ML20137A289

Insp Rept 70-1100/85-07 on 851104-08.No Noncompliance Noted. Major Areas Inspected:Measurement Control & Statistics. Several Followup Items Identified Re Nda Random Measurement Error Require Licensee Mgt Attention

ML20137A289
Person / Time
Site:	07001100
Issue date:	01/03/1986
From:	David Jones, Keimig R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
To:
Shared Package
ML20137A274	List: ML20137A268 ML20137A289
References
70-1100-85-07, 70-1100-85-7, NUDOCS 8601140273
Download: ML20137A289 (4)

Text

- . - ..- ._ . - _ _ _ _ _ _ .. . - - _ .- _. _. _ . - - . .

i

1. pa KtogD UNITED STATES i

( # o NUCLEAR REGULATORY COMMISSION yN o REGION 11 g j 101 MARIETTA STREET.N.W.

• 's

• ATLANTA. GEORGI A 30323

\...../

1 Report No.: 70-1100/85-07 i Docket No.: 70-1100 License No.: SNM-1067 i

l Safeguards Group: III

, Licensee: Combustion Engineering, Inc.

i P. O. Box 500 j Windsor, Connecticut 06095 l

Facility Name: Combustion Engineering - Nuclear Fuel Manufacturing Inspection at: Windsor, Connecticut

Inspection Co ucted: November 4-8, 1985 Inspector: Wuw G $5 O. W. Jo,nes, Statist' ian . NRC Region II Date Signed i

Approved by: ..s =If /- J-# 4 R.JCKeimig,CMieffafeguardsSection Date Signed

SUMMARY

i Scope: Measurement control and statistics. The inspection involved 39 hours4.513889e-4 days <br />0.0108 hours <br />6.448413e-5 weeks <br />1.48395e-5 months <br /> by one NRC inspector and was conducted during regular hours.

Results: The licensee was in compliance with NRC requirements in the areas inspected. iiowever, several Inspector Followup Items were identified which require licensee management attention. -

i

! 0601140273Bf0 00 i

PDR ADOCK O PDR C

~ ._ . _ _

REPORT DETAILS .

1. Persons Contacted Licensee Employees

• H. T. Cohen, Supervisor, Nuclear Material Accountability ,

• 0. R. Masse, Specialist, Nuclear Material Accountability D. L. Parks, Manager, Nuclear Materials ..

• R. E. Sheeran, Manager, Licensing, Safety, Accountability Criticality, & Security  :

The inspector also interviewed several other licensee employees.

• Denotes those present at the exit interview. '

2. Exit Interview (30703)

The inspection scope and findings were summarized on November 8,1985,-with those persons indicated in paragraph 1 ebove. At no time during this inspection was written material provided to the licensee by the inspector.  !

3. Followup on Inspector laentitiea Itw,i (92701) [;

(0 pen) Inspector Followup Item (84-06-01): Verification of N0A random  ;

measurement error by comparison of NDA measured values to actual recovery values. During the inspection conducted on December 3-6, 1984, the licensee agreed to perform a study in which NOA measured values would be compared to  ;

actual recovery values in an effort to verify the licensee's NDA random  ;

measurement error. During this inspection the licensee indicated that the  ;

study had been initiated and would be completed by the end of this year  !

(1985). Therefore, this item will remain open pending NRC review of the i results of the study. ,

4. Measurement Control and Statistics (85209) r The licensee's statistical control program for the evaluation of measurement errors and biases, calculation of limits of error associated with inventory differences and receipts and for monitoring the quality of SNM account-ability measurements was inspected. The inspection consisted of an examination of the licensee's written procedures for performing the statistical calculations, the databases used to evaluate the statistical parameters, and the statistical methods applied in the evaluation of the databases. The performance of the program during the March 19 -

September 17, 1985 material balance period was inspected. The following issues were disclosed during the inspection and are considered as inspector followup items:

r 2

85-07-01 The licensee's written statistical procedures should be revised to include mathematical models of the error structure for each measurement system, methods for testing biases for statistical significance, the mathematical model of the limit of error associated with inventorj differences, computational method for establishing control chert limits, investigation and corrective actions to be taken when control measurements indicate that a measurement system is out of control, and methods of investigation and reconciliation of shipper-receiver differences.

85-07-02 The licensee's FNMC Plan and internal procedures should be revised to incorporate NRC's policy regarding response to ndications that a measurement system is out of control.

85-07-03 The licensee's method of calculating bias, systematic error and random error associated with net weight of U0, powder receipts should be revised to incorporate the uncertaindes associated with the tare weight measurements performed by the licensee's fuel vendor.

85-07-04 The licensee's method of calculating the uncertainty associated with the assigned value of the standard used for controlling uranium concentration analysis should be revised. The standard deviation associated with the mean should be used rather than the standard deviation of an individual measurement.

85-07-05 The licensee's method of calculating the random sampling and analytical error associated with enrichment measurements of U0 pellets should be revised. The licensee routinely selected sik samples from each pellet lot for enrichment analysis. Four samples were analyzed by the Teledyne laboratory and two were analyzed by the Ledoux Laboratory. However, the licensee used only the Ledoux analyses in the random error calculations. Pooled data from both laboratories should have been used in the random error calculations.

85-07-06 The licensee's method of calculating the bias associated with NDA measurements should be revised, lhe measurement control data generated during a material balance period should be used to recalculate the calibration curve and the bias determined from the shif t in the calibration curve at the average SNM value measured during the period. The bias should then be applied to the total quantity of SNM measured during the period. The licensee's current method of calculating the overall mean of the relative difference between measurement of standards and the assigned value of standards results in a disproportionately weighted estimate of the measurement system bias.

85-07-07 The licensee should establish the uncertainty associated with the values assigned the standards used to calibrate and control the NDA measurement system and determine whether that uncertainty

l

~

3  !

l l

should be included in the estimate of the systematic error associated with the measurement system. There was also a second issue pertaining to the NDA measurement uncertainties. The NDA system measures the U-235 content of SNM bearing items and the uranium content is determined by dividing the U-235 content by an estimate of the enrichment of the items. Therefore, the uranium content is uncertain by a combination of the uncertainties in the NDA measurement and the enrichment value. The licensee should determine whether the uncertainty in the enrichment value is a significant contributor to the uncertainty in the uranium value and, if so, include it in the calculation of the uranium values' measurement errors.

85-07-08 The licensee should revise the method of performing LEID calcula-tions such that the LE is calculated around the plant ID instead of around the processing area activity. For the material balance period being inspected, it appeared that, as a result of the licen-see's current method, the coefficient, i.e. the material quantity to which measurement error applies, for U02 powder was overstated by approximately 7,000 kg U.

85-07-09 The licensee should revise the method of establishing control chart limits for scales. The control limits should be set at 0.05 and 0.001 levels of significance rather than at 0.0456 and 0.0026 levels of significance.

85-07-10 The licensee should also revise the method of establishing control chart limits for the NDA system. The control limits should be set at 0.05 and 0.001 levels of significance rather than at the 0.317 level of significance.

85-07-11 The licensee's FNMC Plan should be revised to include a commitment ta establish control charts for monitoring replicate measurements of process materials which are sampled and analyzed.

85-07-12 The licensee should revise the methods and procedures for evaluating shipper receiver differences to include provisions for using the most recent estimates of measurement errors when calculating limits of error for shipments ,and receipts. The procedure should delineate methods for review and evaluation of shipper-receiver differences on an individual container or lot basis, as appropriate, on a shipment basis, and on a cumulative basis for shipments of like material type. The procedure should also include provisions for investigation and reconciliation of statistically significant shipper-receiver differences.

The databases used by the licensee for statistical calculations appeared to be sufficient. Most of the calculations were performed by an on-site computer. Many of the calculations were verified by the inspector and found to be accurate.

1