Bulletin 80-17: Supplement No. 2, Failures Revealed by Testing Subsequent to Failure of Control Rods to Insert During a Scram at a BWR

                                                            SSINS No.:  6820
                                                            Accession No.: 
                                                            8006190032

                                UNITED STATES
                        NUCLEAR REGULATORY COMMISSION
                    OFFICE OF INSPECTION AND ENFORCEMENT
                           WASHINGTON, D.C. 20555

                                July 22, 1980

                                                       IE Bulletin No. 80-17
                                                       Supplement No. 2 

FAILURES REVEALED BY TESTING SUBSEQUENT TO FAILURE OF CONTROL RODS TO INSERT
DURING A SCRAM AT A BWR 

Description of Circumstances: 

At about 3:35 a.m. on July 19, 1980, a manual scram test was initiated at 
Dresden Unit No. 3.  This test was initiated in accordance with the 
requirements of IE Bulletin No. 80-17.  The scram itself was accomplished 
(i.e., control rods inserted).  Following the scram, the Scram Discharge 
Volume (SDV) was monitored by UT in accordance with Item #3 of IE Bulletin 
No. 80-17.  The UT check followed the normal draining of the SDV at a time 
when personnel conducting the test believed the SDV to be empty.  However, 
the UT check revealed the scram discharge west header bank to be 80% filled 
with water (i.e., apparently the drain was not successful). 

Upon investigation, it appeared that the SDV vent system did not function 
due to a stuck ball check valve (i.e., a ball check valve functioning as a 
vacuum breaker).  This ball check valve is installed in a tee connection 
downstream of the vent valve in a one-inch vent header which terminates in 
the reactor building equipment drain tank (RBEDT).  The ball check valve 
provides a vent path to the reactor building atmosphere in the event the 
vent header does not. The vent header itself provides a path to other 
interconnections and extends into the RBEDT under the surface of water 
normally contained there. 

Additional information has also resulted from scram testing at other BWRs as
follows: 

1.   At Duane Arnold, the Scram Discharge Instrument Volume (SDIV) drain 
     valve was found installed so that pressure in the SDIV tended to unseat 
     the drain valve disk.  This resulted in leakage out of the SDIV during 
     the scram.  This was corrected by reversing and reinstalling the valve.  
     The scram tests were performed on July 12 and 13 and the drain valve 
     was corrected before return to power operations on July 17, 1980. 

2.   At the Millstone Unit 1, the scram tests were performed successfully on 
     July 11 through 14.  The function of the 10-second delay on scram reset 
     (per Item #2.(j) of IE Bulletin No. 80-17) was tested separately from 
     the scram tests.  Review of the separate test results by plant 
     personnel established that the scram reset delay feature was not 
     functioning in the scram circuits due to a wiring error on the circuit 
     boards.  This was corrected. 
.

IE Bulletin No. 80-17                                       July 22, 1980 
Supplement No. 2                                            Page 2 of 2 

3.   At Browns Ferry Unit No. 1, a test scram involving two rods was 
     performed on July 19, 1980.  The test showed normal response of level 
     switches in the SDIV.  When proceeding to drain the SDIV, however, the 
     SDV did not empty as required and expected.  A vacuum in the SDV 
     apparently existed which kept the system from draining.  Subsequently, 
     the vacuum was cleared by operator actions and the volume drained 
     properly.  Tests are continuing toward determination of the cause and 
     to measure the vacuum. 

4.   At Nine Mile Point Unit No. 1, one rod failed to scram during the 
     manual scram test on July 14, 1980.  This was due to a failure of the 
     scram pilot valve for that rod. 

In view of the above-described events, the following actions in addition to 
those specified in IE Bulletin No. 80-17, including Supplement No. 1, are to
be taken by BWR licensees. 

1.   Each BWR licensee with a SDV vent system that depends on any component 
     other than the vent valve alone for proper venting must provide an 
     alternate vent path continuously open to building atmosphere on the 
     side of the vent valve piping away from the SDV.  This alternate vent 
     path must be positive in its function at all times (i.e., a vent must 
     be effective regardless of component operability other than the vent 
     valve). 

     Plants have made corrections by providing a vent through a standpipe 
     open to building atmosphere.  This was done at Browns Ferry Unit No. 3 
     and is being done at Dresden Units 2 and 3.  It is noted that due 
     consideration to radiological aspects should be included. 

     An alternate vent path must be provided within 48 hours following the 
     telephone notification on July 19, 1980 in order to continue or 
     commence operations. 

2.   Report in writing within 5 days of the date of this Bulletin Supplement 
     the confirmation of your action in response to the above.  Reports 
     shall be submitted to the Director of the appropriate NRC Regional 
     Office and a copy shall be forwarded to the NRC Office of Inspection 
     and Enforcement, Division of Reactor Operations Inspection, Washington, 
     D.C. 20555. 

More information on the details of the Dresden and Browns Ferry events and 
modifications can be obtained by contacting the Dresden and/or Browns Ferry 
operating supervision.  In addition, we understand that a General Electric 
Company task force is available for consultation at (480) 925-3188. 

For all boiling water power reactor facilities with a construction permit, 
this Bulletin is for information only and no written response is required.  

Approved by GAO, B180225 (R0072); clearance expires 7-31-80.  Approval was 
given under a blanket clearance specifically for identified generic 
problems.