Morning Report for January 30, 2007

Headquarters Daily Report

• N/A

N/A

Failure of Fused Disconnect Switch

Licensee/Facility: BROOKHAVEN NATIONAL LABORATORY Dockets: License No:

Notification: MR Number: H-2007-0002 Date: 01/22/2007

Discussion:

The following information is based on Oak Ridge National Laboratory Lessons Learned Report R-2006-BNL-AD-0001.

On April 14, 2006, at Brookhaven National Laboratory (BNL), an electrical engineer was injured by an arc flash while closing a General Electric (GE) Spectra Series 400 ampere (A) fused-disconnect switch in a 480 volt (V) electrical panel at the Relativistic Heavy Ion Collider (RHIC). The subsequent accident investigation listed the failure of the switch as a possible contributing cause of the arc flash. The switch being operated was destroyed together with an adjacent 400 A switch located in the same panel.

Cracks, broken pieces of the insulating support structure, and dislocated metallic clips were observed, which may have been the cause or the result of the arc flash. These findings prompted an examination of other Spectra series fused disconnect switches throughout the BNL site and life cycle testing of a spare switch identical to the units involved in the arch flash event (480 V and 400 A). The spare was operated approximately 4000 times. Cracks in the plastic support backing were observed that were almost identical to the ones on the failed switches. During this testing, several screws became loose and fell out in the location of the high voltage bus work.

Switches were found to have significant structural defects, including cracked plastic insulating bases (on which the internal switch mechanism is mounted), broken arc chutes, loose mounting screws, and other loose hardware.

The investigation concluded that 32 percent of the switches rated at 400 amperes greater were found to have problems that could lead to an electrical fault. Units rated below 400 amperes were found to have a 0.30 percent problem rate.

In the higher rated switches, there appears to be a serious problem with the screws cracking the plastic material they are fastened to, which allows them to become free within the switch. There appears to be a serious problem with the way the switch jaws are supported by the plastic backing material. Repeated switch operations crack the backing material and/or arc chutes which could eventually get dislodged and permit metal parts to be injected into the high voltage section of the switch.

Preliminary discussion with GE indicates that GE has not dedicated any GE Spectra switches for use in safety related applications in nuclear power plants. Nevertheless, this event and the discovery of degradation of the switches over time with repeated operation prompts a general concern over the inspection and preventive maintenance of fused disconnect switches at operating nuclear power plants.

Contacts:

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