ML20008F629
| ML20008F629 | |
| Person / Time | |
|---|---|
| Site: | Allens Creek File:Houston Lighting and Power Company icon.png |
| Issue date: | 04/20/1981 |
| From: | Hartzman M Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20008F600 | List: |
| References | |
| NUDOCS 8104210406 | |
| Download: ML20008F629 (5) | |
Text
g
)
04/20/81 P00R ORIGINALS
. o)
L UNITED STATES OF A'iERICA MUCLEAR REGJLATORY Co ;ISS10:1 BEFORE_ THE ATO::IC SAFETY A';D LICEf! SING BOARD
- .i t ie : tatter of
)
)
- OJ5T01 LIGiTING A:!D POWER CO:1PA'lY )
Docket flo. 50 466
)
(illens Creek Nuciear Generating i
5tation, Unit 1)
)
- lRC STAFF SUPPLE!1E
- lTAL TESTI'10:1Y OF DR. ItARK HARTZftAN RELATIVE TO C0!1PUTER CODE ERRO4
[TEXPI % AC 25]
q.
Please state your na1e and position with the NRC.
A.
- iy nane is :tark liartznan.
I ar. caployed it the U.S. Nuclear P,egulatsry Colnission as a P-incipal Ilechanical Enginee-in the
':echinical Engincering Branch (ilE3).
'J.
riave you prepared a statenent of educational and professional qualifications?
1 1
A.
Yes.
It is attached to this testimony.
j Q.
What is the purpose of your testimony?
A.
Tne purpose of ay testinony is to respond to TEXPIRG Contention l
AC 23,.hich states as follows:
Computer code error TEXPIRG AC 26 The nuclear piping systens in ACNGS are designed s
using a conputer progran with a subroutine containing an algebraic intranocal force conbination in one of the steps.
Because the tine phasing of the horizontal and vertical c]nponents of earthquake notions are not known, the subroutine should use an absolute sun conbinatior l
l l
8104g3g
]
c PDDR BRIGINAl.
The CJ pJter prog'an used to calculate the strasses on :ne reactor and contain 7ent d2 ring the design basis and safe snutdown earthquakes is defecti.'e ue:aase it subtracts forces when they should be adde 1.
Q.
- ,'het type of conputer progran will be used by tne Applicant tn calculate stress levels on piping or reaction loads on supports during dynar.ic loading conditions sucn as seismic events?
A.
Tne Applicant has speci'ically stated in section 3.7 of the P5fA thet tne cccputar prograns used for design of nuclear piping systens in ACZ5 cont 3in subroJtines ulich Calculate internadal and interdirection31
(" intra nadal") response co,ponents according to toe require combinations as des:ricad in P,egulatory Guide 1.92, " Combining ".odal
% spor.sas and Spatial Co,ponents in seisnic Response Analysis."
Q.
Sciefly, what coes F.egulatory Guide 1.92 spe:ify as an acceptable :etlod of coroining response conoonents?
A.
ine Reg. Gaide specifies tha* for calculations based on tne Response Spe:tran Method, tne interdirectional response corponents shoul d be co cined by the SRSS (sgeare-root-of-the-sun-of-the-squares) algoritnn.
Tais is justified on tne basis that while the tine phasing of the vertical and horizontal naxinu, response conponents ic unkcoan, it is highly unlikely that these response values will occur sinultanec; sly, thJs a acre realistic approach is to combine then by the SRSS algorithn.
Q.
Are any other conbination nethods acceptable?
A.
Yes.
Fiping conpater prograns nay be developed witn a capability to conbine internal interdirectional responses by nany retnods, among rihich are the SRSS, absolute and algebraic combination
P00R OR GINAL 3
ethods.
All ;rograms in use t; aj include at least tne SRSS conaination, as pres: rice; in :ne ansaer above.
dou is it determined anetner an applicant's co puter code is a:r.e ot a ol e?
Tne ::ecoanical Encineering 3rarch of tr.e *D.C nas independently ceseloped a set of piping benchnark nroblens and solutions and has requi ed since 1979 (On a selective basis) that applicants and licensees benc' ar: their progra s against these problens. These pr ble s are
.iestribe: in detail in :.U3E3/Cl 1577, " Piping Eenchmark Proble 3,"
,v
. ez:,:er, et al. (n.u;u s t.n y,
.he pro 51 ens are designe.s.
3
.1, e a us.
i
.u4.
v.
specifical'.. to veriff toat co,p; tar progra s whicn calculate the dynamic response of pi?in; systeis for by tne Response Spectrun "etrod conforn ui tc the reco - <endations in F.egulatory 3; ice 1.92.
inis serificatic-e equire ent is also p-oposed to be in:luded in the upconing revision of l
Jtandar: F.e vi e.: Pian, Section 3.9.1.
i G.
Has toe Allens Creek Appiitant verified i'.s conputer progral i
l against tne :E3 pr:olens?
A.
Ter Aoplicant has stated tnat the various co puter prograns
.c.icn uiil be used for piping analysis have been bencrnarked against l
l l
ciner proge 5, and also hand calculations. Ho.;ever, to conforn with tne s
l
\\
l present verification require ent, it has also co7mitted to bench' ark 1
t.tlese piping prograns against the Ik...s problens.
.il Consist of. tre I n i s
.C Applicant subnitting conplete solutions to these problens (in the forn Of actual computer print-out or equivalent) and the solutions will have to correspond to tne independen.ly calculatud solutions.
If close agreenent I
e<ists between the solutions, the ca puter prograns will te considered
P00R ORIGINA
+
vorified, at least for response spectrun dynanic analysis, and will be so stated in the SER, publisned prior to issuance of the Operating 1.icense.
Q.
Does the NRC Staff du anything else to verify an applicant's piping analysis and design?
Since 1979 the ME3 has also perforaed on selected plants confirnatory analyses of typical conplex piping systens, such as the ResidJal tieat Renoval Systen.
Tnese analyses are perforned independently by the I;RC or one of its contractors.
The ocjective of these analyses is to verify inde?codently t'16t the Applicant nas perforned t1e correct analysis and has satisfied AS!E Code (or any other) requirenents.
Although these analyses 3re perfceled on a limited basis (one or t.vo per plait) they nevertheless provide so e neasure of assurance that piping analysis and design are adequate for the intended service.
11e intend to perfom such an analysis for the AC"GS and report the results prior to issuance of the Operating License.
Q.
Do you agree with tnis contention that the ACNGS conouter progran used to calculate the stresses on piping systens is defective?
A.
10.
As indicated above, the Applicant is using a conputer progran that conforms to the reconnendations set forth in Regulatory Guide 1.92.
Inis assures that the progran conbines response conponents in an acceptable nanner and has been so verified through conparison with other prograns and hand calculations. The NRC Staff will inde?endently f
verify the results of the conpJter progran prior to issuance of the 1
Operating License.
I i
t
DR.11. HARTZi:A!!
Professional Qualifications ity nane is ' lark Hartz,an.
I an a Principal flechanical Engineer on essignnent with the !!echanical Engineering 3 ranch of the Division of Engineering, Office of Nuclear Reactor Regulation.
In this capacity one of ny najor responsibilities is the review and assessnent of piping analysis nethodology.
I received the degree of Bi1E from the City College of New York in 1959, the degree of I!S 1E fron the University of Washington in 1963, and a Pn.D in Engineering liechanics fron the University of California at Davis in 1970.
I have also attended courses and seninars in ny specialty at the University of California at Berkeley and at tne fiass. Institute of Technology.
I have 22 years of experience in static and dynanic analysis of structures in the aerospace and nuclear fields.
This experience was acquired while enployed at The Boeing Co. and the Lawrence Livernore Laboratory.
I have been with the liechanical Engineering Branch since joining the iiRC in 1975 on a pernanent basis.
I an the technical nanager of a research project being conducted at the Brookhaven Natinnal Laboratory to ansess the quality of piping conputer prograns used in the nuclear industry.
I an also a nenber of a working group charged with the revision of faulted conditions allowables in Section III of the AS*1E BSPV Code.
l l
'.