ML19309C043
| ML19309C043 | |
| Person / Time | |
|---|---|
| Site: | Wood River Junction |
| Issue date: | 03/19/1980 |
| From: | Partlow J NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Bowers C UNC RECOVERY SYSTEMS, UNITED NUCLEAR CORP. (SUBS. OF UNC, INC.) |
| References | |
| NUDOCS 8004080072 | |
| Download: ML19309C043 (3) | |
Text
_
~
5)f')g s
- ,.a ne o UNITED STATES
,g y'
g NUCLEAR REGULATORY COMMISSION l
W A$6C NGTON, D. C. 20555
/
MAR 1 S 1930 SGML:CWE 70-820 United Nuclear Corporation Recovery Systems Division ATTH: Mr. C. E. Bowers President One Narragansett Trail Wood River Junction, Rhode Island 02894 Gentlemen:
Our continuing review of your material control and accounting program has identified several deficiencies which should be corrected.
In most cases, these deficiencies have been discussed with members of your staff, and it is our understanding that corrective action has already been taken for several.
The purpose of this letter is to document the deficiencies which have been identified as well as our position on the appropriate corrective actions.
Item one involves your method of controlling the NDA measurement systems used to determine the uranium-235 content of bottled residues, drunmed waste and filters.
For bottled residucs, two standards are measured each time the procedure is applied.
For drummed waste and filters, a single standard is measured twice weekly. The deficiency involves the number of standards, and how the standards data are used in detemining the magnitude of the bias correction. A single standard is clearly insufficient to monitor possible shifts in the slope or intercept of the calibration curve.
Two standards are an improvement but are insufficient to determine the calibration curve.
(Ref. 70.57 Review Criteria, para. 4.2.3.1.b).
In order to adequately monitor the calibration during the period and determine the magnitude of any bias at the end of the period, the standards used to establish the initial calibration should be measured during the period.
This control data can then be used to calculate a new regression line from which a reliable estimate of systematic error and bias can be made.
In connection with this item, you should note that in Section 4.2.4.2.12 of your Fundamental Nuclear Material Control Plan there is a commitment to employ the nethod of weighted least squares for all line and curve calibrations.
8 0 04 0g O yg
.~
f,
The second item involves your procedure for assessing the significance of shipper / receiver differences when the shipper's or receiver's limit of error is not provided. Your practice of doubling the given limit of error is not consistent with acceptable statistical practices and is not acceptable to the NRC. A better approximation would be obtained by multiplying the given limit of error by the square root of two. Even this procedure is somewhat questionable since it assumes that the shipper and receiver have the same limit of error.
In the majority of cases, we would not accept this practice. Ilowever, in view of the fact that one would expect your limit of error on dissolver solution to be significantly smaller than the limit of error on incoming scrap, we can accept this For practice for shipper / receiver difference evaluations on scrap.
other transactions where a single limit of error is available, significance should be judged on that value. The shipper / receiver comparison procedure l
should be described in detail in Chapter 8.0 of your FNMC Plan.
The next item involves your procedure for control of the assay scale.
A single standard approximating the gross weight of a full tank is checkweighed daily, usually at the beginning of a shift.
If the standard is off by more than the smallest scale division, the scale is readjusted to zero. The results for the remainder of the day are assumed to be unbiased.
If a bias is indicated the next day when the checkweighing is done, there is some question as to exactly when the bias developed.
Since the dissolver solutions have been further processed, there is no opportunity to recheck them, in view of the importance of this input measurement, it is our position that the control standard should be checkweighed more frequently. A checkweighing each time an input measure-ment is made would be appropriate. When this point was discussed with members of your staff, they concurred with our assessment and indicated that they did not believe it would require much effort to implement additional checkwelghings.
The last item involves certain aspects of your records system. Two records in particular are of concern to us, the sample control records and the With regard to the former, it appears that insufficient assay scale records.
records exist to properly control and account for the amount of material going to and coming from the laboratory. This item is particularly signi-ficant because of the location of the laboratory outside the material access area, and the fact that the chemist works alone. The assay scale records are a concern because of their limited distribution. The sole copy of this important record is forwarded to the plant chemist for his use in preparing dissolver composites.
If the record is misplaced or damaged, key input information is lost. We suggest that additional copies of this l
record be distributed to the Nuclear Material Control Group to guard against this possibility.
. The items noted above represent needed improvements in your material control and accounting program. Where corrective actions are proposed, they are not intended to represent the only alternative acceptable to us. Should you have any questions on possible alternatives, please contact Mr. C. Emeigh of Iqy staff.
Sincerely,
\\
/s f,
l ul / \\ inLW tw James G. Partl'ow, Chief Material Control and Accountability
' Licensing Branch