Part 21 Report - 1995-020

ACCESSION #: 9411070090

Coltec Industries

Fairbanks Morse Engine Division 701 Lawton Avenue Beloit, Wisconsin 53511-5492 Telephone 608/364-4411 Telecopier 608/364-0382

1 June 1994

Document Control Desk
United States Nuclear Regulatory Commission
Washington, D.C. 20555

To whom it may concern:

Subject: Final Report; Air Start Distributor Cam, part number 16104412 - 10CFR21 Notification 94-002A.

Coltec Industries, Fairbanks Morse Engine Division has completed the technical evaluation into the root cause, and corrective action for the subject part number. The attached Engineering Report, R5.15-0571, has been transmitted to the following affected nuclear power stations: Millstone 1&2, ANO, Prairie-Island 1&2, Vermont Yankee, TMI 1&2, Peach Bottom 1&2, Calvert Cliffs 1 & 2, Crystal River 3, E.I. Hatch Plant, Duane Arnold Energy Center 1, North Anna 1&2, Farley 1&2, Fermi 2, and Limerick 1&2.

If additional information is needed, please contact me at 608-364-8315.

Sincerely, Scott A. Fratianne Quality Assurance Manager Supplier and Systems Quality

(Note: The date of this engineering report is incorrect. According to Jim Golding, QA manager at Coltec, the date should be that of the Part 21 report, June 1, 1994.)

A division of Coltec Industries Inc

 

---

Coltec Industries

	SHEET	PAGE
ENGINEERING REPORT	1 of 4	NO. 1
	FILE NUMBER
	R5.15-0571
Fairbanks Morse Engine Division	DATE
	October 19, 1994
SUBJECT	PREPARED BY
8-1/8 x 10 O.P. AIR START DISTRIBUTOR CAM	D. Beck
REPORT TITLE	APPROVED BY
EVALUATION OF THREE CRACKED CAMS

BACKGROUND

Recently three air start distributor cams at nuclear plants have cracked; two at North Anna and one at Arkansas Nuclear One. The two cracked cams at North Anna were discovered first and resulted in a part 21 notification. As a result of this notification, all such cams at nuclear plants were inspected and a third cracked cam was discovered at Arkansas Nuclear One. The third cracked cam did not follow the scenario described in the part 21 notification and Entergy (the operator of Arkansas One) has asked if there are additional part 21 concerns as a result of this different type of failure.

NORTH ANNA CAMS

NES Materials Engineering Laboratory analyzed the failure of those two cams. They found that the cracks were flat, brittle-like, and intergranular and passed through stress-raisers. In one case, the stress-raiser was a shrinkage crack associated with electric arc marks while in the other case, the stress-raiser was a threaded hole. They inferred that the high stress intensity associated with these stress-raisers exceeded the threshold for stress corrosion and caused the cracks to initiate and/or propagate. The growth of the cracks continued until the remaining cross-section ruptured due to excessive stress. These failed cams were found as a result of normal inspections. In both cases, the engines were starting and running normally although those cams were cracked.

FM briefly examined these came. Two deviations from current manufacture were noted. First, the drawing calls for a machined timing mark and stamped letters and not the use of an electric arc. Second drawing specified C1117, a case hardening material, but a thorough hardening material was used.

ARKANSAS NUCLEAR ONE CAM

In contrast to the North Anna cams, the cracked cam from Arkansas Nuclear One is of typical FM construction. A case-hardened material was used, and all marks were stamped. A hardness transverse radially across the cam shows that the cam perimeter, the tapered inner diameter, and the threaded surfaces of the puller holes are all case hardened. The crack passed radially through a puller hole which is the minimum section. Most likely, the crack initiated at the stress concentration caused by the puller hole. In support of this idea, smaller radial cracks were found in the unbroken side extending from the threaded hole toward the I.D. The cracks were found as a result of the part 21 inspection and not because the engine failed to start.

 

---

Coltec Industries

	SHEET	PAGE
ENGINEERING REPORT	2 of 4	NO. 2
	FILE NUMBER
	R5.15-0571
Fairbanks Morse Engine Division	DATE
	October 19, 1994
SUBJECT	PREPARED BY
8-1/8 x 10 O.P. AIR START DISTRIBUTOR CAM	D. Beck
REPORT TITLE	APPROVED BY
EVALUATION OF THREE CRACKED CAMS

Metallurgic examination of the material shows it contains manganese sulfide stringers and areas with alloy segregation. These two features did not contribute to the failure since they are oriented perpendicular to tho crack surface. In the case hardened areas cementite is present on the grain boundaries and the cracks follow these boundaries. The brittle cementite is more susceptible to crack initiation and growth than the ductile core material and probably contributed to the failure. The case hardened layers were found on the cam O.D. and I.D. as expected and on the I.D. of the puller hole which was not expected. As a result there was little core material between the puller hole and the cam I.D.

The operating stresses are small. The contact stress in the cam due to the 50 lb. load on a follower is about 2500 psi. Other operating stresses are smaller. On the other hand, the assembly stresses are significant. The cam in mounted on a tapered shaft and the necessary axial force is applied by torquing a 1" nut. An a result, significant hoop stresses are generated during assembly. The highest stressed section is the radial plane through the 2 puller holes which is where the cracks were found. Calculations show that a 10 ft-lb torque on the nut produces an average stress of 4200 psi across this section and a peak stress of 6400 psi at the stress concentration next to the hole.

Before 1994, the installation and operating instructions did not specify the required torque for the cam hold down nut. It was expected that the mechanics judgment would suffice. It might be expected that the nut would be torqued to 190 ft-lbs which in the FM standard torque for a 1", SAE-grade-2 nut. The resulting peak stress is 120,000 psi at the puller hole. The case hardened material at the puller hole is more susceptible to brittle type failures such as stress corrosion than the core. It is possible that as in the North Anna cams, the stress intensity threshold for stress corrosion was exceeded and a crack propagated causing failure.

REPLACEMENT HISTORY OF CAMS

Discussions with field service and the O.P. Product Engineer indicate that those came rarely fail. As repair parts, we sold the following numbers of cams:

Year	Number
1990	None
1991	None
1992	1
1993	6 of which 4 were to nuclear plants
1994	31 all to nuclear plants

 

---

Coltec Industries

	SHEET	PAGE
ENGINEERING REPORT	3 of 4	NO. 3
	FILE NUMBER
	R5.15-0571
Fairbanks Morse Engine Division	DATE
	October 19, 1994
SUBJECT	PREPARED BY
8-1/8 x 10 O.P. AIR START DISTRIBUTOR CAM	D. Beck
REPORT TITLE	APPROVED BY
EVALUATION OF THREE CRACKED CAMS

Of those sold to nuclear plants, only 3 were to replace failed cams. Those are the three discussed in this report. Given the large number of engines in the field, it is evident that the air start distributor cam is a reliable part.

EVALUATION OF THE FAILURES

In all three failures, the fracture surfaces are brittle-like, the cracks progressed gradually and fatigue is ruled out due to the low operating stress. This makes stress corrosion the most likely mechanism for crack propagation. We do not know the chemical causing the stress corrosion and in fact the environment around the cam is oily and relatively benign. This leads to the suspicion that the assembly stress was unusually high because of high torque on the cam nut. Another common thread through the three failures is the use of a very high strength material in an area of stress concentration. Normally, this is a good idea, but can be over done when stress corrosion is a possibility. The resistance of steel to stress corrosion cracking decreases with increasing material strength. Thus at some point increasing strength also increases the likelihood of failure. This seems to be the case in these failures.

PRIOR 21 ACTIONS

As a result of the North Anna cam failures, a Part 21 notification was issued. Nuclear plants were instructed:

1.	To dye check all cams
2.	To replace cracked cams and cams with electric arc marks
3.	To lap the cams to the shaft to obtain 70% contact, and
4.	To torque the cam nut to 50-60 ft. lbs.

CAM NUT TORQUE

At the time of the prior Part 21, it was not realized that the cam nut is a prevailing torque nut which on first installation can require up to 90 ft. lbs to turn. Clearly, the current recommendation of 50-60 ft. lbs is not correct. The proper torque specification will adequately clamp the cam to its shaft but not deform the tang of the overspeed governor shaft which engages with the other end of the air-start-cam shaft. Calculations and tests indicate that an adequate clamp load in obtained by a torque of 30 ft. lbs and that the tang can safely withstand a torque of 9.5 ft. lbs. If the torque applied to the cam nut is in the range of 90 to 95 ft. lbs. and the prevailing torque for the nut is in the range of 40 to 60 ft. lbs then the proper clamp load will be applied to the cam. Although new nuts may exceed the acceptable range, their prevailing torque can be reduced by a few cycles of turning the nut on and off the shaft. When doing this, the cam should be removed and the nut turned on past the point of contact with the cam.

 

---

Coltec Industries

	SHEET	PAGE
ENGINEERING REPORT	4 of 4	NO. 1
	FILE NUMBER
	R5.15-0571
Fairbanks Morse Engine Division	DATE
	October 19, 1994
SUBJECT	PREPARED BY
8-1/8 x 10 O.P. AIR START DISTRIBUTOR CAM	D. Beck
REPORT TITLE	APPROVED BY
EVALUATION OF THREE CRACKED CAMS

RECOMMENDED ACTIONS

1.	Replace those cams with marks made using an electric arc.
2.	Replace those cams made with through hardened material or having case hardened material around the puller hole. (One way to check for hardness is a scratch test, making a circumferential scratch through a puller hole, with a tool steel scribe. If the scribe glides on the surface, the material is too hard).
3.	Lap the cam to the tapered shaft to obtain a minimum of 70% contact.
4.	Insure that the clamping torque is at least 30 ft. lbs and the prevailing nut torque is at least 40 ft. lbs. This can be done by tightening the cam nut to between 90 the 95 ft. lbs. while using a nut whose prevailing torque is in the range of 40 to 60 ft. lbs., as described above. This can also be done by holding the shaft while tightening the nut to 30 or more ft. lbs. above the prevailing nut torque. All existing nuclear installations should check the clamp and prevailing nut torques.
5.	Revise the drawing so that it is clear the puller holes are drilled after the cam is carburized. Since April 18, 1975, the manufacturing process has produced puller holes which are not carburized, but this action will make this requirement explicit.

Recommendations 4 and 5 are new. Therefore, a revision of the Part 21 notification is required.

APPENDIX A

Figure 1, "SB-2 CLUTCHING MECHANISM" omitted.

*** END OF DOCUMENT ***