ENGINEERING CHANGE NOTICE

................ ... .......614102 ... ........*

1. E

Supplementa 1 Direct Revision Change ECN Temporary

txl

Supersedure Cancel/Void

12. Description o f Change

Installation o f SASS1 at LANL, validation testing.

13a. J u s t i f i c a t i o n (mark one) As- Found [ x]

C r i t e r i a Change Feci l i t a t e Const.

[I
[]

Design Improvement Const. Error/Cinission

[I [I

Environmental Design E r ror/Omi ss i on

[I [I

13b. J u s t i f i c a t i o n D e t a i l s

Installation at LANL replaces RL Cray installation.

RL Cray was removed 12-93.

14. D i s t r i b u t i o n (include name, W I N , and

no. o f copies)

RELEASE STAMP

See distribution page

A-7900-013-2 (06/94) GEF095

ENGINEERING CHANGE NOTICE

15. Design Verification Required

1. ECN (use no. from pg. 1)

Page CONSTRUCTION

of

614102
17. Schedule Impact (days)
V

16. Cost Inpact

ENGINEERING
A d d i t i ona 1

Savi ngs Savings Delay No f $ 18. Change Impact Review: Indicate the r e l a t e d docunents (other than the engineering docunents i d e n t i f i e d on Side 1) that w i l l be affected by the change described i n Block 12. Enter the affected docunent nunber i n Block 19. SDD/DD Seismic/Stress Analysis Tank Calibration Manual

11 Yes
Ex1

[I [I

Additional

[I [I

Improvement

[I [I

Functional Design C r i t e r i a Operating Specification C r i t i c a l i t y Specification Conceptual Design Report

Equipment Spec.

S t ress/Desi gn Report

Interface Control Drawing Calibration Procedure I n s t a l l a t i o n Procedure Uaintenance Procedure Engineering Procedure Operating I n s t r u c t i o n Operating Procedure Operational Safety Requirement IEFD Drawing Cell Arrangement Drawing Essential Material Specification Fac. Proc. Samp. Schedule Inspection Plan
L J

Health Physics Procedure Spares M u l t i p l e Unit Listing Test Procedures/Specificatio

n
Component Index
ASME C o d e d I t e m

Const. Spec. Procurement Spec. Vendor Information

On Manual

Hunan Factor Consideration Computer Software Electric Circuit Schedu 1e I C R S Procedure Process Control Manual/Plan Process Flow Chart Purchase Requisition Tickler F i l e

FSAR/SAR

Safety Equipnent L i s t Radiation Work Permit Environmental Impact Statement Environmental Report Environmental Permit

Inventory Adjustment Request 19. Other Affected Docunents: (NOTE: Docunents l i s t e d below w i l l not be revised by t h i s ECN.) Signatures below indicate t h a t the signing organization has been n o t i f i e d of other affected documents l i s t e d below. Document Nunber/Revision D o c w n t Nunber/Revision Docunent Number Revision

A A
20. Approvals
Signature
3PERATIONS AND ENGINEERING

Date

Signature
ARCHITECT-ENGINEER

Date

Cog. Eng. Cog. Mgr.

3A

EO Weiner

TJ Conrads

ML Hermason

Safety Envi ron. 3ther

JC van Keuren Environ. Other

DEPARTMENT OF ENERGY

Signature o r a Control Number t h a t tracks the Approval Signature ADDITIONAL

A-7900-013-3 (06/94) GEFO96

DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

RELEASE AUTHORIZATION
Document Number: Document Title: Release Date: WHC-SD-GN-CSWD-30036,
REV. 1

SASS1 System Software C o n f i g u r a t i o n 9/30/94

*************
This document was reviewed following the procedures described in WHC-CM-3-4 and is: APPROVED

FOR PUBLIC RELEASE

*************
~~

WHC Information Release Administration Specialist:

-1
(Signature)

N.L. SOLIS

(Date)

A-6001-400 (07/94) UEF256

SUPPORTING DOCUMENT
2.
Title

I
3.
Wunber

1.

Total Pages

SASSI System Software Configuration

5.
Key Words

WHC-SD-GN-CSWD30036
6.
Author

3 fi 6

4. RevlNo.

software, configuration, SASSI, seismic

NW:EO
S i gnature

Weiner

Organization/Charge Code

8D440/D8023

T h i s document demonstrates compl iance of the SASSI soil-structure interaction computer program w it h WHC- CM-3 - 10
RELEASE STAMP

DISCLAIMER This report was prepared as an account o f work sponsored by an agency o f the United States Goverrment. Neither the U n i t e d States Goverrment nor any agency thereof, nor any o f t h e i r employees, nor any of t h e i r contractors, subcontractors o r t h e i r employees, makes any warranty, express or implied, or assunes any legal L i a b i l i t y o r r e s p o n s i b i l i t y f o r the accuracy, completeness, or any t h i r d p a r t y ' s use o r the r e s u l t s of such use of any information, apparatus, -product, or process disclosed, or represents that i t s use uould not i n f r i n g e p r i v a t e l y owned rights. Reference herein t o any s p e c i f i c comnercial product, process, or service by trade name, trademark, manufacturer, o r otheruise, does not necessarily c o n s t i t u t e o r imply i t s endorsement, recomnendation, or favoring by the United States Goverrment or any agency thereof or i t s contractors or subcontractors. The views and opinions of authors expressed herein do not necessarily s t a t e or r e f l e c t those of the United States Goverrment or any agency thereof.

OFFlCiAL RELEASE
BY WWC

f&-' W

DATE

SEP

9.

Impact Level

SQ

A-6400-073 (11/91) C E O UEF124

I.
(1)
Document Nunber

RECORD OF REVISION

WHC-SD-GN-CSWD-

Page )

30036

A-7320-005 (08/91) UEF168

WHC-SD-GN-CSWD-30036

Rev 1 Page i

SASS1 System Software Configuration August 1994

Prepared by: EO We1 ner Structural Assessments

SECC

Facility Stress Analysis

Approved by: Date Stkuct ural Assessments

Westinghouse Hanford Company Hanford Operations and Engineering Contractor for the Department o f Energy

WHC-SD-GN-CSWD-30036

Rev 1 P a g e ii

TABLE OF CONTENTS INTRODUCTION IMPACT LEVEL DOCUMENTATION SOFTWARE DEVELOPMENT PLAN VERIFICATION VALIDATION CONVERSION CONFIGURATION FPANAGEMENT REV IS IONS REFERENCES

5
7 8

w i.C C 5 D - Gd
c

- CS WB-3DO36
* * <L L

8 e V . I

TYPICAL CHECKLIST FOR INDEPENDENT REVIEW

Oocument Reviewed

5&SS.L
h , qI
L t U

L-4eh.7

~ ~ 4 h C l S Lw ~ P

6 4 4r .k o y

Author

1 4 I/i
[VI

EO &N/A 1 1 1 1 [ 1 1 1
[

Problem completely defined. Necessary assumptions explicitly stated and supported. Computer codes and data files documented, Data used in calculations explicitly stated in document.
-'

[vl [

I I

[ [ [

1 1 1

1.4 I 1

Data checked for consistency with orlginal source information as applicable. Mathematical derfvations checked including dimensional consistency of results.

I I t l t 4

Models appropriate and used within range of validity or use outside range of established validity justified. Hand calculations checked for errors. Code run streams correct and conslstent with analysis documentation. Code output consistent with input and with results reported in analysis documentation.

Id1

Acceptability limits on analytical results applicable and supported. Lfmlts checked agal nst sources. Safety margins consistent with good engineering practices. Conclusions consistent with analytical results and. applicable limits. Results and conclusions address all points required in the problem statement.
1 r

w H c - S P - GN- CS W P - 3 O O s
fevt I TYPICAL CHECKLIST FOR INDEPENDEW REVIEW

Document Rev1 eued

Y e s -No
[

Jf

1
]

1
]

Problem completely defined.

Necessary assumptions explicitly stated and supported, Computer codes and data ffles documented, Data used inAcalculations explicitly stated in document.
-.

[ J] [
[ [ [ [
[

[
[ [

/i

[
[

1
1

1
1

J&ri/Lwi&

1 1 1
[

Data checked for consistency with original source information as applicable. Mathematical derfvations checked Including dimensional consistenc of results.
[~lift&Ln % & ) )

1 [ 1

4 1

-d

1 I 1
[

Models approprqate and used within range o f valfdlty or use outsid range of established validity justifjed, Hand calculations checked for errors,

1 [ 1
[

A 1 .1
Jf
[

[
[

I
I
1

Code run streams correct and consistent with analysis documentation. Code output consistent with input and with results reported i n analysls documentatjon. <&fJf64 dk P A. f ) Acceptability limits on analytica1,results applicable and supported. L i m i t s checked against sources,

[ [

1 1

d[

[ 1 [ 1 I

--/ margins consistent wfth good engineering practices, Safety

[ [

1-4 [
I

/f[ 1

Conclusions Consistent wfth analytical results and applicable limits. Results and conclusions address all points required in the
problem statement.

dReviewer Q&.,

WHC-SD-GN-CSWD-30036 Rev 1 Page 1

SASSI System Software Configuration EO Wei ner August 1994

INTRODUCTION The SASSI (System for Analysis for Soil-Structure Interaction) computer program was obtained by WHC from the University of California at Berkeley for seismic structural analysis of complex embedded building configurations. SASSI was developed in the 1980's by a team of doctoral students under the direction of Prof. J. Lysmer. The program treats three-dimensional soil-structure interaction problems with the flexible volume substructuring method. In the 1 9 7 0 ' ~the same organization developed the FLUSH program, ~ which has achieved widespread international usage in the seismic analysis o f structures. SASSI consists of nine modules, each o f which are to be run as a separate execution. The SASSI source code, dated 1989 and identified as a Cray version, was put u p on the RL Cray XMP/232 Unicos system in 1991. That system was removed at the end of 1993, and SASSI is now installed on the LANL Cray YMP systems.

SHAKE is a much simpler program that solves the free-field soil problem by
iteration with strai n-dependent soi 1 properties. Current practice uses SHAKEderived soil properties as fixed input to SASSI. The SHAKE PC version was obtained from the Earthquake Engineering Research Center (EERC), University of California, Berkeley, in 1990. It was installed on an 80386 PC and the RL Cray, and it is included with the SASSI configuration. An updated version, SHAKE91, was obtained in April 1994 from the same source. Both SHAKE and SHAKE91 have been installed on a 486 PC, the SECC sgi3 workstation, and the LANL Cray YMP systems. A short auxiliary program HOUSEPLT was written to translate the SASSI. HOUSE module input to ANSYS version 5.0 (ANSYS 1993) PREP7 input suitable for mesh plotting. This enables the engineer to obtain a graphic representation of his model that i s not otherwise possible. Filters TRFILTER, HFILTER and SMFILTER were written to limit unneeded volumes of output from ANALYS, HOUSE, STRESS and MOTION. The purpose of this document i s to demonstrate compliance with WHC-CM-3-10 (WHC 1993).

WHC-SD-GN-CSWD-30036 IMPACT LEVEL Impact level 2 was previously assigned t o SASSI. appropriate 1eve1

Rev 1 Page 2

SQ i s considered t h e current

DOCUMENTATION SASSI: User manual (Lysmer 1991) Theoretical manuals (Lysmer 1981, Bechtel 1991) Validation manual (Bechtel 1990) SHAKE: User manual (Schnabel 1972) SHAKE91: User manual ( I d r i ss 1992) HOUSEPLT: Appendix A
SOFTWARE DEVELOPMENT PLAN

SASSI was developed a t t h e University of C a l i f o r n i a over the 1979-1989 timeframe, and i t i s off-the-shelf software. The t h e o r e t i c a l manuals provide development information. VERIFICATION Design v e r i f i c a t i o n i s considered as conducted by t h e developers during the development process. SASSI i s off-the-shelf software obtained a f t e r development

VALIDATION
A v a l i d a t i o n r e p o r t (Bechtel 1990) was made a v a i l a b l e by Savannah River S i t e
(SRS) personnel. I t i s organized i n t o eighteen example problems, each of which may be broken down i n t o several cases. The examples a r e described b r i e f l y a s follows.

Example
1

Description Free-field a n a l y s i s of halfspace; v e r t i c a l l y propagating

SP and P waves.

2
3 4 5 6

7
8 9 10

Containment building model w i t h b r i c k s and beams; uses COMBIN. Rigid d i s k on surface subjected t o a trapezoidal pulse. Impedances f o r c i r c u l a r foundation on layered soil system. S c a t t e r i n g response o f a r i g i d massless c y l i n d e r f u l l y embedded; subjected t o v e r t i c a l SV-wave and horizontal SH-wave. Surface foundation c o n s i s t i n g of both r i g i d and f l e x i b l e components. S c a t t e r i n g response of an embedded massless r i g i d cube w i t h an inclined wave. S e r i e s of simple problems using c u r r e n t l y a v a i l a b l e elements. Displacement s o l u t i o n s . S e r i e s of simple problems using c u r r e n t l y a v a i l a b l e elements. S t r e s s s o l u t i o n s . Coupled building a n a l y s i s , two dimensional.

WHC-SD-GN-CSWD-30036
11

Rev 1 Page 3

12 13
14 15

16
17

18

S c a t t e r i n g response of a r i g i d massless c y l i n d e r f u l l y embedded i n a layered system. Impedance response o f a r i g i d f u l l y embedded cy1 i n d e r . F l e x i b l e p l a t e on s u r f a c e subjected t o machinery loads. Compliance f u n c t i o n s of a s t r i p f o o t i n g on the s u r f a c e of a v i s c o e l a s t i c halfspace and a layered s o i l system. Compliance functions f o r a r i g i d c i r c u l a r s u r f a c e foundation i n h o r i z o n t a l , v e r t i c a l , rocking and t o r s i o n a l modes. Rigid massless c y l i n d e r f u l l y embedded and subjected t o Rayleigh wave input. S c a t t e r i n g response of a r i g i d square s u r f a c e foundation subjected t o SH-, SV-, P-waves and Rayleigh waves. SSI response o f a stick model on a r i g i d d i s k w i t h the control p o i n t a t grade level.

C o l l e c t i v e l y , the examples r e q u i r e 174 executions of the nine SASSI modules. S o l u t i o n s are compared w i t h closed form s o l u t i o n s , open l i t e r a t u r e , and a1 ternate Val i d a t e d program results. The Val i d a t i on probl ems have been run and checked on the DOE Savannah River Cray XMP computer which i s s i m i l a r t o t h e RL and LANL Crays. The extent o f SRS module t e s t i n g i s a s follows: Module SITE POINT2 POINT3 HOUSE
MOTOR

Executions
30

Source Lines
2800

Module Description

Free f i e l d s i t e response s o l u t i o n .
Two dimensional version o f POINTS. Point load s o l u t i o n needed f o r intera c t i o n node f l e x i b i l i t y matrix. S t i f f n e s s and mass m a t r i c e s f o r s t r u c t u r e and excavated s o i l . Load v e c t o r corresponding t o external f o r c e s , e.g., machinery. Formation o f impedance m a t r i c e s and loads; s o l u t i o n f o r t r a n s f e r f u n c t i o n s . Combination of two ANALYS frequency sets i n t o one. Time h i s t o r i e s of accel e r a t i ons , c a l c u l a t i o n of response spectra. Eva1 uation o f maximum stresses and s t r a i n s i n s t r u c t u r a l elements.

4
20

1200
1900 4100

31
26

350
3100 270 1700

ANALYS
COMB IN
MOT ION

49
1
8

STRESS TOTAL

5 174

2000
17400

The scope o f the v a l i d a t i o n e f f o r t shows good balance.

Example 2, Cases 1-5, along w i t h Example 9, Case 5, were adopted as the in-use t e s t i n g set, because they exercise a l l the modules. In a d d i t i o n , Examples 5, 9, 11, and 17 were run on the RL Cray, and the results compared favorably w i t h the v a l i d a t i o n r e p o r t . SASSI i n s t a l l a t i o n on the INEL Cray s t a r t e d i n l a t e 1993 a s a replacement f o r the RL Cray. I t passed t h e in-use test. However, a waste t a n k model produced anomalous results i f a s i n g l e block was used f o r the s t r u c t u r a l s t i f f n e s s

WHC-SD-GN-CSWD-30036 Rev 1 Page 4 matrix in the house module. Testing on other Cray sites and workstations could not reproduce the anomalous results. Installation at INEL was then dropped in favor of LANL. A precise cause of the anomalous results at INEL was never identified. Consequently, other than the extensive testing that followed, no claim is made that a similar incident cannot happen again. At LANL, the in-use tests and two waste tank models (including the one with anomalous results at INEL) proved satisfactory. Considering the INEL experience, however, it was decided to run all 18 test cases from the BechtelSRS validation report and compare with RL output obtained by Kaiser Engineers Hanford (KEH) in an effort following the RL Cray installation (Luo 1994). Testing with these cases was based entirely on computer-differencing of test and reference files and proved satisfactory. Hadjian (1989) reports on an experiment designed to evaluate and synthesize results of several soil-structure interaction analyses based upon measured responses of a 1/4 scale containment model in the quake-prone Lotung region o f Taiwan. Results of this reference indicate that deconvolution analyses using equivalent 1 inear methods and assuming vertically propagating shear waves captured the essential features of variations of ground motion with depth. Considering recorded results with the model containment, several analyses methods were compared by Tang (1989). SASSI was employed by two organizations in this comparison, and it proved to be as good as or better than any other method in its prediction capability. Finally, two in-house projects (Giller 1991, Winkel 1991) used both the SASSI and FLUSH systems for analyzing embedded and underground structures. The reasonable comparisons between the two systems added to the confidence in the SASSI system. SHAKE has been widely used in the industry since its introduction in the early 70's. The test input cases provided by EERC with the source were run and checked satisfactorily with the output provided. In addition, a soil profile typical of the Hanford site was analyzed with the Fc correlation (see Appendix A) to a 0.2 g Newmark-Hall time history appropriate to the response spectrum given in SDC4.1, Rev 11 (Hanford, 1989). The results compared well with FLUSH results (Winkel, 1991). SHAKE91 was given similar testing. HOUSEPLT was satisfactorily tested with nine of the HOUSE module validation problems. Testing included ANSYS plots with all useful plotting options; details may be found in Appendix A. Users should check usage of the filters TRFILTER, HFILTER, and SMFILTER based on normal vs truncated output. The reliability of the LANL/SASSI system demonstrated by the above discussions is considered adequate for impact level 2 software to be used by qualified engineers. CONVERSION The SASSI version as received was for a Cray 1s system. The effort to convert to the Cray XMP under the Unicos operating system was minimal, being restricted to sensing what memory is available to the user and allocating t to blank common. SHAKE and SHAKE91 conversions were also minor, being restricted to character variables, a1 ignment considerations and formatting

WHC-SD-GN-CSWD-30036 Rev 1 Page 5 Details of the conversions are given in Appendix B. CONFIGURATION MANAGEMENT Installation: SASS1 SHAKE91, SHAKE HOUSEPLT

- LANL - LANL

Cray YMP (gamma, zeta machines) Cray, DOS PC, SECC sgi3 w/s LANL Cray, DOS PC, SECC sgi3 w/s

User access: LANL Cray: Normal Cray access plus copy of program file from CFS to temporary directory. See Appendix A. DOS PC: Through the user's PC to server area \\whc240\special. See Appendix A. SECC: Normal SECC access plus copy of program file from CFS to temporary directory. See Appendix A. Custodian: EO Weiner User identi fi cati on : Users should identify themselves to the custodian and indicate the installation to be used. They should also indicate when usage is no 1 onger anticipated. Files: Source files: *.f (Cray, SECC), *.f (PC) and ic1im.c. Main executable fi 1 es : site, point2, point3, house, motor, analys, motion, combin, and stress. Auxiliary executable files (see Appendix A): shake, shake91, houseplt, trfilter, hfilter, smfilter, rdtable, f s Configuration fi 1 es: sassila.res: script - Cray compile and load sources sassila.tst: script - Cray in-use testing sassiws.res: script - sgi3 compile and load sources sassiws.tst: script - sgi3 in-use testing sassi.usr: custodian's list of users sassi .log: custodian's log notes: custodians notes sassires.bat: DOS batch - compile and load sources on PCs sassitst.bat: DOS batch - in-use testing on PCs In-use testing: e2c*d: example 2 inputs, SRS validation e9c*d: example 9, case 5 inputs, SRS validation *.inp, *.out: test i/o files. sassi.out is the concatenation o f representative e[29]c* outputs from tests Other: nsbtest: dynamic memory example Val id/: Val idation problems *.doc: excerpts from user's manuals Primary storage: Source, in-use testing, configuration files (no executables): LANL CFS /eow/sassi .tar.Z

WHC-SD-GN-CSWD-30036 Rev 1 Page 6 Program fi 1 es (executabl es , exampl es, *.doc fi 1 es) : LANL CFS /eow/sassiusr. tar (requires general read access) Backup storage: Source, in-use testing, configuration files (no executables): RL CFS /w91428/sassi .tar.Z Custodian's PC, dir=c:\sassi\ref Program files (executables, examples, *.doc files): Copy and recompile from LANL or RL source files Other storage: Val idation problems: RL CFS /w91428/sassi/bval inp.tar.Z Custodian's PC c:\sassi\valid Error reporting and modifications: All errors and anomalies as well as modification requests should be directed to the custodian in writing. Error reports and program revisions are made by the custodian with the approval of his manager. They are announced by distribution to a list of users maintained by the custodian. Changes to the programs are tested and documented by the custodian, in a manner conforming to WHC-CM-3-10. Custodian log/l i sting The custodian maintains a log of significant events such as error reports, resolutions, modifications, changes in operating environment or configuration. Source listings and test outputs are maintained by the custodian on his PC, dir=c:\sassi. Rest orat i on Restoration (compile and load) is performed by the custodian. For Cray restoration, sources may be copied from the permanent storage (CFS /eow/sassi .tar.Z) , uncompressed, and untarred to a Cray disk directory, typically in the custodian's temporary area. Script sassila.res then compiles and loads the sources, establishing a sassiusr program directory. Sources and other user-unnecessary materi a1 are removed by the custodian with a script cleanup.scr. The sassiusr should be stored on CFS for user access. Access rights for the user should be established with the cfs command: mod sassiusr aval=O/r/-/s. In-use testing can be with sassila.tst. For PC restoration to k:, sassires.bat can be run by the custodian. In-use Testing In-use testing is to be conducted with changes in operating environment or configuration. The Cray and SECC modules are tested by the custodian, and the PC modules are tested by the individual user on his machine. Script and batch files are provided: sassila.tst for the Cray, sassiws.tst for sgi3 and sassitst.bat for the PC. In these scripts, *.inp files are input test files, and *.out files are reference output files. In-use testing consists of running *.inp files with the current modules after restoration and differencing the resulting outputs against the reference outputs. Differences sent to *.dif files may show version number and timing differences, etc. The PC modules typically have reference outputs from the SECC executions, and they may show minor

WHC-SD-GN-CSWD-30036 Rev 1 Page 7 roundoff and timing differences which are acceptable. Testing on the server PC is done initially by the custodian. See Appendix C for details. With satisfactory results, the custodian should replace the reference outputs with the newly generated results on the server. Testing by the user on his own machine by copying the PC server directory k:\eow\sassi to his own directory and executing sassitst. bat should then produce a close match. Scripts Scripts 1 isted in this configuration report provide detail information on how the SASSI system is configured. The custodian insu es that details of the scripts are current. File transfers Moving files such as sassi.tar.2 and sassiusr.tar with FTP should a1 ways be done in the binary mode. Such moves may be necessary w th RL CFS as well as the Internet file transfers offsite.

REV IS IONS
Version cft77 cft77 cft77 cft77 4/3/91 4/26/91 6/7/94 9/7/91 Program SASSI SHAKE SHAKE91 HOUSEPLT Compi 1 er Version cft77 cft77 cft77 cft77 6.0.2.0 6.0.2.0 6.0.2.0 6.0.2.0 Operating System Unicos Unicos Unicos Unicos 7.0
7.0 7.0 7.0

Revision Initial Initial Initial Initial

Version identification is included in the program sources and the printed output of each application. SHAKE, SHAKE91 and HOUSEPLT modules available on the PC are identical to the Cray versions. The PC copies are compiled with Lahey f771 version 4.01 with the DOS 3.31 operating system. REFERENCES WHC, 1993, "Software Practices", WHC-CM-3-10, Re1 0, 1-31-93. DeSalvo, G.J. and R.W. Gorman, 1989, "ANSYS Engineering Analysis System User's Manual", for revision 4.4, Swanson analysis Systems, Inc., Houston PA 15342, 1989. Lysmer, J. et al, 1975, "FLUSH - A Computer Program for Approximate 3-D Analysis of Soil-Structure Interaction Problems", Report No. EERC 75-30, College of Engineering, University of California, Berkeley CAY November 1975, available from NTIC as PB 259 332. Hadjian, A.H. et al, 1989 "Soil-Structure Interaction Lotung Experiment: Prediction Evaluation Basis", SMiRT 10, vol K/1, p 187, August 1989 Idriss, I.M., 1992, "User's Manual for SHAKEgl", program and code avail able from NISEE, Earthquake Engineering Research Center, University of California, Berkeley, CA

J. Lysmer et a1 ., 1991, "SASSI User's Manual", Bechtel Corp., San

WHC-SD-GN-CSWD-30036 Rev 1 Page 8 Francisco, 1991.

J. Lysmer et al., 1981, "SASSI

- A System for Analysis of Soil-Structure Interaction" , Report No. UCB/GT/81-02, Geotechnical Engineering, University of California, Berkeley, April 1981.

Bechtel , 1991, "SASSI Theoretical Manual", Bechtel Corp. , San Francisco, 1991. Bechtel , 1990, "SASSI Val idation Report", prepared for Westinghouse Savannah River Co., Aiken, SC by Bechtel Savannah River, Inc., 1990. Tang, Y.K. , 1989, "Proceedings: EPRI/NRC/TPC Workshop on Seismic SoilStructure Interaction Analysis Techniques Using Data from Lotung, Taiwan", NP-6154, Research Project 2225, EPRI, March 1989. Bathe, K.-J. et al, 1974, "SAP IV - A Structural Analysis Program for Static and Dynamic Response of Linear Systems", Report No. EERC 7311, College of Engineering, University of California, Berkeley CAY April 1974. Schnabel, P.B. et al., 1972, "SHAKE - A Computer Program for Earthquake Response Analysis of Horizontally Layered Sites", Report No. EERC 72-12, Earthquake Engineering Research Center, University of Cal i fornia, Berkeley CAY December 1972. Hanford, 1989, "Arch-Civil Design Criteria, Design Loads for Facilities", HPS-SDC-4.1, Rev 11, May 1989. Giller, R.A. and E.O. Weiner, 3991, "Soil Structure Interaction Analysis for the Hanford Site 241-SY-101 Double-Shell Waste Storage Tanks" , WHC-EP-0504, Westinghouse Hanford Co. , Rich1 and WAY September 1991.

WHC-SO-GN-CSWD-30036

Rev 1 Page A-1

Appendix A.

Runstream Examples and User Information

WHC-SD-GN-CSWD-30036 Rev 1 Page A-2

Appendix A.

Runstream Examples and User Information

SASSI SCRIPT EXAMPLE

The following LANL PROD script runs the example cases 1 and 2 shown in Table A.2-3 of the validation manual (Bechtel, 1990). It demonstrates how the user should obtain the executables from CFS. #!/bin/csh # PROD job assumed. Above line assures C-shell interpretation uname -a # identify platform date da # initiate job accounting # redirected output is forced unset nocl obber # limit memory to < 20 Mw max limit -v -jO -m8000000 # your own tmp directory cd /usr/tmp/yourid # sassi tarred executables from cfs cfs get /eow/sassiusr.tar set p = /usr/tmp/yourid/sassiusr # make program directory mkdir -p $p cd $p tar xvf ../sassiusr.tar # untar program file into $p # back to /usr/tmp/yourid cd mkdi r exampl e # run directory cd example # copy inputs to run directory cp $p/e2c*

..

# # SASSI run script, Example 2 # See User Manual Table 4.6-3 and Validation Manual Table A.2-3
#

# Case 1 Horizontal #

$p/site te2clsd >e2clso

mv fort.1 e2cltl

mv fort.2 e2clt2
#

cp e2clt2 fort.2 $p/point3 <e2clpd >e2clpo mv fort.3 e2clt3

#

$p/house te2clhd >e2clho mv fort.4 e2clt4

#

$p/motor te2clhmd >e2clhmo mv fort.9 e2clht9

#

cp e2clt3 fort.3 cp e2clt4 fort.4 cp e2clht9 fort.9 $p/analys <e2clhad >e2clhao

WHC-SD-GN-CSWD-30036 Rev 1 Page A-3

mv f o r t .95 e 2 c l t 5 mv f o r t . 9 6 e 2 c l t 6
mv f o r t . 8 e 2 c l h t 8 # # Case 1 Rocking # $p/motor te2clrmd >e2clrmo mv f o r t . 9 e 2 c l r t 9
# cp e 2 c l r t 9 f o r t . 9 cp e 2 c l t 6 f o r t . 9 6 $p/analys t e 2 c l r a d >e2clrao mv f o r t . 8 e 2 c l r t 8 # # Case 2 SSI # Warning: house i n p u t e2c2hd must have maxc & maxt i n e2clho o u t p u t above # $p/house te2c2hd >e2c2ho mv f o r t . 4 e2c2t4 # cp e 2 c l t l f o r t . 1 cp e 2 c l t 5 f o r t . 9 5 cp e2c2t4 f o r t . 4 $p/analys te2c2ad >e2c2ao mv f o r t . 9 6 e2c2t6 mv f o r t . 8 e2c2t8 # cp e2c2t8 f o r t . 8 $p/motion te2c2od >e2c200 mv f o r t . 1 2 e2c2t12 # j a -st

The nqs s c r i p t makes no attempt t o save t h e *o o u t p u t o r * t n d a t a f i l e s . T h i s should be done w i t h CFS commands i n t h e s c r i p t . LANL Unicos and o t h e r a p p r o p r i a t e documentation should be consulted f o r d e t a i l s . Assuming t h e above s c r i p t i s j o b l , submit t h e j o b w i t h : gamma% prod Prod> submit j o b l t=180 pr=9 s=* ta=g Prod> end Note t h a t t h e t a r g e t machine (gamma) i s s p e c i f i e d w i t h ta=g and t h a t t h e temporary r u n area i s speci f ied i n t h e s c r i p t as /usr/tmp/yourid/exampl e. The user can expect t o f i n d output f i l e s t h e r e up t o 48 hours l a t e r . I f t h e t a r g e t machine were l e f t open and t h e temporary area n o t s p e c i f i e d , i t i s n o t c l e a r t h a t unsaved f i l e s c o u l d be recovered. SASS1 i s o n l y t e s t e d on gamma and zeta, and usage should be r e s t r i c t e d t o those machines. The t=180 e n t r y denotes a l i m i t o f 180 CP minutes, and pr=9 seems t o be a standard p r i o r i t y . Standard o u t p u t i s placed i n t h e u s e r ' s home d i r e c t o r y . The LANL PROD documentation should be consulted f o r d e t a i l s o f s u b m i t t i n g p r o d u c t i o n runs.
ON-LINE DOCUMENTATION

WHC-SD-GN-CSWD-30036 Rev 1 Page A-4 Excerpts from t h e u s e r ' s manuals have been placed on-line as *.doc f i l e s on the Cray, s g i 3 w/s and PC server w i t h t h e program f i l e s . Some o f t h e seldom used a u x i l i a r i e s such as f s are n o t compiled on some o f t h e p l a t f o r m s . I n such cases t h e *.doc f i l e i s t h e source f i l e which t h e user may compile.

DATA FILES
Data output f i l e s l i k e e 2 c l t 4 appear as f o r t . 4 a f t e r execution o f t h e f i r s t HOUSE module. The exception t o t h e r u l e i s : f i l e s t h a t should show up as f o r t . 5 o r f o r t . 6 (e.g., e 2 c l t 5 from ANALYS) show up as f o r t . 9 5 and fort.96. I t i s t h e u s e r ' s t a s k t o manipulate and save t h e proper f i l e s i n accordance w i t h t h e user manual. Recommended c u r r e n t p r a c t i c e i s t o rename output f 0 r t . n f i l e s needed l a t e r as ( f o r example) eZcltn, and then copy i t back t o f0rt.n when needed as i n p u t . It might be p o s s i b l e t o s k i p t h e rename/copy processes. For example, f o r t . 4 from HOUSE should be good f o r ANALYS, because m o d i f i c a t i o n o f f o r t . 4 by t h e system has n o t been noted unless declared i n t h e user manual.
RESOURCE ESTIMATES

ANALYS CPU, f i l e s i z e and memory estimators are based on a 3 h r run:

CPU (sec) = (NUMGP*NDOF)**3 * NF * SCPU TAPE5 (words) = (NUMGP*NDOF)**Z * NF * S5 MEM (words) = (NUMGP*NDOF)**2 * SMEM t 300000

where

CPU TAPE5 MEM NUMGP NDOF NF SCPU S5 SMEM

= ANALYS cpu estimate, sec

= = =

tape5 s i z e estimate, words number o f i n t e r a c t i o n nodes number o f degrees o f freedom p e r i n t e r a c t i o n node number o f frequencies cpu scale f a c t o r = .4e-6 tape5 scale f a c t o r = 3 memory scale f a c t o r = 2

= HOUSE and ANALYS common b l o c k requirement, words = =

=

TAPE6 s i z e s a r e about l i k e TAPE5, and t h e o t h e r tapes a r e much smaller. The ANALYS module dominates t h e cpu consumption. The CPU estimate i s known t o be low f o r small jobs, by a f a c t o r o f 1.5 f o r a 10 min j o b . The o t h e r modules (SITE, POINT3, HOUSE, e t c ) a r e much quicker. See t h e memory expansion discussion below f o r HOUSE memory. As an example,
NUMGP = 469 NDOF = 3 NF = 10

so

CPU = (469*3)**3

10

.4e-6 = 11140 sec

vs 11100 sec actual

WHC-SD-GN-CSWD-30036 Rev 1 Page A-5 TAPE5 = (469*3)**2 * 10 * 3 = 59 Mword = 480 Mbyte MEMORY = (469*3)**2 * 2 = 4.0 Mword In smaller j o b s , one may consider using the maximum core a v a i l a b l e , e s p e c i a l l y i f i t results i n s i n g l e block s t i f f n e s s s t o r a g e per HOUSE block s t o r a g e information output. This may improve the turnaround, but not l i k e l y the CPU consumption. The l a r g e TAPE5/6 size should be noted. I t may cause permanent f i l e s t o r a g e problems and unneeded c o s t s . A rerun of ANALYS as opposed t o a restart should be considered along w i t h the l i k e l i h o o d of a r e s t a r t need.

MEMORY EXPANSION
Memory expansion i n HOUSE and ANALYS i s f a i r l y s t r a i g h t forward as the programs d e t e c t the maximum a v a i l a b l e (up t o 20 Mw) and back o f f 4.7 Mw with house and 3.7 Mw w i t h analys f o r a maximum blank common. The l i m i t ( ) command may be used t o reduce the maximum memory a v a i l a b l e i n view of t h e memory time i n t e g r a l c o n t r i b u t i o n t o cpu charges o r user requirements t o p l a c e 1 i m i t a t i o n s on the SASSI matrix blocking procedure. In doing r e s t a r t s w i t h HOUSE and ANALYS i t was p o s s i b l e f o r the user t o get trapped i n the RL system. A HOUSE run would block TAPE4 t o t h e maximum s i z e p o s s i b l e . Then ANALYS used t h i s TAPE4 a s i n p u t , and blocked TAPE5 t h e same e way. Then the user changed the s u p e r s t r u c t u r e and d i d another HOUSE run. H knew t h a t he had t o f o r c e HOUSE t o use t h e old blocking w i t h MAXT and MAXC so t h a t the ensuing ANALYS mode 2 run would read TAPE4 and TAPE5 w i t h the same blocking. One would expect the restart HOUSE r u n t o f a i l , because the model i s u s u a l l y l a r g e r . Instead the ANALYS mode 2 run f a i l s w i t h a small shortage o f core. I t i s hoped t h a t the smaller HOUSE blank common (vs ANALYS) w i l l prevent the problem from occurring a t LANL. I f n o t , the user might make i t a p o l i c y t o l i m i t ( ) memory somewhat a s a p r a c t i c e and t h e n i n c r e a s e i t f o r the ANALYS r u n i f the problem occurs.

The need f o r M S i n p u t i n the HOUSE module is not understood. I n p u t of UE MUSE>O appears t o interfere w i t h user i n p u t of MAXT and MAXC required t o insure coordinated block sizes d a t a f i l e s . The d e f a u l t M S input i s UE recommended a t this time.

EMBEDDED BLANKS

The CFT77 options used t o compile SASSI include the d e f a u l t BN i / o option w h i c h condenses o u t embedded blanks i n i n t e g e r and real i n p u t . An input f i e l d w i t h 11 1 1 reads 1111 i n s t e a d o f 110101, so i t i s suggested t h a t the u s e r p u t i n the z e r o s i n c a s e the program reads i t w i t h ( i 6 ) i n s t e a d of ( 6 i l ) . . A completely blank f i e l d is s t i l l read as a zero.
HOUSE GENERATION

Examples of node and element generation i n HOUSE may provide some c l a r i f i c a t i o n t o the user manual.

The node c a r d sequence (start-end c a r d s )

5.0

9.0

0.0 0.0

10.0 14.0

1
2

WHC-SD-GN-CSWD-30036 Rev 1 Page A-6 produces c c s 5 7 9

1 1

1
1

5.0 7.0 9.0

0.0 0.0 0.0

10.0 12.0 14.0

Thus, intermediate nodes are generated using the end card increment and the start card boundary conditions and coordinate system. The last card boundary conditions are not generated from the start card; they are read from the end card. The user will probably always use the same coordinate system on both the start and end cards. Note how the example works when they are different. The beam element sequence
10 13

12 22

14 29

120 120

1 1

1 1

5 2

produces
10 11 12 13 12 17 22 22 14 19 24 29 120 120 120 120

1 1 1 1

1 1 1 1

Thus, intermediate elements are generated from the start card using the start card increment for node numbers, and each i ntermedi ate element is generated. The end card nodes are not generated; they are read from the end card. BASEMENT STRUCTURE AND INTERACTION NODES The topic of interaction nodes and basement structure needs special attention. It was found that an interior basement wall physically not connected to the actual soil did not respond as expected as an oscillator excited at its base. The wall nodes were all identified as interaction nodes. The problem was corrected by creating separate non-interaction nodes for the wall structure away from the base, and this correction was verified as appropriate with one of the developers at Bechtel. Interaction nodes are used to model excavated soil, and their separation requirements are governed by soil properties. Basement structural nodes identified as interaction nodes apparently should have stiff paths to the actual soil. Clearly all exterior wall nodes should be interaction nodes. OTHER THINGS NOTED

Interaction nodes in HOUSE need to be precisely on layer interfaces. The user has to plan his grid accordingly. The recommended way is to use a simple grid with not too many layers. Then SHAKE91 can be run with those same layers, and the resulting data can be fed into HOUSE and

SITE. - SYMMETRY: The user's manual might lead one to believe that simply declaring a symmetry plane will take care of the boundary conditions. Apparently, the program will provide them for the soil impedence matrix,
but not for the structure or excavated soil. The user must definitely

WHC-SD-GN-CSWD-30036 Rev 1 Page A-7 provide them for the structure, and it appears prudent to provide them for the excavated soil. Generally, symmetric boundary conditions zero out one translational and two rotational dofs (degrees of freedom) that produce deformation out of the plane. Asymmetric boundary conditions zero out two translational and one rotational dof that produce deformation wholly in the plane.

SPRINGS: There is no obvious way to model the normal point-to-point truss with the spring element. An alternative is to use beam elements with member end releases (do NOT release BOTH torsional dofs). TRIANGULAR ELEMENTS: P1 ate/shell triangular elements are entered by omitting the fourth node. Quad2d triangular elements have the fourth node equal to the third. manual describe the plate element local axes. The triangle in the figure of H.3.3.2 shows the K- and L-nodes coalesced. Instead, the L-node should be coalesced with the I-node: the L-label should be moved next to the I-label. Then the orientation of the local axes in the Note that figure will be consistent with the definition in C.10.3.3. bending results should be interpreted as occurring at the K-node. Membrane results are constant over the triangle and would be better interpreted at the centroid.

- TRIANGULAR SHELL ELEMENTS: Sections C.10.3.3 and H.3.3.2 of the user's

- NPAR(1) essentially defaults to 1 in

HOUSE

MOTION: Printing several response spectra with MOTION is a good getting a quick overall view of the results. MOTION is cheap. histories are punched (per user manual), they are also printed, just gets in the way of the plotted response spectra. The user consider two separate runs in this case.

way o f If time and this might

- SITE: Aborts with a 90 degree incidence angle. Use 89 degrees. - STRESS with 2D quads will print T-OCT if requested with the fourth key.
The manual does not show this, most likely because the programming (in routine oct4) looks bad: Toct=(SxxtSzztTxz)/3. Asking for Toct with element 4 is definitely not advised. Users are reminded to contact the custodian to insure they are on the user list along with initial and final dates of usage. PC AND WORKSTATION MODULES

In addition to the Cray, SHAKE, SHAKE91 and HOUSEPLT and the filter modules are available on the PC and the SECC sgi3 workstation.

To access the PC versions, key

in

connect k: \\whc240\special copy k:\eow\sassi\*.* c:\sassi The individual programs are run on the user's machine with redirection, e.g.,

WHC-SD-GN-CSWD-30036 Rev 1 Page A-8

C:

cd \sassi shake91 tinput >output

Do not run the program on the shared area of the file server. version runs 20 times as slow as on the workstation.

The PC shake91

It is the responsibility of the user to conduct PC in-use testing to insure the system is working properly on his particular machine. The testing procedure is described in the configuration management section of this report, and the batch file sassitst.bat in k:\eow\sassi is constructed to facilitate this testing. SASS1 auxiliary programs SHAKE91, SHAKE and HOUSEPLT are available on the SECC workstation sgi3. Access to these programs is through the custodian's directory in CFS. The users should store and run them on their own temporary directories in the workstation, e.g., cd /t sgi3/yourid ftp cVsgate [enter id, password, set binary mode] cd /w91428 get sassiusr.tar quit cd /t sgi3/yourid/sassiusr tar xVf ./sassiusr.tar cd /t sgi3/yourid/rundir a1 i as-shake '/t sgi3/yourid/sassiusr/shake' shake tinput >oGtput

SHAKE91 This program was favored over the old SHAKE, because it provided for more soil types and layers, and it provided a more direct way of entering soil properties. Printer plots along with some seldom-used options are no longer found in SHAKE91. It should be noted that some old SHAKE input is still in effect, even thought the SHAKE91 manual does not mention it. Don't throw out the old SHAKE manuals. In particular,

Option 2 line 1 MWL input in cols 11-15 is not in the SHAKE91 manual. Omitting it means that MWL=O and amounts to putting the water table at the top of layer 2. The total pressure in the output has water hydro-pressure removed. It has no other effect on other ouput, however. The user can input MWL to get the total pressure expected. Option 2 line 2+ NLN input in cols 11-15 is also missing in the manual. It will produce sublayers within the one-line layer.

Changes pertinent to the user:

- Usage is like SHAKE; after copying similar to SHAKE described above, the
program is executed with: shake91 tinput >output
# t/h output on file "punch"

WHC-SD-GN-CSWD-30036 Rev 1 Page A-9

- Added code to take Poisson's ratio input in Option 2 line

2 t in cols

66-75 for each layer and produce iterated soil property output suitable for SASSI input. User must insure consistency of units.

Other things noted:

- Option 9 (compute response spectrum) has input for the acceleration of

gravity, which seems unusual since the program has a built-in value of 32.2 ft/sec/sec. The option 9 input seems to affect only the spectral displacement and velocity output units of the same option. If the user wanted the spectral velocity in cm/sec, then g = 981 would be appropri ate. SHAKE The SHAKE program solves the free-field soil problem with iteration on strain-dependent soil properties. Common practice uses SHAKE-derived soil properties as fixed input to SASSI. The FLUSH program has the SHAKE capability built into it; SASSI does not. Punch output is to be found on the named file "punch". The test problems take a little over a minute on a Compaq 386/20e. The PC must have a 80n87 math coprocessor. The time history sequence check was disabled. The treatment of units in the manual is delicate at best, especially for sand. The user should refer to the manual example input and output to insure his understanding of the Option 2 81 8 processes. It is suggested that Option 8 input be used to enter Ks=G/Go and beta vs strain as was the practice with FLUSH. Here, G and Go are the shear moduli at strain and low strain, respectively. The low strain shear modulus is then entered via Fs for sandy soil in Option 2 with Fs = Go / sqrt(Sm) Go = low strain shear modulus, ksf Sm = (1 t 2*sko)/3 * head head = overburden pressure, psf sko = coefficient of earth pressure at rest, default

.45

The (c) note in Option 2 has a factor of 1000 in the Gs formula. It appears that the program internally applies this factor and that the user should not. The user should think in terms of ksf input for the shear modulus. One could view Fs as an correlation tool: a representative value should suffice unless there is evidence of drastic changes in soil properties. See Reference Al. An alternative would be to use the clay correlation Fc=Go which is more .like FLUSH practice. However, the sand correlation tends to produce smoother iterated modulus profiles in large layers with constant Fs values, because the sqrt(Sm) relation tends to soften the soil near the top of the layer, while a constant Go produces increasing iterated G values toward the top of the layer. The user is cautioned about using the sand correlation near the surface where the control motion is specified. Soft soil under the control point tends to attenuate the free field motion at depth, and this could lead to unconservative excitation for embedded structures. Fs input example:

WHC-SD-GN-CSWD-30036 Rev 1
Page A-10

4096 Project t i t l e 2 Read s o i l p r o f i l e data. Note: t o t a l sublayers <= 20 1 9 17 100-N 146.5 1. 0.1350 1 2 2 10.0 3029. 0.05 104.0 1. 0.1350 2 2 2 18.0 4193. 0.05 98.5 1. 0.1370 3 2 3 34.0 6127. 0.05 445.8 1. 0.1400 0.05 4 2 2 21.0 35315. 935.7 1. 0.1400 5 2 1 20.0 84174. 0.05 248.4 1. 0.1250 6 2 2 55.0 26242. 0.05 154.3 1. 0.1290 7 2 3 60.0 19390. 0.05 278.0 1. 0.1200 8 2 1 16.0 38161. 0.05 .12 8000. 9 Assign motion t o surface 3 1 Materi a1 property s t r a i n dependence 8 2 1 10 100. 11 100. CLAY -- N REACTOR IMPELL PHASE I 1 1985, STIFFNESS .000316 .001 .00316 .01 .0316 .1 .316 .OOOl 3.16 10. 1. .99 .98 .94 .88 .80 .69 .60 1 .o .51 .50 .54 11 100. CLAY -- N REACTOR IMPELL PHASE I 1 1985, DAMPING .000316 .001 .00316 .01 .0316 .1 .316 0001 1. 3.16 10. 1.2 1.3 3.1 3.4 5.2 7.6 11. 1.2 5. 15. 15. 11 100. SAND -- SEED & I D R I S S 1970, STIFFNESS .OOOl .000316 .001 .00316 .01 .0316 .1 ,316 1. 3.16 10. .15 1 .o .99 .96 .89 .74 .53 .30 .06 .06 .06 11 100. SAND -- SEED & IDRISS 1970, DAMPING .OOOl .000316 .001 .00316 .01 ,0316 .I .316 1. 3.16 10. 0.5 0.9 1.7 3.1 5.6 9.7 15.5 21. 25. 25. 25. 1 Read input motion 2000 4096 ,005 Newmark-Hall 0.2g SDC4.1 Rev 11 .224 20. 0.014141 0.015267 0.015951 0.016336 0.016758 0.016979 0.016954 0.016629

...

0.097009 0.094396 0.091645 0.089001 0.08675 0.07806 0.069304 0.030445 4 S t r a i n compatible s o i l properties 1 5 5. .65 0 stop

Fc input example (models sand p r o f i l e with Fc):

4096 2 1 1 2

Project t i t l e Read s o i l p r o f i l e data.

Note: t o t a l sublayers <= 20

9 1 1

17 2 2

100-N 10.0 18.0

3029. 4193.

0.05 0.05

0.1350 0.1350

3029. 1. 4193. 1.

WHC-SD-GN-CSWD-30036

Rev 1 Page A-11

6127. 1. 3 1 3 34.0 6127. 0.05 0.1370 35315. 1. 4 1 2 21.0 35315. 0.05 0.1400 84174. 1. 5 1 1 20.0 84174. 0.05 0.1400 26242. 1. 6 1 2 55.0 26242. 0.05 0.1250 19390. 3 . 7 1 3 60.0 19390. 0.05 0.1290 38161. 1. 8 1 1 16.0 38161. 0.05 0.1200 9 .12 8000. 3 Assign motion t o s u r f a c e 1 8 Material property s t r a i n dependence 1 1 10 100. 11 100. SAND -- SEED & I D R I S S 1970, STIFFNESS ,316 -0001 .000316 .001 .00316 .01 .0316 -1 1. 3.16 10. .15 1.o .99 .96 .89 .74 .53 .30 .06 .06 .06 11 100. SAND -- SEED & I D R I S S 1970, DAMPING .316 .OOOl .000316 .001 .00316 .01 .0316 .1 1. 3.16 10. 21. 0.5 0.9 1.7 3.1 5.6 9.7 15.5 25. 25. 25. 1 Read i n m t motion 2000 4096 .005 Newmark-Hall 0.29 SDC4.1 Rev 11 .224 20. 0.014141 0.015267 0.015951 0.016336 0.016758 0.016979 0.016954 0.016629

...

0.097009 0.094396 0.091645 0.089001 0.08675 0.07806 0.069304 0.030445 4 S t r a i n compatible s o i l p r o p e r t i e s 1 5 5. .65 0 stop

Several other comments can be made with r e s p e c t t o Options 2 and 8:

- The d e n s i t y input i n Option 2 i s i n ksf. - Use damping f a c t o r BFAC=l i f Option 8 d e s c r i b e s t h e damping. - The shear moduli and damping values (GMOD and B y cc 26-45) entered i n
Option 2 a r e i n i t i a l guesses used t o s t a r t t h e f r e e f i e l d i t e r a t i o n process c a l l e d for by Option 8. - Note t h a t t h e l a s t Option 2 card is for t h e e l a s t i c halfspace and t h a t one need only e n t e r t h e l a y e r number, d e n s i t y and s t a r t i n g s h e a r velocity - The water l e v e l i s s p e c i f i e d by t h e sublayer number, not the l a y e r number. - Note t h a t Option 8 does not have a material type e n t r y ( l = c l a y , 2=sand, e t c ) . Instead, t h e f i r s t material presented i s for clay c a l l e d o u t f o r i n Option 2, and t h e second material i s f o r sand. I f t h e user has only sand i n the model and shows only one material i n Option 8, t h e run a b o r t s . There i s a f o u r t h material type w h i c h appears t o be a cube r o o t sand correl a t i on.

Program fs determines Fs f a c t o r s f o r SHAKE option 2 from s o i l d e n s i t y and shear v e l o c i t y p r o f i l e s . Coefficient of e a r t h pressure i s .45 (SHAKE

WHC-SD-GN-CSWD-30036 Rev 1 Page A-12 default). Option 8 must be used to enter the iterated to low-strain shear modulus ratio. Initial guesses output are .05 damping and input shear modulus. The latter could be used as a clay Fc by way of an editor. User must review output for validity to his problem. Usage: fs tinput >output Input, one list directed line per layer with: * dbot Depth o f bottom of layer, ft * vs Shear velocity, fps * wt Density, pcf Example input and output:
10 28 62 83

850 1000 1200 2850 1 1 1 1

135 135 137 140

10.0 18.0 34.0

/ comment

1 2 3 4 HOUS EPLT

2 2 2 2

21.0

3029. 4193. 6127. 35315.

0.05
0.05

0.05 0.05

0.1350 0.1350 0.1370 0.1400

146.5 104.0 98.5 445.8

1.

1. 1. 1.

This program is a simple translation of SASSI HOUSE module input to ANSYS version 4.4 PREP7 input suitable for mesh plotting. It is available on both the Cray and PC. User instructions:

Read SASSI HOUSE input and translate for ANSYS 5.0 mesh plotting with boundary conditions included. EO Weiner 6/23/94 Usage: houseplt thouseinput I asa >ansysinput # Cray/ws # PC housepl t thouseinput >ansysinput - Plotting can be done with the ansys system: ansys /show ,2d # or whatever plot device /input,ansysinput epl ot - Since sassi element numbering starts with one in each group, offsets (appearing as ANSYS comments) are added to each group. The element offset is always a multiple of 100. - ANSYS type is HOUSE group in the order of appearance. This permits element group selection with ersel ,type. ANSYS real is HOUSE geometry number for beams. - Lumped masses are treated as Ansys unit mass elements under one group after the sassi element groups. - Viewpoint along the negative y axis is selected with z up. This is suitable for 2D elements. For 3D elements, try /view,1,.7,.5.,5 at plot time. - Boundary condition restraints (id=l) are not removed later in the ANSYS input when id=O is prescribed. The practice of first restraining all dofs followed by removing restraints at nodes in elements will show all dofs restrained. A work-around trick

WHC-SD-GN-CSWD-30036 Rev 1 Page A-13 may be t o remove the i n i t i a l global r e s t r a i n t f o r houseplt processing and then p u t t i n g i t back i n f o r SASSI processing. - There i s no checking f o r HOUSE consistency i n this program. Check d a t a first w i t h the HOUSE d a t a check option. - ANSYS beam k-node d o e s n ' t generate l i k e HOUSE. T r a n s l a t i o n uses the spar element. T r a n s l a t i o n s of r e a l values ( a r e a s , etc) and material property values (Young's modulus, etc) are not attempted. - Node, element, type, material and real s e l e c t i o n s may be made a t p l o t time using ANSYS S e l e c t Logic. - Fortran source n o t a t i o n follows user's manual. - ANSYS input i s generated almost immediately a s SASSI input i s read i n . Houseplt does not remember a l l nodes, elements, boundary conditions, etc. - e t , 1 , 6 3 appears up f r o n t t o a c t i v e a l l 6 d o f s and avoid warnings by d-lines before e-lines i n ANSYS 5.0 - Element type mapping: SASSI: 1 2 3 4 5 6 7 8 9 ANSYS: 45 8 63 42 45 42 8 42 27

Test i n p u t f o r HOUSEPLT i s a c c e s s i b l e t o the user on the PC and Cray program f i l e s . The results a r e shown i n Figures A . l through A.4. TRF I LTER This s h o r t program i s used t o f i l t e r ANALYS transfer function output t o s e l e c t e d nodes. User i n s t r u c t i o n s : F i l t e r out t r a n s f e r function p r i n t from ANALYS output. Usage : trf i 1t e r t a n a l y s .out >output # Cray List nodes f o r d e s i r e d output on f o r t . 1 0 , any order, 1000 max. F i l t e r on/off i s depends on Fortran ANALYS output headings.
HF I LTER

This s h o r t program truncates t h e extensive node and element output from HOUSE and l e a v e s the message "etc, etc, etc".
Usage: h f i l t e r thouse. o u t >output
SMFI LTER

This s h o r t program t r u n c a t e s the extensive time h i s t o r y output from MOTION and STRESS and l e a v e s the message "etc, etc, etc". Usage: smfilter tmotion.out >output RDTABLE This s h o r t program reads i / o t a b l e s provided w i t h the v a l i d a t i o n i n p u t s , and g e n e r a t e s script f i l e s f o r running t h e r e l a t e d SASSI modules. The generated script f i l e s a r e not t o be considered complete. The user must check and u s u a l l y modify them. User i n s t r u c t i o n s :

WHC-SD-GN-CSWD-30036 Rev 1 Page A-14 Read sassi v a l i d a t i o n problem i / o t a b l e s and generate s c r i p t s . Usage: r d t a b l e < t a b l e > s c r i p t Move r i g h t h a l f o f t a b l e under l e f t h a l f and dup module column. I n p u t format i s (4(lx,a7),lx,a8) f o r module, , tapeout. Note r i g h t h a l f i n p u t column i s narrow (6 c o l s ) , make i t 7. Needs underscores i n f i r s t column t o separate out modules. Users checks r e s u l t s . Cannot handle PGMIMP o f Table 12. F i r s t non-blank l i n e i s output as comment. Cols 11-40 i n header a f t e r "MODULE" are output as comment.

...

REFERENCES Al.
~

HB Seed, e t a1 , "Moduli and Damping Factors f o r Dynamic Analyses o f Cohesionless S o i l SI', Report No. UCB/EERC-84/14, Earthquake Engineering Research Center, U n i v e r s i t y o f Cal i f o r n i a, Berkeley CAY September 1984.

WHC-SD-GN-CSWD-30036

Rev 1 Page A-15

Figure A. 1 Housepl t/ANSYS boundary conditions

ANSYS S . 0 A JUN 23 1994 PLOT N O . ELEMENTS TYPE NUN U

16;SO:t4

ROT
IV

xv
XF YF I F

YV

z o . 3

DIST=6.261

.-I :0.3

VUP

= l . S :Z :S

:Z

Figure A.2

Houseplt/ANSYS element outlines

AYSYS S . 0 A JUW 2 3 1 9 9 4

16:Sl:iq

P L O T NO ELEVENTS T Y P E NUY
XY

i o . 3 DlST-S.26l
ZY

YV

:0.3 :-I

XF YF

=I.S
=2 :I

YUP : I EDGE

IF

3 93

WHC-SD-GN-CSWD-30036 Rev 1 Page A-16 Figure A.3 Housepl t/ANSYS element numbers
AYSYS 5 . 0 A JUY 2 3 1 3 9 q 16:51:56 PLOl YO. 3
TYPE

ELEYEWlS
XY YY

IUY

:0.3
= . I
z o . 3

0 6 S T : 6 1 6. 6 5 XF ~ 2 1

IY
IF

:2 =S VU? : I

IF

Figure A.4
1

Houseplt/ANSYS material numbers

5.0 A JUN 23 1994 lS:55:53 PLOT YO 4 ELEMENTS
AYSIS ELEY
YUY

xv
YV Xf

z o . 3

OIS1=6.261
YF

zv

. . I
:2

z o . 3

z l . 5

IF =5 VUP = I

WHC-SD-GN-CSWD-30036

Rev 1 Page B-1

Appendix B.

Conversion

WHC-SD-GN-CSWD-30036 Rev 1 Page 8-2

Appendix B.

Conversion

SASS1 INSTALLATION

Example input e2pd was renamed e2clpd for consistency. Source files csite.f, chouse.f, etc were renamed site.f, house.f, etc, and executables take the same names without extension. All original source lines with cc 73-80 serial numbering were retained, commented out if necessary. Added lines are in lower case without serialization. Module ana1ys.f uses units 5 and 6 for unformatted i/o. This conflicts with stdin and stdout, so the ANALYS streams were redirected to fort.95 and fort.96 with open statements. Users must recognized this when manipulating files between modules. Modules h0use.f and analys.f, as received, utilized dynamic memory expansion in subroutine files2. The COS memory('uc',nnn) call was replaced with the Unicos fortran callable sbreak(nnn). This is the alternate method suggested in Appendix D o f the CFT77 manual. Note that the CFT77 Appendix D primary recommendation of using heap allocation was found to be useless - it passed the small program test but failed in the final application. The original variable NAVA in routine files2 i s a fixed maximum memory size. This was replaced by an upper bound memory resource limit which i s machine and environment dependent. Communication of this limit is done by the routine ic1im.c which calls the Unicos limit(2) system routine. The use of sbreak requires a fixed heap size of 1.2 Mw is identified in the segldr options along with identification of blank common as dynamic. The large heap size was chosen as Cray and INEL operational personal indicated that 32Kw might be needed for each open i/o unit, and analys shows 30 units in the program statement. For the argument in the sbreak call, the maximum resource figure is reduced by the size of the executable (size house, includes the heap) plus a substantial -1-2 Mw cushion to eliminate uncertainties found in different environments. An sbreak error return causes the program to stop. Contrary to the original scheme, sbreak is called only once at the start of the run rather than allow blank common to expand and contract as calls to files2 are made during the execution. A small program nsbtest was written to test th.is memory expansion scheme in both batch and interactive environments.
A user was trapped by HOUSE/ANALYS memory restrictions in a restart attempt as

discussed in the user appendix. Rather than force him to rework the entire problem, the lOOK cushion (see above) in files2 of ANALYS was temporarily removed, and he was able to continue after recompilation. The practice is not encouraged. A larger cushion in HOUSE (t1 Mw over ANALYS) is employed to help avert such problems.

I'

Stress runs aborted when beam time history output was requested. .The compiler warning that varible LL i s used but undefined in beam output led to a small

WHC-SD-GN-CSWD-30036 Rev 1 Page 8-3 correction to get full fft output. OPTIMIZATION Since SASS1 is known to consume computational resources for large embedded basement configurations, a few optimization checks were performed on the RL system. The default compilation options appear to optimize the program as well as can be expected. SHAKE INSTALLATION The i nput/output fi 1 e query feature was rep1 aced by redirect i on by comment i ng out the open statements for units 5 and 6. Option titles written to the console were commented out, as the runs are relatively short (-minute). The seven figure accuracy for shear modulus output during iteration was reduced to six, so that output has a better chance of being repeatable across PCs. Also, time history table columns were move closer together to facilitate printing output. Finally, one instance of an embedded blank in the test input was replaced with a zero. The effect of the embedded blank can be seen in the third property curve in the printed output found in the manual. The proper output, as provided by the digitized test output, has a smoother curve. Installation on the Cray required modification, because character variables were mixed with non-character variables in common blocks. The idnt and title character variables were moved from /EQ/ and /SOILA/ common blocks to their own blocks /EQc/ and /SOILAc/. The FACTOR output field of Option 2 was widened to accommodate larger values that may be desired if one uses the FLUSH practice of entering low strain moduli as Fc. The time history sequence check was disabled - we no longer punch cards. SHAKE91 INSTALLATION SHAKE91 was received as source, executables, and test input and output. The executable failed, with or without the extender provided. The source was installed on pc with f771, on LANL Cray with cft77 and on SECC sgi3 with f77. Timing is 1:2:48 for Cray:sgi3:486/25. The following problems were encountered on sgi3:

sgi3 cannot output certain Hollerith constants in format statements. It seems associated with lack of commas between constants. Commas were inserted in SHAKE91 and SHAKE. - Common block alignment problems associated with character and real variables in one block were corrected as in SHAKE. No alignment problems were noted from complex and real variables. Integers are i * 4 . Retyping reals to r*8 would cause alignment problems. - Acceleration t/h file name read in with Option 3 must not have leading blanks. Introduced code to remove blanks. - Test input option 7 had an extra blank line; removed. - Test last line did not read; replaced. - + ITERATION output caused overprint; added blank lines. - Frequency step in Option 10 different in test input vs output.

WHC-SD-GN-CSWD-30036 Rev 1 Page B-4 Considered a supplier mistake. Additional changes:

Added code to take Poisson's ratio input in Option 2 in cols 66-75 for each layer and produce iterated soil property output suitable for SASS1 input. User must insure consistency of units. - Stress/strain output is limited to 2049 steps. This was increased to 4096 with additional testing. - Usage modified like SHAKE: Usage: shake91 tinput >output # t/h output on file "punch" - Rearranged SHAKE91 routines like SHAKE and differenced the sources. Only one obvious difference: shake91 /time/ has t(9) which clearly should have been t(2049). Changed to t(4096) which corrected frequency output in the amplification option. BLANK LINE HEADER IN PC MODULE OUTPUT Modules SHAKE, HOUSEPLT, TRFILTER and RDTABLE as well as the Cray. Their stdout output (to adds a blank line at the top on a PC, but not as blank lines in HOUSEPLT output are ignored CUSTODIAL NOTES Moving compressed and tarred files such as sassiusr.tar and sassi .tar.Z with FTP should always be done in the binary mode. are avail able on PC/DOS systems files by redirection) always on the Cray. This is immaterial by ANSYS.

WHC-SD-GN-CSWD-30036

Page C-1

Rev 1

Appendix C.

Testing D e t a i l s

WHC-SD-GN-CSWD-30036 Rev 1 Page C-2

Appendix C.

Testing Details

Prior to RL installation, testing was carried out by Bechtel for Savannah River for 18 examples as discussed in the main body of the report. The validation report and example inputs were provided to WHC at the time of the RL Cray installation in 1991. Validation report examples 2, 5, 9, 11, and 17 as received from SRP were tested on the RL Cray and compared with the report results and found acceptable. Example 2 and Example 9 Case 5 were adopted for in-use testing since they exercised all the modules. In-use testing includes differencing with reference output. Example 2 later showed 10-20% changes at a change of operating systems from Unicos 5.1 to 6.1. The problem was traced down to ill-conditioning: beams linking a rigid containment pad to stick models representing containment and internals were much stiffer than necessary. The problem was not with the code, It was with the model which was corrected, and the in-use test was revised accordingly. Installation on the INEL Cray in early 1994 passed the in-use test. However, a waste tank model, identified herein as q8, produced anomalous behavior under certain circumstances. Various tests with the model showed that anomalous results were obtained if the number of structural blocks in the stiffness matrix is 1 in the house output. If more than one block resulted, which could result from a smaller memory allocation, good results were obtained. Bechtel and CEC personnel offered to run the problem on their IBM RISC and Cray installations, and all produced good results. This anomalous behavior was restricted to the q8 model at INEL. An additional problem at INEL, found by fixing the memory and running the q8 1-block model, produced an Operation Range Error in the house module. A quick installation on the gamma machine at LANL proved good with one block in both the variable and fixed memory configurations. For LANL installation, it was decided to run all 18 Bechtel test cases and compare with RL output obtained by KEH in an effort following the RL Cray installation (Luo 1994). Both in-use and q8 tests proved satisfactory. Testing with Bechtel validation cases proceded as follows.

Differenced KEH and SRS inputs, ok Differenced KEH scripts against those generated with SRS tables input to rdtable, ok Ran KEH inputs at LANL with rdtable scripts Differenced KEH/RL and LANL outputs after trfilter, hfilter and smfilter on both outputs at sgi3. ok.

It should be noted that Example 12 needs 30 minutes and should not be aborted prematurely. An MWTF 180 model of a buried waste tank tested ok. This model required three hours to run. The model caused anomalous behavior with the CEC ibm2 RISC demo

WHC-SD-GN-CSWD-30036 Rev 1 Page C-3 v e r s i o n when combin was used t o add 4 more frequenc es t o 12 frequency r e s u l t s . The model was sent t o CEC. An M T q t r model t e s t e d ok. T h i s model r e q u i r e d 2 hours o f r u n t i m e and had WF data a v a i l a b l e from t h e RL Cray as w e l l as t h e INEL Cray. Two v a r i a n t s o f t h e model e x i s t e d : f i l e s r l / i d a t RL/INEL and f i l e s l a n l / i n e l a t LANL/INEL. D i f f e r e n c e o f r l and id(1NEL) o u t p u t s were ok. D i f f e r e n c e o f inel(1NEL) and LANL were ok. T h i s g i v e s a chained ok t o RL data. Users a r e cautioned t h a t t h e source o f t h e anomalous r e s u l t s w i t h t h e INEL i n s t a l l a t i o n was never i d e n t i f i e d , and consequently i t cannot be argued t h a t t h e SASS1 coding was d e f i n i t e l y n o t involved. SHAKE and SHAKE91 were t e s t e d s a t i s f a c t o r i l y w i t h t h e examples p r o v i d e d w i t h t h e programs and described i n t h e u s e r ' s manuals. SHAKE91 punch f i l e s from LANL, SECC and t h e PC p l a t f o r m s were checked a g a i n s t t h e o u t p u t p r o v i d e d by t h e vendor. I n a d d i t i o n , t h e soil p r o f i l e analyzed w i t h FLUSH (Winkel, 1991) was reanalyzed w i t h t h e Fc c o r r e l a t i o n i n SHAKE as shown i n t h e example o f Appendix A. SHAKE91 was a l s o used w i t h t h e data. The comparison o f t h e t h r e e r e s u l t s i n F i g u r e C.4 i s considered e x c e l l e n t . T e s t i n g t h e PC modules by t h e custodian i s somewhat cumbersome. The reference o u t p u t s a r e from t h e SECC Unix systems and show minor d i f f e r e n c e s w i t h t h e PC o u t p u t . To view t h e d i f f e r e n c e s i n a reasonable manner, i t i s suggested t h a t t h e r e f e r e n c e outputs i n t h e PC server k:\eow\sassi d i r e c t o r y have t h e f i r s t column removed a f t e r t h e y a r e copied t h e r e w i t h sassires.bat. This e s s e n t i a l l y converts F o r t r a n c a r r a i g e c o n t r o l t o ASCI I c a r r i a g e c o n t r o l . Also, t h e houseplt.out f i l e has 0.0000E+OO terms which should changed t o 0.0000bbbb (b=blank) w i t h a t e x t e d i t o r . Then t h e PC outputs can be d i f f e r e n c e d w i t h t h e r e f e r e n c e outputs. The m o d i f i c a t i o n s and d i f f e r e n c i n g a r e best done on a Unix system. When found s a t i s f a c t o r y , t h e PC o u t p u t s should be moved t o t h e r e f e r e n c e outputs i n t h e PC s e r v e r k:\eow\sassi d i r e c t o r y t o make i t easy on t h e user t o t e s t w i t h h i s own machine.

WHC-SD-GN-CSWD-30036 Rev 1 Page C-4

Figure C . l
:l

Example 2 Case 3 Transfer Function Magnitudes

Node
Frequency Number = 30 (2.9 Hz)

158

3-

2t
46

Unicos 5.1

Stick-Pac Link Modulus, psf

Figure C . 2
0

Example 2 Case 2 Node 158x Spectra

W.RSpLMs61

15 .

1.o
ui U

05 .

0.0 1

Frequency, Hz

WHC-SD-GN-CSWD-30036

Rev 1 Page C-5

Figure C.3

20 Cycle Sine Response Spectra

Damping 0.02,0.05.0. IO, 0.20 (decreasing response)

Frequency. Hz

Figure C.4 SHAKE/FLUSH Comparison o f Iterated Soil Properties

. m
-0

100 -

= a

1%;
5 -

g
m

150 -

150 -

200 -

8 b
0

200 -

JIF,,

0.02 0.04 0.06 0.08 0.10

Damping Fraction