ML20149H821

Final Deficiency Rept 457/87-01 Re Diesel Engine Generator Rocker Arm Failure.Initially Reported on 870217.Util Took Steps to Determine Strength of Rocker Arms.Low Strength Rocker Arms Replaced W/New Rocker Arms

ML20149H821
Person / Time
Site:	Braidwood
Issue date:	02/11/1988
From:	Hunsader S COMMONWEALTH EDISON CO.
To:	Davis A NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
References
4204K, 457-87-01, 457-87-1, NUDOCS 8802220124
Download: ML20149H821 (6)

Text

,

-. j' N - Commonwealth Edison I ;; / One First National P!aza. CNeago, Illinois of i v (j) Address Reply to. Post CNeago,liiinois Offc3 Box 767 60690 0767

't l

February 11, 1988 I

lI I

I Mr. A. Bert Davis

. Regional Administrator j U.S. Nuclear Regulatory Commission Region III 799 Roosevelt Road Glen Ellyn, IL 60137 I

Subject:

Braidwood Station Unit 2- l 10 CFR 50.55(e) Interim Report l

Diesel Engine-Generator Rocker l Arm Failure NRC Docket No. 50-457 References (a): S.C. Hunsader letter to A.B. Davis j dated March 17, 1987. i (b): S.C. Hunsader letter to A.B. Davis dated April 2, 1987.

(c): S.C. Hunsader letter to A.B. Davis dated May 15, 198'l.

I l

l 1

Dear Mr. Davis:

I On February 17, 1987, Commonwealth Edison Company notified your office of a potential _ deficiency reportable pursuant to 10 For CFRtracking 50.55(e)purposes, regarding the subject failure at our Braidwood Station. Reference (a) provided informa-this deficiency was assigned Number 457/87-01.

tion concerning this matter to fulfill the thirty-day reporting requirement.

Reference (b) provided an interim report regarding an additional rocker arm failure.

Reference (c) provided an interim report regarding our investigation Additionally,. Reference (c) addressed to date on the rocker arm failures. i independent action that commonwealth Edison Company was taking to improve our l confidence in the reliability of the diesel generators. l I

q g33 e

~ ~~_

,,, g N

%1 '

8802220124 880211 PDR ADOCK 05000457 '

, g S DCD

\

_e

% ../ i mn

A.B. Davis February 11, 1988 The purpose of this letter is to provide the final status of the actions that have been taken and our overall conclusions. These are provided in Attachment A. This letter is considered to be the final report.

Should you or your staff have any questions regarding this matter, please contact this office.

Very truly yours,

.4 S. C. Huner. der

c. %s Nuclear Licensing Administrator 1m Attachment cc: NRC Resident Inspector - Braidwood S. Sands - NRR US NRC Document Control Desk 4204K

' ATTACHMENT A Braidwood Station experienced three separate incidents of rocker arm failures on its Unit 2 Emergency Diesel Generators during February and March, 1987. A total of eight rocker arms fractured. Those failures were originally reported pursuant to 10CFR50.55(e) on February 17, 1987 as our item 457/87001.

This letter serves as our final report on this matter. It describes more fully the details of each of the rocker arm failures, and the results and conclusions of both Edison's and the diesel manufacturer's investigations into the root causes of those failures. It further describes the corrective measures Edison took to prevent recurrences of such failures, and our basis for confidence in the continued operational availability of these diesels.

DESCRIPTION OF FAILURES The first failure of rocker arms at Braidwood occurred on February 4, 1987, on the 2A Diesel. Both the intake and exhaust rocker arms on the IR ,

and 3R cylinders failed during the 2A Diesel's 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> pre-operational l testing run. Disassembly and inspection of the valve train found crossheads seized in their guides in the cylinder head. Crossheads are ,

l cam followers positioned ';etween the cam shaft and the pushrod. The seizure of the crosshead within its guide causes impact loadings of the i valve train due to either cam impact on the crosshead or the impact of the {

opened valve with the piston crown. It is believed that this impact l loading of the valve train contributed to this first case of rocker arm l failure. l l

I As a corrective measure, all crosshead-to-crcifiead-guide clearances were checked to assure free crosshead movement. In cases where crossheads were !

found sticking, the crosshead guide bores were honed and the crosshead to l guide clearances were increased. l In spite of these corrective measures, and the completion of corrective measures based on the lessons learned from previous rocker arm failures at Byron, both the intake and the exhaust rocker arms on the SR cylinder of the 2A Diesel failed on February 13, 1987. A check of the crosshead  !

clearances with the Diesel at operating temperature found that all l crossheads moved freely within their guides. No immediate cause was j l

established for these rocker arm failures.

Similarly, on March 28, 1987, the intake and exhaust rocker arms on the SR cylinder of the 2B Diesel failed. Again, a hot check of crosshead clearances found no signs of seizure or distress. No direct cause could be established for these failures.

j

' MATERIAL ANALYSIS Edison performed chemical and physical analyses on each of the failed rockers to determine whether material properties might have contributed to these failures. The rockers are of gray cast iron with a manuafacturer-specified tensile strength of 40 ksi. Test results of the failed rocker arms revealed material tensile strengths that were substantially below the manufacturer's specificacions. Each of the failed exhaust rocker arms had a tensile strength below 18 ksi, with two exhibiting strengths as low as 13 ksi. Each of the failed intake rockers had strengths above 20 ksi.

However, material strength was not considered as a cause of the intake rocker arm failure, as described below.

CAUSES OF FAILURE 6 Based on a physical examination of the Diesel immediately following rocker arm failures, as well as the results of physical and chemical testing of ,

~

each failed rocker arm, Edison concluded that the primary common cause for exhaust rocker arm failure at Braidwood was understrength rocker arm material. In the first case of rocker arm breakage, a contributing factor  !

may have been the seized crossheads. In the latter two cases, no other contributing factors could be identified conclusively by either Edison or the Diesel manufacturer's representatives.

The failure of the intake rocker arms was believed to be a direct result of failure of the exhaust rocker arm in the same cylinder. With the failuure of the exhaust rocker arm, the exhaust valve remains closed.

Puel oil continues to be pumped into the cylinder and, with no exhaust pathway, accumulates in the cylinder with only minimal initial leakage past the piston rings. The impact of the opening intake valve with the fuel oil on the piston at top dead center of the exhaust stroke causes severe loadings on the intake rocker arm, leading to rocker arm failure.

In each case at Braidwood, where an exhaust rocker arm failed, the corresponding intake rocker arm also failed.

The Diesel manufacturer, Cooper Bessemer, cited seizure of crossheads as the cause for the first case of rocker arm failure. However, he could provide no conclusive cause for the failure in the second and third cases. Cooper Bessemer does not concur with Edison's conclusions that below specification tensile strengths contributed to rocker arm failure.

Their static stress calculations and in place static strain measurements show stresses far below even the lowest actual measured strengths of the rocker arms that failed. However, dynsmic strain measurements were not performed. They offer no other plausible reasons for the rocker failures.

Therefore, as a conservative measure, Edison went forward in its decision to replace low strength rocker arms.

t CORRECTIVE MEASURES With Edison's determination that low strength rocker arts were a primary contributor to rocker arm failure, Edison took steps to determine the strength of each rocker arm in Braidwood's four Diesels, and replace all those with strengths below 30 ksi. 30 ksi was selected as a conservative lower strength limit. Although the manufacturer had originally specified 40 ksi for rocker arm material strength, no exhaust rocker arm failures were observed with strengths above 20 ksi. Furthermore, the manufacturer, based on his assembly stress calculations, assures that 20 ski tensile strength is more than sufficient.

Brinell hardness of the rocker arms was selected as a non-destructive indicator of material strength, due to the reasonable correlation between Brinell hardness and material strength for gray cast iron. This correlation was confitned by comparison of destructive tensile strength testing of failed rocker arms from Byron and Braidwood with 3000 kg Brinell hardness measurements.

The correlation between material hardness and tensile strength for cast iron was utilized to non-destructively identify those rocker arms that exhibited tensile strengths below 30 ksi. A two step testing process was employed. Initial in-place hardness measurements were taken with a portable hardness tester. This instrument was conservatively calibrated to identify potentially low strength rocker arms for removal and  ;

subsequently more accurate testing. Such rocker arms identified as being  ;

potentially of low strength were removed from the Diesel Generator and tested on a 3000 kg Brinell hardness tester at our Station Materia) -

- Analysis Department.

These more accurate hardness measurements were used in conjunction with  ;

the known hardness - tensile strength correlation to determine the rocker arm tensile strength. Those rockers with tensile strengths greater than 30 ksi ware reinstalled in the Diesels or were stored as spares. Those found to be less than 30 ksi were tagged as being non-conforming and placed in a hold area in the warehouse. Those low strength rocker arms were replaced with new rocker arms from the manufacturer. All such new rocker arms were confirmed by 3000 kg hardness testing to be of material strength greater than 30 ksi prior to installation. All such testing and rocker arm replacement, as necessary, has been completed for all eight Diesel Generators at Byron and Braidwood. The Diesel manufacturers  ;

service representative was present at Braidwood to witness failure investigation, implementation of corrective measures, and replacement of rocker arms.

d

' _4-SAFETY IMPLICATIONS In previous discussions of the safety implications, the failure of rocker arms in one cylinder was thought not to cause overall Diesel Generator failure due to sufficient reserve power margins in these engines. A question of lubricating oil dilution by fuel oil contributing to bearing failure was taised. Experiences at Byron have shown that the diesel continued to run for at least six hours following rocker arm failure without any apparent bearing damage. It has been observed that dilution by fuel oil is partially offset by evaporation at elevated oil sump temperatures. However, no further investigation into the safety implications of breaking rocker arms has been performed, as the problems contributing to rocker arm breakage have now been adequately addressed.

CONCLUSIONS In conclusion, Edison believes that its emergency Diesel Generators are fully operational. Effective corrective actions have been taken for each potential cause of rocker arm failure postulated by both Edison and the Diesel manufacturer. Corrective measures resulting from lessons learned from previous rocker arm failures at Byron were also fully applied at Braidwood. Specifically, all valve train components were carefully inspected for distress and appropriately adjusted for valve tappet clearances. All parts that appeared deformed or distressed were replaced. The free operation of each crosshead was checked while cach Diesel Generator was in its hot standby condition. Where a crosshead was found to be binding, its bore was honed to allow for free movement. All crossheads were determined to move freely while the Diesel Generators were at operating temperature. These foregoing actions were witnessed by the manufacturers representative.

Moreover, as an additional conservative measure, Edison chose to replace all rocker arms with tensile strengths of less than 30 ksi. Tensile strengths were estimated by correlation to Brinell Hardness measurements taken on each rocker arm.

All of these corrective measures have been completed. To add confidence that each diesel will continue to function reliably, 100 hour0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> reliability test runs were performed on the Unit 2 Diesel Generators. No subsequent rocker arm failures occurred during these reliability runs, or have since occurred. In summary, any and all postulated contributors to rocker arm failure have been appropriately addressed. On this basis, both Edison and the Diesel manufacturer fully expect continued diesel operation to be free from further rocker arm failures.

I l

4204K