ML20086E653
| ML20086E653 | |
| Person / Time | |
|---|---|
| Site: | Pennsylvania State University |
| Issue date: | 07/07/1995 |
| From: | Shirley D PENNSYLVANIA STATE UNIV., UNIVERSITY PARK, PA |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9507120162 | |
| Download: ML20086E653 (5) | |
Text
- - - - - - - - - - - - - - - --- -- _
I PENNSTATE '
A 'fa!L j .- .
Dasid A. Shirley The Pennsylvania State University
~
Senior vice President for Rewarch 207 Old Main and Graduate Education Unis:rsity Ft rk. PA 16802-1$03 i July 7,1995 Nuclear Regulatory Commission Document Comrol Desk Washington, DC 20555
\
Reportable Occurrence for the Penn State Breazeale Reactor License Number R-2, Docket Number 50-05
Dear Sir or Madame:
When comparing control rod wonh data from annual surveillance tests during 1994 and 1995 a discrepancy of greater than one dollar of reactivity was observed. This constitutes a reportable occurrence pursuant to Technical Specification 1.1.33.d. On June 23,1995, it was reported in accordance with Specification 6.6.2.a.(3); this is the 14 day follow-up repon required by that Specification.
Seonence of Events A chronology of relevant events is presented below and also summarized in Figure 1:
5/10/93 -
All control rods were calibrated during the normal annual surveillance measurements. Total control rod wonh was measured to be 511.46. The core,45T, remained in service for another year.
7/5/94 -
All control rods were calibrated, total worth being $10.90. The calibration followed major maintenance, fuel inspection, replacement two suspect elements, and replacement of the three fuel follower control rods (FFCR).
The fourth control rod, the transient rod, does not have a fuel follower and was therefore not replaced. This core,47-8, was detennined to have insufficient excess reactivity so eight more fuel elements were subsequently added to the core periphery.
7/e 4 -
The control rods were calibrated for the new core,47, and found to total
$12.25. The excess reactivity was shown to have been restored to the desired level of 56.03 and the reactor was operated for a year with this core loading.
6/14/95 - Annual control rod calibration was performed for core 47, showing a total rod worth of 513.75 with an excess reactivity of $6.66. This rod worth exceeded historic values so a repeat of the rod calibration was scheduled.
11GG11 A,. Fuimi Onrortunity University f/"
9507120162 950707 PDR ADOCK 05000005 e lj
________...B.. PDR - _ - _ _ - _ _ _ -
i ..
' 6/1995 - The repeat rod calibration for core 47 showed a total rod wonh of $13.74 with excess reactivity of $6.64, extnmely close to the previous measurement. This placed the 1994 measurements in question.
1 i 6/23S 5 - After a thorough review of all of the above data sets, it was hypothesized that the 1995 data was correct despite the unusually high control rod wonhs. On this basis, the 1994 rod worths were low by more than $1.00 and the incident was therefore reported to the NRC by telephone and fax. Additional tests were planned to confirm the hypothesis.
L l 6/26S 5 - The core was reconfigured as core 47-8 to repeat the measurement from 1994. The total rod worth was measured to be $13.49.
l i
6/28S5 - The loading was changed to core 45T (with the exception of the two suspect replacement elements) and total rod wonh was measured to be $11.62.
6/29S 5 - The safeguards committee met to review the situation and found that appropriate action had been taken to return to normal operation. The NRC was infonned by telephone the following morning.
6/30S 5 - Prepare for retum to nomial operation. The new core will have four fewer fuel elements than the core used during the past year. A new set of rod calibration curves will be measured before returning to normal operation.
Analvsis ofIncident At the outset it should be noted that two major misconceptions affet d the analysis. Since they may have been presented incorrectly during discussions they will each be addressed to correct the record before proceeding.
First, the 6/14/95 rod worth data showed a 10-15% increase over the 7/6/94 data, suggesting that all rods were effected proportionately and that the difference was not isolated to the three new FFCRs. Contrary to this, a high degree of consistency was later observed among all data sets except the 1994 data, the conclusion being that the 1994 data is questionable and the remaining data shows a predictable change in the worth of the three FFCRs and no change in the transient control which was not replaced.
Second, from the experience of other T.11GA reactors we expected only a few cents change due to each new FFCR and not the change we observed. Also, our total control rod worth had historically remained below $12. This placed suspicion on the 1995 measurements early in the investigation. While this is . - for replacement FFCRs of the same type, it was found that our old FFCRs were m em :tured in the early 1960s. The boron loading in the standard TRIGA FFCRs was increaed in approximately 1970, making the replacement FFCRs installed in 1994 of a greater total rod worth.
Analysis ofindividual control rod worth shows a high degree of consistency and predictability between data sets from 1993 to each core in 1995 and among each of the three cores measured in 1995. There is also an obvious lack of consistence and predictability between the 1993 and 1994 data sets, between the two 1994 data sets, and between the 1994 and 1995 data sets. While the data taken in 1994 was rationalized to be correct at the time it was taken,in retrospect it should have been challenged and redone. We have concluded that both 1994 data sets contain unexplained variability and show a systematically low total control rod wonh.
The data consists of control rod configurations t= fore and after a rod bump and the power versus time following the reactivity insertion. The increment of reactivity added by the rod bump is related through the Inhour equation to the exponential period of the resulting power increase.
The 1994 raw data has been reviewed but no errors have been identified. The core configurations have been replicated and measurements repeated. Instrument calibration records have been reviewed in search of a source of the error but none has been identified.
A number of hypotheses have been tested such as loss of heavy water into the reactor pool, water intrusion into the control rods, boron washout, etc. Each hypothesis has been refuted by data and observations. A statistical analysis of the raw data is undenvay in seamh of a root cause of the error in the 1994 rod calibration results but to date none has been identified. We may find that there is insufficient data on record to test a hypothesis.
For example, if the reactor operator consistently did the rod bumps from a slightly subcritical condition the reactivity attributed to the section of the control rod withdrawn during the rod bump would be too low; a series oflower than actual rod inemments would result in a low integral rod wonh. The 1994 data does not show how close to critical the reactor was when measurements were initiated.
Until a root cause has been identified or we are convinced that a sufficient history of acceptable measurements has been established, we will verify a sample of all rod wonh measurements with an independent, alternative method. (For example, a compensated ion chamber in the vicinity of the reactor core reading out on an electrometer with time intervals measured by a stop watch.)
Critical control rod positions are recorded weekly. This data has been reviewed and found to be consistent throughout the period in question, showing that there was no change effecting the reactivity of the core. We are confident that no unexplained change in reactivity has occurred during this time; the difference observed is only an anifact in the measurement and not in rod wonh.
Consecuences of the Incident There is a limiting condition for operation which limits the excess reactivity to $7.00. The basis is the consequence of nn inadvertent $3.40 pulse from 115% power. On 7/6/94 the excess reactivity was measured for the existing core (47) to be $6.03 using the rod calibration data taken that day; this is the normally desired excess. However,it has been shown that the rod worths measured that day were low. The rod calibration curves measured for the same core on 6/14/95 showed it to be 56.66. The core is wonh an additional $0.26 beside the heavy water tank. During a cycle the core loses about $0.14 from bumup, making it conceivable that this LCO was violated early in the cycle by operat'ng with an excess reactivity of $7.06. If so, the effect of being $0.06 over the limit would have been inconsequential because the transient rod wonh is 50.40 less than the
$3.40 analyzed.
i Corrective Action ;
Three primary corrective actions will be pursued as a result of the findings of this l occurrence. First, as stated above, until a root cause for the error can be identified or until I i
we develop historical evidence that a recurrence is unlikely, altemate methods of measuring rod worth will be used to verify the integrity of measurements.
l
4
' The second corrective action is to verify that newly acquired pans and materials that are significant hr reactor safety meet appropriate specifications. Even though the control rod worths were verified by reactivity measurement, the knowledge that they had higher boron loading than ne old FFCRs should have triggered an investigation of the erroneous data at the time it was taken.
The third correc tive action is to give greater attention to independent second level reviews.
This will be discussed at a staff meetmg to impress on all personnel the importance of thorough reviews, especially for matters related to reactor safety.
1 Conclusion j After two weeks of dedicated testing and benchmarking of the current and past cores, we l believe that we fully understand the reactivity worth of the new FFCRs and their l relationship to the old FFCRs. The past year of operation has been bounded to show that i operations were conducted in a safe environment, even though better information would have been desirable. Steps have been taken to provide reasonable assurance that a similar situation will not recur. Should you have questions on this matter please refer them to the >
l Director of our Radiation Science and Engineering Center, Dr. Marcus H. Voth.
i Sincerely, -
i M (. -
David A. Shirley i i
Senior Vice President for Research and Graduate Education MHV/DAS/ld14086.95 cc: Region I Administrator v..+ , , y .-e- r ~ w-. - - , .n, - , y-
e Core 45T j Core 47 8 i Core 47 .
(Old FFCRs) 1 (New FFCRs) 8 (New FFCRs and 8 l l additional elements) i i -
1 1 Legend 1 I ~
S0493 I i CRP = Cribcal Rod Positons Plot = Total Rod Reactivi'y l l .
I I Psd - Core Shutdown Reactivity 93 CRP = 713 I I Peu = Core Excess Beactrvity l = 11.46 Pot 1 Year Burnup. Cal Curves Date of Calibration Curves PF3 = 6.10 l O11 replaces 202.
- I Per = fi36 1 33 replaces 205. I Cal Curves - I 3 new FFCRs I I 8 Elements Added 5/1&93 l I I I I I I i I I - _ _ ! _ _
1 I I I l I i i 1 531,94 7/594 7/694 y lg i I 94 CRP = 753 i CRP = 814 I CRP = 721 Pt ot = 11.46 ----- - - - - - . Plot = 10.90 f Ptot = 12.25 Psd = 6.57 I Osd = 6.38 I Psd = 6.22 Pen = 4 89 Pen = 4 52 Pex = 6.03 l l Cat Curves = l Cal Curves = 1 Cal Curves =
5/1093 1 7/594 1 7594 1 1 1 1 1 2 Swapped LWts 1 2 Swapped Elements i 1
_ _ 1 _ _ L_ _ _ _ _ _ _ _
1 1 1 1 1 1 I 1 6/2895 l 6/2695 l 6/19/95 6/14S5 I i 1 l l I I 95 CRP = 813 CRP = 824 CRP = 731 CRP = 731 Itot = 11.62 Ptot = 13.49 : Ptot = 13.74 .g Plot = 13.75 .
Psd = 6.78 Pen = 4 84 1 Psd = 8 12 Pex = 5 37
\\
l Psd = 7.10 Pen = 6.64 Psd = 7.09 fwa - 6 66 l l Cat Curves = l Cat Curves = i Cal Curves - Cal Curves =
6/2895 I 6/2695 1 6/1995 ! 6/14/95 l I u 1 1 2 Swapped Elements 2 Swapped Eierren's 2 Swapped Elements 2 Swapped Elements 3 New F FCRs Replaced No Core Change, weiOld FFCRs. No 8 Elements Removed No Durnup. No Fuel Movement Fuel Movement Rgure 1 Core Compansort Diagram