05000313/LER-2014-002
| Arkansas Nuclear One - Unit 1 | |
| Event date: | 11-25-2014 |
|---|---|
| Report date: | 12-22-2014 |
| Initial Reporting | |
| ENS 50641 | 10 CFR 50.72(b)(3)(ii)(B), Unanalyzed Condition |
| 3132014002R00 - NRC Website | |
plants uses the fuel performance code TACO3. The identified deficiency is in the thermal conductivity model in this computer code. The deficiency is that the code does not adequately represent the reduction in fuel thermal conductivity with burnup. This issue was discussed in the NRC Information Notice (IN) 2009-23 "Nuclear Fuel Thermal Conductivity Degradation". Recent comparisons of the fuel temperatures from this code with fuel temperatures from the code GALILEO (a code that has an adequate fuel thermal conductivity model) indicate that the TACO3 code thermal conductivity model may lead to an under prediction of the Peak Clad Temperature (PCT) during a LOCA.
TACO3 does not model the thermal conductivity degradation (TCD) with burnup explicitly but has adjustments to the methodology and increases in the LOCA fuel temperature inputs. These adjustments were intended to compensate for the non-conservative thermal conductivity model in TACO3.
The continued use of this code was previously evaluated by AREVA in 2009 following the NRC issuance of Information Notice 2009-23. In 2009, it was concluded that sufficient conservatisms in both code predictions and LOCA methodology compensated for a lack of TCD models based, in part, on comparisons to an early version of the code GALILEO. However this conclusion has been invalidated based on recent GALILEO LOCA initialization studies.
Based on these new Large Break LOCA (LBLOCA) initializations, it is concluded that the LOCA EM that uses TACO3 must be modified by application of additional fuel temperature uncertainty to account for the effects of TCD based on COPERNIC2, a code that models TCD adequately.
An evaluation was performed by applying the EM change to a Lower-Loop LBLOCA model with an axial power shaped peaked at core elevation 2.506-feet (ft) with a middle-of-life (MOL) burnup condition. For the representative plant, the 95/95 volume-average fuel temperature from the limiting PCT case was increased by 230°F. The results of the evaluation show that the original limiting MOL case cladding temperatures at the core elevation of 2.506 ft were increased by 481°F for the ruptured node and 288°F for the unruptured node. The results of this evaluation can be generically applied to all B&W plants. These ruptured and unruptured node cladding temperature deltas were applied to the Arkansas Nuclear One, Unit 1 (ANO-1) full spectrum of MOL cases and led to an increase in limiting PCT of 388°F. An evaluation of the cladding temperatures at end-of-life (EOL) has confirmed that the MOL results were limiting. Also it is noted that the cladding temperatures at beginning-of-life (BOL) remain unaffected by TCD. This LBLOCA EM model change results in a significant increase to the calculated PCT. When applying the estimated PCT increases with the revised EM approach, the limiting PCT was estimated to be 2396°F, which is in excess of 2200°F.
In order to reduce the PCT to less than 2200°F, AREVA recommended linear heat rate (LHR) limit reductions on October 21, 2014, and suggested that it was prudent to administratively implement any changes as a compensatory measure. The compensatory measures recommended a 2 kilowatt / foot reduction in the MOL LHRs. Imposition of the compensatory measures reduces the evaluated PCT to be equal to the PCT prior to the EM correction and thus less than 2200°F. As a precautionary measure pending the completed analysis, ANO-1 implemented the compensatory measures on October 20, 2014.
The local oxidation and whole core hydrogen also remain well within the 10 CFR 50.46 acceptance criteria for the LBLOCA scenarios. With the MOL LHR limit reduction, the core geometry remains amenable to cooling and acceptable long-term cooling is unaffected by these changes.
The impact of the EM correction and compensatory measure is summarized in Table 1 for LBLOCA and in Table 2 for Small Break LOCA (SBLOCA). The SBLOCA analyses are not sensitive to the initial fuel temperatures and thus the estimated impact on the SBLOCA peak cladding temperature is zero.
This deficiency will be corrected in a future LOCA analyses on an NRC agreed upon schedule.