i..
............. e
+
. y. 7-. -...
....-....,_r...;.,.-
.._t._
p
_.p
...~...
.s.-..
~
...._.t...........,
.1.'.__...
.. [.
1
...t........
L...
4.....
L [..
u j
1...._..-....
--i--
.. i.
........8
.8
.,. i,.
o
.+.4..-..,...,.
....o...
- 4. -
+....
......c..
.....m e
g.
.1.r..
..a.4
- e. g
.s-..
..*e..
.e.
.]..q...
. ' +..
i
.g
.M.
m
.o c.
_. a...
A..
........ A....U.
s...
.4.
c 1......
2.m.n.5
._:t.
. 4 _......... 1....
u p..... _.__.
.s-p._ m.
~
=
w.(
...~l........A....,...
-._- A 1 m f.
,.w..
.f. f... -
.. = r M.a I. _t... l.
9..
i
.g
... i Pl d r*
.,w
... +
- 1..
$.P..... ' p g i 5....
..i...
=-.
o r. <
i
.)
l......
g j.
4
...,...4
..i.........J.....
i i
7.._-
i 1
..I....
~ ~.
. _J i..
.9..
- l.....
1 e
4..
r.....,
,.m-..
...-J,
.......g.......
..}_....
.). m /u..ajf.=
s.
..]......
. _ :.. 1..........
-H
., o
.~. a
..p.,..
t l
... s.
t.
.i
- o.... -
. 1 4
e e.
l 4
. ~
l i,
2 i:
L'tT. L M. M
- e r.4 T E ')
C c
._ _ _.._..._. ~
T..
.1 1..
p t
- e. -
.p suenct ROTOR Pott A s t C M O L 'r _ / A - 64665*.5"O s o.No. M L c 9 9 6 4 '
.O,
_f. o e mcv e rn f r.w J% Srr io. No.
i
^"
~ PARSONS PEEBLES-ELECTRIC PRODUCTS, l'NC.'
DATE RE VISION
~ '-' ~ -
' ~
- CLEVEl.AND. O. HIO 44112"."--~
SHE.tT NO. -
- OF' I
a
}
.. l
.......a.
..k...
j'
n - - -... - - -
3S A....AC'- V' ~ NT
~
b j lleichhold Chemic a h, Inc.
p 3 cicam own
@b]
twe os T
cea,cun :nny.ana 155o n
7) l V
'glSuf
,yQ
$}$
REICHHOLD
'slK'
"~
~ -
Rs ? -
January 13. 1987
&A.o
'O
- %lq hh l1 Mr. R. A. Rosstr.an MS thnager of t'ateriais
/ "bp flEl Peeblos - Electric Products, litc.
LS Mp 17045 Euclid Ave Cleveland, Cli 44112 y( p J
U**' "
- **M hy$
QR4 m /Erp/C h[fh(
In accordance with e request of your Purchase Order No.14935, we have performed tf ThermalCravamet3 Analysis (TGA), determineo filler content of the samples and TM compared these results with a freshly made batca of D-lllA Polyester Compound, O.fh@!j ii f.e Conclusions of analytical study:
W TGA Index
% Filler mm m s]
Y$!G Standard (0-111A made in Laboratory) 163.5 26.35 ph h Resin Sampic A 1 163.5 29.85 gg:{
Resin Sample B 162.6 33.35 Aba Resin Sample C 164.5 29.90 b}M Qij.
- cp Comparison of the T&A indexes, shape of curve and filler content indicates a very T b' high probability that Samples A 1 B & C are cured specimens of D lllA.
a p :.
(TAf;j TRA Charts are attached for your use and reference.
W@Ih(k;M bW We have enclosed the original and one copy of our Invoice No. 1101-10388, covering
/$
the reported analysis i
if you have any additional questions, do not hesitate to contact us.
w[m Yery truly yours, l'b
% @h U ll&
'A
.d*
onald E. Campbell h;Ih;vl)
AM Product Manager DEC/rmn Enclosures
&y i
g I
20gf
%oJ g
BM Tot (412) ' -4:-
Goo PMbd TWX 71%: J 77 hhh d$.bd
^.
1- - _
' LAB.MADE D-111A WT:
9.9860 mg RATER '
5.00 deg/ min H100.00
.Z 19.90%
gy 85.15%
g 1
-J-FROMs 31.73 1
)
M T0 500.60 I-WT % CHANGE: 73.62
-I W
b FROM '31.73 i
i~
TO. 382.33 WT-I CHANCE, 14.-97 l-
"C FROM. 31.73 4
t.3 De 411.82
,-* 50. 00 --
. to VT % CHANGE: 37.47
,c
.p t
g, y -
26.35%
i
~
7 i
T fo A - 1 6 3, F 4
m 1
I
[
'{h'
~
0*00 40,00 11fi,03 150.00.
250.00 3i0.00 3d0.00 450.00 530.00 6d0.00 610.00 720.00 i
3g ' '
cf EILE: BSAPE.TC TEMPERATURE (C)
TG j
03Te, g7,03e07 T ine, os 24 e
i genx,s.< cysa 1,4, x,,
1, y.
Pag E
~
Ndi e
En rsa,d bjj NEI SAMPLE Al WT 9.3524 mg RATE 5.00 dog / min RW i.tl 2.=i e
?.$-
L.100.DD i '-}'
l9.90%
I'ig BS.08%
MJ
';-t l
1 edit
%v LM
, El FROM: 31.72 kf,k C' I
TO 599.69 Iisj
'sf WT % CHANGE: 70.13 M: b..
a 5.25%
p3 h FROH 31.72 1,
{j?) T TO.-383A2 en WT % CHANGE 13.9 es.s m-wg H
i f&O I
FROM: 31.72 IN O
~
TOs 400.55 I
p@FK ciD. DD --
f WT % CHANGE: 34.73
.d d u
~
3 Nj
~
M if l
N 29.85%
i l
r 4
- ev.
pLd s
IN. I,d
.N t
I smy l
7 g g - /6 3. 7 d
SE 1
M Yh s
- p g,gg L.. _ _,. -
...---s
?
40.00 110.00 189.00 250.00 320.00 393.00 461 03 533.00 600.GG 670.03 740.00 M,
G3<
26 CF FILS DSA VE. TG TEMPERATURE (C)
.s G
%]
5 54 OATE. h/01/06
' TIME 07,28 O
i
_ y_
7.,
......-.-,i. m.. _....
hY 16 7 95 if$
I
- jj lNEI SAMPLE 9 wr,
- o.4001 mg
- aaTE, 5.00 cog / min p:4_.
LU
[
H li:3,00 Ml:j E
'9' * *
- M Lt.1 BB.82%
Bij 2 I
[i" J
FROM: 31.72 O
TO 599.68 Ei T
WT % CHANGE: 66.58 kh h f
S.78%
h? b i
FROM 31.72 f!i l
TO 379.05
(
{p?
l VT % CHANGE: 13.09
)
h t-
@j I
l FROMs 31.72
[$
0 l
TDs 416.19 59,99 r
w T
vr % CHANGE. 33.15 l
3 l
j 33.35%
9>
i k.l, l
1 j
N I
l l
x r 6 A - (6Z.L N
li i
I 1w 0.00 4
- - - '. 00
- '0. 00 25'0.00 320.00 390.00 450.C0 530.00 500.00 670.00 740.00 18
.-.3 1
40.00 110 il n
'g EF FIL6 SSA E 76 TEMPERATURE CD TG DATE:@07/01/06 TIME:
09:51 Q
[ pgggig_gtygn g74,,,,,,,,gl
-9
rw.w- -
wmn - = _ _
^
g 4a y
E4 g
NEI SAMPLE C r:
- .33.;3 vc
> o::
en:. -. :,
i 0
~
!)
LL.l r
- .~. :.
- 1 a
II I""*
l l
2
$9.96%
05.05%
n p.
LL1 1
2 J
,q y
f FROM:
\\
31.72 TO: 593.13 i
a%
O I
.i i
4 WT % CHANGE: 70.OS h
}---
t FROM: 31.72 34 c"-
d q
}4 TO 385.61 l
a h
WT % CHANGE: 14. I I I.
i M
l s
=4 i
\\
d 7
l FROM 31.72 j
3., _.. _
TO: 410.73 j
' " * ~ "
pg t2; WT % CHANGE: 35.36 3
.\\,-
l' y
1 I
I gj 29.9" Ei 3
~
}f[.-/b i
1 S
d n
h,
..J,.,
._.g...
g..,.,._.
- - -, - ~., - -.
J.
r,.,,
en
- a
~,. : :.. S..
._.....)...)._....).
,,)
iu g
a..,m :.;.., u.-
w M
.g. _..
...:...a A
,,,_n,,,,..,_,,,,_g
.y q,
a j
....._r-
.. - 4
AUACMEN7-f
~
SERVICE REPORT F-1063-
- Sketch of the failed rotor pole coli, Serial Number 17609966 Date of service:
11-It;-86 thru 12-8-86 Service Engineer:
F. Lamcha THIS IS FLYWHEEL END
/i V
i I;..,
1 THIS SIDE LOOKS GOOD'- NO MOVEMENT, NO SEPARATION 1 2 9 " L D o.S E
'N TER LAYER NE WIRE THICK) g CAME LOOSE AND N
g_ \\
,g.
4
\\
LIFTED AS A SHEET s
'N N'(
nX ON Tile SPOT
/
.,, y VISUAL INSPECTION g'
~; rec SUGGESTS POOR BONDING
)I BETWEEN THESE TWO LAYERS pQ, TWO OUTER LAYERS WERE CUT HERE AND PEELED OFF COtt -LEADS AFTER REMOVING THE POLE FROM THE ROTOR IN PLACE
. THIS IS OUTBOARD END i
WITHlSLIP RINGS-Note:
Bottom washer was _ removed in G.E. Repair Shop, inspection from the bottom indicates voids and spaces all around, more noticeable on the ends.
Two outer layers were cut and peeled off intact as sheets.. Wires' did not separate.
L 9
sM m
t-m g-m te-
=*1 4
5 g
^'!
Tip!,
1 hg!h k:%
]g RGQ)Q[jffjN'
/
}M(0Qg%I i
e1 5
dl{h E
DQQ)&'h0$$;14kjsl02 l$y 48;.y 1
if c
?)$$ kbb$ fi I
?
$ W 2'Q f
hf w
js pld MdMi$k82) f, A
@ f"s I#
j A
l ' ?!
t
~
$$Y Y.A romnasa?n5 Da kW &
' " &"fEERi% f5"d; %
\\ [ \\,,.;'
l fg.
wjh_3 n,,
R0iiEZ Mj B E E R!CR% % W_i
\\
m_
m dmm
.P 1
- DsETfD$2ft NIi~""G F
V t
l' h Os e
Le}$s$Y;'
f5O y"
n ed i
hh@htye1 -
3 ay "h o$1
- Q "g y
a I{*
=
by 3I L L
l N
Photocyaph No.
- 1. (approximately 1/2 actual size)
General view of the 2nd layer of the wirc-wound rotor pole (received separated s
fiom the 1st layer and from the rest of the rotor pole coil but with individual turns still bonded) showing the beginning of the area of fatture (appearing as a darker spot) undar the i njht side of the picture (towards the center of pole).the letter "C" which becomes wider and wider towards
- y 9
da ENGINEERING LABORATORY RECORD
'"d *. _ _
W 1037 in y
[S" bhsg]M net PEEBLES-ELECTRIC PRODUCTS. INC.
C"
N! EUCLIC AVENUE
- CLEVELANC CHIC 44112
,,, g g, y.
)
q ATAC
- V ~ NT-G j
e=wammawammmmmmmmw=m,
== ~
toers$.
- wru,
')
O f
O 3
h
. = a;;
w q( '
w N,
'g
[,ij ~
A 8
s.fg. af i
NQ97E;=i== M pr.4E E,Ti.ffc-m75,3;,Lg
,ga.,,,,_=__
g, g-----.
.w_
MB
~
~
^
g-_.:.
su p aw 1,qgrg -
4
+sK
- =
p-s'
- i i 1
l I i
O Phoggraph No._1. (approximately - 1/2 actual size)
General view of the 2nd layer of the wire wound rotor from the 1st laver and from the rest of the rotor pole coil but with indi id pole (received separated turns still bonded) showing the beginning of-the area of failure (ap v ual a darker spot) the right side of the picture (towards the center of pole).under the lett "C" which becomes wider and wider towards peanng as L
" ENGINEERING L ABOR ATORY RECORD U""
'"d "9
, e.,c, w ton C#
NEl PEEBLES-ELECTRIC PRODUCTS,INC,
, n"" a '
- 7
. m m m m a. c a m e o c A~ACcf N T G E 5Fge m E3am sm aE N N E n m = 55*= " " "
&$y g, y p$y y d g g j Q' I*
i
/
)hhh$kffff $v4gpAMwq.g' hG{hjkhYlklypy(g ' h 1
f h4f g
y O
- e. g ? e ; h h w kj$h e*"#e$d 8) 1 4'bfll,$y& ~
Q 6
yg # p N
g7 My],?l!?
f:
E h i
t n t.~
p tb$
Y, a
lIfg$
k Y$
y l
Q, q 6%gM}a.y' t#We k,-
R@
'y n
w Y
MR$ $y %F,A He'e i,
1
(,c MMk j$
E e
c:
l ltd p",-
7 h
a i%.J Photograph No. 2. (approximately S times actual size)
Macnified view of a 1 x 1.5 in, area located under the numeral "2" in Photo-graph No.
2nd layer and the inside of the 1st 'ayer (removed).I showing a nearly Note the " smooth" appearance of the fibrous wire covering where the resin was diluted.
note " pinhead" formation in the areas adjacent to poorly bonded contact im-Also prints resulting from driving out volatiles during the cure.
background typified by Photograph No.
- 3. typical of approximately 30% of the total area of failure interspersed in the This view is
"# C '
ENGINEERING LABORATORY RECORD n
.:,.e w lon 9"}
NEl PEEDLES-ELECTRIC PRODUCTS,INC.
Ahma7 17045 EUCLID AVENt.'E + CL EVEL AND OHIO m '2 Cl,'* f _3h fgg g,
q.y. y 2 g y ATTACiN DIT G mgsmanassmWMFewmmmm""""*"~"
i i
i.
fj g,
,,l k d
y
]I;,]l4 hy ig'. f.
h i hkhhi M$h],4.Re;.p,dWhMd.M[kM(c' g
I iY
~,
k<dJn g, "
- e. ;
, AU., L 4 ( ) m ' j ' M$;. --:;
~
gh I
-i
..,8 ;
~
s, l?
[ "
s
.A.g
+ M W g [w,.,.f.1fi} (sp@,V' p!Jyl... u ;-,..
,. 1.._.' l.2?kl.',j,. t ;.
. ;," ', "',.9,.,,,.f,(
+
h
.y m.
g1 b
I,{d f<
0A Y'
}
'.n.' pj l ',.
_ l ' k,,
Nh Ifl.
.. E hNk hhhhbbikf
/.'f b
Y Qgpc.g@htgdWM mjf.g$
<?
b fh
'M 2
',[
b, 'dWfh N.'
I f
h f
/kk
? h:Y l
o
,h Y
P,,hotograph No. 3 (approximately 5 times actual size) i Magnified view of a 1 x 1.5 in, area located under the letters "AY" in Photo-graph No. I showing a nearly total lack of bond between the outside of the
.nd layer and t' e inside of the 1st layer (removed) esen though a relatively uniform contact ver a wide area was established.
Note the " smooth" appear-l l
ance of the fib, as wire covering and " scabbed over" contact. areas which g
did not bond.
Note also the same " pinhead" formation in the vicinity of con tact imp,ints as observed in Photograph No. 2.
- This view is typical of the l
gene,al appearance of the area of failure and forms a background in which spots depicted by Photograph No. 2 are randomly distributed.
ance of the failure area changes into that typified by Photograph No. 4 onCor.eral appear -
both ends of the pole, I
t ss.ac' ENGINEERING LABORATORY RECORD l s n s'-
17609966 and 964 i
~
~
- b. 1037 g
~
ww.
g NEl PEEDLES-ELECTRIC PRODUCTS. INC.
Cri JVP ljl 1 av 1 26 to 1.'045 E /CL 'O M. E Ni.;E
- Ci t. E L A NC CW 4 41 '.*
30 1987 m
L-
,,g 7. y,
3 g q ATTACHMENT-G emessmemamame===r--===
f y;
N'.f i,y % b. '
4*l W#.: f**
. Jy&}wp,g; : % %p
& & WWu&,#W.
A e
O r; <
1g
> +. +
- v y.
? :: s y n.
A..,,
v, % pl.
)f )h. '
{~'%< i; ;NYG.fi.g, qr - -_ j~;,?.E %
.., ; J a ".*
t
- .v
.y.;
l. ~.3 ;g e ), u:f}-( m },j,q..m. m h;, 3. w.'. %.4'.,w,.
j f 5 l;.
- fD(
[.t m' i ;
=N W
a
.; n
,..,3 4%
m[.3.,:.
.:/; f, o.
- r 2 ;,,;3.,.
i.
f U% 0 :;. 'l; % g:. :, ;,q..~..
u-
- ,>ll
~
- y. -
.g :
b
)
v, + 4, ?
.j.
.; i' y.p:.. a.q,q.y y D 3y..
p
.,...,,y,.
j-3 j
_*t l'-
q1 3.. Is
- v..<
u.
1
- y..
a
.a
?r _
7.
i ~
- .. s
.s-1 m.
- S.
k[A.,,, %g f;.
,.x
'm.s M,i.Wh?T @N.STM # # :*
7 fih}%l l ;:g'g.l;M.[k% ' g M.
.5c ~~
(
[
M l
Jl
~
..1 $., ;
. + q.. & n }1 %w!..y f.fyg.c. j,f,,;,::_.Q'-p /[5
)
9 g
g je w
s
. V ?.$ls.t, '; r st.V.,2 f :c.lu..' N.l#.. e.... #@
M4 y,;.
l
\\
l' I'
Photograph ho. 4 (approximately 5 tin es the actuM size) l t
i Magnified view of a 1 x 1.5 in, area located outsid.. the field'of Photograph No. 1 showing the typical appearWe of t.uther enc of the area of failure where l
a more normal bond begins to betoru evident.
Tlwre are st.tl some signs of-I r e s i<) ejilution to the lef t but " brit tic '"ac ture" of-the resin layer is becoming nota.cabic to the right even though 'he bonding still !acks uniformity.
Note the presence of " crate lines" whert. a relatively goed bond was. broken while the 1st layer was being " peeled off" from the coil-Also note a few " pinheads" stil! evident next to the void zones indicating some presence of a sol.
l it before and during the cure.
.,_,,_ g s 'ja'( '
17609M and 969 l ENGINEERING LABORATORY RECORD W 1037 0
.g w-*.-.-.
NEl PEEOLES-ELECTRIC PRODUCTS,INC.
8,,
C P.1 JVP ]' 1 26 to I
r 30 1987
NSEUC:.
A.ENut
- CL E'. EL A NC c-c.na y
u
._.L...
A-.<n O
L o,.
[ggG&gjgMbhhSbW55
cb d W 2)
.9 8 WN 5
7, y 7-m u
k m
s gn-q qg rs.
O o, <3j,'
q e
tr c, t4 u 4
y x m es en Js.
v > h i o
E n4 L L L, 0 1
3 8d M
&s yi m s O
a ey a
X h G
u t
A l y
_en x
m O
L;-- g{l' s
F-d 9
U E o
i tl 5
i 6 %u!
O H 0
W Oe
.b Q
z l
gg o
g g
o nh G, O..l O D e
O u<
o m
a Et i
W N N
N Ut l
3 Yi N M M
M w
m N
c D
Y M
y ye_
~
9,?ii (ojo ;
(
t m
7
. 4
?)
e w
=p t
WN e
e c-
+
s S
b N K
n N d
(t a
[
O 3 h 0
m t,C Q
}.7 S. i b e
>o c e
x u
l
~
w w 2<
u 4
Si 7 3' 6g g
_.1 3
L
,N
.a
= r y
- c..a y
b 5
3 y
e gg r{ e 4
4 x
'N s
--Q $
m y
A
'r.
Q g
r.
4 as 'e.
M i
E n}-y y ph
?
7 e
sV t
1
~;
e S
w s U
?
f Q
m y
e 3 Cf 4,
6 j
u qw _
m 4
u r.
p
.y
- s. h -
U C
s s
iu ts.
- i V
N
'k'Q+,Qt N
e o
g g'
f.1 5
I
-;6- -
i Q
D k
A v$', e% g e i
l :
N N A
ss %
_p% N A
k A(hr-q' m
g t_w n: %
+
?
[ di![ T S i dE4
%2E
'M b y
d}
1 2% %
ie o.
m ~-p g
w e
b h
l bI QvQ
%q ~4
-M s.:w y# =
.t, a
s Q R u
x Q
v f
. k s
k bA N
i g{'s:4 (q w g H%g
,d i a
4 w.
N g
ft 4
d N
6: _hM c' S D E c
't A
k O
9 1 ;.
(J ; <e o g&;
@% 2 N
c:
w b
t; t,
n e
w %i
~
b
%w h
u o
g
?
J:
I e q}
h k N e
t W
w_
-q' N.
'J N.
g%
-J
!.i Q, g' g.
4 p-J h
F C
y 6;
Q q %
g a
A 0
l0
%Y d N 2 4
%f 4 4 S
> %d w
n h,..
c.
v ro e
,q Esl@sEl5EI$f35 M IE M S E E N N ' E
.