ML20033G571
| ML20033G571 | |
| Person / Time | |
|---|---|
| Issue date: | 02/23/1990 |
| From: | Cilimberg R, Cwalina G Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20033G570 | List: |
| References | |
| REF-QA-99901175 99901175-90-01, 99901175-90-1, NUDOCS 9004100301 | |
| Download: ML20033G571 (9) | |
Text
_
)
ENCLOSIFRE ORGANIZA710N:
CRUCIBLE $PECIALTY METALS 5YRACUSE htW YORK i
~
J REPORT INSPECTION Ik$PECTION l
i NO.: 99901175/89 01 DATES:
12/18 21/89 ON 517,t_ HOUR $t 24 1
I i
CORRESPONDENCE ADDR[$$: Mr. Francis J. Petro, President i
Crucible Specialty Metals Division l
Crucible Materials Corporation P. O. Box 977, State Fair Boulevard j
Syracuse, New York 13201 0977 I
i I
i ORGANIZATIONAL CONTACT: William H. Shirtz l
T[LEPHONE NUMBER:
(315) 487 4111 i
l NUCL[AR !NDU$TRY ACTIVITY: Manufacturer of steel bars for use in components i
for townercial nuclear application, j
i I
i I
i i
i l
i AS$1GHED INSPECTOR:
4%%
h</m
- Yfo l
l R. L. CilimbergT Reactive Ighpection 5ection No. 1 Date (Ris1)
OTHER INSPECTOR:
P. Soo, Consultant N34 APPROVED BY:
rr G. C. Cwalina, Acting Chief, R15 1, Vendor inspection ae Branch l
i INSPECTION BA$t$ AND SCOPE:
A.
BASES:
10 CFR Part 21 10 CFR Part 50, Appendix 88 and A$ME 111, Sub.
I section NCA, Article NCA 3800.
l l
B. SCOPE: Obtain information about an allegation that Crucible $pecialty i
i MetaTs TC$M) issued test reports without performing the tests..
i i
l l
i i
PLANT SITE APPLICABILITY: Multi-plant.
l l
j l
i
\\
t "9004100301 900328 PDR COMMS NRCC CORRE$PONDENCE PDC
.,,-.,,._.-,-,_.w.,--
.,n.-.,,-
,n,.
\\
l 4
j j
ORGAN!ZATION:
CRUC15LE SPECI AL11T METAL 5 DIVl$10N I
5YRACU$E, NEW YORK 13201 0977 4
l ELFQM IN5FLCT10h NO.: 99901175/89 01 RESULTS:
PAGE 2 of 9 i
i j
A. v10L AT10Ns:
t None B. NONCONFORMANCES:
1.
Contrary to Criterion V of Appendix B to 10 CFR Part 50 CSM pre-paredaqualitycontiolstandardthatcontainedillegibleinstruc-tional diagrams.
(89 01 01) 2.
Contrary to Criterion XII of Appendix B to 10 CFR Part 50 CSM had 1
an extensometer calibrated without receiving documentation from the calibration laboratory that the calibration standards were traceable to the National Bureau of Standards.
(89 01 02) 3.
Contrary to Criterion Vll of Appendix B to 10 CFR Part 50 CSM did d
j not mark test samples in a rr.anner which avoids the potentia 1 for j
inadvertent mixing.
(890103) i 4
Contrary to Criterion V of Appendix B to 10 CFR Part 50 CSM pre.
pared a quality control standard that was ambiguous and could lead l
]
to inaccurate determinations of the reduction-in area for a tested specimen.
(8901-04)
C.
UNRESOLVED ITEM:
1 l
After reviewing QC 2010, ' Tension Testing of Metallic Materials-Room i
Temperature," Revision 4, dated August 23, 1989, and the procedure for calculating the 0.2 percent offset yield strength, a possible noncon-formance was identified.
The resolution of this item requires addi-tional study by CSM in conjunction with their calibration laboratory.
This item concerns the formula for determining the location of the 0.2 percent offset line on the load displacement graph that is generated during the test. The Formula is provided on page 6 of QC 2010, Revision 4 The offset distance will vary, depending on the specimen gagelen From the procedure table,gth and the speed of the recorder chart.
which gives sample calculations, the most comonly used 2 inch gegelength specimen will require an offset distance equal to 1 inch provided the recorder magnification is set at 250. Mr. Leonard Franty, the person conducting the test, stated that he always uses the offset
s
}
ORGANIZATION:
CRUCitLE $FECIAllTY METALS DIV1510N I
$YRACUSE, NEW YORK 13201 0977
\\
hLiUhl IN5PI,cIlQh NO : 99901175/69 01 RESULT $:
FAGE 3 of 9 I
i i
distances specified in the table which means that the recorder magni.
fication is always assumed to be 250.
However, from additional dis.
l cussions with the QC Oepartment staff, it was not clear whether 4 l
different recorder magnification is automatically selected by the i
tensile machine if the load range for a test is changed by the t
operator.
The particular concern in this respect is that there may be a situation i
where the true recorder magnification is 125.
Following the instrue.
tions in QC.2010 Revision 4, the offset distance for a 2. inch gege-i length would be 0.5 inch rather than the 1 inch value routinely used.
Therefore, the true yield strength obtained through the use of the 1
0.5. inch distance. is lower than the value which would normally be
- obtained, If such a situation is shown to exist in the laboratory then it is possible that some materials were certified as meeting,
l minimum yield strength requirements when in actual fact they may not.
)
D.
STATUS OF PREV 10VS INSPECTION FINDING $:
Not applicable.
E.
INSPECTION FINDINGS AND OTHER COMMENTS:
1.
The NRC staff informed CSM management representatives of the scope of the inspection during the entrance meeting on December 18,1989, and summarized the inspection findings during the exit meeting on Decerber 21, 1989, 2.
Allegation The NRC received a copy of a newspaper article in an October 19, 1989 issue of the Syracuse Herald Journal. The article reported that three ernployees and a former employee at CSM alleged that CSM was falsifying steel strength test reports until 1985.
The falsi-fication was alleged to be the reporting of the results of testing on tensile test reports for steel samples that were not actually tested.
The NRC scheduled this inspection to follow-up on the allegations described in the news article.
The inspectors selected 1983 as a representative year to sample because it preceded by 2 years the alleged end date of the falsification and involved 223 nuclear test reports which were 13.8 percent of the 1612 total test reports issued by CSM during 1983. As discussed in Section 3 below, the review of 223 nuclear tensile test reports for 1983 b
i I
ORGANIZATION: CRUC1BLE $pECIALITY METAL $ DlV1510N SYRACUSE, NEW YORK 13201 0977 htFvhl lh5FLcliQN NO.: 9990117$/89 01 RESULTS:
PAGE 4 of 9 determined that a machine plotted stress strain curve was attached to all tensile test reports which supports that tensile tests were i
performed on samples of material that C$M supplied for nuclear application.
A meeting with Defense Contract Administration 5ervice (DCAS) representatives confirmed that thousands of tensile tests were witnessed by DCAS with no indication that results were i
reported without tensile tests being performed. DCAS experience indicates that no field complaints have been received concerning tensile test results.
Review of all 1983 test reports and actual i
observation of testing by DCA$ indicates that testing was performed t
to provide the results that were reported.
The allegation could not be substantiated.
3.
Document Review j
When a tensile test is performed at C$M, a stress strain curve is plotted by the tensile testing machine. The inspectors hypothe.
i stred that if an original stress strain curve was attached to all of the tensile test reports in the CSM files for a random year l
prior to 1985, then evidence is available that supports a con-clusion that tensile testing was performed by CSM to support i
reported results.
The NRC inspectors reviewed 1612 test reports in the CSM files for material shipped in 1983.
The inspectors deter-i mined that 223 of the 1612 test reports were marked with double j
letter designations which identifies material for nuclear appli-1 cation. The inspectors confirmed that an original machine gen-ersted stress strain curve was attached to all 223 nuclear test reports. In addition machine generated stress strain curves were attached to practically all of the 1389 test reports for conrnercial l
orders. The availability of stress strain curves for the majority of corrnercial orders supports a conclusion that CSM testing prace tice is to perform the test before issuing a test report to docu-ment the results of the test performed. The inspectors were unable to find any incident where CSM supplied a test report on nuclear material without a stress-strain curve which substantiated that the tensile tests were performed.
4.
DCA$ Experience The NRC inspector met with Messrs. Thomas Hearn,lity assurance Richard Moressh, and John Mathews of DCAS.
DCAS monitors CSM qua activities related to the production of material for U. 5. Navy nuclear and other defense applications. The DCAS personnel stated P
. -. ~. - - - -, - -. -,, -., - - - - - -. -.,
m-m
,--...w,-.,r-
i ORGANIZATION:
CRUClBLE $PECIALITY METALS DIYl510N i
j SYRACUSE, HEW YORK 13201 0977 hu uh3 mFLUWh NO.:
99901175/89 01 RESULTS:
PAGE $ of 9 i
i that they had witnessed thousands of tests without developing any l
evidence that would support the allegation discussed above. OCAS i
l personnel stated that they had never received a complaint from the field pertaining to tensile testing of CSM manuf actured material.
Mr. Tom Hearn, Chief of DCA$ Branch A offered to assist the NRC r
with any NRC action concerning CSM activities.
i 5.
Reviews of Quality Assurance Procedures l
Quality control procedures were reviewed as follows: QC 7210,
{
'ldentification Standard for the Metallurgical Laboratory",
l Revision 2, dated April 4,1988, requires that all specimens be stamped with either the heat and bar number or with a letter or i
number code plus the bar number.
QC 2009, ' Obtaining Test Samples for Mechanical Testing *, Revision 0, dated March 27, 1978, details acceptabletechniquesforcuttingrepresentativetestspecimens from production bars. QC 2000, Heat Treatment Procedure for Steel Test Samples in the Metallurgical Laboratory *, Revision 11, dated January 26,1984, gives guidance on the c411bration of heat treet-ment furnaces, the selection and location of thermocouples on the specimens undergoing heat treatment, and the times and rates of cooling for heat treatment. QC 1000, ' Calibration System for Measuring and Test Equipment". Revision 8 dated September 20 1988, describes the system for controlling and calibrating all measuring and test equipment in the QC Department. The calibration
,.of both primary and worting standards must be carried out at 2 year intervals by a commercial laboratory capable of certification with traceability to the National Bureau of Standards (NBS). A certi-ficate of inspection is to be kept on file by the QC Department for each standard or set of standards. QC 1001, ' Calibration Procedure for Micrometers in the Quality Control Department" dated January 29, 1985, outlines the technique for. Revision 4, calibrating a micrometer screw gage, an instrument used to measure lengths or diameters, usually up to 1 inch with an accuracy of 0.0005 inch.
A set of two photocopied diagram,s is given to illustrate the i
I adjustment locations on the instrument. QC-1006, ' Calibration Procedure for Tensile and Stress Rupture Machines in the Metal-lurgical Laboratories', Revision 1, dated October 9,1984, covers i
test machines that must be calibrated annually by a comercial laboratory capable of certification with traceability to NBS. A copy of the certificate of calibration is to be kept on file in the f
r
t j
i ORGANIZATION:
CRUCIBLE SPECIALITY METALS DIYl$10N
$YRACU$E, NEW YORK 13201 0977 Ku o l Ih>FLcTIph I
NO.: 99901175/89 01 RESULTS:
PAGE 6 of 9 i
Metallurgical Laboratory and the date of the calibration noted on I
4 the test machine. QC 2010,
- Tension Testing of Metallic Materials.
Room Temperature", Revision 4, gives details for measuring standard tensile test specimens, loading them into the test machine, i
attaching extensometers to measure deformation (strain), and i
applying the load that eventually causes fracture. QC 2010 also l
describes how the yield strength, tensile strength, and elongation l
j at failure are to be calculated.
i i
j All of the QA procedures reviewed were of acceptable standard and j
should, if followed, be capable of providing accurate and con-i sistent results.
A possible exception is QC 2010 in which there is 1
en ambiguous instruction that could lead to incorrect results as discussed in section 7 below.
j 6.
Observation of Tensile Testing i
The inspectors observed two tensile tests which were carried out by Mr. Leonard Franty, special Inspector.
Also observing the tests were Mr. Clint Johnson, Turn Supervisor and his supervisor Ms. Dodie Korradi.
The first test was on a type 303 stainless r
steel specimen with a 2 inch pagelength.
The material was from mill order 79 11086 9 and was marked with a yellow letter C.
A i
2-inch test section was stamped on the gegelength using a cali-brated punch which gave two indentations 2 inches apart. The identations serve as a baseline for calculating the elongation, t
The diameter of the test specimen was measured by micrometer at the center of the reduced section of the sample. When questioned about i
the precise procedure being carried out, Mr. Franry stated that he measured the diameter many times until he was confident that he had determined the minimum value.
This measurement is used to compute the original cross sectional area. The tensile machine, a Baldwin hydraulle ty(0-24,000pe, was zeroed for load and set for the intermediate i
load range pounds).
The s pecimen was inserted into the i
load train and an extensometer attacied.
Both the machine and extensometer were within the required calibration periods.
i l
Mr. Franry set the dini controlling the speed of deformation at
- 10" until he judged that the 0.2 percent offset yield point had I
been exceeded.
The extensometer was removed and the machine set at 1
- 20* to give a faster deformation rate in order to reduce the remaining test time. When the specimen fractured, the two broken i
halves were held firmly together in a vise and two extra measure.
ments taken. One was a measurement of the new distance between the 1
I j
I i
j
_._,.,...,.,_m__.
-_____,_-m.
)
)
ORGANIZATION:
CRUCIBLE $PECIALITi' METAL $ DlVl$10N SYRACUSE, NEW YORK 13201 0977 KLtVhs 4 W LLIJUN l
NO.: 99901175/89 01 RESULTS:
PAGE 7 of g
{
i two punch marks that would give the percent elongation at fracture, and the other was a series of diametral measurements at the frac.
ture location.
Mr. Franry stated that he was again seeking the minimum diameter to use for calculating the final cross sectional area at the location of failure.
Hr. Franry then removed the load displacement curve from the strip.
i chart recorder and drew a one. inch offset line parallel to the linear elastic part of the trace. This gave the load that was used to calculate the 0.2 percent offset yield strength.
The maximum l
load from the graph was also noted since this would be used to obtain the pitimate tensile strength. When the load and elongation l
information and the original and final diameters were fed into a
{
desk. top computer, the appropriate strength and percent elongation values were printed on a sheet of paper.
This data sheet and the load displacement curve were attached to the paperwork accompanying the specimen, and the test was completed.
A second test was conducted in a manner identical to the first. At all times Mr. Franty displayed a large measure of dexterity and i
efficiency and was clearly very familiar with this type of testing.
in a subsequent conversation he stated that he had been doing the t
l tests for about 6 months. He was trained by another special e
l inspector and was supervised until he was fully qualified to run the test alone.
7.
Nonconformances Observed During Tensile Testing and Review of l
Procedures During the review of the quality control procedures and the obser-vation of tensile testing, four nonconformances were found as j
follows:
The photocopied figures in QC.1001, Revision 4, which describe how adjustments are made to recalibrate the micrometer are of poor quality.
The locations of the adjustment points are not clear because the figures are illegible. An inexperienced person would have difficulty ($ee Nonconformancefollowing the procedure to obtain a succ i
calibration.
890101)
The calibration of the extensometer and the associated strip chart recorder was performed by $atek Systems Incorporated in Grove City, Pennsylvania.
Examination of the calibration sheets showed that
ORGANIZATION:
CRUClBLE SPECIALITY METAL $ DIV1510N
$YRACUSE, NEW YORK 13201 0977 i
KLFVhl lh5FicT10N NO.: 99901175/89 01 RESULTS:
PAGE 8 of 9 l
the extensometer (P55M 1004) was certified fer Class 81, as l
defined in the A$TM E 83 standard.
However, there was no docu-l rentation to show that the c411bration instruments used by $4tek were calibrated to NBS standards.
(SeeNonconformance 89 01 02)
The first tensile specimen tested by Mr. Franry was marked with a yellow *C" on the gagelength.
Mr. Franry stated that it is pos.
I sible that several of the specimens in the batch of 15 that were to i
l be tested could have the same letter code.
If the specimens were l
mixed inadvertently then it would be possible to report incorrect test results to the customer. QC 7210, Revision 2, requires that test pieces be stamped either with heat and bar nue er or with a letter or nu2er code and bar nu2er.
The two samples tested i
during the inspection were not marked with a bar number, and the identification was painted rather than stamped on the samples as required by the procedure.
(SeeNonconformance 89-01 03)
The reduction in area value for a tested specimen is a measure of the material's ductility since it represents the ability of the specimen to neck down before fracture occurs.
It is defined in QC 2010 Revision 4 as follows:
I reduction in area e original area - final area X 100 original area i
A large value for the reduction in area indicates a resistance to brittle fracture. The original diameter of the gegelength and the diameter at the necked fracture location must be measured to calculate the original and final areas. QC 2010. Revision 4 requires that the cross-sectional area be determined by measuring the center of the reduced section and the reduction in area be determined by measuring the minimum cross section of the fractured tensile specimen which has been fitted together.
The staff of the metallurgical laboratory have incorrectly inter-i preted the terms reduced section and minimum cross section to mean j
that minimum diameters should always be measured rather than the average. This is contrary to the requirements of ASTM A 370 which i
states, ' Fit the ends of the fractured specimens together and measure the mean diameter or the width and thickness of the smal-i lest cross section to the same. accuracy as the original dimen.
[
stons.' The CSM practice may lead to a higher value for the 1
reduction in area.
Using the minimum diameter gives a smaller area 1
J i
i
ORGANIZATION:
CRUClBLE $PECIALITY METM.$ O!V1510N
$YRACUSE. NEW YORK 13201 0977 REF0h1 IN5PECTION NO.: 99901175/89 01 RESULTS:
PAGE 9 of 9 which results in a higher than actual reduction in area value.
This could occur whenever post test specimens have an oval shaped Cross section rather than a circular one at the failure point.
However, little error results from measuring the minipum diameter for the pretest specimen since tolerances during the machining of the specimens are tightly controlled and the minimum and everage diameters are almost identical.
(See Nonconformance 89 01 04)
F.
PERSONS CONTACTED:
L. Franty J. Funsch T. Hearn C. Johnson i
J. Kelahan D. Korradi J. Mathews R. Moreash W. Nesspor F. Petro W. Shirt 2 J. Wright I
l l
1
. -.