ML20040G408
| ML20040G408 | |
| Person / Time | |
|---|---|
| Issue date: | 06/25/1981 |
| From: | Mendelsohn B NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | |
| References | |
| NUDOCS 8202120283 | |
| Download: ML20040G408 (6) | |
Text
m -
1 FAX to: Ronald A. Harlan I
ANS - Tech. Comm.
SUMMARY
COVER SHEE B.TS rIbeNoIt FROM:
i S
IKNRP A'
d CONTRIQUTED PAPER O INVITED PAPER b i
67R
~
h ORIGINAL AND THREE COPIES REQUIRED h GETTING SAFEGUARDS CREDIT FOR GOOD PROCESS M0tilTORING
.rl rIE
?
AUTHOR (S):(List authors in the proper order and exactly as they are to be published. PLACE AN ASTERISK AFTER EACll AUTIIOR WiiO 15 AN ANS MEMBER; AN "S" AFTER STUDENT AUTHOR.)
Barry T. Mendelsohn, Senior Systems Analyst, Division of Safeguards 4, - -
L N
2.
f Y
UCC'y g g-) _ $
3-(
f AFFILIATION (S): (List corresponding author's af61iation and complete maihng address. @
C cc E
U.S. Nuclear Regulatory Commission, Washington,,D.C., 20555 l 7982h Q 1.
h f
b-2.
a k
3 Indicate number of author to whom correspondence should be addressed
_ and con
, d$
r NRC Division of Accounting, CON
(
To whom should the bage char.ge be bilked? Station L-3 6 Washington, DC 20555 Mai L
i Dilling address:
Preferred: Attach purchase order with appropriate purdtase order number to original copy of the summary.
f Number of: Pages Tables Figures Q
Word Count: Text
+ (No. of figures plus tables) x 150
+ (No. oflines of equations x 10) t' Total
$1.
d=
'Originalline drawings or glossy black-and-whit prints of each Ogure must be attached to original.
+
h4 FOR CONTRIBtTTED SU5f M ARY:
3 Identify ANS Division or Technical Group having cognizance of your subject l
g.
In which subject category (category number from p. 3) do you feel this summary belongs?
t r
N Alternative Category Number:
4 lias the substance of this summary tren presented or published previously (including U.S. DOE or equivalent
[
reports)?
E{
YES NO Give details j
Has she, paper been submitted for publication in a technical journal?
YES NO Give details Have you presented related papers?
YES NO Give details e
Has this summary been approved for publication by your institution or company?
o l
o YES NO Give details t
G FOR INVITFD SUMM ARY-Which ANS Division or Tecimical Group invited you? echnical Committee on Nuclear Safeguards
- =-} -
l g
T Ronald A. Harlan l
gg Persori who ;nvited you Session No. 8.5
\\
a,-,
I wr A COMPLETED
SUMMARY
COVER SIIEET, TOGETilER WITil Tile INFORMATION REQUESTED
$g ON PAGE 4. MUST DE ATTACliED TO EACH OF Tile FOUR SETS OF TIIE
SUMMARY
. Please have No copics made to complete your four sets.
l GL
~2-l
m TITLE:
GETTING: SAFEGUARDS CREDIT FOR GOOD PROCESS MONITORING' AUTHOR: Barry T. Mendelsohn i
Submitted to Transactions of the haerican Nuclear Society for publication.
This preprint (draft) summary is not.to be cited or repro-duced.
This 'is a preprinti (draft) of a summary intended for pubit-cation in a journal or proceedings.
Since changes may k made before publication, this preprint'(draft) is made a-vailable with the understanding that it will not be cited or_ reproduced without permission of the author.
DRAFT 6/25/81 DRAFT
1 For n.any, years it has been recognized that a significant loss of special nu-clear material (SN!!) during processing of nuclear fuel would likely be noticed by the personnel who monitor the yield of the process line.
It has also been recognized that if a diverter attempted to conceal a loss by substitution of inert paterial for SNM, or SNM of different isotopic content for SW of greater strategic value, then this would probably be detected during quality control monitoring of the product.
In 1978 an NRC special task force on NClA wrote:
"A major new thrust for safeguards recomirended in this report is the monitort g of the status of material during intervals between physi-cal inventories. This would be achieved through the development of a monitoring program to track material through a process to detect irregularities which may have safeguards significance. The pur-pose of this is to provide increased vigilance over material status between physical inventories... NRC assessment reviews have indi-cated that considerable process control infonration exists that could be used to detect irregularities in material status between physical inventories."Il)
As a result of studies made for and by the NRC to see how to use these process monitoring capabilities as either an addition to or a substitute for other safeguards regulations, it became obvious that with a few modifications the existing process monitoring systems could be credited with good safeguards value.
One of the modifications needed to allow process monitoring to fulfill its safe-guards potential has to do with what action is taken when an irregularity in material status is discovered.
If production yield is abnormally low one
2 day, the production manager's normal response may be to reschedule some work so that he ca'tches up'th' next day.
If it is low over an extended period of time,
e he may have a study made to see what the problem is.
The study may consider worker performance, attitude and training.
It may consider changes to proce-dures or equipment that promise higher yield.
But it would not ordinarily consider the possibility of deliberate diversion.
For a process monitoring system to be a good safeguards system, it must include plans and procedures to i
respond to anomalies in a manner that investigates the likelihood that the anomaly is a symptom caused by a diversion. The investigative response l
procedures ideally should be able to lead to a conclusive determination of l
whether an anomaly was due to theft or to innocent cause.
Pragmatically, this l
1s too much to ask of a facility's operator.
It is sufficient for the facili-ty's response system to be able to provide evidence to substantiate conclusions l
that the anomaly was caused by excessive. holdup, bookkeeping error, material handling.. error, etc., and to be able to provide information to help focus any criminal. investigation that might be called for in the event of an enresolved anorraly.
A second modification involves the decision criteria used for deciding if an irre-gularity is large enough to warrant an investigative response. A goal of a quality control program is to keep the process in a state of " statistical control " i.e., to limit variability to a well-characterized stochastic behavior.
This is accomplished by setting control limits, which, if exceeded, trigger a check for sources of " assignable variation."
It is common industrial practice to use three-sigma values for these control limits. This is because with three-sigma limits the facility manager may be highly confident that the process will not be de-clared out of control when, if fact, it is actually in control. In hypothesis testing
m i
3 1
0 t
t terminology, the null hypothesis is that the process is behaving normally, and the critical values of the hypothesis test are set to keep alpha, the proba-bility of a Type I error, small.
But this does not address the likelihood that the process will be declared in control when it is actually out of control.
For safeguards it is important to have a low value of beta, the probability of a Type 11 error. Control limits must be so established to give a high " power of the test."
Because existing process monitoring systems were established to turn out a quality product in an efficient manner, their emphasis is on the intermediate and final forms of the product. This creates a third area where a modification is needed. Namely additional moritoring of sidestream materials. Although rejected because of being "off-spec," scrap materials may be as attractive as product to a thief.
Using either for fabrication of a crube bomb would require about the same amount of skill and resources.
Therefore some measurement points for monitoring of certain sidestream materials will often need to be added.
The fourth modification needed involves separation of functions and overchecks.
If foreen or operators report measurement data to management that indicate poor yield, and as a result get " leaned on" too hard, they may develop a tendency to falsify those measurements.
It would be even more likely to expect such falsification if the poor yield was due to their intential diverting some of the material.
It is necessary to either separate the responsibilities for collection and reporting of information that could cause an investigation, or to provide overchecks that would eventually disclose falsified reports.
, t[
^ REFERENCES (1) NUREG-0450 Report of the Material Control' ~ an'd Material Accounting' Task ~
force, April.1978. Vol. 3, " Blueprint for the future," pp V-14 V,
4
~_