ML20247H071

Description & Verification Summary of Computer Program, Gappipe

ML20247H071
Person / Time
Site:	Byron, Braidwood, 05000000
Issue date:	06/30/1989
From:	ROBERT L. CLOUD ASSOCIATES, INC.
To:
Shared Package
ML20247H068	List: ML20247H061 ML20247H071 ML20247H079
References
RLCA-P182-01-89, RLCA-P182-01-89-001, RLCA-P182-1-89, RLCA-P182-1-89-1, NUDOCS 8907280268
Download: ML20247H071 (59)

Text

- _ - _ _ _ _ _ _. . _ _ _ . _____

i Robert L. Cloudand Associates, Inc t

DESCRIPTION AND VERIFICATION

SUMMARY

OF COMPUTER PROGRAM, GAPPIPE RLCA/P182/01-89/001 June 30,1989 i

i Submitted to:

Office of Nuclear Regulatory Regulation United States Nuclear Regulatory Commission By:

Commonwealth Edison Company for the application at Byron and Braidwood Nuclear Facilities f Prepared by:  ;

i Robert L Cloud & Associates,Inc. )!

1 125 University Avenue 105 Summerwood Way, Suite B i Berkeley, California 94710 Aiken, South Carolina 29801 J

(803) 648-8801 (415) 841-9296 l

1

}

8907280268 8907j4 PDR I ADOCK 0D000454 F' FDC

,. 3 -* %_

~

[

jl Communwtalth Edison ? .

- 72 West Adams Street, Chicago, Illinois L v(

1 V t'l . Address Reply to: Post Office Box 767 :

s Chicago, lilinois 60690 - 0767 ~

July.14,-1989

's

'Dr.tThomas.E.'Murley, Director Office of' Nuclear Reactor, Regulation U. S. Nuclear. Regulatory Commission Washington,LDCf 20555 Subject Byron Station' Units 1 and 2 Braidwood' Station Units 1 and 2 Non-Linear Piping Analyses to Justify Removal of Snubbers and Pipe Stpports

'NRC Docket Nos. 50-454/455 and 50-456/452 Dear Dr Murleys

-A meeting was held between Commonwealth Edison and members of your staff on May 2,.1989,'in Rockville, Maryland to. discuss the Robert L.' Cloud-Associates non-linear piping.enalyses that would be used to justify removal'of snubbers and pipe supports at Byron and Braidwood Stations. This letter.and the attachments. provide the formal submittal.to satisfy the requirements of-the Standand Review Plan Section 3 9.1 titled."Special Topics for Mechanical Components". Part II.2' delineates the specific requirements for computer

~

procrams to demonstrate thej r applicability and validity. Attach:aehts 1 and 2 to.this letter provide the required information in the prescribed format.

Commonwealth Edison requests review and. approval of.the analyses.by February 1, 1990, to permit modifications on-Byron Unit 2 during..the September: .

1990 refueling outage. ~

Please direct any questions on this matter to this office.

Very truly yours, :j

'I l R. A. Chrzanowski Wl Nuclear Licensing Administrator Attachments I

LE cct Byron Resident Inspector L. N. Olshan - NRR S. P. Sands - NRR l ~

[ Region III Office j Office of Nuclear Facility Safety - IDNS '

[

0200T:10

j

)

)f l I

) l

)-

f l

1 1

i l l

'l 4 AIIACIR$DII_1

)

DISfRU'TlDtlAND_XERIFICATIQU EUt9LARY OF CQldPJJJER PROGRAM, GAPPIPE I

1 l

l i

l l

l 1

! 02007:11 I

I  !

1 i

{-

l Robert L. Cloud and Associates, Inc

.RLGA l

DESCRIPTION AND VERIFICATION

SUMMARY

OF COMPUTER PROGRAM, GAPPIPE RLCA/P182/01-89/001 June 30,1989 Submitted to:

Office of Nuclear Regulatory Regulation United States Nuclear Regulatory Commission By:

Commonwealth Edison Company for the application at Byron and Braidwood Nuclear Facilities Prepared by:

Robert L Cloud & Associates,Inc.

125 University Avenue 105 Summerwood Way, Suite B Berkeley, California 94710 Aiken, South Carolina 29801 (415) 841-9296 (803) 648-8801

t ,

-l L ']

I TABLE OF CONTENTS

):

.)

-Pace No.

~

l

1.0 INTRODUCTION

1 j

)  !

2.0 ' COMPUTER; PROGRAM, GAPPIPE- 2 2.1 Program Overview- 2 2.2 Program Organization 4 3.0 VERIFICATION OF GAPPIPE 6 3 .1' Comparison with NRC Benchmark 7 Problems

'3. 2 Correlation.with Shake Table. 12 Test. Data and ANSYS 3.3 Correlation with HDR Experimental 15 Tests 3.4 . Comparison with Literature 17 1 Analytical Results 4.0

SUMMARY

AND CONCLUSIONS 20

5.0 REFERENCES

21 FIGURES / TABLES APPENDICES:

L A. Input Listing of Uniform Benchmark Problems B. Input Listing of. ISM Benchmark Problems C. Comparison Summary of Uniform Benchmark Problems D. Comparison Summary of ISM Benchmark Problems l

l

)

i

)

f

R LIST OF FIGURES 1 i i

Floure No. Title Pace No.  !

2.1 Flow Chart of GAPPIPE-Program 22 1 l 1

3.1 UNI - Benchmark Problem No. 1 23 3

3.2 UNI - Benchmark Problem No. 2 24 j i a i

3.3 UNI - Benchmark Problem No. 3 25 1

3.4 UNI - Benchmark Problem No. 4 26 U

3.5 UNI - Benchmark Problem No. 5 27 3.6 UNI - Benchmark Problem No. 6 28 3.7 UNI - Benchmark Problem No. 7 29 L 3.8 ISM - Benchmark Problem No. 1 30 3.9 ISM . Benchmark Problem No. 2 31 3.10 ISM - Benchmark Problem No. 3 -32 3.11- . ISM - Benchmark Problem No. 4 33 3.12 Single Span Test Configuration 34 3.13 3-D Hovgaard Bend Test Configuration 34 i

3.14 Single Span Dynamic Test Configuration 35 3.15 3-D Hovgaard Bend Test Configuration 36  ;

3.16 Time History and Response Spectra 37 Plots (0.82 G ZPA) 3.17 Time History and Response Spectra 38 Plots (1.33 G ZPA)  ;

3.18 Comparison of Pipe Bending Stress 39

@ 1/3 Location Input Level =

0.82 G ZPA 3.19 Comparison of Pipe Bending Stress 40 0 1/3 Location Input Level =

1.33 G ZPA 11

i p-LIST OF FIGURES (continued)

~i Flaure No. Title Pace No. j L

L 3.20 Comparison of. Impact Forces 41: 1 I

Input Level =.O.82 G ZPA l

l 3.21' Comparison'of Impact Forces 42 {

Input Level = 1.33 G ZPA 3.22 Analysis Model.of the 3-D 43 Hovgaard Bend Test Configuration 3.23 HDR SHAG Experiment Test 44 I Correlation Test Configuration 3.24 HDR. SHAG Experiment Test. 45 Correlation Comparison of l

Maximum Pipe Accelerations 3.25 .HDR SHAG Experiment Test 46 o- Correlation Comparison of Maximum Pipe' Stresses.

3.26 HDR SHAG Experiment Test , 47 Correlation Comparison of Maximum Support Loads.

L 3.27 Description of STARDYNE 48 Verification Problem 30 j 3.28 Geometry and Forcing Functions 49 of the Molnar Verification Case i

i l

1 l

1 l' iii L

i

l,'

i- 1 LIST OF TABLES i; Table No. Title- P_aae No.

3.1 -Comparison of Maximum Pipe :50-

! Bending Stress for the h;. 3-D Hovgaard Bend Tests u 3.2 Comparison Summary of GAPPIPE 48 L- and' Theoretical Results 3.3 Comparison Summary of GAPPIPE' 51 and Molnar Results V

l~

l

.{

l i

i l

i, iv i

1

[

t ll g

1.0 INTRODUCTION

l

The GAPPIPE computer' program.is a general purpose piping analysis program developed by Robert L. Cloud &

l- Associates, Inc. (RLCA) and. sponsored in part by the I . Electric Power Research Institute (EPRI). GAPPIPE performs both linear and nonlinear elastic' analyses of i' 1 three-dimensional piping systems subject to' thermal expansion, imposed displacements, internal pressure,-

externally applied loads, seismic and fluid transient loads or motions. In addition, GAPPIPE contains a'postprocessor capable of performing stress evaluation of piping l-components in accordance with the ASME Boiler and Pressure Vessel Code, Section.III requirements.

i. GAPPIPE differs from other piping computer programs in that it has the' capability to analyze piping systems containing gaps.- GAPPIPE has two. analysis methods to

' compute the dynamic responses of such systems. 'The first method is nonlinear time history analysis by modal

i superposition and pseudoforce-representation of gap responses. .This method is most suitable for the simulation of piping responses induced by fluid transient loads or excitations where the input cannot be easily or adequately characterized by response spectra.

'For excitations defined by response spectra, GAPPIPE offers a second analysis method that uses the response spectrum analysis technique and the method of equivalent L

linearization [Ref. 1) to account.for the nonlinear behavior of gaps. In this method, GAPPIPE can use either

-uniform enveloped response spectra or different spectra at different supports using the independent support motion technique.

This report provides a summary description of GAPPIPE j and presents verification results to demonstrate its i 1

applicability and validity as a new piping design analysis computer program. The information provided in this report are required under Section 3.9.1 of the NUREG-0800 Standard Review Plan for the specific purpose of including the ,

, computer program GAPPIPE into the FSAR of the Byron and Braidwood nuclear power plants. GAPPIPE is intended to be (

used for improving the performance and reliability of the '

Byron and Braidwood piping systems and mechanical components. I 1

l:

L- ,

b:

2.0'. COMPUTER PROGRAM, GAPPIPE b -

Program Developer: . . Robert L. Cloud'& Associates, Inc.

125 University Avenue Berkeley, California 94710

!=

. (415) 841-9296.

i b Program Language: Fortran 1 :, . .

j: Program, Version:' ' Version 2.0, dated March 31, 1989; compiled under DEC/VMS operating system 2.1 PROGRAM OVERVIEW GAPPIPE is developed based on the public domain code SAPIV [Ref. 2). However, extensive modifications.have been L made by deleting the unnecessary elements and functions, changing and improving existing subroutines, and adding new logic,-elements,' subroutines, and capabilities. The L

current version of GAPPIPE-is capable of performing a wide range of analyses as described below.

Static Analysis GAPPIPE analyzes the following piping static loading conditions:

1. . Thermal expansion

2. Gravity loading in any of the 3 global directions

3. Applied-loads at arbitrary pipe locations

4. Support movements

5. Internal pressure effects

'6. ' Residual modes' effects Dynamic Analysis GAPPIPE performs the following dynamic analysis options:

1. Eigenvalue solution (frequency and mode shape determination using either the Determinant Search or the Subspace Iteration methods).

2. Response spectrum analysis of a system subjected ,

to either uniform support excitation or independent support motions. Directional responses may be combined by either absolute summation (ABS) or the square-root-of-sum-of-square (SRSS) methods. The modal combination options are:

2

+

.. I

, i p

a

["4 a.' .

Square-root-of-sum-of-squares (SRSS) b '. - Absolute summation c.. .NRCJ10% method.

d. :NRC. grouping method.

L ,

3. Equivalent linearization: analysis of a. system g subjected to either uniform-support excitation or independent _ support motions._ This analysis option offers the same modal and directional L combination methods as in the response spectrum analysis.

-When independent support motions are used,.~either?

'the SRSS or the ABS method may'be used to" combine the results. associated with different support

, groups.

4. Seismic anchor movement. analysis. The SRSSfor the ABS method may be used to combine the results associated with different support groups.

Time History Analysis using either arbitrary'

~

5.

nodal forcing functions or anchor acceleration time histories.. Time responses are determined' using the modal superposition method. Both linear and nonlinear analyses can be performed.

Code Comoliance Evaluation GAPPIPE is capable of generating load combinations.of f existing-load' cases and creating ~new load cases based on

.the existing ones. GAPPIPE performs support summary and code' compliance evaluation in accordance with the ASME Section III Subsections NB-3600 and NC-3600 rules ~for classes 1,':2, and.3. piping systems.-

The capacity of GAPPIPE depends primarily on the total.

. number of nodal points in'the piping model, the number of vibration ~ modes' requested:in the dynamic analysis, and the computer system used. There is practically no restriction on the number of elements used, the number of load cases or the order and bandwidth of the system stiffness matrix.

For systems with nodes arbitrarily numbered continuously from the first node to the last node, GAPPIPE performs internal renumbering of nodes to minimize the memory requirements. .

1 The piping systems to be analyzed may be composed of  !

combinations of the following elements currently available in GAPPIPE:

3

1

{ n' ,

N 1 1

i j

' 1 <. Pipe elements (straight and curved segments)

2. Boundary elements (used to model pipe supports j including rigid anchors, springs, struts,  !

snubbers and Seismic. Stops) 3.: Three-dimensional truss. elements L 4. Three-dimensional beam elements.

L 2.2 PROGRAM ORGANIZATION _

~

1

.A GAPPIPE analysis typically consists'of three phases: .model. generation, analysis execution, and output, ]a

-processing. A flow diagram.of these' phases is illustrated in Figure 2-l'.

The first analysis phase is model generation where model data are read and system stiffness and mass matrices are formulated. A user generated-input data file specifies the geometry information of the piping system. The input-data file can be used to general mode 1' geometry plots.

U The next analysis phase performed by GAPPIPE is the execution of-the various analysis options. For static analysis,'even.though multiple load: cases can be executed' in a single run, a single-load-case run is recommended.

For dynamic analysis, only one analysis is allowed in each run. Restarts can be performed'for the Equivalent

'Linearization Analysis (NDYN=5,-5) or the Equivalent Linearization Analysis using the Independent Support Motion-input 1(NDYN=6,-6) by using.the restart files generated by

.these analysis options. The'eigenvalue solutions are L calculated.in any of the dynamic analysis options.

The final' phase of a GAPPIPE analysis is the processing of analysis results. GAPPIPE generates a number-of files after each execution run. These are the Plot Files, the Analysis Print. File, the Restart Files, temporary files, and the Postprocessing File.

The Plot Files contain the plot data of the model geometry and the mode shapes. . Plots of either the model or  ;

the mode shapes are generated by executing an auxiliary program, called GAPPLOT.

The Analysis Print File is created'for each GAPPIPE execution run. The amount of information to be printed is controlled by print control parameters requested by the user in the input data.  ;

The Restart Files are created and used primarily for ,

continuing the analysis solution when performing the Equivalent Linearization Analysis and the Independent 4

q l

-_ - - - - 1

[ . .

Support' Motion AnalysisL(NDYN=5,-5,6,-6). -Since'both analysis options involve an iteration solution procedure, 1the restart. capability provides an efficient means of using the existingLfiles.and mode shapes 7 to save computation time.

l L .GAPPIPE creates and uses a number;of temporary files

~during.each execution run.- Most of these' files are used as

( part of the program operation and are discarded after the-

~ cexecution. They contain no'useful information'for the.

user. .Only one file, called the Trace File,.may be-

' reviewed:by the user. The. Trace' File contains a procedural record of a GAPPIPE run. It would be of interest if.an abnormal' termination of'a GAPPIPE run occurs.

The Postprocessing File contains all analysis-results. It is formatted for easy access by the user.using either the GAPPIPE postprocessing program, GAPPOST,'or custom postprocessing programs' created by.the user.

l

'GAPPOST performs load case combinations and. evaluates; code compliance in accordance to the requirements of ASME Section.III'NB-3600'and NC-3600 for Classes'1, 2 & 3' piping L components.. It generates a. stress summary report.for the piping. components.and supports. In addition, GAPPOST can' also create new load cases based on the combination of existing' load cases.

j.

i I

4 1

5 i

'l 4

5 L

i

{ ,' t.

L t

I[ ' .3.0L VERIFICATION.0F GAPPIPE F

The purpose of'the effort. presented in this section is'

.to1 verify the adequacy'of the computer program GAPPIPE for

.use in'the dynamic analysis andLdesign-of nuclear. piping

[ systems. The verification. effort consists of four-Q ' independent. sources of comparison:

1.

y 1.. Comparison with.the NRC benchmark problems L

described in NUREG/CR-1677, Volumes ~ILand II

[Ref. 3,4),

2.- Correlation with laboratory shake table test data-

.[Ref. 5) and the ANSYS computer program [Ref. 6),

3.- ' Correlation with the<in-situ HDR Experimental Tests (Ref. 7),

4. Comparison with literature analytical results.

b 'The comparison using the NRC benchmark solutions is a mandatory verification procedure specified in Section 3.9.1-of.the NUREG-0800 Standard' Review Plan to meet:the L requirements of 10CFR Part-50, Appendix B and GDC 1. This comparison of the NRC benchmark 1 solutions.with GAPPIPE

'results is intended to validate the linear: response spectrum analysis option and-the associated programming.

structure and logic of GAPPIPE.- These: include the validation of element formulation,: solution algorithms, eigensolution techniques, modal combination. methods, and i

element load and stress calculations.

The second source of. verification is to use the shake-L table' test' data which wera obtained by RLCA as part of the GAPPIPE research and development effort. The intent is to validate the equivalent"linearization analysis optionLof-GAPPIPE by correlating the GAPPIPE solutions with actual test measurements. An alternate comparison is also made' with nonlinear time history solutions calculated using the l ANSYS computer program. This comparison shows the accuracy j of GAPPIPE solutions, which are based.on the response spectrum technique, relative to the ANSYS nonlinear time  !

history results which are generally considered as'" exact" analytical solutions.

The third verification source is the in-situ HDR _ ,

L. experiment sponsored in part by the U.S. Nuclear Regulatory ,

Commission Office of Research. In-situ piping and equipment dynamic responses due to seismic-like excitations j i were recorded for both snubber and gapped support piping ]

designs. The verification performed here compares the 6

r 1

-____2______________.. ')

GAPPIPE results with the recorded test data. The intent is I to show that the analytical solutions of gapped support piping system designs obtained by the GAPPIPE equivalent linearization method are valid design solutions and are comparable to the current industry piping analysis of snubber support system designs. This verification source supplements the preceding efforts in that the HDR test data are realistic in-situ responses of actual hardware and physical conditions. Furthermore, identical tests were performed for snubber and gapped support system designs. j The last source of comparison is the use of analytical results published in the literature. Two examples are used. One is reported and documented in the STARDYNE Verification Manual (Ref. 8], and the other is taken from an ASME technical paper by Molnar, et al (Ref. 9]. The intent of the comparison is to verify the time history analysis option of GAPPIPE. j In the following subsections, the description and results of each of these four sources of comparison are discussed and summarized.

3.1 COMPARISON WITH NRC BENCHMARK PROBLEMS A total of eleven Benchmark Problems are provided in NUREG/CR-1677, Volumes I and II [Ref. 3,4] for the purposes l of verifying the adequacy of any computer programs used for dynamic analysis and design of nuclear piping systems.

There are seven problems in Volume I for analysis using the -

Uniform Support Motion Response Spectra method, which will i be referred to as the UNI Benchmark problems from here on. l In Volume II, there are four problems for analysis using i

! Independent Support Motion Response Spectra Method, which I will be referred to as the ISM Benchmark problems from here on.

For the UNI Benchmark problems, the seven problems range from simple to complex configurations which are .

assumed to experience linear elastic behavior. The i solutions provided include: (1) frequencies, (2) modal participation factors, (3) nodal displacements, and (4) y element stresses. The solutions were determined by

( application of Uniform Support Motion Response Spectrum Method of seismic analysis, based on interspatial combination (SRSS) and then intermodal combination (GROUPING) described in Regulatory Guide 1.92, Rev. 1, February 1976. For Problem Nos. 2, 4, 6 and 7, alternate solutions based on performing intermodal first and then

) followed by interspatial combinations are also provided in

) NUREG/CR-1677, Volume I. For verification of GAPPIPE, only

}

f 5

~

l

[

solutions' based onJperforming the interspatial combination

'first,and:then intermodal combinations are used'for-comparison.

. . For the. ISM Benchmark problems, the.four. problems F include.a simple-two anchor problem,.a simple three branch problem, and two large problems; simulating piping from actual nuclear power plants. The dynamic loadings applied to the four' problems are represented by distinct sets of-support excitation spectra' assumed to be induced by non-uniform excitation in the three spatial directions.

The solutions provided' include: (1) predicted natural frequencies, (2) modal participation factors, '(3) nodal

' displacements, and-(4) element stresses. For each problem, three sets.of solutions from different. combinations are

presented;;the'different combinations are

(1) enveloped spectra excitation, (2) independent support: excitation with SRSS combination between support group contributions, and (3) independent support 1 excitation with ABSOLUTE: H combination between support group _ contributions. .In all solutions,sthe combination over~ group contributions was.. ,

performed first, followed by SRSS interspatial. combination, t . followed by SRSS'intermodal combination without the consideration of closely spaced frequencies-(which is consistent with present NRC guidelines). For purposes of GAPPIPE verification, the solutions from independent .

support excitation with ABSOLUTE combinations are used in the comparison between GAPPIPE and the-NRC Benchmark solutions.

- 3.1.1 Procedure Used for Verification q l The procedure used for verification of the linear i portion of GAPPIPE program is as follows: (1),model all eleven Benchmark Problems by using-GAPPIPE with'all parameters identical in NUREG/CR-1677, Vol. I and II, (2) a fictitious gap with a very large gap size is added to each-problem, with the intent of verifying the program subroutines involving gapped supports in the GAPPIPE program. Since the large gap does not close upon loading, it will not affect the results of the original problem, (3) run all eleven problems and tabulate the results, (4) compare the results from GAPPIPE to the results in NUREG/CR-1677, Vol. I and II.

)' All eleven Benchmark Problems are modelled and run by using GAPPIPE, and their results are tabulated and compared to NUREG/CR-1677 results. Input listings of the seven UNI E Benchmark problems are given in Appendix A; input listings of the four ISM Benchmark Problems are given in Appendix B.

) 8

l.  ;

)

= =_=--__-- _ _ _ .

1

{'

R-

}' . .

The results of the comparison between GAPPIPE and

) NUREG/CR-1677, Vol. I, are' summarized in Appendix C. The R results of the comparison between GAPPIPE and NUREG/CR-1677, Vol. II are summarized in Appendix D for ABSOLUTE support group combinations.. All comparisons

) include natural frequencies, modal participation factors, L nodal displacements, and element' stresses.

)

e For the UNI Benchmark problems, the results of all

sevenLproblems except Benchmark Problem No. 5 are'obtained by performing the interspatial combination.first in accordance with the SRSS method followed by intermodal combinations using.the GROUPING method;.for. Problem No. 5, the SRSS method is used for both interspatial and' intermodal' combinations, which are the' identical procedures used in NUREG/CR-1677, Vol. I.

l 3.1.2 Uniform Suncort Motion Benchmark Problems UNI Benchmark Problem No. 1 The model is a simple, three-dimensional piping bend made up of straight and bent pipe elements between two

> fixed anchors (Figure 3.1).

UNI Benchmark Problem No. 2 The model is a multi-branched configuration resembling a four legged platform consisting of all straight pipe elements (Figure'3.2). The problem has symmetric and

)

antisymmetric modes which allow for quick check on the symmetry of the deformation of the model.

l UNI Benchmark Problem No.'3 This problem is primarily an extended version of the first Benchmark Problem No. 1 (Figure 3.3), with several anchors and a branch connection. It also includes intermediate spring supports, used to simulate hangers and snubbers, and a flexible anchor.

For this Benchmark Problem, the results presented in NUREG/CR-1677, Vol. I were determined to be in error from page 84 to page 111.(Ref. 10]. The correct results of natural frequencies and modal participation factors have subsequently been prepared by the authors of NUREG/CR-1677. '

) and presented as Problem No. 2 in NUREG/CR-1677, Vol. II.

The corrected results of this problem are used in the  ;

comparisons.

UNI Benchmark Problem No. 4 9

l I

lt; L

h p :This model: simulates the primary system of a r hypothetical.two loop reactor plant'(Figure 9.4). It consists'of an. elastically supported;. reactor-vessel, two-steam generators, and four primary. pumps: connected by three

[ <:and'four. foot diameter. piping. The reactor, steam

generators, Land. pumps were modelled with massless pipe t elements dimensioned to simulate'the. stiffness cf these

, components. This model is very significant because it i

incorporates most of the features found in:true piping h systems in a realistic configuration.

UNI Benchmark Problem No. 5 L

This model is an in-line system between two fixed anchors (Figure 3.5). This problem,-which was taken from actual nuclear power plant piping systems, has two unique-features:- one feature is a. transition between'two materials,.and the other feature is the inclusion of valves

) which.were.modelled with thick walled, stiffened piping L elements-by increasing the modulus of elasticity of valve elements _by a factor of three. The' method of modelling valves is similar to present industry practice.

L-

. UNI Benchmark Problem No. 6 The'model is primarily one large sweeping bend between two fixed points (Figure 3.6). This problem was also-derived from an actual' piping system which has a unique and continuous curve geometry.

UNI Benchmark Problem No. 7

} The model is a multi-branched structure which contains four anchor points (Figure 3.7). This problem, also derived from an actual piping system, is the largest Benchmark Problem, and.thus permits checking.of most I

analysis features including multiple branches, multiple anchors, intermediate' supports and hangers, valves and

! multiple excitation.

l 3.1.3 Independent succort Motion Benchmark Problems i s-i'i- ISM Benchmark Problem No. 1 The first ISM Benchmark Problem simulates a 3-1/2 inch l diameter water line running between two elevations. It represents a simple configuration joining the anchors and has numerous intermediate supports (Figure 3.8). The i excitation consists of two individual single direction spectra corresponding to the two elevations.

10 j

p , <

$1 ISM Benchmark-Problem No.-2 a- ~

The second ISM Benchmark Problem-is a three branch configuration. originally used'as a Benchmark for the

. Uniform. Support Motion. analysis method (UNI Benchmark p Problem No. 3). The support elements are divided into excitation spectra sets (Figure 3.9). The four excitation.

1; spectra correspond ~to actual spectra developed for a real 1J reactor structure and show variations with elevation.

t .

'),

ISM Benchmark Problem No.-3 q The third' ISM Benchmark Problem is a two anchor configuration simulating safety injection piping in a nuclear power plant. .It is comprised of 12-inch ~ diameter

- Schedule 40 stainless steel pipe between two elevations L: (Figure 3.10). The input excitation consists of four spectra sets; the vertical components of excitation: varying from. set to set while the. horizontal components of l

[ excitation are identical for all sets.

ISM Benchmark' Problem No. 4

>' The fourth ISM Benchmark Problem.is a three branch,.

three anchor piping subsystem from an actual nuclear power plant. .It contains numerous section. changes and complex geometry associated with real systems (Figure 3.11) . ' The input excitation consists of four distinct excitation-spectra sets developed for the actual system and show variationsffor elevations. This problem represents 4a I

benchmark.having the size and diversity-to-fully exercise proposed analysis methods.

3.1.4 Summarv of Comoarison The comparisons between GAPPIPE and NUREG/CR-1677, g Vol. I~and II, provide the following conclusions:

(a) Natural Frequencies The results between GAPPIPE and NUREG/CR-1677 are identical for all eleven problems.

(b) Modal Participation Factors For all major modes in all eleven problems, the L results from GAPPIPE are nearly identical to NUREG/CR-1677; for minor modes, there are some larger differences, but the differences are due to the use of r different computer hardware. The original 11 1

L

K s

L NUREG/CR-1677 problems were run on a CDC-7600 machine i B which is'a 64-bit machine, whereas the GAPPIPE problems were run on a VAX-11/750, which is a 32-bit machine'._'This. produces. differences when dealing with small numbers as in the 4: cases of minor modal-L participation factors. Also, the round-off error has contributed'somewhat.to the percentage differences.

Overall, the differences between GAPPIPE'and NUREG/CR-1677 are considered negligible.

(c)- Nodal Displacements For all eleven problems, comparisons between GAPPIPE-and NUREG/CR-1677 showed very good agreement.

(d) Element' Stresses

)-

Based on the comparisons of-element stresses for the eleven Benchmark Problems, the. differences between

'GAPPIPE and-NUREG/CR-1677 are negligible.

It is concluded that the GAPPIPE program can predict

-and calculate accurate results as compared to'NUREG/CR-1677

> for linear piping system under both (1) Uniform Support Motion excitation,-and (2) Independent Support Motion.

excitation.-

3.2 CORRELATION-WITH SHAKE TABLE TEST DATA AND ANSYS'

~

Seismic testing was performed to provide test data.in I

the' development of computer program GAPPIPE. The' tests were; performed using full' scale pipe specimens on a' shake table. located at the University-of California Earthquake l

i Engineering Research Center. Two' pipe geometry configurations /were' tested, each involving a variety of support, gap size, and input amplitude parameter c combipitions. Both configurations used portions of full size J-inch Schedule 80 pipe with simulated. gapped supports. One configuration used a straight pipe span excited only in the transverse direction. The second configuration used a three dimensional Hovgaard Bend which produced multi-axis response with ir.put excitation in only one direction. The two test configurations, as installed on the shake table, are illustrated in' Figures 3.12 and 3.13. The geometry of the two test configurations are shown individually in Figures 3.14 and 3.15 respectively.

The test configurations described above were instrumented and monitored so that all the pertinent h parameters of the tests were recorded. The instrumentation includes: (1) table motions including displacements, ,

l 12 l

Y

)

L velocities, and accelerations, which were measured by the F ' internal instrumentation of the shake table system, (2) l support accelerations and loads for both rigid supports and gapped supports (the accelerations were measured by mounting accelerometers at appropriate locations on L supports, and support loads.were measured by installing L load cells at support connections and by strain gages i mounted on supports), (3) piping lateral accelerations at

, various points, particularly at gapped support connections,

' were measured by accelerometers mounted on the pipe, (5) displacements of the pipe were measured.by potentiometers connected between the pipe and rigid supports, (6) pipe strains at various points along both systems were measured by strain gages mounted on both inside and outsida surfaces of the pipe.

3 3.2.1 Correlation Procedure For each test configuration, numerous sei<smic' tests t

were performed by varying gap sizes and input excitation I amplitudes. The recorded test data were then compared to analytical solutions determined using' computer codes GAPPIPE and ANSYS. The GAPPIPE analyses were performed to e evaluate the' accuracy of the equivalent linearization method for predicting nonlinear responses. ANSYS was used to perform corresponding nonlinear time history analyses as reference basis for accuracy. The method of nonlinear time- j history analysis, as employed within ANSYS, is an accepted i analytical technique for solving nonlinear dynamic j problems. 1 1

1 Two simulated building filtered El Centro earthquake

( motions were used in these tests as input excitations to 1

/ the shake table. The two earthquake motions correspond to )

0.82g and 1.33g ZPA excitation levels. The recorded shake table motions were used as time history inputs for conducting the ANSYS analyses. The same inputs were also used to generate the response spectra employed in the corresponding GAPPIPE analyses. The time history data and i response spectra for two earthquake excitation levels are shown in Figures 3.16 and 3.17, respectively.

3.2.2 Correlation of Sincie Scan Test Configuration I Figures 3.18 and 3.19 show the comparison of the pipe bending stresses for the single span test configuration.

) The measured and calculated stress values are plotted versus the average gap sizes. The comparison shows that I the equivalent linearization method employed by GAPPIPE is I

>_ as accurate as the ANSYS nonlinear time history analysis in I predicting the nonlinear piping response due to gapped pipe  !

l l

13 l 1

I' \

+

"~

u

$ a-supports.. This agreement between GAPPIPEland the time U ' history solutions is expected because the single span Ldynamic responses are first mode' dominant.- For complex piping. systems'with multi-mode participation, ituLs expected that GAPPIPE will calculate more conservative solutions asiwill all linear response spectrum analysis computer. programs in ocneral..

V Similar'results sre also found when comparing the gap 1 1

' impact force results calculated by GAPPIPE and ANSYS. The '

comparisons are illustrated by Figures 3.20 and 3.21 for.

the_two~ earthquake excitation levels. It is noted!that both. analytical solutions are conservative in calculating .'i T

the gap impact: forces as compared to the actual measured

. responses.

3.2.3 Correlation of the 3-D Hovaaard Bend Tests

_The~ shake table' earthquake' inputs for the 3-D Hovgaard

[L Bend. Test configuration were identical to those used for the single span dynamic tests. The same earthquake input levels ofl0.82 g and 1.33g ZPA were used..

l TheJ3-D Hovgaard Bend Test piping l system was observed -

to. exhibit significant response coupling as expected. .The first mode resonant frequency was found to be approximately 4.9 Hz:in the horizontal direction orthogonal to<the-directioniof table notion.

A number of tests were performed with various l .

combinations of gap sizes-and earthquake inputilevels.

Analyticalfresults were calculated using ANSYS and 1

'GAPPIPE. The: ANSYS analysis employed was. nonlinear time history. analysis and used'the recorded acceleration data at the; anchor points as-input motions. The same acceleration data were used to generate response spectra which were'then

h. utilized as input for the corresponding GAPPIPE analysis.

Figure 3.22 shows the analysis model of the 3-D.Hovgaard Bend Test configuration used in.both analysis types.

The maximum pipe bending stresses of the 3-D Hovgaard Bend test configuration are' summarized in Table 3.1. The

> first two' columns in the' table state the. gap conditions L used at each of the two gapped supports shown in Figuref .

3.15. Each gap condition is described by two values that are corresponding the gap sizes on the two sides of the i pipe. The value, "open", means a sufficiently large gap f.- was used so that;no impact occurred on that side of the gap.

V

! The last three columns in Table 3.1 are_the maximum 14 r

0 pipe bending stresses corresponding to the recorded test g data, the ANSYS and GAPPIPE analysis results respectively.

In all tests, the comparison shows that both ANSYS and GAPPIPE results are conservative with respect to the actual responses. The GAPPIPE results are more conservative than L ~

the ANSYS results. This is expected since the 3-D Hovgaard Bend test configuration was observed to have multi-mode o response. The GAPPIPE analysis results were determined by the equivalent linearization analysis option using the p response spectrum method.

3.3 CORRELATION WITH HDR EXPERIMENTAL TESTS 3.3.1 Backaround A major structural dynamic test program, known as the

}' SHAG experiments, was conducted at the HDR decommissioned experimental reactor facility of Kernforschungszentrum Karlsruhe (KfK), Federal Republic of Germany (FRG) during

! 1986. These tests were cosponsored by the West German government, the U.S. Nuclear Regulatory Commission Office of Research (NRC/RES) and the Electric Power Research Institute (EPRI). The overall objective of these tests was to generate data on structural response, soil / structure interaction, and piping and equipment response for a full scale reactor under strong excitation conditions. A detailed description of the SHAG test program was presented by Kot, et al., [Ref. 7] at the 15th Water Reactor Safety a Research Information Meeting.

The principal objectives of the piping tests in the SHAG program were to provide full scale in-situ test data and to demonstrate the feasibility of alternate piping  !

I support designs to be used in place of snubbers. In addition, the test data would serve to qualify the methodologies needed for acceptance of the alternate piping ,

support designs for implementation into power plants. One {

such alternate pipe support is gapped supports, known as  ;

the HDR SHAG Seismic Stop design. Figure 3.23 shows the  !

HDR piping system and the support designs. In this i section, the HDR SHAG test data are used to correlate with j the analysis results obtained by GAPPIPE.

1) 3.3.2 SHAG Test Description The SHAG test program was designed such that the building dynamic excitation was provided by a large 1 mechanical coast-down shaker on the operating floor of the  !

HDR reactor containment building. The shaker was  !

configured with two opposing concentrated weights and spun l in the balanced condition to the desired circular i frequency. Once the desired speed was obtained, one of the j

) 15

)

l rotating arms was released, allowing it to pivot and couple

) with the other arm. This configuration created an

) unbalanced force as a function of the magnitude of the concentrated weights and the initial rotational frequency 1 at release. After release from the initial balanced ,

condition, the shaker slowly coasted down with the

! frequency of rotation and the amplitu<l.e of the unbalanced

) force excitation decaying with time. The shaker transmitted the eccentric loading to the building

? structure, thus exciting the piping and components in a

" building filtered" manner similar to the dynamic loading of a seismic event. j Several types of piping system response data were recorded. These data included accelerations at support bases, pipe, equipment and a moter operated valve. Strains l were recorded at selected pipe locations, at the motor operated valve, and at components or discontinuities (reducers, tees, and nozzles) in the system. Strains

) converted to reaction forces were available for the rigid supports, snubbers, seismic Stop supports, and spring hangers. Support impact forces at the Seismic Stop support locations were also recorded. System data such as temperature, pressure, mass flow rate, and the valve position were measured.

3.3.3 Correlation Analysis Post test analysis using GAPPIPE were performed for both the snubber and the seismic stop supported test 1 configurations. Recorded accelerometer data at the HDR reactor building, the pipe support anchors, the HDU pressurizer, and the DF-16 accumulator were used to

)

generate response spectra as input to the analysis. These spectra were calculated using the Code Case N411 damping values and were enveloped for each of the following three structures groups:

Group 1: Reactor building accelerations at the base of each support.

Group 2: Equipment accelerations at the nozzles of the DF-16 accumulator.

)

Group 3: Equipment acceleration at the nozzles of the

'HDU pressurizer. l

}

The GAPPIPE analyses were performed with the Independent Support Motion analysis option using the above ,

three structural groups as three independent groups. In i the case of the Seismic Stop supported tests, the equivalent linearization method is also used to model the 16 l

f 4

gap' responses.

L The Regulatory Guide 1.92 groupino modal combination

. method is used in all analyses. :The spatial. directions were combined by the SRSS method. The ISM groups were combined by absolute summation.

l 3.3.4 correlation Summary Figure 3.24 shows a. summary' comparison of'the maximum '

pipe. accelerations for the snubber and1the' seismic stop test configurations. . For each configuration, a comparison was made between the test resultsLand the corresponding GAPPIPE analysis results. Five piping locations were compared.

The comparison.in. Figure 3.24 shows the GAPPIPE analysis results are higher than the actual responses in all casec.- This finding is consistent.with the analytical assumption that the response spectrum' solutions provide-conservative designs. As expected, the degree of; conservatism, measured by the. relative amplitudes of the:

test and analysis results in Figure 3.24,-varies-from pipe location to location.

'An'important characteristics demonstrated by the-results in Figu's 3.24 is the. similarity.of responses.for

~

the two pipe sukysrt configurations. >It is'noted that'the GAPPIPE aralysis using the equivalent linearization method-L for the. Seismic Stop support configuration retain the same degree of conservatism as the analysis for the snubber configuration. This correlation supports and confirms the use of the equivalent linearization' method employed by_

GAPPIPE.

Similar results'were also determined for other piping response parameters. Figure 3.25 shows a summary comparison of the maximum pipe stresses _at five pipe locations. Figure 3.26 is a summary. comparison of the maximum support loads. In each case, it was found'that the' GAPPIPE analysis results for Seismic Stop provide similar degrees of conservatism as the snubber analysis results.-

-3. 4 COMPARISON WITH' LITERATURE ANALYTICAL RESULTS The purpose of this comparison effort is to validate the time history analysis capability of GAPPIPE. Two sets f- of literature solutions are chosen to compare-to GAPPIPE

,- results. The first case is used to verify the linear time

) history analysis logic of GAPPIPE. The second set is used

! to test the nonlinear time history analysis method in GAPPIPE for the analysis of gapped supports.

17 f

L

L 3.4.1 GAPPIPE Linear Time History Analysis 3

This verification case is taken from the STARDYNE Verification Manual Example 30 [Ref. 8). It is a cantilever beam subjected to a sine pulse forcing function applied at the tip as shown in Figure 3.27. Assume the case where the sine pulse has a period of T = 0.14352 L second and the cantilever beam has the following properties:

l E = ' 1.0 x 10 6 psi I = 1.3333 x 10~4 in4 A = 0.04 in 2 l p =

0.1 lb/in 2 1 = 30 inches 1

The cantilever beam is modeled by six straight pipe elements in GAPPIPE. Using a time step of 0.004784 second,

, the vertical displacement response at the tip of the

) cantilever beam is determined. Table 3.2 summarizes the GAPPIPE results at four time intervals as compared with the the theoretical solution. It is dotermined that the l GAPPIPE results differ from the theoretical values by less than 0.5%.

p 3.4.2 GAPPIPE Nonlinear Time History Analysis This verification case is taken from the technical paper by Molnar, O al. [Ref. 9).. Molnar presented the

> methodology and CMuple results for the dynamic analysis of piping systems vita gaps. The Molnar method has been used in the design and analysis of Westinghouse PWR piping systems.

Figure 3.28 shows the piping model presented by Molnar. It consists of nine straight pipe elements, two elbows, and three gapped supports. The piping system is fixed at the ends. The gap sizes and stiffnesses at the i

three gapped supports are:

[

Gao No. Gao Size Gap Stiffness (lb/in) 1 0.250 2.0 x 10 6 2 0.125 3.0 x 10 6 3 0.062 1.5 x 10 6 i

h 18

I I The forcing functions applied to the piping system are also shown in Figure 3.28. The GAPPIPE analysis is 1

I performed using 30 modes and a time step of 0.0000625 seconds. The maximum impact forces at the three gapped supports and the time of occurrence are computed by GAPPIPE and compared with the Molnar results in Table 3.3. The I differences in the comparison are found to be less than 4%.

1 I

I I

I I

I I '

l l ,

I E

I  !

I 1e I ,

I 4.O

SUMMARY

AND CONCLUSIONS I

I The analysis solutions of computer program GAPPIPE have been compared with the NRC Benchmark Piping Problems, Shake Table test data, ANSYS analysis results, the HDR

'E Experimental Tests, and analytical results published in the 5 literature. The summary of results presented in the i preceeding section shows:

o GAPPIPE linear solutions are nearly identical to the NRC Benchmark Solutions in NUREG/CR-1677.

o GAPPIPE nonlinear solutions are comparable to I. ANSYS results and in many cases more accurate l when compared to test data.

o GAPPIPE nonlinear solutions provide the same degree of conservatism for piping analysis of gapped supports as in. current industry practice of piping analysis of snubber supports.

o GAPPIPE time history analysis solutions are nearly identical to literature results.

These' comparisons have demonstrated the accuracy,

< applicability and validity of GAPPIPE in accordance with I' Section 3.9.1 of NUREG-0800. It is concluded GAPPIPE can be applied for the analysis of nuclear piping systems.

E E

Ll I

I I

I g 2e I

l

5.0 REFERENCES

1. Iwan, W.D., " Application of Non-Linear Analysis Techniques," Applied Mechanics in Earthquake Engineering, AMD-Vol. 3, ASME, NY, pp. 135-162 (1974).

2. " SAP-IV, A Structural Analysis Program for Static and Dynamic Response of Linear Systems,"

Earthquake Engineering Research Center Report No.

EERC 73-11, University of California, Berkeley, June 1973, Revised April 1974.

3. " Piping Benchmark Problems, Dynamic Analysis Uniform Support Motion Response Spectrum Method",

NUREG/CR-1677, Vol. I, Brookhaven National Laboratory, August 1980.

4. " Piping Benchmark Problems, Dynamic Analysis Independent Support Motion Response Spectrum Method", NUREG/CR-1677, Vol. II, Brookhaven National Laboratory, August 1985.

5. Cloud, R.L., P.H. Anderson, and J. Leung,

" Seismic Stops vs. Snubbers, A Reliable i

Alternative", Nuclear Engineering and Design, Vol. 107, North-Holland, Amsterdam, pp. 205-213, l'

1988.

. 6. ANSYS Structural Analysis System, User Manual,

! Swanson Analysis Systems, Houston, Pennsylvania.

l 7 Kot, C.A., L. Malcher, H. Steinhilber, i " Vibrational Experiments at the HDR: SHAG Results and Planning for SHAM", presented at the 15th Water Reactor Safety Research Information Meeting, Gaithersburg, MD, October 26-29, 1987.

8. STARDYNE Verification Manual, System Development Corporation, Santa Monica, California.

9. Molnar, A.J., K.M. Vashi, and C.W. Gay,

" Application of Normal Mode Theory and Pseudoforce Methods to Solve Problems with Nonlinearities," Journal of Pressure Vessel Technology, ASME, May 1976, pp. 151-156.

10. Record of November 16, 1989 telecon between Dr.

Paul Bezler of Brookhaven National Laboratory and Kenneth Wu of Robert L. Cloud & Associates, Inc.

21

jQ hi

[.

)-

GAPPIPE .

-1 k..

L 4 MODEL GENERATION {

! INPUT DATA 1 h (NODAL & ELEMENT DATA) i

.; o-ANALYSIS EXECUTION l t- 'NDYN = 0, 2. 3, 5. -5  !

(. 6.-6 NDYN =

2. 3. 5, -i 6. -6 NDYN=0 p NDYN = 11 ir STATIC EIGENVALUE  : DYNAMIC

=

ANALYSIS SOLUTION ANALYSIS i

.t s,-s 2 RESPONSE SPECTRUM '

I ANALYS$

I . EQUIVALENT UNEARIZATON -

ANALYS6 INDEPENDENT MOTON . . _ _ ,

ANALYSIS

NONUNEAR WODAL _

• THE HSTORY ANALYSIS Ir OUTPUT FILES '

PLOT ANALYSIS RESTART TEMPORARY POSTPROCESSING FILE PRINT FILE FILES FILES FILE I I PRINTED DISCARD p AFTER EXECUTION AFTER EXECUTION f

GAPPLOT GAPPOST

, t t NEW POST- POSTPROCESSING PROCESSOR FILE PRINT FILE NDYN = 4 -6 I~

Figure 2-1 Flow Chart of GAPPIPE Program 22

i ny 1

[-

I

)I l

l Z x

}

3 .

) q f

I j

. g,s l

3 e

@ 7

@e 9@

'l

>g to 9 il n

G norE.

w.g i CIRCLED NUMBERS ARE-ELEMENT NUMBERS.*

l i

I 1

l l

l UNI-BENCHMARK PROBLEM No.1 I

l l

)

1 i

Figure 3-1 23 f

6 dY L'

I l

L s

Z x .

I

)

i .

S 7 4 8 '

I 4 6 9 '

S' g 9 11 10 l 18 86 G l' &

@ i

'4 , ;2 l

0 3 q

%R 1 2g

• & 1 gg wor => .

l' l CIRCLED NUMBERS ARE ELC.E:llT NUMBERS.

l i

UNI-BENCHMARK PROBLEM NO.2 l l

1 l

l i Figure 3-2 24 1

uY Z \X 4

4 4 , s .

d5

, s 4 . :.

~, ,

d, i W '

y T '

S g

• 6

$$Y

< < w ,,

i 9 g to a.

-h .

St.

19 0 3 4

• m a e .

31 M so i

i l

l UNI-BENCHMARK PROBLEM NO.3 Figure 3-3 25

r L

e g t%,7 g4 0

l

.R .h L 33- _. _.

y

(( <o *

!E

[ . akM e %...?

ef T]y i E .9 . 9 A \ /

j

,. E i 41g -

b

. .m

.w .

9 ,

l],

5 s= # .

A g *

, ~_ a

r. .

3 g

&R a

~

Z UNI-BENCHMARK PROBLEM NO.4 Figure 3-4

' 26 1

. i

j' '

j

.'. x e ::

1 f a

).  !

A A ,

7 f Th l I @

N 3 1

I ,

T- 1 '.

l.

• a

.l l i c.

g ,

~!

l E .

, g -

i e

l. g 1! -Q

< d g e. e 5 @ . 4 t

t f

1

. 1 4

e UNI-BENCHMARK PROBLEM NO.5 l

t Figure 3-5 l 27 ,

I

i'

-I dY 7 .

)I i

I _'

2 X

\

I .

)

. c.

59 31 .-

4 0,44 34, 4 33 o $7

'}

St i y

g, ' %

, es 28 2

.I i

ft,6 84 gg gy 80 , 27 6

78 &

gg  : - ?

A9 g5 ,g, i

7 4 y .s NOTE:

CIRCLED NUMBERS m N- ___ .

i UNI-BENCHMARX PROBLEM NO.6 I

Pigure 3-6 28 I

-I 1

[.

en Q

• r 1

e

• N

-=

2

~d

• I k'

- % e 4 i -

3 k*

K s

~

g..S*9 . .

8 i'

h k  !

'! Ik

~q

,o AD 4

.,L /

4 a

1 1

AD t

1 l

1 UNI-BENCHMARK PROBLEM NO.7 i

j 1

l 1

Figure 3-7 29

)

i l

2 eb% J' i

.~

S l g< 'g o

@o lo

$< ,gg

@, ,a . :. 1 13 TYPICAL. INTERMEDIATE .l SUPPORT .

'15 -

'W ,gs X ol7

!!8 p

i

@2z- l; 23 j

. q gg a

l

@ 24 6'

i27 ", g 30 g O ,

O li ISM-BENCHMARK PROBLEM NO.1 Figure 3-8 30 l i

o i

i ylL

! I I

I E

e 3 z/Nx M g 4 .

GL , .

O 4

,,e , .

. c ,

. .. sg ,

=

tb

& 3* M

$ , e6 t M g . R **

. n l

i I

ISM-BENCHMARK PROBLEM NO.2 Figure 3-9 31

I I

^

CYe

8

.g, ,

as m, j bI 8

I E O T l '

W  % g, ."

i

/lxy g 8 .

w a E 3 .

.. ... g E

l 5 s &

I ,

- I# 4 ,

)

V w I

I ISM-BENCHMARK PROBLEM NO.3 I

I Figure 3-10

)

\?..

x. I 16' p6 g

usa f* y n =%

i W)"m 81 ss gj s 146  ; rig ,4 fM

/. ..

2 19 83 s)'

,47 l

l@

't r F

<c9 40 10 0 03 10 5 l

6 l

t ISM-BENCHMARK PROBLEM NO.4

)

Figure 3-11 l

33

) i

.I '

,g Va -. :

i .

I 3

z .,

~ :_:.

Figure 3.12 ~

Single Span Test Configuration l

t i

g I

E9LE.k b ti I ~ " d 2 -- . . . ,

J - -

. s 7

/ . . ..

7 .

I L

Figure 3.13 3-D Hovegard Bend Test Configuration l

i

(1 E

)' .

i l

i lf; 1

1 I

'l

]

I

-1 72" 72" 72" Wt.1 lu. Wt 2 h

h F ]

: t  ::

108.5" 107.5" 15" GAP SUPPORT 15" i LOCATION 1

l SINGLE SPAN DYNAMIC TESTS J l

PIPE DIAMETER PIPiSCHEDULE SPEC EN 86 I 1 3" l

i-

)

• NOTE: )

Supplemental 30 lb. lead weights designated as Wt.1 and Wt.2 were included in test specimen 1 (3"SCH 80) only.  !

0 l

> l l

7 1

Figure 3.14 Single Span Dynamic Test Configuration i

l s

l

)

35

)

.lj q

.q Y  !

I ~

g) SUPPORT S3 8

69.73" 4

61.75" v

\

% 84.25" GAP SUPPORT 67.875" NO.2 .

1 Z

3*3I~'

GROUND MOTION DIRECTION GAP NO.1 SUPPORT g /

6. 5" )

.X l

SUPPORT S2

t. 106.25" 136.25" p SUPPORT.

51 35.0" n /ge r

11.25" Figure 3.15 3-D Hovegard-Bend Test Configuration 1

36 i'

4 '

}

941E of 1($18 717.1885 tist no. 7, y laid (asss(L at.: 3.

. 118( lattte(RIs .003ee Pfut * .31956 4 er IllE

• 8.9120 l s MIN * .72197 '

- ni n,e e.em

,8

.:- lun = . m d.

g*

i

\

m b

za d 1

- j; "o

?

4 s.se (.se d.se ib.ee l's.ee de.es A.so 28.ee I T1HE (SECJ l 2500 -

2250-2000 - ....0.251 Damping 1750 -

"o /

f 1500- [

) $

g 1250-U .

I h 1000- *1.0%

g Damping U

• 750-I 500-250-0 . . . . . . . . .

l- 0 4 8 12 16 20 24 28 32 36 40 j Frequency (Hz) 1 i

Figure 3.16 Time History and Response Spectra Plots l 7 (0.82 G ZPA) (0.82 G ZPA)

)

l 37

)

[f-

) 881f er 1($18 717.1985 Test C4.8 8.

G Salt CNAIR(L 88.8 3.

. litt laCAtatatt .ee3ee NW

• 8.31718 I AT fine - 4.e540 g MIN * *.S$955

. .: n c i..u5.

), j. INIT R .see73

. N'.

b l

)

(-

( - iA

.z."

)

E.

0 .

)

8.00 4'.88 d.88 lb.00 l'6.00 d8.80 d4.00 28.80 TIME (SEC) 4000 3600 -

\ .

3200 -

^ 2800- .... 0.2 W Damping i

$ . .. . . -l . 0*. Damp ing

$u 2000-O

{u 1600-E 1200-i 800-

- l 400-0 4 8 12 l6 20 25 22 32 $6 40 y

' Frequency (Hz)

Figure 3.17 Time History and Response Spectra Plots l l (1.33 G ZPA) (1.33 G ZPA) l

)

38 )

l

)

_____ J

4 2 -

4 E 8 P 4 S I 1 t Y P -

s S P e N A i

T A G .

- - 6

- - - 1 1

O + o .

I i 4 1

=

)

_ i n

i

~(

i 2 p 1

a G

i e

g a

r I 1 e v

A .

4

- t 8

0

_ i 6

i 0 i

- 4 i 0 I

+ 2

- - - - - - - - - - - - - - - _ 0 1

0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 2 2 1 1 1 1 1 1 1 1 1 1 e

t

_.j _ r, r, E a m t 3. z @ *

- uO

{

. 2 i

i 8.

1 i

i 6 1

_ i E , 4 P 1 S I t Y P )

s S P n e N A ,

i T A G -

(

- - - , 2 p

- - - 1 a

O + 0 G

e g

a r

, 1 e v

A 8

, 0 i

i 6 0

i 4

, 1

, 0 i

- 2

- _ - - - - - - - - - O 8 6 4 2 0 8 6 4 2 0 6 6 4 2 2 2 2 2 1 1 1 1 1

_ c n7u mmoLy s3tgj

_ a l

,. e

. ^

E i 2 S P I .

Y. P t s

S. -

P e i

_ N A T A G -

i 8

~- - -

1

+ O

o. i i 16 i

i 4 s 1 e c'

rA i

oP FZ

)

1 n tG i 2 i

(

c' 1 a2 p p8 a m.

I0 G

~ - i e f=..

g o' i 1 a l r ne e ov v se

, A iL r-at 8 pu mp i 0 on CI i

0-6 2

- i 0 3 e

, i r

- u g

i 4 i F

-l - 0 i

2

- - - _ - - - 0 5 4 5 3 5 2 5 1 1

4 3 2

~

T %_~oEo' tcQE_.

O

411I ) ii 'i l1Jl1jl'i]1

. . l!ili119 s

i-- -

, 2 7 ,.

8 1

E P i

_ S I

t. YP C .

s S P . 6 eNA e 1

T A G

- - i

.O + o s

4 e

8. c 1 rA oP

) FZ

'. n .

i tG

( c a3

. 2

' p p3 1 a m.

G I1

' j e

g f5 o

a l r ne i 1 e ov v se A iL I.

r at pu mp I

8 on 0 CI 1

2 6

3 0 .

e _

r

• u g

i F

• 4

- < 0

- 2

- - - - - - 0

_ 8 7 6 r, 4 3 2 3

- e D . 3. + 0yO.

. . k # v c D C': .

• N lllilil1

-i 1

) d o IM E .<'a'M

>! 3:. so t

I " 27

. 3 I i- 5  !

, u n=

1. og

) a og

)- 3g Gap Support 2 15 Gap Support 1 i

13 , ,

12

'64 sc 10

)

9

) 8 7

t, s Z gg 'A 33 Maximum Stress location u

i is 8 8 80

)- x  : 7 3D HOVCRARD BEND TEST MODEL

)

Figure 3.22 Analysis Model of the 3-D Hovgaard Bend Test Configuration I

43

)

)

)

O L S L T P A R D P (

O L U S D;O R YP S E BPU S D EC S

- P BD S BE R E EG ZT I

RA TI N

O O UC D TAHNESEHD .UDOT T

S NA S PP E LP LOH TST DN I

C 8ES

- C I N X EA LN vRENO MG SRG AU sPTAC I

SIS l

- EE 4 e .

SD 1

H .

m.

/ 'a n

e

' E m

" n~ - a

, & s

. ~ ,,I -

a U

N' '

D

' - s fp .

H A M H

h e

1 W1 ,

p -n g n

e d

e t

d e iWy p

8 r

=

- l H -

dl I 7 H

miagt I eis tRn ll t

a =

s . I e ,

ny d Sy S. m ha w

.Ut e

" Y o

Y n Key s l - o

- h V htS -

ehr t htio T

Wp p u

S a^

t a n rr eu o bg ll bf i a

un r no u g

.. Sc f

_ n

/ s n

o i

nt oa ir t e al l e gE e e

ec rc

" T rA

• o Ce s "" ti p

t sP l

u^

s

• n e Tm e -

l a o u R c i tm t Ws i r

e M ni eR s c ex T H h

p o ma S

iM L r

/ ef t po n x En e o n Gs e o Ai l

v p Hr a Sa V n r

p Rm

- o Do C HC e

_ r p e i 4 c

u P 2 d 3 e e R r

- u g

- i F

e p

i P

- 2

, e a

- 5%bmOC E:E.

3

_ t a n rr p eu o.i o bg tt bfi Sa r

un cu no i g

- SC mi f

/ s in

! /eo SC B g i

s s

y l

s a t n u A l

s e E s , e R P t l

I l z

- t Pu z o n s Ps e Ae N i o

T GR t a

c

- l o a L i

c /

r eee tn

_ hTp e S n o

- p e m l o Uz Dzo C

- HN i

e p

P

- s r

/ ec

- u

- d e

R 7

/ / '6e

_ l 1

- zz Fo It 5

2 0

2 5

1 0

1 r

C e n 8b )L oU O W2 Q '-@ ? 5tW uo

bg o o i i t t bf Sa un r no SC c ug imi s f

- in eo SC E eum /

_ s 3 n i

s 1 os y H id l ta s a f

ao t n lL e

u A l

/ rt s E 1 rr e Pst oo

~

RI t Pu l d1H Cp p

t u s Ps sS e Ae e T CR n Tm 0 o u

- 1 H

i t

t m ni a ex c ma i iM f r i ef po t

9 n x M En e o d Gs I Ai

> Hr t Sa

/ r Rm p

S o Do H p HC p

_ u S 6 2

3 4

H e r

f u g

i

- F 3

H 8 5 2 9 O 1 1 1 9aE n e D ha. Yoa u"

l l

lS p(gy 3' s  !

n '. YCt) l g 7 4 \

i l fa) - .

01 SIN ([)t o<tgy L

0 'tk r I

t 0,f ......

_ g I'

E W

Figure 3.27 Description of STARDYNE Verification Problem 30 E

I i COMPARISON

SUMMARY

TIME, t, Y(t), TIP DISPLACEMENT (IN)

(SEC) GAPPIPE THEORETICAL 0.04784 -0.396 -0.395 0.09568 -1.150 5 0.17701 0.872

-1.151 0.868 I

0.24877 4.887 0.871 I Table 3.2 Comparison Summary of GAPPIPE and Theoretical Results 48

l' E 7 8' y-x  !

/~j-g' Y .. ,a ,

S . v'

,i l

.g "

A i s I

r

^

Y 3 l W HmMmm h )

3 I

I

• 45 SEC l k a w .- -- , i I

e 1 05 SEC gg .------ .--

5 00 3

Q2 45 SEC Applied Forcing Functions E- Figure 3.28 Geometry and Forcing Functions of the Molnar Verification Case b

Es Pt Il Pu 30 78 94 91 49 Ps .. . . .. .. ..

_ s Ae 28 82 39 82 92 s GR 1 11 1 11 11 e

. r t

S g

n s i t d) Sl ni Yu 00 42 52 91 23 es Ss .. . . .. .. ..

Bk Ne 27 71 96 27 59

-( AR 1 11 1 1 e 1 p

i s P t

s m ge nT u i . m .

~ dd i ta 01 85 19 61 25 nn x st .. . .. .. ..

ee a ea 96 38 95 17 38 BB M TD 1 1 1

_ ed pr ia Pa n g o 1

mv gi 3

uo nt mH ic i

e xD de l

a- nr zy zy zy zy zy b M3 ei

~ a BD T fe oh t

n or so

)

A if tlP r ueZ 2 3 2 3 3 pv 8 3 8 3 3 as pe neg . . . . .

ms IL( 0 1 0 1 1 os Ce r

t n n n n n S e e e e e

) p p p p p n o o o o o 2i / / / / /

n / n 5 4 4 5 o Pn e

_ i t

Ai G(

p o

7 0

4 0

4 0

7 0

i d

n

_ C o

n n 2 2 7 e e e 6 6 8 z ) p p . . .

i n o o 0 0 0 S

- 1i / / / / /

/ n n n n n p Pn e e e e e a Ai p p p p p G G( o o o o o g

' l l

0 r

f a

)

OOMPARISON

SUMMARY

GAP NO. MAXIMUM GAP FORCES l

, TIME OF MAXIMA VALUE OF MAXIMA GAPPIPE 1 0.2301 636.1 MOLNAR 1 0.2295 612.6 f

GAPPIPE 2 0.2839 606.1 r

MOLNAR 2 0.2836 606.1

GAPPIPE 3 0.2773 678.0 MOLNAR 3 0.27/0 664.5 I

)

Table 3.3 Comparison Summary of GAPPIPE and Molnar Results t

i 51

)

- _ --- _