ML20211F284

Ltr Contract:Task Order 3, Developmental Assessment Numerical Techniques Under Contract NRC-04-97-046

ML20211F284
Person / Time
Issue date:	09/25/1997
From:	Mace M NRC OFFICE OF ADMINISTRATION (ADM)
To:	Smithson C PURDUE RESEARCH FOUNDATION
References
CON-FIN-W-6749, CON-NRC-04-97-046, CON-NRC-4-97-46 NUDOCS 9710010005
Download: ML20211F284 (6)

Text

.

p itt j

UNITED STATES 5

E NUCLEAR REGULATORY COMMISSION f

WASulNGTON, D.C. 20565-0001

n,,,,*

SEP 171997, Purdue Research Foundation ATTN: Chris Smithson-

-1021 Hovde Hall. Purdue University West Lafayette.- IN 47907-1021

Dear Mr. Smithson:

l SUBJECT.

TASK ORDER NO. 3 ENTITLED *0EVELOPMENTAL ASSESSMENT NUMERICAL TECHNIQUES" UNDER CONTRACT NO. NRC-04-97-046 i

l In accordance with Section G.4, Task Order Procedures, of the subject contract, this letter definitizes Task Order No. 3.

This effort shall be performed in accordance with the enclosed Statement of Work.

The period of perforeance for Task Order No. 3 is September 30, 1997 through June 30, 1998. The total estimated cost for full performance of this task order is $103,704.00.

Funds in the amount of $68,000 are hereby obligated for performance of this task order. The Contractor shall not incur costs for this task order which exceed this obligated amount.

Accounting data for Task Order No. 3 is as follows:

B&R No.: 760-15-11-500-5 Job Code: W-6749 BOC Code: 252A RES ID: RES-C97-088 Appropriation No.: 31X0200 Obligated Amount: $68,000.00 The following individuals are considered by the Government to be essential to the successful perfctmance of the work hereunder:

John Mahaffey

-The contractor agrees that such personnel shall not be removed from the effort under this task order without compliance with Contract Clause H.2-Key Personnel i

DFbc3 t

c,:, o.3 -

FoA c82n *

• lllllllltl1Hll11lil!IllElli NRC-04-97-046 PDR SEP 2 21997-

Purdue

-Contract No. NRC-04-97-046' Task Order No. 3 Page 2 of 2 The issuance of this task order does not amend any terms or conditions of the

-subject contract.

Your contacts during the course of this tast order are:

. Technical Matters:

Jennifer Uhle, Project Officer (301) 415-6023 J ontractual Matters:

. Stephen Pool, Contract Specialist C

(301) 415-8168 Please indicate _your acceptance of this task order by having an official,

. authorized to bind your organization, execute three copies of this document in L

the space provided and return two copies to the Contract Specialist. You-should retain the third copy.for your records.

Sincerely, pf

/

A-,

.\\j

^

Mary H. ' Ma e. Contracting.0fficer

-Contract, nagement Branch No. 1 Division of Contracts and Property

. Management Office of Administration

Enclosure:

As stated-ACCEPTED: TASK ORDER NO. 3 Q

[k%

SEP 15 t%7 MAME a,,,, w, N*r w

'TITLt

c 7- _.. _

~ _... _ _ _. _ - _... _ _ -. _ _ _ _ _... _ _ _ -. _. _.

4 4

4 STATEMENT OF WORK TASK ORDER #3,

• DEVELOPMENTAL ASSESSMENT NUMERICAL TECHNIQUES" L

Background

- Code assessment plays a vital role in NRC thermal-hydraulic code improvement efforts.' A j-great deal of effort is spent comparing the results of the code predictions to test data to determ!ne the accuracy of the code. Currently, 'he evaluation is described with terms such as iL

. excellent, re'asonable, minimal and insufficient and although some guidelines are used to define these ratings, the evaluation is up to the judgement of the individual doing the assessment (see

)

attachment from Adequacy Evaluation of RELAP5/ Mod 3, Version 3.2.1.2 for Simulating AP600 Small Break Loss-of-Coolant Accidents). Peer review enhances the consistency of the ratings, p

but human opinion determines the assessment results. This methodology does not allow the i

comparison to completely 191y on automation, where computer scripting language will run the test cases with the code, compare the results to the date, and yield a numerical assessment j

result. Presently, scripts can be used to run the test cases and plot the test results with the -

data, but human effort is spent determining the fidelity of the simulation.

i in order to provide full automation, NRC can use a variety of numerical techniques to rate code assessment results. For example, these techniques may include; root mean squared comparison and Fast Fourier Transform techniques to evaluate the frequency components of

}

' the shape of the data. Therefore, rating the simulation fdelity of the code as excellent, L

reasonable, minimal and insufficent will no longer be subjective but will be standardized based

[

on the results of the numerical techniques. To set numerical standards that correspond to the i

qualitative labels (excellent, ressmable,...), a peer review process will be used. This l

numerically based rating system will ensure consistency throughout the project and will facilitate full autcmation of the developmental assessment activities.

l Minimizing the subjectivity of assessment is essential for successful completion of the code -

consolidation effort. Throughout the effort as capabilities are recovered from the current suite

of codes and incorporated into consolidated coda, assessment will be used to determine if the simulatxm fidelity of the,wovered models in the consolidated code is comparable or better p

than that of the current codes. Basing the rating on numencal results will eraure that the l

accuracy of the models is consistent throughou+ the 5 year project ard will provide a

[

_ quantitative measure of the code's perfo Tnance so that progress can be tracked.

I' Additionally, analysts can devote more time to evaluating code predicted system response to heighten understanding of model interactions and less time in peer review sessions rating j

simulate fidelity. Essentially, peer review will be used to correlate numerical results to the

{

qualitative labels once, precluding the need for peer review sessions every time a model is modified. However, analysts will be responsible for ensuring that the numerical rating system -

i

!L yields valid results in each instance but will'do so without full peer review.

i-L

__ Objective The objective of this contract is to:

I j

Develop and document a set of numerical techniques and algorithms which can be used 1-to rate the simulation fidelity of NRC thermal-hydraulic systems and kinetics codes to be

2 used in the automation of developmental assessment activities. Since the techniques will be used to ensure that the code predicts all the dominating phenomena, the.

techniques should provide quantitative measures of the simulation fidelity by corpparing the magnitudes of the parameter, the slopes and inflection points of the curves that represent major changes in the trends of the parameters and timing of events. The use of a numerically based rating system will ensure that all models are rreviewed with a consistent set of criteria throughout the consolidation effort.-

Work Requirements Task 1:- Literature Review; L

The contractor will review the literature to identify techniques that can be used to determine the

_ similitude of experimental data and code calculations. The review should address all work currently used in the thermal hydraulics community (e.g., the FFT method of D'Auria) and in addition provide a brief review of approximation theory of mathematics (Lp-norms), methods

used in electrical engineering, and methods used in image processing (wavelets). This literature review will make a clear distinction between methods that are applicable to steady --

state data (i.e., a variable as a function of distance) and transient data (i.e., a variable as a 4

function of time). All of this information, which represents the current state in the thermal-hydraulics field and a brief review of related fields, will be consolidated into a letter report and

- submitted to NRC in both text and electronic format. This report will provide NRC with a means by which the completeness of the literature review can be reviewed Estimated duration: 1 m Task 2f Selection of Numerical Techniques Based on the literature review performed in Task 1, the methods appropriate for use will be selected. A letter report shall be submitted to the NRC including a detailed description of the rationale for the selection process, a description of the chosen techniques and a few simple example calculations demonstrating the behavior of the selected methods. In essence, what the individual techniques measure and how the combination rate the simulation fidelity of all prototypic sets of data and code resuks shall be explained.

The NRC will review this report before Task 3 is initiated. Note, NRC will require 2 weeks for this action, so work should be scheduled accordingly. The deliverable shall be submitted to NRC in both text and electronic format.

(

Estimated duretion: 1.5 m Task 3r Development Plan and Implementatin

- This task is comprised of 6 Subtasks.

Subtask 3.1: Test Plan

3 To ensure that the selected methods can be used to quantify the simulation fidelity of the NRC codes, the techniques must be applied to combinations of data and code results prototypic'of comparisons performed in developmetnal assessment, such as depicted in the "Adequa_cy Evaluation of RELAP5Niod3, Version 3.2.1.2 for Simulating AP600 Small Break Loss-of -

Coolant Accidents'. This document will be provided by NRC at a later oate to assist in the selection and/or generation of curves against which the numerical techniques selected for implementation in Task 2 will be run.

.Therefore, the contractor shall develop a plan that will outline how the techniques will be testad and also the means of verifying that the numerical methods are coded properly (to be used when the coding is developed). Therefore, the test plan shall describe:

1) The method to verify that the techniques have been coded exactly as described in the letter

-report of Task 2; L

-- 2) The testing procedure, including which experimental data and which code results will be g

compared to each other to demonstrate that the methods are working properly and are an adequate measure of simulation fidelity.

This test plan shcIl be submitted to the NRC as a letter report. This test plan must be approved

' by the NRC before coding and testing begins. Note. NRC will require 1 week for this action, so work should be scheduled accordingly. All experimental data and code calculations delineated -

in the plan shall be gathered before the plan is submitted.

Estimated duration: 1 m Subtask 3.2: Software Design and implementation Document A software design and imp!ementation document (SDID) shall be prepared outlining how the techniques will be coded The following information shall be included:

1) The rationale for code language selection;

11) A description of the programs input and output;

2) A desenption of the data flow through the code;

-_3) A brief description of each subroutine / function.

The SDID shall'be submitted to NRC in the form of a letter report in both text and electronic format.

The SDID must be reviewed and accepted by NRC before work under Subtask 3.3 begins. Note, NRC will require 1 week for this action, so work should be scheduled accordingly.

Estimated duration: 1 m Subtask 3.3: Code Development The p' rogram described in the SDID of Subtask 3.2 shall be written.

Estimated duration: 1.5 m

4 Subtask 3.4: = Testing The test plan described in Subtask 3.1 shall be implemented. If the chosen techniques,are found deficient for the objective of this task order, then attematives will be selected and all work repeated.

- Letter reports can be modified accordingly.

Estimated duration: 1.5 m -

- Subtask 3.5: Completion Letter Report

- A letter report that presents all results from the test plan shall be written. The report shall depict the curves compared and the corresponding code ' calculation results. Additionally, based on the code -.

results and the subjective rating system outlined in the attachment, suggestions on how the numerical results correlate to the rating system shall be given. At this time the program shall be delivered along with all experimental data, code calculation results, and all code inputs. That is, all files, scripts, and programs will be delivered to the NRC, so that NRC personnel can reprcduce the results presented in the completion report. This deliverable shall be submitted to NRC in coth text and electronic format.

Estimated duration: 1 m Task 4 ~

The results of this work shall be submitted as a paper to a conference to be agreed upon by the NRC. This conference shall peer review the paper and publish it in the conference proceedings.

Estimated duration:. 0.5 m Meetings and Travel 4

For successful completion of the work required in this contract, it is anticipated that the contractor shall travel to the NRC office in Rockville, MD at the times specified below. The contractor may propose additional travel deemed necessary for the successful completion of this effort. /J1 travel shall be approved in advance by the Project Manager, be specifically performed in support of this

- contract and be in compliance with Federal Transportation Regulations.

Tasks 1,2,3. Upon completion of Tasks 1,2and 3, one two-day trip for one person.

Attachment:

SBLOCA PIRT high ranked phenomena will be supplied by NRC PO upon request.

l 1