ML20207H555
| ML20207H555 | |
| Person / Time | |
|---|---|
| Site: | Fermi  |
| Issue date: | 07/16/1986 |
| From: | NRC |
| To: | |
| Shared Package | |
| ML20207H552 | List: |
| References | |
| NUDOCS 8607240363 | |
| Download: ML20207H555 (8) | |
Text
!{
g
+ o g
o a
UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20E55
...../
SAFETY EVALUATION REPORT FERMI-2 EMERGENCY DIESEL GENERATORS DOCKET NO. 50-341
1.0 INTRODUCTION
In January 1985, and during November and December 1985 Fermi-? experienced a series of emergency diesel generator (EDG) failures. Specifically, the EDGs exhibited an abnormal number of distressed or failed bearings which resulted in extensive engine damage. The damaae included failed piston liners and two crankshafts.
By letter dated March 18, 1986, which superseded, in its entirety, a previous February 5,1986, letter by the licensee on this matter, Detroit Edison described its program to demonstrate the reliability of the Fermi-? EDGs through various engine tests. The licensee's submittal also included a summary of the results of those tests; a discussion of the possible causes of recent engine failures; and requirements for future inspections. In addition to the information sup-plied by the licensee, our evaluation considered information regarding the EDG's gained during site inspections, including visits to other licensees with similar cngines; meetings with the licensee and their consultants on January ?4 and February 14, 1986; follow-up telephone conferences; and information provided by our technical consultants.
2.0 RACKGROUND The standby AC power system for the Fermi-? facility consists of four EOGs iden-tified as Nos. 11 and 12 which provide power to Division I, and Nos. 13 and 14 which provide power to Division II. The EDGs are manufactured by the Fairbanks-Morse Engine Division of Colt Industries (Model 3808 1/8b and are 12-cylinder, opposed-piston, turbocharged engines. These EDGs are rated for continuous duty at 3967 horsepower.
Each EDG is equipped with a lube oil keep-warm system, a prelube system, and a lube oil booster for the upper crankline. In the keep-warm systen, an electric heater maintains the lube oil at about 130 F to achieve.the desired viscosity.
The keep-warm circulating pumo sends the warmed lube oil through the lower crankline, maintaining lubrication of bearings in the lower portion of the engine and back to the oil sump. Due to the possibility of getting lube oil into the firing chambers of these opposed-piston engines, the keep-warm system is designed to oot circulate lube oil to the upper crankline. ,
In the prelube system, a separate oil pump is actuated for manual EDG starts .
(i.e., a non-automatic start) and run for about two minutes prior to the actual starting of the enoine. The prelube system provides lube oil to both the lower and upper cranklines. For all surveillance test starts of the EDGs at Fermi-2, the prelube system is actuated prior to the actual diesel start; i.e., for those simulated engineered safety feature (ESF) actuation / loss-of-power tests, the 8607240363 860716 PDR ADOCK 05000341 P PDR
prelube systent is started manually prior to actually simulating the need for an emergency automatic start. In the event of an emergency start, the diesels are started immediately without the use of the prelube system described above.
The lube oil booster system includes an oil supply in an accumulator which has a piston operated by the engine starting air. This system delivers about 1.25 gal-lons of oil to the upper crankline during the engine start sequence.
As discussed in Supplement No. 5 to the Fermi-2 SER, we believe that the lube oil booster system is inadequate to properly lubricate the upper crankline.
During 1984, most planned starts were conducted relying only on the lube oil booster system; i.e., the manual prelube system was not used. In January 1985, a number of failed main bearings were discovered; we believe these bearing failures were caused crimarily by inadequate prelubrication of the upper crank-line. Subsequently, the licensee reinstituted the use of the prelube system into its procedures. With the advent of the recent EDG main hearina failures in November 1985, it became apparent that there may be other unidentified fac-tors contributing to these main bearing failures.
The Fermi-? EDGs are not a unique design and are used for similiar purposes at other nuclear power plants. In general, Fairbanks-Morse diesel engines are a significant fraction of the total number of the diesel engines at nuclear plant sites, and are used in numerous non-nuclear applications. The licensee has indicated there are 49 nuclear applications of this type of Fairbanks-Forse diesel. These engines have exhibited a long and apparently favorable operat-ing history within various industries. We have investigated the performance of other nuclear power plants which use similar Fairbanks Morse engines and have reviewed various NRC documents (i.e., NUREG/CR-4440, -4347, and -7989) which consider EDG performance and reliability. Our review and investigations did not reveal a significant frequency of main bearing failures for the Fairbanks-Morse type diesel engines. Therefore, in the staff's view, the operating prob-lems of the Femi-2 EDGs are ,iudged to be plant specific.
3.0 RECENT FERMI-2 EDG FAILURES On November 13,1985, EDG No.13 was undergoing its required 18-month surveillance run (i.e., the 24-hour run) in accordance with Fermi-2 Technical Specification 4.8.1.1.2. At about 21 hours2.430556e-4 days <br />0.00583 hours <br />3.472222e-5 weeks <br />7.9905e-6 months <br /> into this run, EDG No. 13 was shut down when excessive noise was heard. Subsequently, inspection of EDG No. 13 revealed severe engine damage which consisted of the following: No. 3 upper connecting rod bearing was completely destroyed; bearing material was completely fragmented and pieces deposited throughout the engine; the lower portion of the No. 3 upper piston skirt and lower piston rings were broken; and the No. 3 cylinder liner was cracked.
The licensee concluded that the No. 3 upper connecting rod bearing failed due ~
to fatigue. 'The licensee deduced that there was movement of the bearing cap relative to the connecting rod, thus suggesting to the licensee that the fa-tigue resulted from an inadequate clamp on the bearing cap. As a consequence,
the licensee stated that an excessive clearance developed which resulted in pounding. The licensee concluded that the continuous pounding, in turn, ex-truded the bearing, thereby further increasing clearances and causing even greater pounding. This "see-saw" sequence continued until the bearing frac-tured and was forced out of the connecting rod.
After examinino the connecting rod bearing fragments, the licensee concluded that the bearing failure was not related to inadequate lubrication; i.e., there were no signs of high temperature effects associated with the failure. The licensee asserts that fretting on the connecting rod and on the cap nut faces indicates relative movement between the connecting rod and cap. Therefore, the licensee concluded that either the cap was not as tight as it should have been, (e.g., possible misassembly during an earlier inspection), or that the failure of the bearing caused the relative movement. The licensee also adds that the root cause could not be determined conclusively from the inspection of the failed components. The staff agrees with the licensee on this latter point and believes that it is unclear whether the root cause of this bearing failure in November 1985, has been found.
Because of the damage that occurred in EDG No. 13, the licensee inspected EDGs Nos. 11, 12, and 14. " Distressed" bearings were found mostly in EDG No.11.
Subsequently, the licensee replaced all upper crankline main bearings in EDG No. 11. The licensee claims that the enoines exhibited less than adequate bearing "run-in" following the February 1985, maintenance outage. During the "
process of returning EDG No. 11 to service, several main bearings failed on the upper crankline. Further investigation by the licensee revealed an unaccep-table upper crankline bearing cap alignment pattern. Subsequently, the licen-see replaced the failed bearings and performed a detailed realignment of the upper crankline bearino caps. Some additional abnormalities were also dis-covered: an oil ring segment was found in EDG No. 12; foaming of the lube oil occurred in EDG No. 11; a small quantity of fuel oil was found in the EDG No. 14 lube oil system; and a fuel cil leak into the lubricating oil system occurred in EDG No. 12.
4.0 LICENSEE'S EVALUATION The licensee, with assistance from the EDG manufacturer (Fairbanks-Morse) and its independent consultant, Failure Analysis Associates (FaAA), has investi-gated the EDG failures. In its letter dated March 18, 1986, the licensee sum-marized its findings and identified the factors which they believe caused the recent EDG failures which have been discussed above; i.e., nossible misassembly, misalignment of bearing caps and foreign material in the lubrication system.
As part of its investigation into other factors which could have contributed to the unreliable performance of the EDGs, the licensee surveyed other owners of Fairbanks , Morse engines to identify the brands of lube oil in common usage; ,
the licensee found that a Mobil brand of lube oil is used predominantly. Con-sequently, the licensee elected to change the oil in all the EDGs from their current brand to Mobil.
I i
Another factor which may have influenced the engine performance is the relatively long on-site storage period of the Fermi-2 EDGs (i.e., about 5 to 7 years). It is unclear how the long-term engine storage period contributed to the Fermi-2 problems with the upper crankline main bearings. Fairbanks-Morse stated that inadequate hearing "run-in" or " seasoning" may have also contributed to the bearing failures. However, the experience at other nuclear plants with similar engines appears to rule out this possible cause.
In addition, the licensee asserts that numerous bearings were replaced in each of the Fermi-2 EDGs due to heightened NRC concerns about the main bearing surface condition. The diesel vendor concluded that replacement of some bearings were not necessary. In the staff's view, there was insufficient information during the inspections conducted in the periods of November to December 1985, and early January 1986, to determine how the condition of the bearing surfaces could be used in predicting imminent hearing failure. Therefore, any question-able condition of a main bearing surface resulted in the replacement of the bearing at the staff's request. Subsequently, on January 24, 1986, the staff and its consultants met with Detroit Edison and their consultants at the Fermi-2 site. As a result of the meeting, the staff concluded that aluminum transfer from a nain bearing to the .iournal would most probably be an indication of a potentially shortened bearing life. Thus, some of the earlier bearing replace-ments were probably not necessary since a number of main bearings were lightly scored and there.was little evidence of a metal transfer to the journal.
Considering the unknown effects of the factors mentioned above and the inherent uncertainties associated with diagnosing the cause of the recent EDG failures, the staff is not convinced that the root cause(s) has been definitively deter-mined by the licensee. The staff recognizes the inherent difficulty in deter-
~ mining the principal contributors to engine bearing failure, or evaluating the potential of a combination of adverse factors. As a result we have depended on
, the outcome of the licensee's Demonstration Test as described in the letter of March 18, 1986.
5.0 EVALUATION OF THE LICENSEE'S CORRECTIVE ACTION PLAN As part of its March 18th letter, the licensee has provided a description of their Crankshaft Bearino Peliability Demonstration Test Program or simply, the Demonstration Test Program. The purpose of the program is to restore NPC con-fidence in the reliability of the Fermi-2 EDGs. Detroit Edison proposed that EDGs Nos. 11 and 13 be selected for the demonstration tests. These selections were based on the unfavorable operating history exhibited by these two engines.
Another consideration was that these EDGs are in separate power divisions. The licensee's test program is intended to simulate the anticipated number of starts, with some margin, that would be expected of an EDG over an 18-month fuel cycle and the run-time which may be needed to assure safe shutdown of the plant if EDG operation is required in the event of a design basis loss-of-coolant accident
~
~
(LOCA).
Initially, Detroit Edison proposed to rely on an inspection technique which checked the gaps of the main bearings to determine bearing integrity. Speci-fically, a gap check would be used to determine the clearance or gap between the bearing shell and the bearing cap on the bearing cap side of the parting line and at the parting lines of the two shells. If the measured gap is 0.002 inch or greater, it may be conservatively concluded that the bearing has experienced severe thermal distortion indicating overheating. On this basis, the licensee would declare the bearing to be failed. The licensee's consultants and the EDG vendor support this method which is an industry-wide practice.
However, the staff does not agree that the gap check method is reliable based on observed failures of the Fermi-2 EDGs. We conclude that the failures and bearing abnormalities observed at Fermi-2 warrant a more effective measure; e.g., the inclusion of a limited visual inspection program.
In discussions with the licensee, the staff expressed its concern regarding the licensee's approach for relyina solely on the gap check method in determining the acceptance of the upper crankline main bearings. Our interpretation of the operating experience at Ferni-2 regarding the EDG problems indicates that the gao check method is unreliable in determinino incipient bearing failure as dis-cussed above. This method only indicates that a main bearing has already experienced a gross failure and cannot be relied upon to detennine if. an other-wise " good" bearing will fail within a short period. Consequently, the staff required the licensee to perform a limited visual examination of the upper crankline main bearings upon completion of the demonstration test run for EDGs Nos. 11 and 13.
We concluded after a number of telephone conferences with the licensee that a limited visual inspection of a sample of the upper crankline main bearings (i.e., EDG No. 11 bearing numbers 3, 5, 6, 8, 9, 13 and EDG No. 13 bearing numbers 3,4,7,10,13) was acceptable. The basis for selecting these bear-ings was predicated primarily on their past failure history. The March 18, 1986, submittal by the licensee reflects our telephone conferences on this matter. The demonstration tests for EDGs Nos.11 and 13 have been completed and the results of these tests are also included in the licensee's submittal of l
March 18, 1986.
Before starting the Demonstration Tests, the licensee completed the following steps on EDGs Nos. 11 and 13:
o a new brand of lube oil was installed, I
l 0 a flush of the engine lube oil system was performed to reduce foreign j material which may have been introduced during maintenance of the engines and, o the EDG's were run for about 100 hours0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> to " season" the new bearings as '
the vendor recommended.
1 1
These EDGs pas 3ed the bearing gap tests. Subsequently, the Demonstration Test was performed consisting of the following elements:
o 20 prelubed " slow" starts; for each slow start, the EDG was run at least two hours under load in compliance with the applicable Tech Specs; o a gap check of the upper crankline main bearings; o 10 prelubed " fast" starts; for each fast start, the EDG was run at least two hours under load in compliance with the applicable Tech Specs; o a cap check of the upper crankline main bearings; o a 7-day continuous run under load; o a gap check on the upper crankline main bearings.
As a result of the bearing run-in period and the Demonstration Test Program, there was a total accumulation of 30 slow starts and 10 fast starts by EDGs Nos. 11 and 13, and about 416 run hours and 504 run hours, respectively. All bearings on both EDGs passed the gap checks. As discussed above, selected upper crankline main bearings were visually inspected. The bearings which were visually inspected were in good condition, except for number 13 bearing on EDG '
No. 13. This bearing had a trace of surface scoring with a very thin hand of aluminum, about i-inch wide, transferred to the journal. ,
The vendor determined that the bearing was in satisfactory condition; our technical consultant, who also inspected the affected bearing, concurred with the vendor's assessment on the basis that the aluminum would be removed from the journal. Accordingly, the licensee removed the thin band of aluminum from the ,iournal prior to reinstalling the bearing.
l Based on the results of the Demonstration Test, Detroit Edison asserts that there is a sufficient basis to conclude that the Fermi-2 EDGs are a reliable on-site power source. Although there was a trace of aluminum on one fournal of EDG No.13's upper crankline, the staff finds that an operability envelope has been satisfactorily established.
l 6.0 SPECIAL EDG BEARING INSPECTIONS The licensee proposed in its letter dated March 18, 1986, that a special visual
( inspection of EDG No. 13, number 13 bearing, be performed at the next scheduled l 18-month Technical Specification surveillance inspection. The Demonstration Test Program established an envelope in which confidence in the reliability of i the Fenni 2 EDGs to supply on-site power was demonstrated. However, there is l no quarantee,that the established envelope of engine starts will not be exceeded ,
! prior to the special visual inspection of the upper main bearing No. 13 on EDG No. 13 as proposed by the licensee. Therefore, our acceptance of the licensee's Demonstration Test Program is based on the licensee performing the special 1
\
l l _
visual inspection of upper main bearing No. 13 at the next scheduled 18-month Technical Specification surveillance requirement or at the time the envelope of engine starts established by the Demonstration Test Program is exceeded, whichever occurs first.
In addition, Detroit Edison proposed to perform a gap check inspection of the upper and lower crankline main bearing every 6 months, or after three non-manually prelubed engine starts, whichever occurs first. All planned EDG engine starts are manually prelubed. However, automatic starts are not. A limited number of automatic starts are expected to occur during plant operation. No
" dry starts" (i.e., starts without prelubel were conducted during the Demon-stration Tests since it was the licensee's and the staff's position that such starts should be minimized.
Accordingly, to address the staff's concern regarding " dry" starts, Detroit Edison added the provision to gap check after three non-manually prelubed starts to its proposed inspection program for the EDGs. However, the staff does not concur with the licensee's basis in proposing three " dry" starts in triggering a gap check inspection.
Detroit Edison' defines a " wet" start as one in which an EDG is' started within one hour of its prior operation, with'out manually prelubricating the EDG. Prior to January 1985, each of the Femi-?. EDGs experienced about 60 fast starts with-out any manual prelube. Considering a 95 percent EDG reliability, to be the acceptance criterion, a maximum of 3 failures in 60 starts would be minimally ,
acceptable. The staff's consultant has indicated that some lubricating oil may exist on the upper crankline for several days and some trace of oil from the previous EDG operation can be ex^pected beyond one hour. When considering a protracted shutdown period longer than several days, the " dry" start sample size is reduced to about 20 starts. (Refer to the March 6, 1985, DECO submittal).
Based on a 95 percent EDG reliability factor, we conclude that there should be a gap check inspection of the hearings after each dry start. Our acceptance of the licensee's Demonstration Test Program is based on this feature of the EDG inspection prooram.
7.0 EDG MAINTENANCE ACTIVITIES In the staff's view, the dominant failure modes as cited by the licensee (i.e.,
system)possible bearing misassembly and foreign material in the lubricationappe been avoided. The staff has been informed that a variety of individuals have worked on the four Fermi-2 EDGs; i.e., DECO personnel, the manufacturer and Bechtel personnel. The specific qualifications of these individuals are unknown. Our concerns on this matter were conveyed to the licensee in our telephone conferences.
~
Accordingly,"the staff recommends that DECO employ care in the future in maintaining the Fermi-2 EDGs during plant operation. The staff further recommends that DECO assess their EDG procedures and maintenance pi;ctices to determine any weaknesses and to assure the previous operating problems will not recur.
i l
To emphasize that good maintenance practices are critical to assure EDG reli-ability, the staff cites a utility sponsored report prepared by the Nuclear Safety Analysis Center for EPRI, NSAC-79, "A Limited Performance Review of Fairbanks-Morse and General Motors Diesel Generators at Nuclear Plant," April 1984. The first three conclusions of this report are:
- 1. A comprehensive preventive maintenance should be performed under a well-documented program;
- 2. A thorough initial training of maintenance personnel should be accomplished and periodic refresher courses should be given by the diesel manufacturer or other qualified organizations;
- 3. There should be good communication between the maintenance crews and the personnel at plants with similar diesel generators; e.g., through INP0.
8.0 CONCLUSION
As a result of the successful completion of the Demonstration Test, we find that there is reasonable assurance that the Fermi-2 EDGs will function as a reliable onsite power source; i.e., an envelope of operability has been estab-lished which is expected to cover the 18-month fuel cycle, including an al-lowance for an extended startup test phase. In addition, the staff concurs with the licensee's proposed special bearing inspections which should provide satisfactory future monitoring, provided the envelope established by the Demon-stration Test Program is not exceeded. Our acceptance of the licensee's Demon-stration Test Program is based on the licensee performing 1) a bearing gap check inspection after each automatic start and 2) the special visual inspection described in its letter dated March 18, 1986, within the envelope of engine starts established by the Demonstration Tests. For the special visual inspection to be considered within the envelope of the Demonstration Tests, it should be conducted after the number of engine starts is comparable to the number of engine starts in the Demonstration Test Program. Lastly, we recommend the licensee assess its EDG operation and maintenance practices to ensure that qualified and well-trained personnel are assigned to maintenance work on the Fermi-2 EDGs.
Principal Contributors: A. Notafrancesco and F. Maura l
+
i .
t L _ _ _ _ _ _ . _ . - _ __ _ . _ . . - - - -