IR 05000329/1980013
| ML19332B279 | |
| Person / Time | |
|---|---|
| Site: | Midland |
| Issue date: | 08/13/1980 |
| From: | Erb C, Foster J, Norelius C NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML19332B277 | List: |
| References | |
| 50-329-80-13, 50-330-80-14, NUDOCS 8009260415 | |
| Download: ML19332B279 (23) | |
Text
U.S. NUCIEAR REGULATORY COMMISSION
,
OFFICE OF INSPECTION AND ENFORCEMENT
REGION III
~
Report No. 50-329/80-13; 50-330/80-14 Docket No. 50-329; 50-330 License No. CPPR-81; CPPR-82 Licensee:
Consumers Power Company 1945 West Parnall Road Jackson, MI 49201 Facility Name: Midland, Units 1 and 2 Investigation At: Midland Site, Midland Southern Bolt Company, Shreveport, LA J. W. Rex Company Lansdale, PA Mississippi Valley Structural Steel Co.
St. Louis, M0 Investigation Conducted: February 27-29, March 5-6, 12-13, 20, April 18, and May 2, 1980 Investigator:
h[
M f!// 7C J
Foster
~
/ Dalfe Inspector:
// 70
'
C. M. Erb tat 6 I 3 Reviewed by:
-
-
C. E. Norelius Date Assistant to the Director
D. H. Danielson,' Chief Date Engineering Support Section 2 Investigation Summary Investigation on February 27-29, March 5-(,
12-13, 20, April 18, May 2, 1980 (Report Nos. 50-329/80-13; 50-330/80-14) ~
8009260 i k
_.... _.. _...
.
-
__
_ _ _ _ _ _ _ _ _ _ _, _ _ _ _
-
_
_
.
Areas Investigated:
Special, announced investigation concerning manufacture
,
and installation of reactor pressure vessel holddown studs utilized in Midland Unit 1.
The investigation required 150 inspector hours by two NRC personnel.
Results: us the areas investigated, 3 items of noncompliance were identified:
(Infraction - Inadequate Procurement Document Control - Details'section, Para-graph 6b; Infraction - Inadequate Control of Special Processes - Details Sec-tion, Paragraphs 6d, 6e, 6f; Infraction - Inadequate Control of Purchased Material, Equipment and Services - Details section, Paragraph 6f).
!
-2-
-..
. -..,..
_. _ _..... _ _
. - _.. _ _.. _. _
.
_
_ _ _ _
.
REASON FOR INVESTIGATION
.
On September 34, 1979, Consumers Power Company (CPCo) personnel, notified NRC Region III, by telephone, of the discovery of a broken reactor vessel i
holddown stud on the Midland Unit I reactor vessel. This condition was subsequently reported under the requirements of 10 CFR 50.55(e) on October 12, 1979, with interim status reports on December 14, 1979 and
)
March 3, 1980. Two other studs were subsequently found to be broken. As J
l this condition reflected a significant deficiency, an NRC investigation j
was initiated to review the materials, manufacture, and installation of
the studs.
SUMMARY OF FACTS Region III (RIII) inspectors visited the Midland site on February 4-5, 1980,
,
l and also attended a meeting at the supplier's facility on February 14, 1980.
The results of this inspection and meeting are reported in IE Inspection
,
Report No. 50-329/80-05, 50-330/80-05.
The investigation into the causes of the stud failures was initiated by a site visit during February 27-29, 1980. Subsequently, visits were made to the principal contractor (Mississippi Valley Structural Steel), the supplier (Southern Bolt and Fastener Corporation), the heat treating
)
'
facility (J.W. Rex Co.), and the Architect-Engineer (Bechtel Power i
Co rporation). During these visits, pertinent files were reviewed, and
personnel were interviewed. Materials gathered during these visits were intensively reviewed.
l The investigation findings indicate that the root cause of the anchor stud failures was the failure to characterize the studs as American Society of Mechanical Engineers (ASME)Section III, Class 1, Component Supports (Divi-j sion NF). This failure allowed use of an American Society of Testing and j
Materials (ASTM) standard specification which would not be alloved under
Division NF.
Among contributing factors were:
1.
Th ASTM specification utilized (ASTM A-354) allowed use of American Iron and Steel Institute (AISI) 4140 and 4145 steel in stud manufacture.
This material is very difficult to properly heat treat in the diameter
,
required for these studs. Difficulties in through-hardening of the
'
steel in the larger diameters may produce a hard surface and softer center.
2.
The heat treater had extreme difficulty treating the material and ob-taining acceptable hardness and tensile test results. Finally, hard-ness tests taken from halfway between the surface and center locations provided acceptable hardness results, but did not indicate the unac-ceptably hard surface (44-48 Rockwell Company). Two reported tests Were from test pieces which did not receive the same treatment as the
,
production run of studs.
-3-
_. -. _ _ _.. _
_ _ _ _.. _ _ _..... ~
.
..
.
j
.
3.
Charpy impact tests were obtained for the studs, and test results provided
.
indications of questionable properties. However, these impact tests had been performed "for information only" and the results were not reviewed.
Previously reported manufacturing problems had not triggere_d any concern which would cause a review of the Charpy tests.
Several Quality Assurance deficiencies were noted; (1) lack of licensee involvement; (2) failure to advise the heat treater of different heats of material; (3) inadequate document review; (4) failure to respond to indications that the studs were deficient; (5) failure to review ma-terials previously purchased, when the purchase specification was revised; and (6) uiscalculation of the stud stress area resulting in a slight over-specification stressing of the studs (this item was licensee identified).
The stud failure mechaniso has been identified as stress-assisted corrosion cracking, resulting from properties of the stud naterial.
The licensee is in the process of de-tensioning the Unit I studs and evaluating their use.
Tests indicate that some studs utilized in Unit 2, although of different material and heat treatment, have above-specification surface hardness readings.
Some steam generator bolts are also questionable and are
~
under review.
An unresolved item was identified during file reviews. A Bechtel memo-randa indicated that it had been project practice not to include refer-ence to ASME III in design documents.
It is not known if other items were procured without reference to ASME III. An unresolved item is one where more information is needed to determine if noncompliance exists.
l l-4-
.
. -.. _., _.
. __..-..
__.
.
_---
.
DETAILS
,
1.
Personnel Contacted
,
Consumers Power W. Bird, Manager, Midland QA J. Cook, Vice President, Midland T. Cooke, Preject Superintendent J. Corley, ';ection Head, IE&TV S. Howell, Senior Vice President H. Hudson, Procurem(nt D. Feating, QA Group Supervisor B. Peck, Construction Supervisor H. Slager, Materials Section, Design R. Wheeler, PND-Civil Section J. Wood, Quality Assurance Group Supervisor Bechtel Power Corporation J. Barbee, Supervisor, Codes and Standards W. Barclay, PFQCE A. Boos, Project Field Engineer C. Boyak, Project Engineer R. Brown, Attorney P. Corcoran, Resident Assistant Project Engineer L. Davis, Construction L. Dreisbach, PQA Engineer M. Elgaaly, Project Engineer P. Goguen, Field Engineer H. Hudson, Procurement J. Russell, QC J. Rutgers, Project Manager R. Sevo, QA Ingineer E. Smith, QA T. Suplee, Ptaject Engineer D. Yuan, Project Engineer Mississippi Valley Structural Steel M. Cohn, Engineer J. Pantukhoff, Vice President
-
Southern Bolt and Fastener Corporation R. Alexander, Vice President K. Day, QC Administrator T. Goin, Field Sales Representative E. Nelson, President D. Sibley, Quality Assurance J. Williams, Shipping J. Wood, Purchasing
.
-5-
...
.,.. _ _..... -,. _..,. - _ - _. - - - -
-. ~...
- _ _.
-.
.
J. W. Rex Company
,
G. Derstine, Director, Quality Control K. Krewson, Division Superintendent
~
F. Vasso, Sales Manager 2.
Introduction The Midland Nuclear Power Plant, Unit I and 2, licensed to Consumers Power Company, is under construccion on a site approximately one mile south of Midland, Michigan.
Bechtel Power Corporation is the Architect-Engineer and Constructor for the plant, designed to utilize a Babcock and Wilcox Nuclear Steam Supply System. Unit 1 is designed to supply process steam to nearby Dow Chemical Corporation in addition to pro-ducing electric power.
The reactor pressure vessels for these units are supported by a reactor vessel skirt, which rests on a sole plate in the reactor pedestal. Two rows of reactor holddown studs (48 inner, 48 outer) secure the reactor skirt to the sole plate. These studs are 2 1/2 inches in diameter, 7 feet 4 inches in length, weigh approximately 124 pounds each, and are secured to an embedded anchor plate. By design, the studs were to be pretensioned to 75 KSI (See Exhibit I).
These studs are designed to accomodate postulated accident loadings (vessel tip and uplift) and perform no critical function during normal reactor operation.
While the reactor holddown studs are studs by definition (no bolt head is present) the terms stud and bolt have both been used to des-cribe this equipment.
3.
Scope This investigation was conducted to review the history of the reactor pressure vessel studs at the Midland Plant as to their specification, materials, fabrication, heat treatment, testing and installation. The investigation focused on the studs utilized for Unit 1.
The chronology of the NRC investigation i_ attached as Exhibit II, and a chronology of bolt manufacture is attached as Exhibit VI.
4.
Technical Background The hardenability of an alloy is defined as its ability to trans-form to a fully hardened structure (martensite) throughout a cross section from the austenitizing temperature in the quench medium used. Statements from the bolting section of the 1978 Metals Handbook indicate that (1) "As strength increases and section size increases, hardenability becomes the most important factor in choosing a bolting material," and (2) following an oil quench, the center section of a bolt should be 90% martensite.
The choice of AISI 4140/4 % 5 steel for studs 2\\ inches in diameter by 7 feet 4 inches in length, weighing approximately 124 lbs. each, makes meeting this important metallurgical requirement extremely difficult Test results indicate that the studs hav,e varying pro-
-6-
._..,7....._m.__.-
..
-.
._.
_
.
.
perties, indicating that the heat treatment did not produce uniform re-
,
sults. However, due to the properties of the steel itself, it is questionable whether AISI 4140/4145 steel could have been adequately heat treated in this size range witho st high rejection rate,s.
AISI 4140-4145 steel is a commonly utilized bolting material, found in many applications.
It is recognized by the ASTM Code ar an ac-ceptable material in smaller diameter bolting, in a range from 1/2 -
1 3/4 inches.
In this size range, the material can be heat treated with relative ease.
In larger sizes the material is very difficult to through harden, with the center of the material being several points Rc (Rockwell Hardness) softer than the surface.
As a consequence of the material properties and heat treatment, the surface of the studs is extremely hard, while the mid radius proper-ties barely meet or are below the hardness and mechanical require-ments of the stud specification.
Certain anomolous indications raised questions about the stud material.
In addition to hardness gradients across the studs, test record indi-cate some locations along the length of the studs are harder than other locations.
The bar stock utilized for reactor holddown studs did not receive any of the special treatments commonly utilized for critical nuclear grade bolts.
Such bolts are typically purchased as vacuum-degassed steel, and purchased oversize. The material is then machined to the needed size, climinating surface defects which could be a cause for rejec-tion when magnetic testing is done.
The application of the studs is as important as the material in judging suitability. The studs are considerably stressed, and embedded in con-crete, conditions conducive to stress assisted corrosion cracking. The threaded areas provide a notch area where this failure mechanism is most likely to occur.
5.
Review of FSAR The Midland Final Safety Analysis Report refers to the reactor vessel anchor bolts specifically in several sections, and by inference in other sections.
Section 3.8.1.6.4, " Containment Liner Plate," in Paragraph 3.8.1.6.4.1,
" Materials," notes that the bolts are to be to ASTM 354, grade BD (modi-fied).
Paragraph 3.8.3.1.1 describes the bolts, but does not discuss their design.
Paragraph 3.8.3.4.1 addresses Reactor Coolant Equipment Supports, and on Page 3.'-49, refers to design standards for bolts utilized in Seismic Category I structural support.s. This section was added as part of
.
-7-
...,....,
-.
.
-.....
--
... -.
_ _ _ _ _ _ _
_
_
_________-___
-
Revision 17, dated January, 1979, and was in response to NRC questions
,
on FSAR statements. This section appears to commit the licensee to ASME Section III.
~
NRC question 110.51(3.9.3) resulted in the revision of Section 3.8.3.4.1, noted above. The question dealt with anchor bolts, and support designs.
NRC question 110.57(3.9.3) requested further clarifying information after the initial response to question 110.51, and also applies to bolting.
The licensee stated that this response had been mislocated in the FSAR, and was not meant to pertain to reactor vessel support bolting.
Table 3.8-32 appears to apply to the bolts, again describing their material as ASTM A-354, Grade BD.
Figure 3.8-30 is the drawing in the FSAR reflecting stud location and arrangement.
None of the FSAR sections appear to specifically commit to ASME Sec-tion III for reactor support holddown bolts.
6.
Manufacture of Holddown Studs a.
Material purchase. AISI 4140 ata 4145 (low alloy) hot rolled steel rods, 2 1/2 inches in diameter, were utilized for stud manufacture.
The steel was purchased from Shill Steel (heat "0000," and not utilized), Armco Steel (heat "00") and Bethlehem Steel (heats
"0" and "000") during February 1973 to March 1974. No special require-ments were imposed on the material, such as vacuum degassing or machining to reduce surface defects. Chemical analyses supplied by the suppliers showed typical values for these steels. As the rods were purchased well prior to issue of the stud specification or purchase order, Southern Bolt and Fastener (Southern Bolt) did not know how this material would be utilized, and was simply stocking steel rod. Southern Bolt personnel advised that this material was utilized due to unavailability of other grades of steel or larger diameter material.
Discussions indicated that, at this time, Southern Bolt and Fastener was a relatively small firm which manufactured bolts and studs by cutting and threading steel rods and forging heads for bolts. This was their first significant nuclear order.
b.
Specification. Requirements for reactor vessel anchor studs were included in Bechtel Specification No. 7220-C-233(Q), " Technical Specifications for Purchase of Miscellaneous Metal for Consumers Power Company."
The specification, in Revision No. 3, dated December 5, 1974, l
and later revisions, included in Section 5.10 the notation that l
"These anchor bolts and nuts will be utilized as A3ME Section III l
l
\\
-8-
..
.
. - _ _ _, _... - _. _..
._
..
.. -
T
.
Division 1, Class I, Component Supports." File information,
,
(See Exhibit III) indicates that this notation is not an error, and ASME Section III we.s intended to govern the pro-curement of reactor vessel anchor bolts.
j
,
While component Supports were described in ASME Section III, 1968, a separate Subsection, NF-Component Supports, was added to the 1973 Winter Addenda of the ASME Code, and was required for ma-terials purchased to ASME III specifications six months later. As the purchase order for the reactor vessel holddown bolts was issued on September 16, 1974, the studs should have been characterized as ASME Section III, Class 1, Subsection-NF materials (the reactor pressure vessel code dated 1968 is not applicable to these bolts as they were not a part of the reactor vessel contract). File documents indicate that attempts were made to specify the studs to the equivalent of NF requirements. However, the specification does not meet NF r2quirements in several significant areas, in-cluding ASTM specifications, materials, and testing requirements.
failure to properly characterize the studs is contrary to 10 CFR 50, Appendix B, Criterion IV, and the Procurement Specification No. 7220-C-233(Q).
(50-329/80-13-01, 50-330/80-14-01).
Included in file documentation was a memo (See Exhibit IV) in-
!
dicating that it was a project practice to refrain from citing ASME Section III in purchase specifications.
It is not known if other items were procured without reference to ASME Section i
III. This is an unresolved item (50-329/80-13-01U, 50-330/80-14-01U).
As originally issued for procurement on May 3, 1974, Bechtel Specification No. 7220-C-233(Q), Revision 2, required anchor studs to ASTM A-490-1971 requirements.
ASTM A-490(1970) " Quenched and Tempered Alloy Steel Bolts for Structural Steel Joints" included a range of 1/2 inch to 4 inch diameter bolts in its scope. Thir was changed in 1971 to allow a range of only 1/2 - 1 1/2 inch diameter bolts under the speci-fication. The vendor, Mississippi Valley Structural Steel (MVSS)
advised Bechtel that ASTM A-490 (1971) did not apply to bolts 2 1/2 inches in diameter, and following discussion, the speci-fication was revised to require ASTM A-354-1966 (Quenched and Tempered Alloy Steal Bolts, Studs, and other Externally Threaded
'
Fasteners). ASTM A-354 is not acceptable under ASME Section III.
When ASTM designations were changed from ASTM A-490 to ASTM A-354, a requirement for Charpy impact test (a measure of ductility)
to show a minimum lateral expansion of 25 mils was deleted. The revised specification required Charpy impact test results "for information only." Bechtel personnel advised that this require-ment was deleted on the basis of an engineering decision.
ASTM A-354-1966 Grade BD allowed the use of a number of steels, as long as they met the chemical, tensile, and hardness require-
.
-9-
.. _...
_ _ _.. - _
_
_
_
.
ments specified. When the vendor proffered AISI 4140-4145 material,
.
Bechtel advised them that it would be acceptable if it met the speci-fication requirements. However, Bechtel file memos indicate a recog-nition that AISI 4140-4145 material was " marginal" for the application, and suggestions were made by Bechtel personnel to purchase additional bolts because of expected test failures. No action was taken in response to these comments.
As originally issued, the stud specification did not contain testing requirements. Specification Change Notices (SCNs) added these re-quirements SCN 4004 dated September 27, 1974, SCN 4005 dated October 11, 1974). Following these changes, the purchase order was modified to include the testing requirements. The specification provided values for minimum yield, and minimum but not maximum, tensile strength, (See Exhibit V, two pages of the Specification).
c.
Fabrication. The AISI 4140-4145 rods were cut to size and threaded at each end. This was apparently completed in early December, 1974.
d.
Heat treatment. The studs were shipped to the J. W. Rex Company (REX), Lansdale, PA, sometime during December 1974-January 1973.
Southern Bolt personnel indicated that REX was selected due to availability and size of rod they could accommodate.
J. W. Rex personnel indicated that they were not initially noti-fied that there were four heats contained in the stud order, and for several months treated the studs indiscriminately as though all material was one heat. This is contrary to 10 CFR 50 Ap-pendix B, Criterion IX, and material traceability requi:ements contained in Mississippi Valley Structural Steel, P. O. Box
,
24844 and J. W. Rex Heat Treat Procedure #1.
(50-329/80-13-02, 50-330/80-14-02).
REX documents indicate the first full heat treatment (austenitizing and then tempering) was performed during late January 1975. The REX Laboratory Mechanical Property Test Report for this treatment (tests performed on reduced size mechanical specimens), dated
,
January 28, 1975, indicates tensile strength values of 144,500-
'
158,000 PSI, yield strengths of 116,200-130,800 PSI, and Rockwell
,
hardness of Rc 37-42.
Twelve of the values reported do not meet requirements, including those pertaining to hardness. These results were reported to Southern Bolt, Mississippi Valley Structural Steel, and Bechtel.
Mississippi Valley inquired if the specification could be changed to ASTM A-354-74, Grade BC, or if hardness requirements could be
,
relaxed.
When questioned by Bechtel as to the amount of relaxation on hardness tpecifications necessary, they requested an allowable Rockwell hardness of Rc 45. Bechtel advised that the test results were unacceptable and hardness requirements could not be relaxed.
Southern Bolt was advised of this via telecon on March 21, 1975.
This information was passed on to REX.
- 10 -
.
_............
. _..
.
-
.
Several tests were made=at the REX facility in attempts to ascertain
.
a heat treatment procedure which would yield acceptable studs. A request to lower the tempering temperature was made, partially as a result of these tests. A letter from Southern Bolt to MVSS, dated April 13, 1975, was used as partial basis for the requ'est to lower tempering temperature.
It reflects a resultant hardness of Rc 37 from a tempering run at 850 F.
However, the REX file test for this run indicates a hardness value of Rc 41 (all other reported values were correct). As the tempering temperature requested was within the allowable range per ASTM A-354, the change from a tem-pering temperature of 900 -1000 F to a tempering tenperature of 850 F was approved by Bechtel.
Further heat' treating was performed at REX, and 21 mechanical property tests were run between April 22-May 16,1975. These results were given to Southern Bolt and transmitted to MVSS by letter of May 28, 1975. Five of the reported values did not meet minimum yield value requirements. These results apparently were not reported to Bechtel.
Further heat treatments were run at REX, utilizing 850 F as a tempering temperature. Results from tests run on June 27, 1975
,
and July 2, 1975 (heat "00" at 925 F) were reported for formal documentation. Test reports were to be on pieces accompanying production runs. However, records indicate that two tests run on July 2, 1975 were for test pieces which did not accompany the production pieces, and one hardness value appears to have been reduced from Rc 39 to Rc 38 on the REX file report. This is contrary to 10 CFR 50, Appendix B, Criterion IX and the J. W.
Rex Heat Treat Procedure #1.
(50-329/80-13-02,50-330/80-14-02).
Steel from Heat "0000" could not meet specification requirements, and it was apparently scrapped. No information concerning the disposition of this material could be developed.
There are some indications that the heat treatment was improper as to temperature actually induced in the studs during tempering.
Furnace heat charts for most furnace runs were from wall thermo-couple readings, and for heat "00" the thermocouple placed on the studs was utilized. A comparison of the furnace charts in-dicates that the studs did not heat as rapidly as the furnace wall, and may not have reached tempering temperatures for the desired length of time.
There are also indications that the presence of a suspending nut as part of the heat treatment fixture may have caused that por-tion of the stud covered by the nut to heat more slowly than other sections, and hence be tempered to a lesser degree.
Consumers Power personnel have obtained flow rates for the oil bath quench, and have indicated their belief that flow rates are low for a sufficiently rapid quench following stud austenitizing.
This would affect the hardening of the studs.
'
- 11 -
_-
_..
. _
-..
-
_
.
From a review of test records, the dates on the furnace heat charts
,
(date of heat treatment) supplied for the formal documentation package are in error (only month and year were noted on these records).
Actual dates e-re determined from dates on test records and pen-ciled datec on furnace chart margins.
In some cases t'he date is nearly one month in e--'r.
Furnace charts submitted for documentation indicated that 38-39 studs were tempered per furnace load. This is contrary to 10 CFR 50, Appendix B, Criterion IX and J. W. Rex Heat Treat Procedure
- 1, which required a maximum furnace load of 10 pieces plus test bars.
(50-329/80-13-02, 50-330/80-14-02).
REX personnel stated that the heat treatment of the Midland studs was possibly the most frustrating order that they had taken. They noted that the studs were in their facility over six months, when a routine order is processed in approximately two weeks.
e.
Testing. Tensile, yield, and hardness testing was performed at J. W. Rex Company following heat treatments. As allowed, tensile and yield tests were performed on reduced specimens. No test pieces were preserved.
REX personnel stated that hardness tasts were performed on the stud surface for the initial hardness test. The tests performed subsequent to June 1975, were subsurface tests done on the tensile specimens themselves at the mid radius of the bolt.
Correspondence indicated that there was discussion of ASTM A-354, Paragraph 4.3, which states " Acceptance on the basis of the tensile requirements shall take precedence where minimum requirements are subject to controversy." It was indicated that a part of ASTM A-370, which gives hardness testing guidance, was also discussed.
This part provides for an " arbitration point" in the threaded area of a bolt, and mid radius hardness testing in the thraed areas.
This portion of the specification is intended for use when the readings are in dispute.
Hardness tests are non-destructive examinations, often done on each piece of critical equipment. Many standards (such as ASTM A-490, ASTM A-540) specify such surface hardness tests be per-fo rmed. The sections of ASTM discussing subsurface tests, men-tioned above, were apparently intended to be utilized in case of controversy over requirements, not in case of unacceptable results from surface hardness tests. Therefore, the subsurface tests do not meet the requirements of the rtud specification. This is in noncompliance 10 CFR 50, Appendix B, Criterion IX, and ASTM Code requirements (ASTM A-354, A-370).
(50-329/80-13-02, 50-330/80-14-02).
Charpy impact testing was performed on the studs and nuts following heat treatment, by a laboratory at Standard Pressed Steel (SPS).
Charpy acceptance criteria of 25 mils lateral expansion had not been removed from the purchase order to Southern Bolt, and the SPS lab noted this requirement on their nut Charpy Impact Test
.
- 12 -
...
._,. _
_
_
. - _ _ _ _ _ _ _
.
E l
Report. When reviewed by Bechtel, they we e advised to delete
.
,
this statement from the test form. STS did not place it on the stud Charpy Impact Test Report. Values reported for lateral expansion on studs range from 1.5-9 mils and would not,have met
'
the lateral expansion requirement for the studs had it been im-posed.
It was noted that the notarized Charpy Impact Test Report in the site documentation file contained the statement "Charpy test I
specimens on studs were taken longitudinally, more than one inch below the surface and from the mid tan inches of the seven foot four inch stud. Tests were run after heat treatment." This i
statement is not contained on the SPS file copy cf the report and was apparently added following notarization ci the document.
Magnetic particle inspection was performeu on the studs by Peabody Testing. On August 8, 1975, the Bechtel shop inspector witnessed this testing, and observed unacceptable linear indications (ex-ceeding one inch in length).
It was found that Peabody was using a less strict standard than specified, and all of the tested studs were rejected by the Bechtel shop inspector.
The studs were then returned to Southern Bolt, and actions were taken to remove the indications. The studs were variously hand ground and some 20 were machined to 2.257 inches in diameter.
During the period September 30, through October 3, 1975, actions
.
were taken by Southern Bolt to procure alternate bar material (AISI 4340) and to begin stud manufacture again. File memos in-dicated that this action was apparently begun on the belief that the studs could not be acceptable due to difficulty in meeting magnetic particle test criteria. Due to withdrawal of material suppliers, this course of action was abandoned.
Records indicate that on Janu.y 6, 1976, the Bechtel shop in-spector witnessed magnetic partP.ie testing at Southern Bolt and approved 97 studs for shipment 9 Midland. These studs were utilized in Unit 1.
Unit 2 bolts were subsequently manufactured of AISI 4340 steel and heat treated at a different facility.
f.
Quality assurance review. During this investigation, aspects of quality assurance related to studs were reviewed. File reviews indicated that Consumers Power personnel had no active involvement, beyond approval for financial expenditures, in stud procurement or document review.
No Bechtel shop inspection was performed until after the material
'
had been procured, the studs manufactured, heat treated, and magnetic particle examined. Shop inspection points are at the discretion of the purchaser and inspection prior to final shipment was chosen.
.
- 13 -
,.. -..
....... - -.... -.
.
By memo dated July 17, 1975, (BCBE 604) Bechtel personnel accepted
,
testing on the basis of heat numbers, but required the number of Charpy impact tests to be as specified in Section 5.10.4(c) of Bechtel Specification No. 7220-C-233(Q).
This required at least two Charpy tests for heat "0" (approximately 6,325 pounds), and one test for heats "00" and "000."
However, only one test was supplied for each heat, and this was not identified during docu-ment reviews. This is contrary to 10 CFR 50, Appendix B, Criterion VII and Procurement Specification No. 7220-C-233(Q).
(50-329/80-13-03, 50-330/80-14-03).
As the Charpy impact test had been required "for information only" no technically knowledgeable personnel reviewed the test results.
Bechtel personnel indicated that tests "for information" are not reviewed unless manufacturing problems are identified.
The following indications of manufacturing problems, did not re-sult in further review:
(1) Questionability of material.
(2) Early failing tests.
(3) Request for relaxation of hardness requirements.
(4) Magnetic particle examination failures.
(5) Length of time to successfully heat treat the material.
J (6) Total length of time for stud manufacture.
Review of the Southern Bolt Quality Assurance Manual indicated that it contained requirements for the content of the beat treat-ment purchase order (Dou. ment sent to Heat Treating Compsny des-cribing treatment).
Section 10.0 of Revision 4 (February 27, 1972)
in Paragraph 2.M.,
requires that the purchase orders state "where the heat treater is to Brinell (hardness test) pieces."
Southern Bolt personnel indicated that they could not locate a copy of the heat treatment purchase order for the Unit I studs, but provided a copy of the heat treatment nurchase order for the Unit 2 studs. The required information on hardoess tests location was not provided on this purchase order, and there is no blank provided for recording this information on the standard heat treat form. This is contrary to 10 CFR 50, Appendix B, Cri-terion IV and the Southern Bolt Quality Assurance Manual.
(50-330/
80-14-02).
Bechtel Specification No. 7220-233(Q) was revised by Specification Change Notice 6007 on November 8, 1976. This change added Charpy impact acceptance criteria to the section of the specification per-taining to reactor vessel anchor bolts. However, no review of materials procured prior to this change was made to ascertain whether the change affected their statt..
Bechtel personnel stated that their review of the specification, done when bolt failures were identified, determined that this revision had been intended for another part of the specification.
.
- 14 -
...
-.. --
.--
....
.
A review of pertinent codes indicates that, in the diameters and
.
strength ranges specified, Cuarpy impact tests have no acceptance criteria.
7.
Installation
~
There were no indications of shipping deficiencies or receipt inspection problems other than the failure of the document review to note that a sufficient number of Charpy impact tests were not provided, and identify that furnace loads had exceeded those set by the Heat Treatment Procedure.
The studs for Unit I were embedded in concrete during April 1977, and tensioned during the period July 23-30, 1979.
The licensee advised RIII that the studs were over-tensioned due to mis-calculation of the effective stress area. The studs were preloaded to an initial stress of 75 KSI in the shank area, but should have been preloaded to this figure in the thread area. The effect of this mis-calculation was to prestress the bolts to approximately 92 KSI versus the specified 75 KSI in the thread area.
A review of Resion III records of inspections pertaining to reactor vessel anchor bolts revealed that during an inspection on November 16-19, j
1976, a citation was issued to the licensee for failure to protect some
)
of the threads in embedded bolts for Unit 2.
There were no other inspec-tion reports relevant to reactor vessel holddown bolts.
8.
Identification of Problems On September 14, 1979, workmen placing jam nuts on the tensioned studs found that a stud (with a nut attached) had failed, and could not be
)
located. This stud was subsequently retrieved from a scrap pile.
Consumers Power advised RIII by telephone of this discovery on September 14, 1979, and followed with a formal letter under the requirements of 10 CFR 50.55(e). Status reports dated Octob.,r 12, 1979, December 14, 1979, and March 3, 1980 advised of the status of their review. Two additional studs were subsequently found to have failed.
9.
Consultant Review Consumers Power contracted with Teledyne Enginer.-ing Services to per-form a failure analysis of the Unit I studs, and a review of Unit 2 studs.
Their initial report " Investigation of Preservice Failure of Midland RPV Anchor Studs," (TR-3887-1), dated January 25, 1980, indicates that the studs have a severe hardness gradient, and indicates the failure mechanism as stress corrosion cracking.
10.
Management Meetite A management meeting with representatives of Consumers Power Co., and Bechtel Power Corporation was held at the RIII office on May 2,1980.
- 15 -
_...,.. _.,.. _. _
_
._
.
During this meeting, the findings of the investigatics were discussed,
,
including matters which were being conaidered as items of noncompliance (no delineation of noncompliance items was made at that time).
Consumers personnel indicated that they disagreed with the RIII position regarding ASME Set +. ion III applicability.
Consumers and Bechtel personnel discussed possible modifications being considered to compensate for the identified stud deficiencies. Any engineering changes formally proposed will be referred to the Office of Nuclear Reactor Regulation for review and acceptance.
The licensee advised that Unit I studs were in the process of being detensioned, and detensioning of Unit 2 studs was planned for the near future.
11.
Unresolved Items Unresolved items are matters about which more information is required in order to ascertain whether they are acceptable items, items of noncompliance, or deviations. Unreso' rd items disclosed during the inspection are discussed in Paragraph 6.B.
Attachments:
Exhibit I. Reactor Vessel Support Diagram (Bolts)
Exhibit II. Investigation Chronology Exhibit III. File Information Related to ASME III Exhibit IV. Memorandum on ASME III Usage Exhibit V. Stud Specification (2 pages)
Exhibit VI. Stud Manufacture Chronology
- 16 -
.
_ _..
._..
..
l i
'
ee
'IY..***.I.=
- f'
y
._ _
.
e.-4 a ig"..h*= i-*u*.*."* * '*** I' M-f.
h
.
~.=
.q
.
v,,,., :.5.,4~.e
.
...x
..~
g, y
.-
Y
.
.
.....,
t 3,,;p;, ',,,,,
8 + P **!.
'
,,,
2nAp.tq y
re:=."
edw;:.r....,
,s.:r.
', :..? e :..
u m. t..:.
'
.-
.
.
ec..vg -
.
.
y..
y.m
%.1
....h....._,,,4..
9 (c
-.~<
,
.
,
e s. 1 : -..
./
Q,,.,.
.
y
.s m
.
M
.!..e_s
-
._
-
-
.,e,, y.
43 - n
.
,
_ __ _...,
....
....
.
,
w.
g 3. >&
.
._ =..=..
d%t~g{,c.. d.
~~
.
,
- ~
.
' '
-
=":
- Lj A.
-[~
-w~
g
._
<
/ h;6
-
e-. m -
y /.q
~,.
c. -
~....
.c p
,
a
-... s
. ~7 d
,,.. /,
sid,.
.
)
y
e,=
,
c,
-.
.
,
.
xc
.s.-.
..
-
4.? j !. f,Q N -
.s@
t (
..f
..
.
?
-
s
...,
p., ~
..
...
- gs @f
- **
--
_,
w c.n.;:
secTsou M
+
a
_...w... w t
nuo
-
..,e.,....;..._,,....,_,.
!
.......
..,. _
q t
_..
...
..
,
d I
W=='
i CONSUMERS POW' R COMPANY j
d MIDLAND PLANT UNITS 1 & 2
- )
F FINAL SAFETY ANALYSIS REPORT (5'8
>
'
(_
'
Reactor Vessel Support I
M k
g l
(C-376. Rev 9)
FSAR Figure 3.8-30 (E t
!,
2/79 Revision 18 I
i
.
.
__ _
-
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _.
.
NRC INVESTIGATION CHRONOLOGY
.
9/14/79 Licensee reports stud failure.
,
10/12/79 50.55(e) report from licensee.
12/14/79 interim report on 50.55(e).
2/4-5/80 Inspection of studs on site.
2/5/80 Third stud found broken.
2/14/80 Meeting at Southern Bolt, NRC, CP, SB personnel.
2/27-29/80 NRC Investigation initiated, Midland site.
3/3/80 second interim report from licensee.
3/5-6/80 NRC investigation at Southern Bolt and Fastener.
3/20/80 Inspection Report 80-05 transmitted (2/4-5/80 inspection report).
3/12-13/80 NRC Investigation at J. W. Rex Company.
3/18-19/80 NRC Investigation at Bechtel office, Ann Arbor, Michigan.
3/20/80 NRC Investigation at Mississippi Valley Structural Steel.
4/2/80 Call to Consumers passes on issues for resolution.
4/15/80 Phone call to clarify issues for resolution.
4/18/80 NRC Investigation at Bechtel, Ann Arbor (answers to questions).
'
5/2/80 Meeting with Consumers Power.
.
Exhibit II
.
-
..
.
FILE INFORMATION RELATED TO ASME III APPLICATION TO HOLDDOWN STUDS
.
(excerpts)
,
9/23/74, telephone call memo by R. Grote to R. Ryden/D. Koski: "proj ect en-gineering added to the magnetic particle inspection of the nuts so to be in accordance with ASME Section III - NF," "the subject bolts are classified as ASME Section III Class I component supports."
"
9/27/74, Specification Change Notice (SCN) C-223-4004:
(Note:
these an-chor bolts will be utilized as ASME Section III Division I Class I compo-nent supports)."
10/1/74, memo BCBE 436, by R. E. Felton to R. L. Castleberry (pg. 2):
" Pro-ject engineering has af firmed the magnetic particle examination requirement on nuts, the reason being that ASME Section III governs the procurement of reactor anchor bolts."
4/11/75, unsigned notes identified as having been made by Mr. John Hink:
"the RVAB (reactor vessel anchor bolts) are classified as component sup-ports in Section NF, Section NF is not mandatory," " design appears to be fairly close to the design requirements of NF."
i
Exhibit III
-
.
4 e e
.
.
Bechtel Memorandum
.
To:
R. L. Castleberry Location: A2-6A From:
G. Tuveson Date: 8/30/76 Subj ect:
Midland Units 1 & 2 Job No. 7220 application of ASME File: C-2135 B&PV Code Section Ill Division I Subsection NF Requirements to Component Support Structure The above mentioned subject was discussed between M. Rothwell and M. Elgaaly, A. Desai and B. Dhar of civil group on August 19, 1976.
It was agreed that to be consistent with Midland project position, the ASME code would not be directly referred to in the design documents.
But the de-sign, fabrication and construction would meet, to the extent possible, the ASME code requirements within the applicable boundaries.
Accordingly, to meet the intent of the code, civil group will add a section to the specifications C-38 and C-233.
When required, the design drawings will call out the applicability of this section for a particular structure.
typed copy of handwritten memorandum Exhibit IV
.
- P
-
p..
"
5.9 Sheat Stu shall be in accordance with At J1.1 S:ctica 4 Part VI.cnd the following. The material chall ccnform te either AS1H A 307 or ASTE
,
-
A 108 an applicabic, and shall meet the tensile requirements cc +2ined
/
'
in AWS D.1.1.
5.10 Reactor Vessel Anchor Bolts and Nuts
.
(NOTE:
These anchor bolts and nuts will be utilized as ASME Section III Division 1 Class 1 component supports.)
~5.10.1 Bolts shall be ASTM A 354 Grade BD, with the following addi-tional requirements:
ASTM A 614 as specified in Section 5.10.3 below.
a.
b.
ASTM A 354 Section 4.4 and Table 3 - Mechanical tests on machined specimens from the Grade BD 2-1/2 inch diameter bolts shall have a minimum yield strength of 130,000 psi, minimum clongation of 14 percent and minimum reduction of area of 35%.
ASTM A 354 Table 2 - The Grade BD 21/2 inch diameter bolts c.
.shall have a minimum tensile strength of 150,000 psi, a minimum proof load of 120,000 psi and a minimum yield strength of 130,000 psi.
,
.
d.
The bolt material shall be subjected to impact testing as specified in Section 5.10.4 below.
5.10.2 Nuts chall be in accordance with ASTM A 194 Grade 2 or 2H, with the following additional requirements:
a.
ASTM A 614 as specified in Section 5.10.3 below.
b.
ASTM A 194 Section 5.1 - Certified Material Test Reports shall be in accordance with ASTM A 614 Section 8.
If ladle analysis is not availabl.e a check analysis may be substituted.
.
c.
ASTM A-194 Section 9 - The Cone stripping test is not required.
,
d.
ASTM A 194 Section 14.1 applies.
e.
ASTM A 194 Section 14.3 - Certification shall be in accor-dance with ASTM A 614 Section 8.
f.
The nut material shall be subjected to impact testing as
.
specified by Section 5.10.4 below.
5.10.3 The materials, testing and documentatic-nf cha cubjcet nuts and bolts shall be in accordance with ASIM A 614 with the following additional requirements:
ASTM A 614 Section 9.1.2 - The written procedure shall be a.
submitted to the Buyer.
3A
.
EXHIBIT V page 1 of 2
,... _..... -. _ _
-._ _.-._ __
_
,
,
,
.
._ __ - ___ ____ _ _______ _ _ _ _ _ _ _ _ _ _
-
.
-
.
'
.
.
b.
ASTM A 614 Secti n 9.1.4 - Tha writtsn proccdura ch:11
,
be submitted to th,e Buyer.
.e
-
~
c.- ASTM A 614 Section 10 is required.
-
d.
ASTM A 614 Section 11 is required.
.
ASTM A 614 Curtion 12 is required.
'
e.
f.
ASTM A 614 Section 13 is a Seller's option to ASTM A 614 Section 12.
.
5.10.4 The Charpy V-notch test (Cv) shall be required for the bolts
and nuts in accordance with the following:
i
I
.
'
Testing Procedure - Test procedures shall be in a.
accordance with ASTM A 370-72a.
!
g
-
b.
Location and Orientation of Test Specicens - The Cy i=7act test specimens shall be prepared with the longitudinc1 axis
.
of the specimen located it least 1/2 radius or 1 inch below the surface plus the machining allowance per side, whi.ch-ever is the lesser.
The fracture plane of the specim_a
.
shall be at 1 cast i diameter or thickness from the heat i
- I l
-
treated end.
-
.
Sampling Frequency - One test shall be made for each lot
'
~
c.
of material where a lot is defined as one heat of material
-
heat treated in one charge or as one continuous operation,
-
not to exceed 3,000 lbs by weight.
.
.
d.
Condition of Material - The test specimens shall be taken after heat treatment.
,
Test Temperature - The ftpact specimens shall be tested
.
e.
at 40'F.
..
f.
Certified 16terial Test Report - The test temperature, lateral expansion, absorbed energy and percent sheer fracture as well as the orientation and location of all
.
tests shall be reported for information in accordance with ASTM A 614 Section 8.
.
5.10.5 Handling, shipping and storage shall be in a manner that shall
-
avoid damage to the material. The Seller shall submit written procedures for handling and shipping for approval by the Buyer prior to shipment.
-
t
.
EXHIBIT V 3B page 2 of 2
.........-.
., - -.....
.
.
t STUD MANUFACTURE CHRONOLOGY
.
11/6/73 Specification 7220-C-233(Q) issued for client review.
3/25/74 steel heats "0" and "000" received, "00" and "0000" received
'
previously.
6/28/74 bid requests sent.
8/5/74 Decisions made as to applicable nondestructive tesing requirements.
8/8/74 TWX adds nondestructive testing requirements, original supplier withdraws.
8/21/74 TWX, MVSS-Bechtel, ASTM A-490 is not right specification.
8/23/74 TWX, MVSS-Bechtel, need specification, trying to find material.
8/23/74 Bechtel response TWX, A-490 is correct, 4140/5 steel not approved.
8/27/74 Memo, test requirements, tensile values, 25 mils expansion for Charpy test.
9/3/74 TWX, MVSS-Bechtel, proposal, ASTM-354, 4140 steel, 25 mils expansion.
9/10/74 Memo BEBC 527, approves use of ASTM A-354 as specification.
9/16/74 Contract date.
10/1/74 Memo, history of studs to date.
12/20/74 Rex heat treatment procedure #1, revision O.
12/74-1/75 Studs shipped from SB to Rex.
1/28/75 Rex material test report, specimens #1, 2, 6, 8, 9, 11, 12.
2/4/75 letter, SB-MVSS, material cannot meet requirements.
2/6/75 TWX, MVSS-Bechtel, provides Rex test results.
2/12/75 Phone call memo, SB requests relaxation of hardness to Rc 45.
2/18/75 Memo, discusses six tests, hardness relaxation request.
3/21/75 TWX, test results unacceptable, not relax hardness requirements.
4/3/75 letter, SB-MVSS, justifying 850 degree temper (reported hardness is wrong).
4/11/75 Notes, 4140 marginal, excessive hardness, where was hardness tested?
4/18/75 Memo, hardness, tempering, material is marginal.
4/22/75 Rex material test, test 1-4 of 19 finally made.
4/25/75 Rex material test, tes ts 5-11.
5/1/75 Bechtel approval of Rex heat treatment procedure, revision #3.
5/5/75 SB Quality Control manager visits Rex.
5/16/75 Rex material test, tests 11-19.
I 5/28/75 19 test reports sent with " dummy" documentation package for review.
6/3/75 Rex test, " machined from 2' of end of bar."
'
6/9/75 Rex test, " machined from 7' from end of bar."
6/16/75 Rex test, stud #1 from heat "000".
6/18/75 Rex test, stud #8 from heat "000", 850 degree temper.
7/2/75 Rex test, " machined from center of bar," heat "00".
7/15/75 Midland meeting, Bechtel and MVSS determine allowable number of tests.
7/17/75 Memo, BCBE 604, physical and mechanical tests to be by heat number.
7/21/75 date on thermocouple furnace chart for heat "00".
7/24/75 Revision #4 of Rex heat treatment procedure approved.
7/29/75 date of material properties report supplied for documentation.
8/18/75 Rex surveillance report, all studs rejected for linear indications.
8/20/75 SB Quality Control manager visits Rex.
8/27/75 TWX on reducing diameter of shank of stud by.060 inches.
9/30/75 TWX, history of studs, start again, new material suppliers withdraw.
11/11/75 TWX, SB proposes turning some studs to 2.257 inches in diameter.
11/20/75 TWX approves turning to 2.257 inches in diameter.
1/4/76 97 studs pass examination, are released for shipment, 96 shipped.
1/22/76 studs received at Midland site.
11/8/76 SCN 6007 adds 25 mil expansion criteria to stud section, possibly in j
error.
4/77 Unit I studs embedded in concrete at Midland.
7/23-30/79 Unit I studs tensioned.
9/14/79 first stud found to have failed.
-
1/25/79 Teledyne Engineering report on stud failure mechanisms.
Exhibit VI
--.
.