ML20197G657
| ML20197G657 | |
| Person / Time | |
|---|---|
| Site: | Portsmouth Gaseous Diffusion Plant |
| Issue date: | 12/03/1998 |
| From: | Jonathan Brown UNITED STATES ENRICHMENT CORP. (USEC) |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| GDP-98-2047, NUDOCS 9812100003 | |
| Download: ML20197G657 (5) | |
Text
-
I d USEC A Global Energy Company i
l t
December 3,1998
[
GDP 98-2047 i
United States Nuclear Regulatory Commission 1
Attention: Document Control Desk l
Washington, D.C. 20555-0001 i
Portsmouth Gaseous Diffusion Pisnt (PORTS) l Docket No. 70-7002 Event Report 98-03, Revision 1 l
Pursuant to Safety Analysis Report (SAR), Section 6.9, Table 6.9-1, J (2), Enclosure i provides a revised 30 day written Event Report for an event involving an actuation of the Cascade Automatic Data Processing (CADP) Smoke Detection System in the X-330 Building at the Portsmouth Gaseous Diffusion Plant. The revised event report includes the root cause and corrective actions. Changes from the previous report are marked with a vertical line in the right margin. There are no new commitments contained in the report.
Should you required additional infonnation regarding this event, please contact Scott Scholl at (740) 897-2373.
Sincerely,
~4S+k-
/
J. Morris Brown General Manager Portsmouth Gaseous Diffusion Plant
Enclosures:
As Stated cc:
NRC Region III Office NRC Resident Inspectors - PORTS hp 9812100003 981203 PDR ADOCK 07007002l C
PDR j
EO. Box 800, Portsmouth, OH 45661 Telephone 614-897-2255 Fax 614-897-2644 http://www.usec.com Ollices in Livermore, CA Paducah, KY Portsmouth, OH Washington, DC i
I
\\
i GDP 98-2047 Page1of4 4
i Event Report 98-03, Revision 1 j
Description of Event On February 11,1998, at 1643 hours0.019 days <br />0.456 hours <br />0.00272 weeks <br />6.251615e-4 months <br />, the Cascade Automatic Data Processing (CADP) smoke detector S64, which monitors cell 31-2-6 alarmed in the X-330 Process Building Area Control Room
- 2. Following the alarm at approximately 1646 hours0.0191 days <br />0.457 hours <br />0.00272 weeks <br />6.26303e-4 months <br />, aperations personnel investigating the alarm observed a smokey haze above the cell housing where the smoke detector head was located.
l 1
Operations personnel in the area followed the "See & Flee" policy and evacuated the affected area.
Building " recall" was sounded at 1646 hours0.0191 days <br />0.457 hours <br />0.00272 weeks <br />6.26303e-4 months <br /> and the Fire Department and Plant Shift Superintendent j
(PSS) were notified. At 1647 hours0.0191 days <br />0.458 hours <br />0.00272 weeks <br />6.266835e-4 months <br />, cell 31-2-6 motors were stopped which reduced the cell pressure to below atmosphere. Due to a valve sequence timer malfunction, the required valves did not travel and isolate the cell from the cascade. As a result, the cell pressure rose to slightly above atmospheric pressure (approximately 15 psia). At 1650 hours0.0191 days <br />0.458 hours <br />0.00273 weeks <br />6.27825e-4 months <br />, the PSS responded to the X-330 Process Building. At 1700 hours0.0197 days <br />0.472 hours <br />0.00281 weeks <br />6.4685e-4 months <br />, the PSS gave permission for an operator wearing a respirator to go to the 31-2-6 Local Control Console (LCC) and operate the required valves using alternate means.
l The operator took the cell off stream by operating the valves using the manual switches at the LCC.
At 1708 hours0.0198 days <br />0.474 hours <br />0.00282 weeks <br />6.49894e-4 months <br />, the cell pressure was reduced to below atmosphere (approximately 13.75 psia). At l
1733 hours0.0201 days <br />0.481 hours <br />0.00287 weeks <br />6.594065e-4 months <br />, personnel using self-contained breathing apparatus and dressed in appropriate personal protective equipment (PPE) entered the cell floor. The area was sampled for airborne radioactivity i
and all samples were below detection levels. An all clear was given by the PSS at 1834 hours0.0212 days <br />0.509 hours <br />0.00303 weeks <br />6.97837e-4 months <br />.
l Operators noted no buffer system alarm indications prior to or immediately following the UF.
l l
release.
According to the Safety Analysis Report (SAR) for the Portsmouth Gaseous Diffusion Plant, the i
CADP smoke detectors are part of the UF. detection alarm system and are classified as Q safety systems when monitoring equipment operating above atmospheric pressure, in this instance, the monitored equipment (cell 31-2-6) was operating above atmospheric pressure. This CADP smoke detector actuation is reportable in accordance with S AR, Section 6.9 Table 6.9-1, J (2).
Cause of Event l
l The direct cause for the CADP UF, Smoke Detection System actuation was the occurrence of a UF.
l l
leak from a brazejoint on a buffer line supplying air to an X-165 expansion joint on cell 31-2-6, l
stage 9. The buffer line supplies dry air between the inner and outer expansion joint walls at l
t pressures greater than the cell operating pressure so that if a leak should develop, dry air, rather than l
atmospheric air will enter the cascade. The UF, is believed to have come from an inner wall leak l
)
on this X-165 expansion joint bellows. Although the exact sequence of events leading to the UF l
6 release could not be determined, it is believed that a small leak first developed in the buffer air l
supply line brazejoint. A small leak likely resulted in a pressure drop between the associated buffer l
l GDP 98-2047 Page 2 of 4 Event Report 98-03, Revision I air supply panel and the buffered expansion joint..A reduced pressure in the buffer supply line and I
the buffered expansionjoint allowed UF to migrate from a breach of the inner bellows wall into the l
buffer line where a small quantity escaped to atmosphere through the braze joint leak.
l l
As part of the investigation into this event, a section of the leaking buffer line was removed and l
analyzed to determine the failure mode. The analysis determined that the failure occurred at a j
sleevedjoint fitting as a result ofpoor brazejoint quality combined with minor corrosive attack. The l
poor brazejoint quality was attributed to excessive heat and inadequate brazejoint preparation which l
l resulted in gas porosity and non-continuous filler metal along thejoint. Corrosion of the brazejoint l
over time completed and exposed a network of flaws in the joint, creating a leak path through the l
joint.
l The root cause for the poor braze joint quality could not be determined. The original buffered I
expansionjoints were installed between 1978 and 1982, when the cascade was upgraded to operate l
at pressures above atmosphere. The buffer system is a non-safety system that was installed with the I
standards and practices that existed at that time. Currently, plant personnel are required to pass a l
certification test prior to performing instrument line brazing. This certification will minimize the l
potential for a poor quality brazejoint to be placed in service. A review of event reports for the past l
two years did not reveal any other safety system actuations caused by poor buffer system braze joint l
quality. As a result, this failure mode is considered an isolated occurrence and the likelihood for a l
l recurrence of this failure mechanism is considered to be low. Personnel protection in the event of l l
small releases of this type is assured through administrative controls such as the "See & Flee" policy l
and use of PPE.
l l
The Safety Analysis Report (SAR) Section 4.1.1.3.1, discusses the likelihood and safety significance l
of small UF leaks originating from copper tubing and through buffer systems. The SAR states "Small amounts of UF, and other toxic materials such as HF, CLIj, and UQ F can be released during sampling operations, seal changes, failure to obtain sufficient cell and piping UF negatives, l
ruptured copper tubing, and possibly through buffer systems. These releases are prevented primarily by using engineering and administrative controls. To protect the operator in the unlikely event a I
release does occur, operating specifications require that personnel performing operations and l
maintenance, where the possibility of release exist, wear protective equipment such as individually l
fitted gas mask. The releases in these cases may mean that a few grams of UF will escape to the atmosphere. There are no TSR systems (Safety Systems) to prevent the release of UF while performing these types of operations. Engineering and administrative controls are considered adequate."
l The double walled expansionjoints are buffered to reduce the likelihood that an expansionjoint leak l
l will result in a UF. release to the atmosphere. The buffer system design also includes alarms that l
l l
l l
l
\\
i i
~
GDP 98-2047 Page 3 of 4 Event Report 98-03, Revision I will alert operators if there is a significant change in buffer system pressure or flow. These design features provide engineering controls to help prevent minor releases of UF and to detect a i
significant bellows leak.
l An engineering evaluation was also conducted to determine why a buffer system high flow alarm l
or low pressure alarm was not received in response to the system leakage. These alarms provide a l
means of detecting a large breach in the UF. pressure boundary.
l In this installation, the buffer system supplies air to other expansion joints and to control valve I
i bonnets located in the cell. Since the design of the control valve bonnets allows some air to leak l
through the valve bonnets, there is normally some bufTer air flow. As a result, the buffer system flow l
alarm is set to alarm at a change in flow rate corresponding to an instrument reading of.5 inches of l water pressure differential above the normal operating differential pressure. The most likely cause l
for the lack of a flow alarm was that the increase in buffer air flow resulting from the leak in the l
expansion joint was small compared to the total flow rate.
l The low pressure alarm system is designed to detect a significant failure of a buffered component.
l The low pressure alarm pressure sensing instrumentation is located in the buffer cabinet immediately l
downstream of the pressure regulator that controls the bufTer pressure. Since the instrumentation l
is located close to the regulated air supply source, a significant leak in a buffered component located l
downstream would have to occur before the pressure at the alarm location would drop enough to l
cause an alarm. Since this leak was small, the pressure drop was not large enough to cause an alarm.
l The engineering evaluation also recommended enhancements to butler system testing that would l
lead to an improvement in the overall condition of the expansionjoint buffer system. This testing l
could help identify small leaks for correction and subsequent improved buffer system performance.
l Since this leak testing would require isolating certain sections of buffer systems, the testing would l
only be performed when a cell is shutdown and evacuated to minimize the possibility of a UF.
l release from an already existing breached internal expansion joint bellows or bufTer line. Plans are l
being made to implement this testing, when appropriate, prior to returning shutdown and evacuated l
cells to service.
I l
Follmving removal of the defective bufTer line it was also discovered that a plug, consisting of a l
uranium compound, existed in the line near the expansion joint. A plug in this location would l
prevent the buffer system from supplying air to the expansion joint and from detecting a bellows l
leak. It cannot be determined whether the plug existed before the release or developed as a result l
of the release. However, it is likely that the plug formed after the release began as UF migrated into l
the buffer line and reacted with wet air from the failed buffer line when the cell was shutdown.
l l
i r
i GDP 98-2047 Page 4 of 4 i
Event Report 98-03, Revision 1 l
i Corrective Actions 1.
The cell was shut down on February 11,1998, and will remain out of service until the l
expansion joint and buffer line are replaced.
l 2.
On May 18,1998, Engineering performed an evaluation to determine the cause for the buffer l
system failure to alarm. No corrective actions were determined to be necessary.
l l
Extent of Exposure ofIndividuals to Radiation or Radioactive Materials Nine individuals were required to submit a urine sample as a result of the out-leakage. Eight individuals were below 5 pg/ liter for soluble uranium (flag level) and one individual was above flag level. The intake for the one individual above the flag level was approximately 120 pg of soluble uranium. The limit for occupational workers is 10,000 pg of soluble uranium per week.
l Lessons Learned There were no lessons leamed from this event. Operators responded to this event in accordance with l
l plant policies and procedures.
l l
l l
i i
i
,