United States Nuclear Regulatory Commission - Protecting People and the Environment

Morning Report for March 31, 2008

Headquarters Daily Report

To top of page

REGION III

Fermi - Failure of Barton Differential Pressure Switch

Licensee/Facility:

DETROIT EDISON CO.
Fermi 2
NEWPORT, Michigan
Dockets: 05000341
[2] GE-4
License No:
Notification:

MR Number: H-2008-0003
Date: 03/13/2008

Discussion:

On Thursday, March 13, 2008, the Fermi licensee reported that the differential pressure (dP) switches, Barton model 581A-2, that operate to open the suppression chamber to reactor building (torus-RB) vacuum breaker isolation valves would not perform their design function (EN 44062). The condition affected both trains of these valves. The licensee opened one isolation valve to maintain the function of the vacuum breaker in that line and took actions to correct the deficiency within the 72-hour limiting condition of operation.

The dP switch is designed to actuate the vacuum breaker isolation valve when pressure in the reactor building exceeds pressure in the torus, creating a relative vacuum in the torus. The torus is normally at a higher pressure. Opening the vacuum breaker isolation valve equalizes pressure between the reactor building and the torus to prevent structural damage to the torus. The torus pressure is connected to the low side of the instrument, so the instrument is normally pegged on the low stop. While normal torus-RB dP is between 7 and 15 inches water column (equivalent to 0.3-0.6 psid), the switch has a working dP range of 0-10 inches water column (equivalent to 0-0.4 psid).

In troubleshooting this failure, the licensee determined in bench test that an applied dP of 15 inches of water column, pegging the switch on the low stop (under-ranged condition), for 22 hours caused the switch to lock. After the pressure was relieved, the switch remained frozen and would not respond to any pressure input for about 6 hours. When the applied dP was within the span of 0-10 inches water column, the switch consistently worked as designed.

The NRC contacted the vendor, who is now Cameron Measurement Systems Division. The vendor stated that qualification testing demonstrates functionality for pressure differences within the normal factory-calibrated range but not for pressures outside this range. The vendor tests the instrument in over-range and under-range conditions but only to verify functionality. The vendor's standard production testing consists of applying full safe working pressure (SWP) of 3000 psig to either or both sides of the instrument for a short time (about 1 minute).

The vendor suspects that the "lockup" condition reported for the Fermi instrument was being caused by either (1) long term leakage through the differential pressure unit (DPU) internal over-range protection feature or (2) thermal expansion because of plant heat-up. The instrument was never designed to be tight against this long-term leakage. The thermal expansion results in an internal hydraulic dP. After the over-range condition is cleared, the internal hydraulic dP must decay before the instrument can operate normally. The vendor has not validated these suspicions.

In a previous evaluation of this sticking phenomenon for a customer's Model 580A and 581A instruments, the vendor advised against applications where the instruments would be over-ranged for a significant time.

There has been some operating experience related to this and similar instruments. In 1995, NRC received a series of Part 21 reports on ITT Barton Models 288A and 289A dealing with switch contact chatter. Also in 1995, NRC received a Part 21 report on a high radiation test failure of B-fill (ethylene glycol) bellows fluid for Models No. 200, 227, 288A and 289A. In 1998, NRC received a Part 21 report on ITT Barton differential pressure switch labels, dealing with switch contact chatter of Models 580A, 581A, 583A, and 583B.

Contacts:

Name Office Abbrev Phone No E-Mail
HODGE, VERNON NRR (301) 415-1861 CVH@nrc.gov

REGION III

Part 21 2008-0004 - Palisades - Failures of EDG Snubber Valves

Licensee/Facility:

ENTERGY NUCLEAR OPERATIONS, INC.
Palisades
COVERT, Michigan
Dockets: 05000255
[1] CE
License No:
Notification:

MR Number: H-2008-0004
Date: 03/13/2008

Discussion:

On March 13, 2008, the Palisades licensee submitted a Part 21 report on snubber valves (Part Number 2402466) used in the fuel oil distribution system on their ALCO Model 251F emergency diesel generator (EDG; EN 44059). The purpose of the snubber valves is to dampen pulsations created by the fuel injection system and to serve as the fuel oil pressure boundary. This report discusses two failures of these snubber valves since November 2005. The failed snubber valves were made from AISI E 52100 (also denoted ES52100) material but not all the installed snubber valves were constructed of this material. The licensee determined through destructive metallurgical analysis that incorrect selection of AISI E 52100 material during manufacture had resulted in snubber valves that were improperly through-hardened rather than case-hardened, as desired. This condition may allow the snubber valves to crack under operation, resulting in fuel oil leakage and loss of fuel oil supply to the associated engine cylinder. The licensee is replacing affected diesel fuel injection system snubber valves.

The first failure occurred on November 20, 2005, when a replacement snubber valve associated with EDG 1-2 cylinder 5R cracked axially in 3 separate locations, causing a fuel oil leak. At the time of failure, the EDG had completed post maintenance runs totaling 11 hours and had completed 1.5 hours of a monthly test (LER 50-255/2005-07). After additional analysis of past operability, the licensee concluded that EDG 1-2 could carry design basis loads and that the leaking fuel oil did not present a fire hazard. Therefore, the licensee concluded that the EDG was operable and cancelled LER 2005-07. The EDG vendor performed an engineering analysis that, in agreement with the licensee, concluded that a substantial safety hazard did not exist.

The second failure occurred on February 22, 2007, when another snubber valve, this one associated with cylinder 5L of the same EDG, cracked during a test (LER 50-255/2007-06). This failure occurred after an estimated 135 hours of operation. After discussions with Fairbanks Morse, the manufacturer of this model EDG, the licensee concluded that both failed snubber valves had come from incorrect material lots manufactured prior to 1995. The licensee attributed the presence of the snubber valves in their supply system to inventory practices that did not maintain traceability to purchase orders. The licensee subsequently ensured that all installed snubber valves were from lots manufactured after 1995. In addition, the licensee implemented a receipt inspection technique capable of determining if the snubber valves are made from an incorrect material.

The original assessment by Palisades and Fairbanks Morse concluded that the failed snubber valves had been procured prior to 1995 and that the incorrect material was isolated to Palisades. However in January 2008, while the site was replenishing their stock for snubber valves, Palisades receipt inspectors discovered 5 snubber valves from Indian Point that were possibly made from AISI E 52100 material. The licensee sent the snubber valves to an off-site lab and confirmed through destructive testing of one snubber valve that the material from Indian Point was AISI E 52100. Since the Indian Point stock had never been mixed with Palisades stock, Palisades concluded that use of improper material could be a current issue. The licensee provided a snubber valve to Fairbanks Morse for their analysis but that analysis is not complete.

NRC staff contacted Fairbanks Morse about these snubber valves. Fairbanks Morse stated that this problem is restricted to the snubber valves, construction material (AISI ES52100), and EDG Model ALCO 251F discussed here. Potentially affected domestic nuclear power plants include Diablo Canyon Units 1 and 2, Ginna, Indian Point Units 2 and 3, Palisades, Pilgrim, and Salem Units 1 and 2.

Palisades found that these snubber valves were purchased in 2004. Fairbanks Morse reviewed its shipping and receiving records dated 2004 or later and found none that involved snubber valves constructed of AISI ES52100 alloy. The vendor performs both shipping and receiving tests but not on all valves because material identification requires valve destruction.

No similar recent operating experience involving cracked snubber valves was found.

Contacts:

Name Office Abbrev Phone No E-Mail
HODGE, VERNON NRR (301) 415-1861 CVH@nrc.gov
Page Last Reviewed/Updated Thursday, March 29, 2012