ML20093C440
Text
.__
g(
4,h; #.
. ?
[p u%qfo UNITED STATES s
J(/
g y
g NUCLEAR REGULATORY COMMISSION As e
/j
,I a
s, WASHINGTON, D. C. 20555
/
iN /,jk,,
NOV 2 5 1983
(-
MEMORANDUM FOR: Thomas M. Novak, Assistant Director for
([
L/
Licensing, DL i
hrc 0 F 19P,'
FROM:
James P. Knight, Assistant Director for Components & Structures Engineering, DE
SUBJECT:
TECHNICAL EVALUATION OF MIDLAND HVAC STRUCTURALDESIGNADEQUACY(TAC #52311)
)
References:
1.
Memo from RLSpessard ta DGEisenhut, dated 8/4/83 2.
Memo from DTerao to RBosnak, dated 10/28/83 k
3.
Memo from DTerao to RBosnak, dated 11/17/83
\\\\
\\
In Reference 1 above, NRR was requested by Region III to review the structural design adequacy of the Clinton and Midland HVAC systems.
In References 2 and 3 above, the Mechanical Engineering Branch provided its trip report summaries of the design review and audit of the Midland HVAC f
systems. Attached to this memorandum is the technical evaluation of the Midland HVAC structural design adequacy based on the findings from the MEB design review and audit.
It is our understanding that DL will l/
coordinate the technical evaluation inputs from ASB, MTEB, and MEB and transmit the combined results of NRR findings to RIII.
The HVAC review for the Clinton facility is planned to be perfonned by MEB in early December.
It is anticipated that the Clinton HVAC review will be similar to the LaSalle review effort because of Sargent &
Lundy's design role in both facilities.
j
/
/
W ff g
n' t, Assistant Director for Compone.,& Structures Engineering Division of Engineering,
Attachment:
As stated f
g cc:
R. Vollmer, DE E. Eisenhut, DL n
ph E. Sullivan, DE R. Bosnak, DE il y
yE. Adensam, DL W. Little, RIII
\\J D. Danielson, RIII F. Hawkins, RIII' t
\\
D. Hood, DL M. Miller, DL
- 0. Parr, DSI W. Lefave, DSI s
--f' ~
\\
(
/
h k '/
p.-
- o
/q nDP t N 1-
{
~
=
o' i
i TECHNICAL EVALUATION OF MIDLAND HVAC STRUCTURAL DESIGN ADEQUACY I.
Applicable Codes and Standards for HVAC Ductwork and Support Preseatly, there are no national codes or standards which provide specific requirements for the overall design, fabrication, and installation of HVAC systems in nuclear facilities. The only national standard which addresses the design and construction of duct systems in a limited manner is ANSI-N509,
" Nuclear Power Plant Air Cleaning Units and Components." ANSI-N510 covers the functional system testing aspects. The ANSI-N509 standard does not require specific material documentation.
Typically, the HVAC systems of nuclear facilities have been designed according to the guidelines showa in Sheet Metal and Air Conditioning Contractors National Association (SMACNA) publications, " Low Velocity Duct Construet{on Standards,"
l which is applicable to duct pressures up to 2 inches wate gage and, "High Velocity Duct Construct 1 n Standards," which is applicable duct pressure up
( w to 10 inches waterf gafje. These design standards are based on performance only and are not based on-the stress and deflection considerations associated with seismic Category I structuras.
The American Iron and Steel Institute (AISI) code was adopted by the applicant for the Midland facility to govern t ign of seismic Category I ductwork because of its applicability to thi gage heet metal.
The American Institute of Steel Construction (AISC) Code, " Specification for the Design, Fabrication and Erection of Structural Steel for Buildings," was used for the design of the HVAC ductwork supports.
The supports and ductwork were welded in accordance with the American Welding Society (AWS) Structural Welding Code (AWS 01.1), Specification for Welding Sheet Steel in Structures (AWS D1.3), and Specification for Welding of Sheet Metal (AWS 09.1).
11/22/83 1
MIDLAND HVAC STRUCT DESIGN
.--...-..p.
s' II.
Documentation-Because there are no national codes nor standards which specify the documentation required, the documentation requirements become the responsibility of the Utility (or its architect-engineer) to define.
For Midland Plant Units 1 & 2, the architect-engineer, Bechtel Power Corporation provided a technical specification for HVAC Work (Spec. No. 7220-M-151A1) which specifies the documentation requirements.
The technical specification requires that a certificate of conformance (C of C) is necessary for all requirements of the technical specification.
A certificate of conformance is a written statement signed by a qualified party certifying that the items or services comply with the technical specification requirements.
A material certificate of compliance is required by the technical specification to be provided for subcontractor-supplied (Zack) construction materials includ-ing dampers, diffusers, grilles, registers, air flow measuring units, ductwork, hangers, supports, and miscellaneous materials specifically identified by the technical specification. A material certificate of compliance is a written statement signed by a qualified party certifying that the materials are in accordance with a particular material specification.
When required by the referenced codes or material specifications, the material certificate of compliance is required to be accompanied by a certified material test report (CMTR). When the requirements of the technical specification are more stringent than the referenced code or material specification, the material certificate of compliance is required to be accompanied by a CMTR which demon-strates compliance with the more stringent criterion.
For example, ASTM specification A526 does not require a mechanical strength test for the sheet steel and, thus, no minimum yield strength is specified.
However, the techni-cal specification M-151A requires a minimum yield strength of 30 ksi for A526 1
and A527 sheet steel. The CMTR includes all chemical, physical, mechanical, and electrical property test data required by the material specification, applicable codes, and procurement documents.
The CMTR includes a statement of conformance that the material meets the technical specification requirements.
11/22/83 2
MIDLAND HVAC STRUCT DESIGN
.o i
s.-
III. Materials
)
The technical specification (M-151) for HVAC ductwork specifies the materials for the HVAC ducting, stiffeners, fasteners, and supports.
For the typical duct details, the materials used are standard commercial grade materials.
The sheet steel is typically galvanized carbon sheet steel conforming to ASTM A526-71 or ASTM A-527 with a coating designation G-90 and a minimum yield strength of 30 ksi.
Carbon steel sheet includes ASTM A366-72 (minimum yield strength of 30 ksi) and ASTM A607-75, Grade 50.
An austenitic stainless steel sheet or plate (Type 304-28, ASTM A240-75A) with a minimum yield strength of 30 ksi is also specified.
For support steel, the technical specification requires that carbon steel structural shapes, bar s'zes, and plate conform to ASTM A36-75, ASTM A572-77A, (Grade 50), and ASTM A284 (Grade A) with minimum yield strength of 36 ksi.
Structural tubing conforms to ASTM A500-77 (Grade B) and angles 2\\ inches by 2 inches by % inch and smaller conform to ASTM A575 (Grade 1020) with a minimum yield strength of 36 ksi.
Carbon steel fasteners (including Huck bolts and sheet metal screws) conform to ASTM A325 galvanized and ASTM A307-74 galvanized.
The only acceptable substitute permitted by the Midland technical specification for ASTM A325 is ASTM A490-76a. Acceptable substitutes for ASTM A307 are ASTM A193-76, ASTM A354-766, ASTM A449-76c, ASTM A490-76a, ANSI B18.2.1-65 with C W R, or ASTM A325.
IV.
Structural Desian Margins In order to determine the structural adequacy of the HVAC system (supports, stiffeners, and ducting), it is necessary to ask ourselves the following question, "Is the structural design of the HVAC system adequate if the mater-ials used are questionable?"
It logically follows that if the design margin to failure is large and if the range or possible variation in material proper-ties in question (e.g., mechanical strength) is small, then we can reasonably 11/22/83 3
MIDLAND HVAC STRUCT DESIGN l
v.
conclude that the design is adequate.
The adequacy or design margin can be expressed in the form:
all wable stress design margin = calculated stress For the components to be acceptable the design margin must be greater than 1.0.
The larger the value, the more design margin is available.
If the design margin is less than 1.0, then the question arises, "Will the component fail?"
In order to answer the question, it is necessary to define what is meant by " failure".
It is also important to understand what the basis is for the allowable stress.
In the following sections, we will be comparing the potential reduction in material strength due to substitute materials with the typical design margin for the various structural components in the HVAC system.
The structural components that will be covered include the following:
A.
Structural Steel Supports and Welds B.
Ductwork and Stiffeners C.
Ducting Companion Flange Bolts D.
Concrete Expansion Anchor Bolts A.
Structural Steel Supports and Welds 2
For the Midland HVAC supports, the design specification requires the use of carbon steel structural shapes, Dar sizes, and plate to conform to ASTM A-36, A-572 Grade 50, and A-284 Grade A, structural tubing to conform to A-500 Grade B, and angles to conform to A575 Grade M 1020.
The material minimum l
yield strengths and minimum tensile strengths of the HVAC support steel are provided in Table 1.
The structural steel used for the Midland HVAC support member is designed in l
accordance with the AISC, " Specification for the Design, Fabrication, and Erection of Structural Steel for Buildings."
i l
11/22/83 4
HIDLAND HVAC STRUCT DESIGN
d' 3:
A 2
The applicant specified in its design guide, the allowable stresses for the structural steel and tube sections as follows:
Allowable stress in accident conditions:
bending and torsion = 0.9 Fy shear = 0.5 Fy where Fy is the material yield strength.
In the calculations reviewed by the staff, it was found that the Idaterial yield strength used for the support steel was assumed to be 36 ksi.
It was noted by the staff that the applicant prudently used a 36 ksi yield strength for a structural tube steel (A500) which actually had a a minimum yield strength of 46 ksi. Typically, the applicant used 36 ksi yield strength for all struc-tural steel in the support calculations.
It should be noted that because the tube steel is welded in construction, the use of 36 ksi is prudent since its higher tensile strength resulting from coldwork will be annealled out in welding.
For the A284 Grade A plate material, the minimum yield strength is only 25 ksi.
Although, the staff did not review specific calculations for the A284 material, it was concluded by the staff that the design margin for plates is large, thus, if the applicant had used 36 ksi instead of 25 ksi for the plate material, it is unlikely that the actual stresses would be near yield. The design margin to the allowable stress in accident conditions for a plate was found to be 7.7.
The design margin to failure is greater than 10.0.
For A575 (M1020) material used as angles in the Midland HVAC supports, the ASTM specification does not require mechanical tensile tests. However, the i
Midland technical specificationi does require a minimum yield stress of 36 ksi for A575 material.
Because several grades of A575 are available with lesser carbon content (and thus lesser strength) than Grade M1020, the strength properties of the lesser grades needed to be determined to evaluate whether the design adequacy could have been compromised. The staff obtained typical test results from Northwestern Steel and Wire Company for various grades of A575 material. The values are shown in Table 2.
Thus, it appears that the lowest grade (M1008) of A575 material could exhibit strength properties approx-I imately 10% less than that required by the design specification.
11/22/83 5
MIDLAND HVAC STRUCT DESIGN
,c
^1l 30l};
L r,
v
(:
The typical design margins for HVAC supports are provided in Table 3 of this report. As can be seen, the support steel (wide flanges, angles, plates, and tube steel) exhibit substantial design margin to the allowable stress at acci-dent conditions.
It should be noted that the staff found other conservatisms in the HVAC support design. One conservatism is the damping values specified for the seismic building response spectra used in the HVAC support analyses.
The supports (welded structures) are designed using a damping value of 2% for both OBE and SSE loads.
Regulatory Guide 1.61 allows for welded steel structures 2% for OBE and 4% for SSE.
The ratio of the maximum peak acceleration for the SSE at 2% to the maximum peak acceleration for the SSE at 4% is approximately 1.4.
Thus, at the maximum peak acceleration, the use of the 2% damping results in an additional design margin of approximately 1.4 for welded steel structures.
f It should be noted that the HVAC duct is more rigid than the HVAC supports Y'
because of the conservative 8-ft span criteria. Typically, the HVAC duct
's
- Y fundamental beam bending freyency between support spans of 8 ft is approxi-
_y mately 150 hertz (with the lowest frequency approximately 55 hertz) whereas the fundamental frequency of HVAC supports are typically less than 33 hertz.
M e walds for HVAC supports are governed by AWS D1.1-72.
Weld tensile strength 16 is assumed to be 60 ksi for E60 electrode.
For a 3/16" fillet weld the allow-able weld strength is equal to:
gis>, -
4y..n,.wv)(.3 x n
7
- n //c?
w 0.3)(60,000 h 2386 Rs/inc D M
g)vv 6 g,,t 6
pp c~
For accident conditions, a 50% incraase in the design allowable is used resulting iy anjlge, s,tgengg of 1.5 x 2386 = 3579 lbs/ inch.V The design margin to tensMilitire is, thus, 48,060/27,000 = 1.78 at the accident condition allowable weld strength. *
]
A WJ;.Dl.l dieuzbI s-sh/sI -
la !E'*f"
- ' 7'
\\
c v
r
' g a y a g j, t..s' s 1 9, o c o a 2 7,, o c o pi. AW S.D I. / m,G G J % s
)
v r
l
/*lhe ultimate breaking strength of fillet welds and partial joint penetration groove welds shall be computed at 2.67 times the basic allowable stress for l
60 ksi tensile strength p r ".i$ Ol d #-*c y cg 2,4 *J x / 9, ooo pt, o go l
t 11/22/83 6
MIDLAND HVAC STRUCT DESIGN
s.
As shown in Table 3, the design margin to the allowable weld strength at accident condition varies from 1.3 to 33.3 and is in addition to the 1.78 margin described above. Thus, the staff concludes that welds have a sub-stantial design margin to failure.
B.-
HVAC Ductwork and Stiffeners For HVAC ductwork, the staff found that typically A526 or A527 sheet steel is used.
However, the design specification also stipulates the use of carbon 1
steel sheet material A366 and A607 Grade 50 and austenitic stainless steel sheet (or plate) Type 304-28, ASTM A240.
The material minimum yield strengths and minimum tensile strengths of the HVAC ductwork are shown in Table 4.
In order to understand the design margins in the HVAC ductwork, it is important to clarify the analytical and testing methods used by the applicant in quali-fying the ductwork.
The applicant does not follow the design guidelines of the SMACNA standards but rather uses the generic design guidelines as depicted in their HVAC drawings C-842 through C-849.
The staff has compared the differences between the SMACNA standard and the Midland HVAC drawings and has found that the Midland sheet metal thicknesses and stiffener sizes tend to be larger than those specified by SMACNA for the corresponding duct sizes and is, thus, conservative.
The SMACNA stiffener spacing tends to be closer than the spacing used at Midland.
However, because the stiffener is primarily used to prevent buckling of the sheet metal, the additional thickness of the sheet metal compensates for the increased stiffener spacing.
In 1977, the architect-engineer for the Midland facility (Bechtel Power Corp-oration) sponsored testing of the HVAC duct specimens for the Limerick plant.
The test results were used to develop a Bechtel generic HVAC duct design guide which was used for the Midland plant. The main goals of the Duct Test 3
Program
- were:
I 11/22/83 7
MIDLAND HVAC STRUCT DESIGN
e
- s.,
a)
To substantiate the use of width to thickness (w/t) and height to thickness (h/t) ratios of up to 1500 while maintaining the AISI specification as the basis for design.
b)
To justify stiffener design.
c)
To obtain a rational design method for the structural design of HVAC ducts by correlation between theoretical prediction and experimental results.
d)
To assure that the duct details and materials used would not cause any fabrication problems when full scale production began.
The testing was performed by Hales Testing Laboratories of Oakland, California.
The testing was based on A526 and A527 ductwork material with a minimum yield strength of 36 ksi. The significant conclusions of the testing included the following results.
I Failure modes of the ducts were not catastrophic and there was a great reserve stren..h after failure.
Pressure loading was the most important loading.
Live load and seismic loads were less important.
Effect of seismic loads can be simulated by pressure loads.
The primary failure modes of rectangular ducts were by corner crippling of sheet and by stiffener buckling.
Live load stresses'in the sheet and stiffeners were low.
The Bechtel generic HVAC duct design guide was used to qualify the ductwork spans in the Midland plant. The calculations assumed a minimum yield strength of the duct material to be 30 ksi. Thus, the ductwork materials specified in the design specification all meet or exceed the 30 ksi value.
It should be noted that the ASTM Specification for A526 and A527 material does not require i
11/22/83 8
MIDLAND HVAC STRUCT DESIGN C
I*
mechanical tensile strength tests.
The Midland design specification 1 does require that the sheet metal (where there are no ASTM tensile test requirements) be purchased with a minimum of 30 ksi yield stress.
The staff reviewed several purchase orders and confirmed that for the A526 and A527 material, the yield strength and ultimate-tensile strengths were specified by the supplementary i
. test requ rements.
All purchase orders reviewed showed that the yield strengths for safety-related duct material were greater than 30 ksi. With regard to material substitution, the staff has found that drawing quality sheet steel can have a yield stress as low as 25 ksi.
However, the staff concluded that approximately 20% decrease in yield stress (25 ksi vs. 30 ksi) is not a sign-ificant concern because of the adequate design margins in the HVAC ductwork.
The HVAC ductwork design margins are shown in Table 5 of this report.
4 i
C.
HVAC Ductwork Companion Flange Bolts The standard bolts used in the HVAC ductwork companion flanges are 3/8 inch l
diameter and made of A307 low carbon steel.
The generic design detail is shown on Midland Dwg No. C-844 (Q) and specifies a 6-inch maximum spacing between the bolts in the companion angle flange connections. The calculation 5 l
of the 3/8-inch bolt loads was performed for the worst case loadings and i
included many conservatisms.
The calculation was based on A307 bolt material with an allowable design stress of 20 ksi (per AISC Manual of Steel Construc-tion). A307 bolts (Grades A and B) are required by the ASTM specification to l
have a minimum tensile strength of 60 ksi.
The allowable tension was calcul-ated as follows:
l Allowable tension load = (20 ksi)(0.078 in )(1.5) 2 l
(accident condition)
= 2340 lbs.
The ASTM (A307) tensile strength requirement for 3/8 inch oiameter bolts is 4650 lbs.
Thus, there is a design margin of 2 to failure at the allowable tension load at accident conditions.
The staff found that assuming one bolt l
is effective in each corner of the flange, the bolt has adequate strength to accommodate the applicable loads and load combinations. The staff found the bolt calculation to be based on conservative assumptions and the results show 11/22/83 9
MIDLAND HVAC STRUCT DESIGN n,-,
-,n v.,- -..~.,
1 an adequate design margin.
It should be noted that prying action (steel-to-steel) was considered in the calculation per AISC (8th Edition).
A summary of the bolt design margin from the calculated load to the allowable bolt load at accident condition (2340 lbs) for several duct sizes are shown in Table 7.
D.
Concrete Expansion Anchor-Bolts The HVAC ductwork supports are generally anchored to reinforced concrete foundations with expansion anchor bolts.
The drilled-in concrete expansion
~
anchor bolts are supplied by Hilti Fastening System for all sizes except for 7/8 inch nominal diameter bolts. The 7/8 inch bolts are supplied by Phillips Drill Company. The material properties are shown in Table 8.
1 In reviewing the design margins in Table 3 of this report, it can be seen that the anchor bolt tends to be the controlling component in the HVAC support design (i.e., the anchor bolts have the least design margin).
Anchor bolts are designed with a margin of safety of four to its ultimate tensile load capacity as published in manufacturers' catalogs.
The ultimate tensile load capacity is based on the failure of the anchor bolt in concrete due to static loadings.
IE Bulletin 79-02 also accounts for bolt slippage in its safety factor of four.
Thus, the staff concludes that although the expansion anchor bolts have the least design margin to the allowable design load, there is a design margin of at least 4.0 to the anchor bolt failure due to static loads.
t To provide additional verification of the accuracy of the catalog data presented by the anchor bolt manufacturers, Teledyne Engineering Services (TES) has performed both experimental and analytical work on anchor bolts made by diff-8 erent manufacturers including Hilti and Phillips. This work was done for a l
group of 14 utilities, in response to IE Bulletin 79-02.
The TES report is l
-discussed in detail in Appendix B of NUREG/CR-2137. The TES test data for Hilti and Phillips wedge anchors showed relatively close correlation with the catalog loads.
The maximum ratio of catalog loads to TES average test loads i
for Hilti and Phillips was 1.3.
l
~
l 11/22/83 10 MIDLAND HVAC STRUCT DESIGN l
l The available test data (s) indicates that by using a safety factor of four to the average strength of the expansion bolt, the probability of failure at the design load is less than 0.001. The probability of failure at two times the design load is about 0.023(7).
The ultimate strength of drilled-in concrete expansion anchor bolts for dynamic and vibratory loadings was investigated by the staff.
The safety factor of four as recommended by anchor bolts manufacturers is applicable to static load-ings.
The design margin to failure for seismic loadings which are dynamic and vibratory in nature is a function of both load magnitude and the number of cycles. A report on an investigation by Bechtel Power Corporation to justify the use of expansion anchor bolts in the Fast Flux Test Facility (Richland, Washington) was prepared for the Hanford Engineering Development Laboratory in January 1975.8 The objective of this investigation was to establish the allow-able design loads (tension, shear, and combined load) for expansion bolts to be installed in various mixes of concrete.
The test loads included static loads and alternating loads which simulated the dynamic earthquake loads.
The expansion bolts included the stud type wedge anchors manufactured by Hilti Fastening Systems.
The seismic loading was simulated by about 6000 cycles of a sine wave which varied from zero to 0.25 (where S is the static load capaci-ty of the anchor bolt). The test found that all expansion bolts which were tested successfully withstood 6000 cycles of 0 to 0.2 S alternating load as designated f'or seismic qualification. The dynamic load capacities of the ex-pansion bolts were found to be the same as their corresponding static load capacity.
It was further discovered that at 6000 to 7800 load cycles when the dynamic test load sequence was increased to 0.6 S subsequent alternating load-ing caused appreciable wedge movement (or " walking").
If the bolt did not fail in a brittle mode due to pull-out or in some other premature failure made i
i (e.g., poor installation), the " walking" ceased after a certain number of load cycles.
Extensive dynamic testing of expansion anchor bolts was also discussed in NUREG/CR-2999 8) by Hanford Engineering Development Laboratory under contract I
i with the NRC.
Prior to the testing, a survey was performed to determine the adequacy of existing concrete expansion anchor test data. Based on the survey findings, it was concluded that there was a lack of testing to assess the effect of bolt preload under dynamic loadings. Thus, exploratory dynamic 11/22/83 11 MIDLAND HVAC STRUCT DESIGN l
t.
testing was performed on typical wedge and shell anchors.
It was found that, when the installation torque is properly applied, residual preload does not significantly affect anchor load displacement characteristics until the pre-load drops to less than 50% of the full installation preload.
It was con-cluded that this must be considered in design situations where support stiff-ness is an important factor. Table 9 presents the dynamic test results for typical wedge anchor bolts.
It can be seen from the ultimate dynamic load capacity and the number of cycles to failure, that there is a large design margin (a minimum of 2.4 for test number DW-SR). The number of cycles exceeds the number of seismic cycles recommended in the Standard Review Plan (10 SSE and 50 OBE) by approximately a factor of three.
It should be noted that 3 out of 20 tests did experience 1/4 inch bolt pullout at a load less than the static design load (which is based on a safety factor of four).
The 1/4 inch pullout occurred at approximately 80 percent of the static design load.
Thus, the staff finds that the dynamic testing performed by Bechtel and Hanford Engineering Development Laboratory provide similar results.
Both testing results appear to indicate that a safety factor of four for dynamic vibratory loads is adequate for the number of peak cycles associated with seismic events, and that the ultimate anchorage capacity is not completely lost although some degree of bolt slippage might occur. Thus, the staff concludes that based on the dynamic testing discussed above, the wedge-type expansion anchor bolts when designed with a safety factor of fcur to the static anchor capacity and when properly installed is capable of withstanding the dynamic loads associated with a design basis seismic event.
The staff discussed the effect of the prying action of the support baseplates on the anchor bolts. The applicant does not account for prying effects in its anchor bolt design for non piping supports.
The AISC, ACI-318, and ACI-349 criteria does not address the prying action of baseplates on bolt loads. How-ever, ACI and AISC does address the steel-to-steel prying action.
Bechtel concluded that because the concrete is relatively soft compared to steel, the j
effects of the baseplate prying action will be small.
In addition, Bechtel believes that the slippage of the bolt does not degrade the ultimate anchorage capacity. The staff review of responses to IE Bulleting 79-02 found similar conclusions. A test report summary by Sargent & Lundy(
) found that for a flexible baseplate with four expansion anchors, the prying action is of the i
11/22/83 12 MIDLAND HVAC STRUCT DESIGN
order of 15-20 percer.t of the applied load.
The S&L report also concluded that the small increase was much lower than the expected increase in an assembly with embedded steel bolts where the prying action was calculated to be 110 per-cent because of the effective lower stiffness of expansion anchors in concrete.
Thus, based on the consistency in the results of the prying action of base-plates on concrete anchor bolts as discussed above, the staff concludes that the prying action will not cause a significant increase in the expansion bolt loads.
With regards to the use of lesser grade materials, the staff believes that it is unlikely that material substitution is a significant concern for expansion anchor bolts because of their unique application and configuration.
Use of low strength bolts or bolts made of poor quality materials would likely become evident during bolt installation when the bolt preload torque is applied.
A low-strength or poor quality bolt would likely yield or break before the required preload torque could be achieved.
If an expansion anchor bolt were made with a substitute material of a lesser quality (e.g., A307 material) and remained undetected following application of the preload, high shear strengths given in the manufacturer's catalogs could be unconservative.
However, the staff believes that the safety factor of four when applied to the manufacturer's ultimate shear loads provides an adequate margin of safety to account for substitute materials.
The ultimate anchor pullout load is not likely to be affected because the ultimate anchor pullout load is in all cases less than the tensile requirements for A307 bolts.
A comparison of the bolt preload values with ASTM A307 tensile strength require-ments is shown in Table 10.
The staff has found that use of lesser grade ma-terials could be a potential concern with the ITT Phillips Wedge Anchors (7/8 inch diameter only).
ITT Phillips supplies both a nuclear grade and a non-nuclear (commercial) grade expansion anchor bolt.
For Midland, the pro-curement specification specifies an NWS-7880 (nuclear grade) wedge anchor.
The difference in the nuclear grade and the non-nuclear grade bolts is in material and traceability.
The nuclear grade bolt material is AISI 1144 grade with an f
average tensile strengi.l. of 100-120 ksi and a yield strength of 90-110 ksi.
I The nuclear grade is stamped "NWS" and has a " gold" chromate finish.
The com-mercial grade bolt is 1213 to 1215 carbon steel (no traceability) with a tensile strength of 80-95 ksi and a yield strength of 70-80 ksi. The 11/22/83 13 MIDLAND HVAC STRUCT DESIGN
l commercial grade is stamped "WS" and has a silver finish.
In accordance with the manufacturer's recommendations, the nuclear grade bolt for 7/8 inch diameter has a pullout ultimate load capacity of 14 ksi (vs 11.85 ksi for commercial) and a shear capacity of 22.5 ksi through the threads and 30.0 ksi through the shank (vs. 24.9 ksi for commercial). Thus, the use of a commercial grade bolt instead of a nuclear grade bolt could reduce the design capacity by 15-20 per-cent. Based on a review of the dynamic test data, the staff concl.udes that a reduction of 15-20 percent of the anchor capacity, or in equivalent terms, a reduction of the safety factor from 4.0 to 3.2 appears to be acceptable.
V.
Conclusions A significant effort has been expended by the staff on the subject of expansion anchor bolts largely because of the many uncertainties involved in the actual strength of the installed anchor bolt. The conclusions of the tests, performed on the expansion bolts were based on properly installed bolts and under con-trolled loadings.
Some uncertainties which could affect the overall findings of the staff include 1) improperly installed expansion anchor bolts, 2) the dynamic effects of a seismic event on the anchorage capacity of floors and walls in which the expansion anchor bolts are installed, 3) the long-term aging effects on the anchot strength, and 4) the uncertainties in the dynamic load-ings itself. The staff has found that the most limiting component in the HVAC structural design is the expansion anchor bolt assembly.
Although the factor of safety used in the design of the anchor bolt capacity appears to be adequate to account for the static and dynamic loads associated with normal and design basis accidents, there is a certain degree of uncertainty involved with as-installed expansion anchor bolts and the actual loading conditions which could occur that remains as potential concerns of the staff.
These concerns extend beyond the scope of this evaluation and into the areas identified above where further generic development should be performed. Thus, our findings on the design margins do not take into account the above uncertainties, except in a qualitative manner.
Based on a detailed review of the typical design margins available in the structural design of the HVAC ductwork and supports, the staff has concluded that there is an adequate margin between the stress or load level that would result under normal and design basis accident conditions and the stress or 11/22/83 14 MIDLAND HVAC STRUCT DESIGN r.----,
~,,n,---~
.n,
s to load level that would result in structural failure of the HVAC ductwork and support systems. The staff further concludes that the available design margin provides adequate compensation for potential degradations in the structural integrity that could result from substitution of lesser quality or lesser grade materials. Therefore, the staff finds that the overall structural design of the Midland HVAC systems is adequate and provides a sufficient s
margin of safety to failure under normal and design basis accident conditions.
1 11/22/83 15 MIDLAND HVAC STRUCT DESIGN o
i VI.
References 1)
" Technical Specification for Seismic Class I Heating, Ventilating and Air Conditioning Equipment and Ductwork Installation," for the Midland Plant Units 1 & 2, 7220-M-151A(Q), Rev. 15, 2)
" Design Guide for HVAC Supports," (DRAFT) Calc. No. 3471(Q).
3)
" Design Guide for Nuclear Power Plant Seismic Category I Rectangular HVAC Ducts (DRAFT)," dated April 15, 1978.
l 4)
" Report on Testing of Class I Seismic HVAC Duct Specimens for the Limerick Generating Station, Units 1 & 2," April 1976.
5)
Calculation No.34-323(Q), Revision 0, dated 10-11-83.
6)
Teledyne Engineering Service, Summary Report, " Generic Response to USNRC I&E Bulletin Number 79-02, Base Plate / Concrete Expansion Anchor i.
Bolts," August 1979.
1 7)
NUREG/CR-2137, " Realistic Seismic Design Margins of Pumps, Valves, and Piping," June 1981.
8)
FFTF Report, " Drilled-in Expansion Bolts Under Static and Alternating Load," January 1975 (BR-5853-C-4).
9)
" Final Report USNRC Anchor Bolt Study Data Survey And Dynamic Testing,"
4 NUREG/CR-2999, dated December 1982.
- 10) Sargent & Lundy report, " Summary Report on Static and Dynamic Relaxation Testing on Expansion Anchors in Response to I&E Bulletin 79-02," dated July 20, 1981.
t 11/22/83 16 MIDLAND HVAC STRUCT DESIGN
\\_ y;
(
Af 4 %
9y k 4 I
t<
. Table 1 v'
i e,
tNs--
g
- ' t HVAC Support Material N
ASTM-ASTM Minimum
,A.3TM Minimum-M-151 Minimum
- r..
Mateial,\\
Yield Strergth Teasij,e Strength Yield Strength Specifikatici (ksi)
"( k's i )
(ksi)
Notes v
/
's, t
. I,.
A36'"7 s
[ d:'@6
,' 58-80 '
same as ASTM t4 c, 50 6E same as ASTM
'A 572 Gr. 50 s
v same as ASTM plate 50
,A 284 Gr. A 25 4
m.
4 same as ASTM tube A 500 Gr. B '
46' 58 1
4 ]
t-l steel 1
7A'575 (M1020) not required not required 36 angle
%' 4, gg
,..c,~
y sf
' g xg i
\\
1, e
j y
s g
,c s,
- (
's.
- S i il
~%
s s. ;-
t
\\.,
1 i
\\,LW
\\
j t
i
,s
'i 1
1.
s s
i*
s 4,
\\
~ ' '
A.
g s
4 t
\\
f i
T
[N.
w t
-s tr j
11/22/83
17 MIDLAND HVAC STRUCT DESIGN s
g
'O f
, / 't i.
. - - - D
,.e,
~....,.,,
-..-. m
- c..
s.
Table 2 HVAC Support Material Properties (A575)
I ASTM-A575 Minimum Yield St;'ength Grade M1008 34.0 ksi Grade M1010 35.7 ksi Grade M1015 36.1 ksi Grade M1020 37.2 ksi 11/22/83 18 MIDLAND HVAC STRUCT DESIGN wo--.
F
.y.
3 Table 3 4
HVAC SUPPORTS Tabulation of Calculated vs. Allowable Stress i
Reference d.
Calculated Stress Design Location Calc. No.
Description Allowable Stress Margin
~
Control Room 21 G (4.4143) W 6 x 12 0.23 4.3 L3x3xk 0.19: 0 5.3 L2x2x r'
- 0.13 r
es7.7 L2x2x%
0.13 7.7 L3 x3 xk 0.05
-20.0 weld 0.76 1.3 weld 0.10 10.0 weld 0.61 1.6 weld 0.51 2.0 Control Room 21 G (4.146) all structural members 0.48 2.1 weld 0.03 33.3 anchor bolt 0.50 2.0 Control Room
'29 0 276 L 3 x 3 x % (all) 0.33 3.0 W 6 x 12
- 0.04 25.0 TS 2 x 2 x %
0.04 25.0 weld 0.42' 2.4 weld 0.73 1.4 weld 0.57
- 1. 8 i
Service Water Bldg 648-5126 TS 3 x 3 x %
0.15 6.7 TS 2 x 2 x k 0.09 11.1 L2x2x%
0.13 7.7 weld 0.03 33.3 weld 0.12 8.3 weld 0.68
anchor bolt -
0.80 1.3 Auxiliary Bldg 21 F (3.136)
L2x2xk 0.13 7.7 TS 2 x 2 x %
0.14 7.1 weld 0.04 25.0 weld 0.20 5.0 weld 0.15 6.7 weld 0.04 25.0 anchor bolt 0.58 1.7 anchor bolt 0.34 2.9 11/22/83
- 19 '
MIDLAND HVAC STRUCT DESIGN J
f.
s.
r Table 3 (continued) 1 Reference Calculated Stress Design Location Calc. No.
Description Allowable Stress Margin Auxiliary Bldg 21 I (6.95)
TS 4 x 4 x %
0.32 3.1 TS 2 x 2 x %
0.48 2.1 i
L2x2x%
0.36 2.8 PL % x 18 0.13
- 7. 7 i
- 2. 9 weld 0.15 6.7 weld 0.24 4.2 weld 0.29 3.4 weld 0.25 4.0 weld 0.10 10.0 weld 0.23 4.3 weld 0.32 3.1 L4x4x 0.44 (shear 2.3 controlling) t:
o9Ah rs = n h!
PL
- M 11/22/83 20 MIDLAND HVAC STRUCT DESIGN e *-
I~
~
Table 4 HVAC Ductwork Material ASTM ASTM Minimum ASTM Minimum M-151 Minimum Material Yield Strength Tensile Strength Yield Strength Specification (ksi)
(ksi)
(ksi)
A526 not required not required 30 A526 not required not required 30 A366 not required not required 30 A607 Gr. 50 50 65 same as ASTM A240 Type 304 30 75 same as ASTM 4
11/22/83 21 MIDLAND HVAC STRUCT DESIGN
r t-Table 5 Summary of HVAC Duct Analysis Results 3) f Sheet
[g Allowable Governing Calculated Duct Size Metal Stiffener Pressure (psi)
Allowable Worst Loading Desig Sheet (inches)(1)
Gauge Metal Stiffener Pressure (psi)
(psi)(2)
Margi Control Room (Aux Bldg) 60x26 18 _
L2x2x3/16 0.86 0.69 0.69 0.294 2.39 36x26 16 L1 x1\\x1/8 1.40 1.40 1.40 0.301 4.65 Diesel Generator Bldg 60x60 16 L2x2x3/16 1.086 0.691 0.69 0.253 2.73 30x40 16 Ll\\x1\\x1/8 1.322 1.40 1.32 0.253 5.22 Service Water Pump Structure 72x44 16 L3x3x3/16 1.064 1.102 1.102 0.230 4.79 72x24 18 L3x3x3/16 0.865 1.102 0.865 0.223 3.88 52x44 16 L2x2x1/16 1.237 0.98 0.98 0.230 4.26 42x26 18 L1 x1\\x1/8 1.111 0.94 0.94 0.223 4.22 28x26 18 L1 x1\\x1/8 1.408 1.04 1.04 0.223 4.66 Auxiliary Building 108x 16 14 C 3x5.0 1.14 0.47 0.47 0.335 1.40-108x16 14 C 5x6.7 1.14 1.25 1.14 0.628 1.75 60x32 18 L2x2x3/16 1.15 0.69 0.69 0.326 2.1%
3.7g%
38x38 16 L1 x1\\x3/16 1.44 1.22 1.22 0.330 3.8 76x40 16 L3x3x3/16 1.04 0.97 0.97 0.254 50x40 16 L2x2x3/16 1.25 1.08 1.08 0.259 4.1 54x36 18 L2x2x3/16 0.98 0.89 0.89 0.320 2.7g 4.4 28x14 18 L1x1x1/8 1.41 1.05 1.05 0.234 24x24 18 L1x1x1/8 1.56 1.59 1.56 0.223 7.0:ge 12x6 18 L1x1x1/8 2.59 11.10 2.59 0.234 11.0$
60x36 16 L3x3x3/16 1.15 1.70 1.15 0.593 1.95 (1) Largest duct size for the same gauge sheet metal and stiffener.
(2) Worse case loading is Dead Load + P + W where P - operating iressure W - wind load.
The worst case loading bounds seismic load combinations.
(3) Summary of results from Bechtel Calc. No. SQ-180(q) dated 5/16/83.
Stresses due to dead load, seismic load, wind and internal pressures are converted to equivalent internal pressure loads for comparison.
M) h* Y C5 cA<--/
1 11/22/83 22 MIDLAND HVAC STRUCT DESIGN O
e
p.
9 Safe Shutdown Earthquake Max. Tension Operating In Bolt of Max. Calculated Load Sheet Pressure Companion Allowable Design Duct Size Thickness. in W.G.
Flange Tension Allowable Load Margin 60" x 26" 16 Gage 13" 1200 lb 2340 lb 0.51 1.96 60" x 60" 14 Gage 13" 1900 lb 2340 lb 0.81 1.23 30" x 30" 18 Gage 13" 586 lb 2340 lb
-0.25 4.00 60" x 60" 16 Gage 4"
840 lb 2340 lb 0.36 2.78 i
i E
F
[
1 l
11/22/83 23 MIDLAND HVAC STRUCT DESIGN
y*
c e-("I_
Table 8 Concrete Expansfon Anchor Bolt Material Properties i
Type Size Material Requirements (inches)
Properties Met Stud'(bolt) 1/4-1/2 AISI1[L41 ASTM A108 5/8-1 AISI 1144 ASTM A108 Expansion ANSI 1050 Wedges spring steel Nuts commercial ASTM A307 manufacture Washers SAE material ASA B27.2-1949 0
11/22/83 24 MIDLAND HVAC STRUCT DESIGN 1
/
i e-M DYNAMIC TEST RESULTS (FROM REFERENCE 9) i Ultimate Static Test Results Strength No. of Ult. Load Load at 1/4" Displ.
Test No.
Anchor Type Load Type (Kips)
Preload**
Cycles Kips Note Kips DW-1 Wedge Tension (25.3)
Full 845 25.3 1, 2 15.2 DW-1R Full 141 20.2 1, 2 15.2 DW-2 Full 255 25.3 1, 2 10.1 DW-3 Half 239 25.3 1, 2 15.2 DW-4 Half 181 25.3 1, 2 10.2 DW-5 Zero 133 20.2 2, 3 5.0 DW-5R Zero 105 15.2 2, 3 5.0 DW-6 Zero 179 25.3 2, 3 10.2 DW-7 Shear (24.0)
Full.
208 28.8 2, 4 24.0 g
DW-8 Full 179 24.0 2, 4 14.4 DW-9 Half 176 24.0 2, 4 14.4 DW-10 Half 165 24.0 2, 4 14.4 DW-11 Zero 163 24.0 2, 4 14.4 DW-12 Zero 167 24.0 2, 4 14.4 DW-13 Combined
- Full 161 25.3 2, 5 10.1 DW-14 Full 135 20.2 2, 5 15.2 x
DW-15 Half 139 20.2 2, 4 5.0
(
DW-16 Half 161 25.3 2, 4 10.1 5
DW-17 Zero 161 25.3 2, 4 15.2 DW-18 Zero 140 20.2 2, 4 15.2 y
- Tension = 1.732 NOTES
- =
Shear 1.
Anchor pullout, no concrete failure g
2.
Test stopped at 1" displacement
- Full preload:
125-175 foot pounds 3.
Anchor pullout and local concrete failure m
Half preload:
62-88 ft,ot pounds 4.
Anchor shear failure Zero preload:
Finger tight 5.
Anchor shear and local concrete failure i
,*f a
Table 10 Comparison of Anchor Bolt load Requirements Minimum A307 Bolt Preload Minimum Ultimate Anchor Requirement Torque Anchor Bolt Pullout Load for Tensile Bolt Diameter (ft-1bs)(a) Preload (1bs)(b) Capacity (lbs)(d) Strength (lbs)(c) 1/2" 35 2,800 5,510 8,500 5/8" 130 8,320 9,100 13,550 3/4" 240 12,800 13,400 20,050 7/8" 275 12,571 14,000 27,700 1"
425 17,000 18,900 36,350 (a) per Specification 7220-C-305(Q) Rev. 17 (b) Calculated using the equation:
T = KDL where:
T = preload torque applied K = assume 0.3 for unlubricated threads D = nominal bolt diameter L = bolt preload force (c) per ASTM Specification, " Standard Specification for Carbon Steel Externally and Internally Threaded Standard Fasteners," A307-76b.
(d) per Hilti Fastening Systems and ITT Phillips Drill Company Catalogs Based on 3500-4000 psi strength concrete 11/22/83 26 MIDLAND HVAC STRUCT DESIGN
?*
.m Table 10 Comparison of Anchor Bolt Load Requirements Minimum A307 Bolt Preload Minimum Ultimate Anchor Requirement Torque Anchor Bol Pullout Load f r TensH e Bolt Diameter (ft-lbs)(,) Preload (1bs)g) Cipacity (1bs)(d) Strength (1bs)(c) 1/2" 35 2,800 5,510 8,500 5/8" 130 8,320 9,100 13,550 3/4" 240-12,800 13,400 20,050 7/8" 275 12,571 14,000 27,700 1"
425 17,000 18,900 36,350 (a) per Specification 7220-C-305(Q) Rev. 17 (b) Calculated using the equation:
T = KDL where:
T = preload torque applied K = assume 0.3 for unlubricated threads D = nominal bolt diameter L = bolt preload force (c) per ASTM Specification, " Standard Specification for Carbon Steel Externally and Internally Threaded Standard Fasteners," A307-76b.
(d) per Hilti Fastening Systems and ITT Phillips Drill Company Catalogs Based on 3500-4000 psi strength concrete 11/22/83 26 MIDLAND HVAC STRUCT DESIGN
--- n-w w..