ML20153A971
| ML20153A971 | |
| Person / Time | |
|---|---|
| Issue date: | 11/14/1978 |
| From: | NRC OFFICE OF STANDARDS DEVELOPMENT |
| To: | |
| Shared Package | |
| ML19220A223 | List: |
| References | |
| REGGD-05.058, REGGD-5.058, NUDOCS 7811290027 | |
| Download: ML20153A971 (12) | |
Text
._
NOV 141978 VALUE IMPACT ANALYSIS FOR A REGULATORY GUIDE ON THE CONCEPT OF MEASUREMENT TRACEABILITY RG 5.58
. Introduction Implicit in the provisions of 10 CFR 70.57, Measurement Control Program for Special Nuclear Material Control and Accounting, is the requirement to establish traceability of measurement results to the " National Measurement System." Within the measurement control program, calibrations must be made relative to national standards or to nationally accepted measurement systems. A regulatory guide is needed to clarify the intent of these provisions in 10 CFR 70.57. There is a general lack of industrial understanding and agreement on wFat the " National Measurement System" is and what " National Standards" and " Nationally Accepted Measurement Systems" are. Even within the National Bureau of Standards, there is disagreement over the use of the term " traceability" and its implications. Within the nuclear industry there exists a need to establish a position on how such a common base is to be established for measurements made on nuclear materials. The proposed guide will address this issue. Once the concept is established, NUREG reports on methods for preparation of " Working Reference Materials" can be issued.
The guide would provide the general rationale for the detailed procedures in the reports.
It is essential that every measurer of a given type of material be able to make a statement about the uncertainty of his results relative to a common base.
This concept provides the basis of shipper-receiver comparisons, whether between plant areas, between plants, between companies, or between nations. The common base must, therefore, be some agreed upon standard. The concept of traceability, then, involves the ability to relate a measurer's results to such standards 7 8112 9 oc3:1;7
through an unbroken chain of comparisons, the uncertainties of which are cumula-tive. For many industries, there are typical reference materials available for which the cumulative uncertainties are known. These materials are used for the I
calibration of measurement systems. The problem in the nuclear industry is that i i
there are no reference materials available which are typical of some of the mate-rials forms encountered. Licensees must prepare their own reference materials .
t but in such a way as to maintain a chain of comparisons to some acceptable refer-ence materials. The methodology for preparing in-house nuclear reference materials has been developed but has not been well publicized. Consequently, a guide which describes this methodology in general terms is needed. NUREG reports will provide specific procedures for specific material forms.
Value A. Improvements in Effectiveness of Safeguards Systems
~
A person entrusted with valuable material normally takes an inventory periodi-cally to determine whether all the material can be accounted for. If it cannot, he can initiate efforts to recover what is missing, determine the cause of the '
loss, and take steps to correct any deficiencies found in his material protection procedures. One of the principal limitations of the materials accounting system in a nuclear fuel processing facility is that some disparity between the accounting records and a physical inventory is inevitable because of the limits of accuracy and precision of the process flow (inputs and outputs) and inventory measurements.
The overall effect of uncertainties in the various measurements of volume, mass, and material composition is an uncertainty in the material balance figure. In response to this reality, NRC has imposed regulations on the nuclear industry that require licensees to provide quantitative evidence that their observed material 2
i balance discrepancies, designated material unaccounted for (MUF), are not greater than the limiting value that could be expected because of measurement uncertainty.*
(That value is designated the limit of error of MUF (LEMUF)). If a MUF value exceeds that limit, it must be assumed that either a significant measurement or accounting error has been made or that a loss of material, possibly by theft, has occurred. The statistically sound way to test MUF values is to:
- 1. Calculate LEMUF (the statistical uncertainty of MUF) by propagation of the uncertainties of the measurements that affect the material balance.
- 2. Test MUF against LEMUF on the hypothesis that a loss of SNM has not occurred (i.e., that the MUF value is the result of measurement uncertainty only). ,
In order to perform the test, licensees must know the uncertainties, both systematic and random, associated with the SNM measurements that affect their material balances. If the SNM measurements used in the licensee facility are not being continually monitored for systematic errors, a measurement bias can occur resulting in LEMUFs that are invalid. An unknown measurement bias can persist for an unacceptably long time if the measurement systems are not constantly monitored.
Instances of measurement bias that were not known to the user are common and many such cases in nuclear facilities have been documented.** Furthermore, it has been observed that many licensees do not maintain close enough surveillance over their SNM measurement methods to insure the continuing reliability of calibration. On the other hand, it is known that some facilities have a good CRM program that o
See 10 CFR Part 70, Section 70.51.
"See, for example, (1) SALE Report 1, First Report on Error Components in the Analysis of Uranium and Plutonium Product Materials, F. H. Tingey et al.,
Aerojet Nuclear Co. and Allied Chemical Corp., April 1972, Idaho Falls, Idaho; (2) " Qualification of Umpire Laboratories for the Analysis of Uranium and Plutonium Materials, " Report XV, Third Report on Uranyl Nitrate, F. H.
Tingey et al., Aerojet Nuclear Co., Idaho Falls, Idaho, June 1971; and (3)
Personal Communication from Stanley Yamamura, Idaho Fr11s, Idaho, August 19.75.
3
maintains traceability which, along with other controls, helps identify biases.
The proposed regulatory guide will present conditions and procedural approaches acceptable to NRC for establishing and maintaining traceability of SNM control and accounting measurements as part of a measurement control program to assure freedom l
l from measurement bias complying with the material control requirements of 10 CFR Part70, paragraphs 70.51(e)and70.58(a),
- 8. Improved Safeguards Procedures i
In addition to meeting the need discussed above, the proposed measurement i
traceability guide will assist in maintaining a measurement control system that will:
- 1. Insure that control of MUF is on a consistently and quantitatively sound basis throughout the industry.
- 2. Detect biases promptly so that they can be corrected or eliminated, thereby reducing the potential for masking thefts and diversions.
- 3. Reduce the number of false alarms (indicating losses) caused by positive biases, thereby reducing the costs associated with extra inventories, cleanouts, etc.
C. Licersing or Inspection Benefit This guide will provide a clear benefit to licensing and inspection person-nel. In the case of licensing personnel, the guide provides a body of knowledge useful for comparison with submit't ed applications for SNM licenses in partial fulfillment of the requirements of 10 CFR 70.57, Measurement Control Program for SNM Control and Accounting.
This guide, where adopted by the licensee, will provide a clear statement of requirements in the area of personnel, and, if not adopted,will aid licensing 4
9 e
personnel in reviewing applications to insure that inspection personnel are pro-vided license requirements with which to ascertain the adequacy of certified reference materials (CRMs) and techniques.
Moreover, in cases where improved calibration results, greater reliance on this phase of the material ba' lance will allow inspection personnel to better localize any apparent unaccounted for losses and thus seek appropriate recertification.
D. . Reduction in Regulatory Control No increase in regulatory control is anticipated as a result of this guide.
E. Policy Implementation and Definition Paragraph 70.57(c) requires certain licensees to submit plans describing their Measurement Control Program for Special Nuclear Material Control and Accounting, required by paragraph 70.57(b). This guide provides assistance to licensees in implementing paragraph 70.57(b)(5)(iv) with regard to CRM traceability.
Impact A. NRC Staff Resources The information used in writing this guide has already been provided by the contractor. As such, principal expenditure of resources has occurred. Expendi-tures to publish form should be limited to expenses of review and actual publish-ing expense, which is only a fraction of costs already incurred.
B. Impact on Licensing Process Impact on licensing process should be to facilitate review effort in this area and thus reduce licensing effort.
C. Impact on Operating Facilities This guide provides comprehensive information to licensees in an area that has lacked such information. Reduction in costs of preparing licensing submittals 5
w l
l and procedures should result. Additionally, avoidance of possible pitfalls should ;
occur, whereby costs incident to poor measurements such as significant MUF requir-ing an investigation and reinventory should be avoided.
D. Impact on Employment and Labor No impact on employment and labor is anticipated.
E. Societary Impact The primary impact on society is increased protection provided by improved material containment through improved loss detection capability. -
F. Cost to Licensees
- 1. Summary The estimated average incremental cost of carrying out the proposed considerations for establishing traceability of SNM accounting measurements is 20 cents per kg of LWR fuel manufactured in a typical 1 MT/ day facility. The approxi-mate annual cost would be $61,530. Some licensees will incur little additional cost beyond that of formalizing and documenting their existing measurement trace-ability programs but those whose current programs are very limited will incur higher
~
incremental costs.
- 2. Discussion The base case for the cost estimate was fuel fabrication for light water ..!
reactors.
- a. Reference Process j To provide a basis for the cost estimate, a reference process and a
^
conceptual p: ant for manufacturing uranium oxide fuel at approximately three weight percent U02 was selected. The design capacity is one metric ton per day, and it was assumed that the plant operates three shifts / day seven days a week with about a 300 MT annual production (approximately 80 percent operating efficiency).
I 6
The input to the process is enriched UO2 . After some physical preparation steps (drying, grinding, and screening, if necessary) this material is dry blended in batches of about 250 kg each. Fuel pellets are made from the blend by slugging, granulating, pelleting, sintering, and grinding of fixed dimensions. The final product is sintered-pellet UO 2 fuel clad in zirconium alloy tubes and assembled into fuel elements in accordance with customers' specifications.
The plant is assumed to have a wet process for recovering uranium from dirty scrap. The recovered uranium is converted to UO2 by the standard diuranate precipitation-calcination / reduction process. The recovered material is then recycled to the uranium feed points in the fabrication process. The average scrap recycle rate is assumed to be 10 percent of the plant feed rate,
- b. SNM Control and Accounting System Material balance areas (MBAs) and key measurement points were selected in accordance with Regulatory Guide 5.26.
Five MBAs (Blending-Pelleting, Fuel Assembling, Scrap Recovery, Waste Treatment, and Analytical Laboratory) and three item control areas (ICAs)
(UO2 Feed Storage, Finished Pellet Storage, and Finished Assemblies Storage) were designated. Conversion of recovered uranium to the oxide is done in the scrap recovery MBA. SNM accountability measurements points, 4 MBA product output points, 4 MBA scrap removal points and 7 MBA waste removal points.
Weighing stations were assumed to be located at 7 key measurement points, volume measurement instrumentation at 4 points, sampling stations at 8 points, and nondestructive assay (NOA) measurement capability at 9 points.
7
- c. The Proposed Measurement Traceability Program The measurement of standards allows the calculation of bias correc-tion factors and the uncertainties of calibrations. Periodic special tests of sampling and measurement procedures, recalibrations of volume measurement vessels and in-process holdups, and tests and recalibrations of scales are recommended.
The assumptions for these measurement traceability activities within the measurement control program are:
. NDA, weight, and volume measurements are done for each measurement system.
. Each scale, NDA instrument, and mass spectrometer is monitored with a Reference Material (RM) that simulates (to the extent possible) the appropriate process material in form and composition.
. Each chemical assay procedure is monitored with an RM that simulates (to the extent possible) the appropriate process material in form and composition.
. Volume measurement instrumentation is checked with appropriate stand-ards. The tank calibrations are also checked.
. Special measurement method tests involving additional sampling and analysis and comparative analysis by independent methods are made to evaluate the accuracy of sampling, chemical assay, and NDA methods.
The assigned frequencies of calibrations and analyses of RMs will normally supply enough measurement data to control bias and estimate systematic error uncertainties of the various measurements independently in each inventory period. This schedule assumes a stable situation with measurement quality under good control. It should be noted, however, that much higher frequencies may be 8
required during startup of a new facil.ity, or during testing of new measurement equipment and methods, or when a conshierable fraction of the personnel perform-ing measurements have limited experiet.ces. Additional program efforts during such periods were not included in this estimate.
It was estimated that about 0.9 technical man-year annually is required .for methods evaluation, preparation of working standards, and liaison with the process function,
- d. Current Measurement Traceabilty Practice .
l In order to estimate the incremental cost of the proposed measure-ment traceability program over and above current measurement traceability prac-tices, it is necessary to define " average current practice." 'In reality, the extent of measurement traceability in effect at the present time varies widely from facility to facility, and essentially no factual information is available on the current practices in industry. Therefore, it was necessary to make a somewhat arbitrary judgment as to the features of the average current program. The assumed features of that program are:
Sampling error is generally not monitored but is considered to be either
! negligible, relative to assay error, or constant at the value determined when the sampling procedure was developed and tested.
Reevaluations of sampling error are done only wnen troubles traceable to measurement data arise.
Scales, NDA instruments, and volumetric instruments are tested with RMs.
SNM assay methods are tested twice a week.
The standards used for monitoring assay procedures are usually pure mate-rials and are not true working RMs in the sense that they do not adequately )
9
simulate orocess materials. (Therecommendedmeasurementtraceabilityprogram emphasized the need for matching working RMs). Also, current practices do not usually include a formal method testing and comparative measurements program for continuous monitoring for biases. However, licensees are required to test for bia's and correct any bias that exceeds 10% of its standard deviation,
- e. . Incremental Measurement Traceability Costs for the Referenced Plant Using the assumptions in the foregoing sections the incremental annual operating cost for a normal measurement control program was estimated to be 1
$61,530 or $0.20 per keg of fuel. A detailing of the estimated costs of the various measurement control program elements and the assumed frequencies of measurements, tests and studies are summarized in Table 1. These costs reflect i
only the incremental measurement control activities over and above those which are assumed to be carried out in the average facility for either process and product quality control or measurement control activities to meet current regulations.
10 ,
TABLE 1. Summary of Costs for Additional Measurement Traceability Required to meet the Proposed Guidelines
- Number of Measurements Annual or Tests Cost Activity Per Year ($/ Year)
Routine measurements of standards for determination and control of biases (Cost of RMs)
- a. Assays and isotopic analyses 490 2,880
- b. Bulk mass measurements 2600 1 ,040
- c. Bulk volume measurements 86 260 Periodic calibrations of scales and 11 600 l volume measurement equipment Periodic evaluation of sampling 20 tests 750 quality Evaluation of measurement quality 5-20 studies 11,000 through special studies including interlaboratory exchange programs Evaluation of NDA by cr,mcarative 200 10,000 analysis (including n ";OfRMs)
Preparation and main. . nance of working 10 batches 35,000 reference materials TOTAL COSTS $61,530 o
For the reference 300 MT/ year uranium oxide fuel fabrication plant.
11
The unit cost for measurements and calibrations were estimated from private information obtained from various ERDA contractors and from a private laboratory which performs measurement services for licensees. The costs include all over-l . heads, facility and equipment charges, and material necessary for full-cost recovery.
The cost for various test equipment items and primary standards is estimated to be about $10,000.
Summary Statement of Conclusions The estimated average incremental cost of carrying out the proposed considera-tions for establishing traceability of SNM accounting measurements is 20 cents per kg of LWR fuel manufactured in a typical 1 MT/ day facility.
The average cost per kg of manufactured LWR fuel is $850/kg. Therefore the incremental cost would be approximately 0.002% of the average fuel cost. The establishment of traceability as part of a measurement control program will result in fewer biases and consequently a savings in effort and cost by reducing the frequency of false alarms and cleanouts. It can therefore be concluded that the smallincrementalcost(approximately$200/ day)isworththeincreaseincapability of detecting possible theft or diversion of SNM and provides possible savings in cost and effort by reducing the frequency of false alarms and cleanout.
12
- _ - _ _ _ _ _ _ _ _ _ _ _ _