ML18065B011

Part 21 Rept Re Application of Certain Aspects of ABB-CE Safety Analysis Methodology

ML18065B011
Person / Time
Site:	Millstone, Calvert Cliffs, Palisades, Palo Verde, Saint Lucie, Arkansas Nuclear, Waterford, San Onofre, Maine Yankee
Issue date:	10/18/1996
From:	Rickard I ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY, ASEA BROWN BOVERI, INC.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
REF-PT21-96 LD-96-045, LD-96-45, NUDOCS 9610240239
Download: ML18065B011 (6)

Text

t

.. +--~:*~*-*::*

SOv.5'5

• t. 301 3(3

,.,1*1*

J]l 1111 317 31&

33.S/33/,p 31Pt / 3b.Z, 3btb October 18, 1996 3'8;;..

LD-96-045

.381 s-ol./3/91 JS~t>

Document Control Desk U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Subject:

Report of a Defect Pursuant to 10 CFR 21 Regarding the Application of Certain Aspects of ABB-CE Safety Analysis Methodology

Dear Sir:

The purpose of this letter is to notify the Nuclear Regulatory Commission of a defect, as defined in 10 CFR 21 - "Reporting of Defects and Noncompliance". The identified defect involves the application of certain aspects of the ABB Combustion Engineering (ABB-CE) reload safety analysis methodology.

Specifically, the identified defect concerns the screening methodology used by ABB-CE to assess the continued conservatism and applicability of the DNBR probability distribution function (pdf) from cycle-to-cycle. This assessment in turn provides assurance that the fuel DNBR Specified Accept.able ~uel Design Limit (SAFDL) will not be violated. It is important to point out that the c~ ABB-CE thermal-hydraulic Statistical Combination of Uncertainties (SCU) methodology usea1o explicitly determine the pdf input to the calculation of the DNBR SAFDL is not in question. That i~. if the SCU analysis methods are explicitly applied to determine the pdf input fo the DNBR SAFDL calculation, the SCU pdf input to the DNBR SAFDL calculation is correct. Rather, the defect involves the evaluation criterion used during cycle specific screening to determine whether or not an explicit SCU pdf re-analysis is required to verify the continued conservatism during the upcoming cycle for the DNBR SAFDL of record.

Based upon an evaluation, ABB-CE has concluded that the above described situation

- -- - --represents-a-defect-as-defi*m* ed-in-10-CER2t.3,_i.e. ,_::_A__ cor:iditio_n__Q_~_c~i.~IJm_~ta~-~!~-~~~~1-~ _

contribute to the exceeding of a safety limit ... ". In accordance with 10 CFR 21.21 (c)(4), the I

- --- \ /

Enclosure to this letter summarizes the information available to ABB-CE at this time.

I '

--~tA 0. 2-fg~~i-6gA8A~s5 . /*: ----\ . . ~bl 0

s* ,. ' PDR :

1 y

• • .... *- - -ABB Comb ustion Engineering Nuclear Power Combustion Engineering, Inc. P.O. Box 500 Telephone (860) 688-1911 2000 Day Hill Rd. Fax (860) 285-5203 Windsor, CT 06095-0500

'* Document Control Desk LD-96-045 18 0<?tober, 1996 Page 2 If you have any questions, please feel free to contact me or Mr. Chuck Molnar of my staff at (203) 285-5205.

• ard, Director rations Licensing

Enclosure:

As stated cc: M.A. Barnoski (ABB-CE)

-*-- - - - - - - -- - - - - - --- -- - --- - ~-

- -~ -~- --- ~

• Enclosure to LD-96-045 Page 1 of 4 ABB Combustion Engineering Nuclear Operations 1-0-CFR-21 .Report of a. Defect or Failure to Comply The following information is provided pursuant to the requirements identified in 10 CFR 21.21 (c)(4):

(i) Name and address of the individual(s) informing the Commission:

Dr. I. C. Rickard, Director Operations Licensing ABB Combustion Engineering Nuclear Operations 2000 Day Hill Road Windsor, CT 06095-0500 (ii) Identification of the facility, the activity, or the basic component supplied or such facility or such activity within the United States which fails to comply or contains a defect:

The activity for which this report is being filed is the DNBR probability distribution function (pdf) safety analysis screening methodology used by ABB-CE to verify the continued conservatism of the analysis of record for those plants for which ABB-CE performs reload safety analyses.

The current ABB-CE thermal-hydraulic Statistical Combination of Uncertainties (SCU) methodology used to determine the pdf input to the calculation of the DNBR Specified Acceptable Fuel Design Limit (SAFDL) is not in question. If the SCU analysis methods are applied to determine the pdf input to the DNBR SAFDL calculation, the SCU pdf input to the DNBR SAFDL calculation would have no safety concern. Rather, the reported defect involves the evaluation criterion used during cycle specific screening to determine whether explicit SCU pdf re-analysis is required to verify the continued conservatism of the DNBR SAFDL of record for the upcoming cycle.

The SAFDL on DNBR currently contains an allowance which accounts for uncertainties in parameters (e.g. inlet flow, rod diameter, calculation uncertainty, rod pitch, etc.) which are not explicitly included in the DNBR on-line calculation for digital plants and the setpoints for analog plants. The effects of all these uncertainties are statistically combined according to SCU methodology to produce a pdf which is used to determine

-- -- - - -------- ---the-E>NBRSAFDl.::--The-pdf-is calculated-using-the sensitiv.iti.es of _DNBR tQ_ various parameters. The most significant component sensitivity is the sensitivity of DNB-Rtothe--

inlet flow in the limiting assembly.

The current screening methodology for assessing the continued conservatism of the DNBR pdf is based on a two stage screening process. In the first stage, the limiting assembly inlet flow factor (and the flow factors in the four adjacent assemblies to the limiting assembly) and pin-to-node power ratios for the upcoming cycle are compared with the corresponding values from the Analysis Of Record (AOR). If the limiting

i.

• Enclosure to LD-96-045 Page 2 of 4 assembly inlet flow factors and pin-to-node power ratios are higher for the upcoming cycle than the corresponding values from the AOR, no further DNBR pdf analysis is

- performed*and*the AOR DNBR-pdf is concluded to remain applicable for the upcoming cycle. If the limiting assembly inlet flow factors and/or pin-to-node power ratios for the upcoming cycle are lower than the corresponding values from the AOR, the second screening criterion is applied. The second criterion compares the CETOP/TORC penalty factors derived as part of the 4-pump TORC analysis. If the reference cycle CETOP/TORC penalty factors are lower than those for the upcoming cycle over the range of interest for the upcoming cycle, then the pdf calculated using the reference cycle TORC model is concluded to c.ontinue to be bounding. If the above described screening process concludes that the AOR DNBR pdf may not remain applicable, explicit TORC analyses are performed to either establish the applicability of the AOR pdf to the upcoming cycle or calculate a pdf that does apply to the upcoming cycle.

Recent preliminary calculations for Palo Verde-1 Cycle 7 cast doubt on the sufficiency of this screening methodology for confirming the applicability of the DNBR pdf for all core designs. Specifically, explicit TORC calculations for a core design which was not ultimately chosen for the upcoming cycle demonstrated that the pdf, and hence DNBR SAFDL, was not bounded by the reference case, even though the reference CETOP/TORC penalty factors were demonstrated to be conservative using the screening methodology described above.

Further evaluation has shown that the screening methodology may not be sufficient for more recent fuel managements that exhibit the characteristic of a very flat assembly power distribution, which forces the limiting or "hot" subchannel more interior to the assembly. A review of the thermal-hydraulic models and SCU methodology, particularly for the more recent fuel managements exhibiting the characteristics identified above, has shown that the models and methodology are sound, except for the screening methodology. As such, this situation represents a condition of circumstance that could contribute to exceeding the DNBR safety limit defined by plant technical specifications.

It should be noted that although the identified defect may impact the DNBR SAFDL for specific core designs, it does not imply that the margin to the SAFDL is non-conservative in all cases. For instance, conservatism in the CETOP/TORC penalties in the AOR may compensate and provide conservative margin to the DNBR SAFDL.

(iii) Identification of the firm constructing the facility or supplying the basic component which fails to comply or contains a defect:

-- - - - - - - - - - - - - - - - ~- - - - - - - - - - - - - - - - - - - - - - - - - - - -

- - - --- ----------- - - ---- ----~---

Combustion Engineering, Inc.

2000 Day Hill Road Windsor, CT 06095-0500 (iv)* Nature of the defect or failure to comply and the safety hazard which is created or could be created by such defect or failure to comply:

Enclosure to LD-96-045 Page 3 of 4 The defect identified involves a deficiency in the screening methodology used to verify the continued conservatism of the DNBR pdf for the safety analysis of record such that a non-conservative conclusion may be reached. As such, this situation represents "a condition or circumstance ... that could contribute to the exceeding of a safety limit ... ", as defined in the defect definition of 10 CFR 21.3.

(v) The date on which the information of such defect or failure to comply was obtained:

ABB-CE determined on 17 October, 1996 that a defect as defined in 10 CFR 21.3 did in fact exist.

(vi) In the case of a basic component which contains a defect or fails to comply, the number and location of all such components in use at, supplied for, or being supplied for one or more facilities or activities subject to the regulations in this part:

The defect applies to the DNBR pdf safety analysis screening methodology used by ABB-CE to verify the continued conservatism of the analysis of record for those plants for which ABB-CE performs reload safety analyses. This screening methodology was applied to ABB-CE designed nuclear power plants for which ABB-CE continues to provide fuel and thermal-hydraulic safety analyses utilizing the SCU methodology.

Specifically, the nuclear power plants potentially affected by this defect include; Arkansas Nuclear One Unit 2 Calvert Cliffs Units 1 and 2 Palo Verde Units 1,2, and 3 San Onofre Units 2 and 3 St. Lucie Unit 2, and Waterford Steam Electric Station Unit 3 For these plants, ABB-CE has determined that the AOR for all except Palo Verde Unit 3, Cycle 6 remain bounding and no compensatory measures are required to avoid violation of the DNBR safety limit. For Palo Verde Unit 3, Cycle 6 operation it was ascertained that the screening methodology did in fact result in a potentially nonconservative DNBR SAFDL value. Compensatory measures were identified and provided to the utility (Arizona Public Services) to avoid exceeding the DNBR safety limit. It is ABB-CE's understanding that Arizona Public Services has taken the steps necessary to implement the compensatory measures provided.

For the ABB-CE designed Omaha Public Power District (OPPD) Ft. Calhoun nuclear

--- ---- ---- - - -powerstation~-although-ABB-GE-is not-currently-the-fuel-and/or: safety--analysis _vendor, ____ _

ABB-CE safety analysis methodologies are employed by OPPD to perform its own safety analyses in-house. ABB-CE has insufficient information available to make a meaningful determination of applicability in this case. As such, OPPD has been informed of the condition reported herein advising that an applicability evaluation of this situation is warranted pursuant to 10 CFR 21.

I*

Enclosure to LD-96-045 Page 4 of 4 ABB-CE designed nuclear power plants, for which ABB-CE is not currently the fuel and/or safety analysis vendor and, therefore, which are not affected by the identified defect include; Millstone Unit 2 Maine Yankee Palisades St. Lucie Unit 1 (vii) The corrective action which has been, is being, or will be taken; the name of the individual or organization responsible for the action; and the length of time that has been or will be taken to complete the action:

The defective screening methodology has been replaced by a revised screening methodology which reduces the degree of engineering judgment applied and imposes specific guidelines which must be satisfied prior to use of the revised screening methodology. If the specified guidelines cannot be satisfied, the SCU analysis methods will be explicitly applied to determine the pdf input to the DNBR SAFDL calculation.

(viii) Any advice related to the defect or failure to comply about the facility, activity, or basic component that has been, is being, or will he given to purchasers or licensees:

As mentioned in item (vi), ABB-CE has informed Arizona Public Services of compensatory actions to assure that the Palo Verde Unit 3, Cycle 6 plant operation does not exceed the DNBR SAFDL. Additionally, ABB-CE has notified OPPD of its inability to evaluate this situation for the Ft. Calhoun nuclear power station and advising that they evaluate their situation pursuant to 10 CFR 21.