ML18136A223

Submits Followup Questions from Subcommittee on Nuclear Regulation 790316 Hearing Re Shutdown of Five Nuclear Power Plants Due to Error in Seismic Design Analysis.Requests Comment

ML18136A223
Person / Time
Site:	Beaver Valley, Surry, Maine Yankee, FitzPatrick
Issue date:	06/26/1979
From:	Hart G SENATE, ENVIRONMENT & PUBLIC WORKS
To:	Hendrie J NRC COMMISSION (OCM)
Shared Package
ML18136A222	List: ML18136A221 ML18136A223
References
NUDOCS 7911270208
Download: ML18136A223 (63)

Text

. *~...

t 1

• }

• ~*-~

--~-- ---*

e RESPONSES TO QUESTIONS IN JUNE 26, 1979 LETTER FROM SENATOR GARY HART

• -*--*-*~* -------- --------*--*--

e

---

~--..

1-',.

e e

QUESTION l:

ANSWER:

When perfonning cost/benefit analyses of alternatives in NEPA reviews, how does NRC factor into those analyses costs such as those entailed in shutdowns (whether voluntary or by order or license conditions) of reactors because of safety concerns?

The cos~ associated with unscheduled shutdowns, whether voluntary or by order or license conditions, is factored into NRC's cost/benefit analyses through the forced outage rates. For generic purposes planned outage rates (POR) of 12%

to 15% and forced outage*rates (FOR), including shutdown to remedy safety con-cerns, of 9% to,~r 1re representative for nuclear units.

POR of 10% to 12%

and FOR of 10% tt* ~4;; are **:,';,:--esentative of large coc.1 units with sulfur removal equipment. These are equivalent to about a 75% to 80% availability factor for nuclear and about 76% to 81% availability factor for coal units. Because of distribution system reliability and other considerations, capacity factors are generally a few percentage points less than the availability factors. Thus a capacity factor of about 60% is reasonable for comparing the economics of coal and nuclear. This is consistent with the historical capacity factor for large base loaded coal and nuclear plants.

Generally, the unit costs of electricity generation for nuclear and coal in NRC's NEPA reviews are calculated for a range of capacity factors.

Figure 1-1 shows the sensitivity of generation cost as a function of capacity factor for both coal and nuclear in the New England and North Central (MT, ND, SD, WY, CO and UT) regions. These two regions represent the extremes for the contiguous United States. The capacity factor at which the cost of generation is equal for coal and nuclear is about 60% in the North Central region and 40% in the New England region.

J:. s:

.,t -

u,

..I

..J

i:

J:. ;:

.,t ui

..I..,

§ 10:l so 60 40 20 e

NEW ENGLAND

\\ ' ' '

e-COAL NUCLEAR o.__ ____...._ ____ _._ _ __.'---_...__. ___..._ ____ _._ ____ _

20 40 50 60 65 80 100 CAPACITY FACTOR%

100 so 60 40

• -~..... -

~'~--- COAL

~NUCLEAR NORTH CENTRAL 20 0.__ ____

2_._0 _____

4_._o __

_.s_o ____

s'-o-ss..._ _____

a_o ____,... 0-0-----

cAPAcJTY FACTOR%

FICiURE 1-l TI,e EH.*cts of Capa::ity Factor on Unit Con (lnilizl Vear of O;:,cr~tion

• 19!)0 Mill~/kWhl

• ... { *--------------- *- *-*-*..

QUESTION 2:

How has NRC assured that the codes being used in the reanalysis of seismic design produce valid results?

ANSWER:

The Office of Nuclear Reactor Regulation has instituted a code verification and confinnatory analysis program whereby the licensees and/or their contractors were required to solve a set of piping benchmark problems devised by the NRC staff. These problems consist of representative piping structures of varying complexity subjected to seismic loading, for which solutions were detennined in-dependently by an NRC consultant, the Brookhaven National Laboratory. The licensee-generated solutions have been compared with the bencl'lnark solutions and acceptable agreement has been found between them.

In addition to the benchmark problems, the licensees also provided to the NRC a representative piping problem from each of the affected plants, together with their corresponding solutions. These problems were in turn solved indepen-dently_by the NRC consultant, who confinned (by comparison of the solutions) that the licensees' results were correct. This constituted the confinnatory analysis portion of the program.

As a preliminary step to the analysis program described above, the NRC staff has also reviewed the FORTRAN code listings of portions of the codes used for re-analysis and has confinned that the analytical algorithms as progranmed in these codes conform to presently acceptable methods of seismic analysis of piping structures.

These three steps {i.e. licensee verification analysis, independent confinnatory analysis, and code listing review) provide reasonable assurance-that the codes used for reanalysis provide valid results.

e*

QUESTION 3:

What steps have been taken to assure other computer codes currently being used for reactor designs do not contain errQrs?

ANSWER:

The code verification and confinnatory analysis program described in the res-ponse to Question 2 is being extended and applied to codes used for seismic analysis of piping structures by other licensees and their contractors. ;n addition, a previously instituted research program at the BNL for generating benchmark problems and solutions is also being extended to obtain benchmarks for a broad variety of codes, by both analytical and exp~r~men~a1 ~echniques.

The use of benchmark pl!Oblems and solutions for code ver1f1cat1on 1s des-cribed in item (b) below.

Although computer codes used in the analysis of structures and s~ste'!ls.other than piping are not spec~fically reviewed by the staff, the appl1cab1l1ty and validity of*these computer programs have been demonstrated by one of the following criteria or procedures.

(a) The computer program is a recognized program in the public domain, and has had sufficient history of use to justify its applicability and validity without further demonstration. The dated program ver-sion that will be used, the software or operating system, and the computer hardware configuration must be specified to be accepted by virtue of its history of use.

(b) The computer program's solutions to a series of test problems, with accepted results, have been demonstrated to be substantially identi-cal to those obtained by a similar, independently written program in the public domain.

The test problems should be demonstrated to be similar to or within the range of applicability for the problems analy-zed by the computer program to justify acceptance of the program.

(c) *The program's solutions to a series of test problems are substantially identical to those obtained by hand calculations or from accepted ex-perimental test or analytical results published in a technical litera-ture. The test problems should be demonstrated to be similar to the problems analyzed to justify acceptable of the program.

I

• (

~

QUESTION 4:

ANSWER:

e e

Please list each reactor which has been found since March 13, including the five reactors which were the subject of the hearing, to have had an error in the seismic analyses of plant design.

In your response, please include:

(a) whether the reactor was shutdown because of theerror; (b) whether the shutdown was voluntary or by order; (c) the systems involved; (d) whether the systems are safety related or non-safety related; and (e) the resulting corrective measure if any.

At the time of the original safety review of the plants in question, specific NRC {then AEC~ guidance on acceptable methods for combining seismic forces did not exist. Nuclear industry practice to combine seismic forces for piping systems varied; some design organizations used algebraic sulllilation, others used square root sum of the squares (SRSS) and others used absolute surmnation methods.

It has thus developed that a number of plants were designed using analysis techniques, which were accepted practice to a portion of the nuclear industry at the time (i.e., were state of the art) and are clearly unacceptable today.

In December 1974, when Regulatory Guide 1.92, "Combining Modal Responses and Spatial Components In Seismic Response Analysis", was issued providing specific guidance on acceptable methods, the staff did not review earlier plants to deter-mine if un~cceptable methods had been employed.

Our efforts to reevaluate the seismic analyses and design of piping systems have been directed at only safety related systems since these are the systems which are of importance to assure the protection of the public health and safety. The

.list of the plants which have been found thus far to have used the algebraic su11111ation technique for the combination of codirectional responses to multiple earthquake input components is contained in the accompanying table, including

-whether or not the reactor was shut down, whether the shut down was*voluntary or by order, a general description of the system involved, and any corrective measures.

• TABLE FOR RESPONSE 4 (page l of.3) (as of,_.10/5/79)

~,

PLANT SHUTDOWN EXTENT OF SYSTEMS i

REQUIRED ANALYZED US ING ALGEBRAIC SUMMATION

~-

ORDER

• OTHER i

TECHNIQUE CORRECTIVE MEASURES I

Beaver Valley l Yes Extensive Complete and Order terminated 8/f!/79 i

\\.

Brunswick 1.2 No Voluntary Extensive Complete I

' i i

I Cook 1. 2 No No Main Reactor Coolant Loop Complete and some lines inside e l containment Complete Cooper -

No No SRV lines Fitzpatrick Yes Extensive Complete and Order terminated 8/14/79. *

• i Ginna No No Main Steam and RHR lines Complete Indian Point 2 No No 10 Lines Completed by licensee. Staff. SER _in preparation.

l i

Indian Point 3 No No Extensive Shufgown for retuf}in~

All work* to1 be comgleted

~lt 1c~nse~ & s a S ~ ~r1tt6n pr1or to st rt up.

1ma es art up - m1 ecem er.

Maine Yankee Yes 19lines (Initially thought Complete and Order terminated 5/24/79 e:*

to be extensive)

Millstone l No No 2 systems (Control Rod Complete Drive Exhaust and cu.a Bypass)

Millstone 2 6 systems (Volume Control No No Complete Tank Changing Bypass, I

Nitrogen Addition, Charging.

I Diesel Generator Exhaust, RCP Top Root Valve Instru-ment, SI and Containment Spray Test Line) i I

TABLE FOR RESPOHSE 4 (eage 2 of 3}

(as of 10/5/79)

~-

PLANT SHUTDOWN EXTENT OF SYSTEMS REQUIRED ANALYZED USING ORDER OTHER ALGEBRAIC SUMMATION TE CUN I QUE CORRECTIVE !MEASURES

. 0 Nine Mile Pt. l No No 7 systems (Reactor Recir-Complete culation, Shutdown Cooling, Emergency Condenser Returns, Reactor Cleanup, Reactor Drain, Reactor Feedwater CRD).

Pilgrim l No Recirculation and Main Complete Tech Spec..

e Steam 1 ines Pt.Beach 1~2 No No 2 CCW and 2 SW lines

_Complete in radwaste system Robinson 2 Ho No Main Reactor Coolant Complete Loop Salem l No Immediate Extensive Reanalyses and implementation of modifications in Action Letti r progress.

I Surry 1,2 Yes Extensive Order permittin9 operation of Surry 1 issued 8/22/79.

I Surry 2 shutdown for steam generator repair

• Turkey Pt. 3,4 No

- No Main Reactor Coolant Complete Loop Zion 1, 2 NO No.

. Main Reactor Coolant Loop Complete

• During the a gebraic sum revi~w, the licensee 1denttf1ed 11a,s built'! problems

. With a numbe ~ of. snul~bcrs.

Tech i pees rcqui red pl ant shutdown unde,. these.

conditions.

e I

' i

TABLE FOR RESPONSE 4 (page 3 of 3)

(as of l0/5/79).

.),;

---

~___,..,...........,........,..._................ ~---r-,_...,... ________________________________,

- i PLANT ' 1 (Under*construction)

EXTENT OF SYSTEMS ANALYZED USING.

ALGEBRAIC SUMMATION TECHNIQUES CORRECTIVE MEASURES

• }

I

---

t-----------------+-------------~-----'

Salem 2 Forked R1 ver WNP 1. 4 Extensive (Reactor Coolant System excluded)

Containment Spray ASME Code Class 1 Reactor Coolant System Branch Lines I(

~:.

Reanalyses and implementation of any required modifications prior to criticality..

1 Reanalyses and implementation of any required modifications prior to receipt of*

operating license.

~

Reanalyses and implementation of any required modifications prior to receipt of operating license.

l* *,

ia:'

..,;* I

e I

'I QUESTION 5:

(a)* What technical standards/methods are being used to detennine the a~equacy of design seismic events - those existing at the t1me the 5 plants were licensed or those in existance ANSWER:

• at the time? If the fonner, please describe:

(b) The differences.

(c)

The rationale for not applying modern standards, and

{d) A_brief assessment of the relation between the existing seis-m1c designs for the 5 plants and the existing standards.

The analytical methods.used in the reassessment of the three soil supported pl ants (Surry l & 2 and Beaver Valley) were the same standards used to assess plants applying for a license today, but the seismic inputs and other accep-

• , tance criteria were those approved in the Final Safety Analysis Reports (see attached Table). The response of the structure {buildings) to an earthquake in the original analytical method was overly conservative, therefore current and more realistic techniques were used to model soil-structure interactions.

The seismic inputs, which included t'he original design earthquake and the assoc-

~ated damping values for structures and piping systems, were analyzed and compared to an analysis which used current design earthquake and the corresponding damping values. This comparison showed that the response of the structure and its equip-ment were essentially the same.

The original damping value used for the piping is less than that required today resulting in a higher seismic load on the piping.

Therefore, the original design earthquake together with the originally assigned damping values for structures and piping systems is acceptable when compared to that which-is required today.

Based on the assessment of the three soil-supported plant design earthquakes and their damping values, the original design earthquakes and damping values were detennined to be adequate and conservative in comparision with those which would be used today.

• The other two plants (Fitzpatrick and Maine Yankee) are founded on bedrock.

The reanalysis of these plants was limited to a reanalysis of the piping systems and did not include a reevaluation of building structures response to earthquakes.

However, in both cases the NRC staff reviewed the adequacy of the original design earthquake and structural damping values and detennined that the seismic input to the piping reanalysis was acceptabl_e.

The results of the Systematic Evaluation Program':s seismic review will be used as a basis for further seismic analysis at Maine Yankee.

-~*-* r

. TABLE FOR RESPONSE 5 TECI-INICAL DATA CURRENT 1

SURRY MA*INE BEAVER UNIT:*

DESIGU FITZPATRICK 1 & 2 YANKEE VALLEv**1, PRACTICE ORGINAL DESrnl'I ORIGINAi nF<::T~rJ

'ORIGINAL DESIGN ORIGINAi nF<.Jr.N i.**

EARTHQUAKE:.

REGULATORY OBE GUIDE 1.60

.07 g

.05 g

.08 g

.06 g DBE

.15 g

.IO g

.15 g

• . 125 g -

VERTICAL 2/3 HORIZ.

2/3 HORIZ.

2/3 HORIZ.

2/3 HORIZ.

COMPONENTS 2

2 2

2 DAMPING:

REGULATORY STRUCTURES GUIDE a'.'61 "*

CONCRETE.Jt..!!:.

CONCRE!~

..L OBE 5.,.

2 0/.

2 °/o 2°/o 1°/o 2.,.

5°/o o*aE' 10.,.

5.,.

5 °lo 30/o l~/o 2.,.

1 *1.

PIPING OBE 0.5.,.

1.0.,.

• 0.6 °/o O. 6 °lo DBE I.O */.

• 2.0., **

I.O 0/o 1.0 °/o I

~

I COMPUTER IN GENERAL ALL

PROGRAMS SAFETY RELATED ALL > 6 11 ALL > 6 11 II

> 6"

'. USED FOR SEISMIC CATEGORY l SOME< 6 11 SOME< 6 11 ALL*>&

ALL I

& ASME SECTION III

• 1 PIPE DIAMETER:

PIPING ARE ANALYZED I

USING COMPUTER CODES

• Verification done

• 0.5/ 1.0 for

• Steel frame

• Total 1011*

with 0.6 °lo welded 1teel bolted/ riveted containment

( Ref. 1 S0l1mic low-1trea1

,tructure oe,ign Rovlew piping

• • welded 1y1t1m Report)

I b:,tween rigid

support, l

I \\;

... *.--..... _ ~-----**-*- ***-... *_.... --..,....

~ 'QUESTION 6:

(a)

. e How do the perceived risks associated wl!h the error in the seismic design of the 5 plants compare with those associated with the Babcock and Wilcox plants during the first five weeks following the accident at Three Mile Island?

(b) What factors led to the shutdown of all of the former within

• a few days of learning of the shortcomings, while.some Babcock and Wilcox plants never were shutdown?

ANSWER:

At the time the decision was made to require irm1ediate shutdown of five plants for seismic reasons, the perceived risk was as high or higher than in the case of the other B&W plants af~er the Three Mile Island accident, some of which were allowed to continue operation while modifications were made.

• -*-~

The error in seismic design appeared significant in that a single event (of a seismic nature) could damage the integrity of the reactor coolant system there-by causing a LOCA and also preclude operation of the ECCS which is designed to protect against the LOCA.

An accident and the disabling of required protective systems could occur as a result of not meeting a fundamental design criteria.

On the basis of early recalculations by Stone and Webster for the Beaver Valley facility. it appeared that the problem was widespread.

In the judgment of the Director of Nuclear Reactor Regulation, the problem was significant enough to reco1t111end shutting down the affected units.

On the other hand, a single event. including a seismic one, was not known to endanger the safe operation of a B&W plant after Three Mile Island. It is true that a B&W plant did experience a real problem while a seismic event is h.YPO-thetical. but a TMI type of event was the result of several _actions occurring in a particular sequence. A repeat or a similar event was judged unlikely in the very short term.

The confidence that there was no undue risk in the short tenn (few weeks) from B&W reactors while additional modifications were made included:

l. The high state of readiness and training of operators to cope with feedwater transients as a result of bulletins which were issued shortly after the TMI accident.

2.

The lowered likelihood of relief or safety valves lifting on feedwater transients because of the reduc~d scram pressure setting and higher power operated relief valve setting reconmended by B&W and required by the NRC.

3.

The low likelihood of failure *of initiation of auxiliary feedwa,ter *

4.

Evaluations performed by B&W which were stated to show prediction of the lMI voiding sequence and good cooling for several analyzed tran-sients with failure of feedwater where high pressure safety injection systems would need to be relied on.

(Note: Although Commissione: ~radford agreeds at the times that thes7 and other.specific modifications were prudent and provided a considerably enhanced level of assurnace he rese:ved fina~ judg~ent until the completion of the 5then-ongoing generic.review of feedwater transients in B&W rea~tor and plant systems: Following completion of that review,.staff recommendat10ns resulted in the temporary shu~d~wn ~f a 11 other B&W nuclear power pl ants for additional modif1cat1ons.)

"II I

(

• -**-*~-

e e

QUESTION 7a:

What are the recurrence frequency and magnitude of the design b~sis and operating basis earthquakes at each of the 5 plants?

ANSWER:

The Design Basis Earthq*uake or the Safe Shutdown Earthquake and the Operating Basis Earthquake (OBE) are defined in detail in Appendix A to 10 CFR Part 100.

The design requirements for the OBE are such that the plant structures, sys-tems and components necessary for continued operation, without undue risk to the health and safety of the public, are designed to remain functional.

In the event of the occurrence of an earthquake, up to and including the QBE level, no regulatory action would be required. If the QBE level were to be exceeded, NRC Regulations require plant shutdown.

Prior to resuming operation following this shutdown the licensee wculd be required to demonstrate to the Commission that no functional damage h~s occurred to those plant features necessary for continued operation without undue risk to the health and safety of the public.

For earthquakes up to and including the SSE, it is required that the structures, systems and components necessary to assure integrity of the reactor coolant pressure boundary, the capability to shutdown the reactor and maintain a safe shutdown condition and the capability to prevent or mitigate accidents leading to unacceptable offsite exposure all be designed to remain functional. There-fore, the SSE, not the OBE, is the important earthquake level on which to focus attention from the standpoint of safety in the evaluation of the capability of a plant to withstand a seismic event and safe shutdown.

In his testimony of March 27, 1979, Mr. Denton presented some estimates of recurrence frequency of the Design and Operating Basis Earthquakes*at the five plants.

He indicated that the Design Basis (Safe Shutdown) Earthquake had a chance of being exceeded at each of the four sites that was of the order of 10-3 to 10-4 per year.

He also indicated that the chance_ of the Operating Basis Earthquake being exceeded was roughly estimated to be on the order of five times that of the Design Basis Earthquake.

These numbers were based upon pr!:!vious estimates of earthquake ground motion exceeding a given peak acceleration at various locations throughout the eastern United States. Because of the lower design acceleration and higher local seis-micity, the Maine Yankee site appeared to be at the higher end of the r_isk of exceedance range.

In these estimates no attempt was made to expand upon the applicability of or the uncertainty associated with these values.

In response to your question we will supplement Mr. Denton's original testimony with a dis-cussion of these factors and provide an updated and more site specific estimate of recurrence frequencies.

It should be pointed out first that probabilistic estimates of earthquake hazard (recurruence frequency} were not used in defining the original earthquake re-*

sistant design at the five plants. Those numbers presented previously by.

Mr. Denton and in this response represent _g_ oosteriori estimates of exce7d1~g original design ground motion parameters which were chosen in a detennin1st1c manner.

While probabilistic estimates of seismic hazard can be made, insight and great care must be exercised in utilizing these estimates in the decision making pro-cess.. Our experience indicates that absolute estimates of these hazards for a site can vary by more than an order of magnitude, depending upon the choice of

e.,

f I

input parameters and assumptions. The choice of parameters and assumptions will vary among expert seismologists. A thorough estimate of seismic hazard should systematically include these varying opinions and should account for the related uncertainty. This type of estimate could require a lengthy re-search program which is at or possibly beyond the state-of-the-art.

In order to answer your question at this time we can only examine those readily available studies that have included the different site regi~ns in their estimates of seismic hazard and. by interpolation and extrapolation, provide gross ranges of return periods (recurrence intervals) for the different design and operating basis earthquakes (see Table 7-1). These studies include those perfonned by individual members of the U.S. Geological Survey, the Canadian Department of Energy, Mines and Resources, the Applied Technology Council and other seismologists.

The most important data base upon which all of these estimates ultimately

• rest is the historic (non-instrumental) recorj of the feit effects of earth-quakes.

Converting these felt effects (earthquake intensity) into instrumen-tally determined earthquake magnitudes or ground accelerations that may be useful in design is itself a complex and often controversial task. This is in great part due to the shortage of appropriately measured earthquake motion in the eastern United States. The magnitude scale utilized in the table below is that developed by Professor Otto Nuttli of St. Louis University and is roughly equivalent to the Richter Magnitude {developed for California earth-quakes) in the magnitude range of interest.

The return periods listed below are for earthquake ground motions used in the design of the five plants. They do not represent return periods for exceeding s.. ""uctural design limits or for failure of any plant component.

The ground

.. ~ion at each site is specified in terms of two parameters:

l) Peak Ground Acceleration (PGA) - the most comnon description of earthquake ground motion.

This is the parameter used in most of the studies and can, therefore, be detennined directly.

In our March 27, 1979 sutrnittal to this corrmittee our initial judgement of earthquake recurrence was based solely upon chances of exceeding the peak ground acceleration.

2)

Response Spectrum (RS) - a method of characterizing the variation of level of ground motion as a function of frequency. It can be shown that at very short periods (high frequency) the value of the response spectrum is about the same as the peak ground acceleration. The spectra used for design are usually standardized shapes developed from studies of actual earthquakes.

Over the years, as the number of earthquakes recorded has increased, these standardized shapes have changed.

When the 5 plants were originally de-signed, the response spectrum shape used at that time was different than that which our present regulatory guides specify.

In relating earthquake response spectra to either intensity, magnitude or peak acceleration, we have assumed that the level of ground motion indicated in the current re-gulatory guide spectrum is appropriate. As a result the estimated return periods associated with the response spectra shown on Table 7-1 differ from those relying solely upon peak acceleration.

~ -- :... -* -~ ---** **: *--* -*-

e TABLE 7-1 Estimated Magnitudes and Return Periodsl for Design Earthquakes

. Corresponding to the FSAR Peak Ground Acceleration (PGA) and FSAR.. Response-Spectra (RS) for the Four Reactor Sites (From Extrapolations and_ Interpolations of Readily Available Studies)

DESIGN BASIS EARTHQUAKE (SSE)

OPERATING BASIS EARTHQUAKE Plant Site SurCY Magnitude roughly 5.3 roughly4.8 Return Period (yrs) greater than 2500 yrs roughly 500 to 2500 Magnitude roughly 4.8 roughly 4.4 Return Period (yr~

roughly 500 to 2500 roughly 80 to 600 PGA2 RS Beaver Valley*

PGA RS

• NOTE -

roughly 5.2 roughly 5.0 roughly 1000 to 10,000 roughly 4.8 roughly 800 to 7000 roughly 4.6 roughly 250 to 2500 roughly 150 to 1500 the Beaver Valley site has soil conditions which,when properly accounted for, could amplify the ground surface accelerations resulting from bed-rock motion.

This could lower the magnitude of the design earthquake and lead therefore to shortened return periods.

Fitzpatrick PGA roughly 5. 3 greater than 1000 roughly 4.9 roughly 400 to 6000 RS roughly 4. 8

• roughly 300 to 4000 roughly 4.3 roughly 100 to 800 Maine Yankee PGA roughly 5. 0 roughly 250 to 3500 roughly 4.5 roughly 50 to 700 RS roughly 4. 5 roughly 50 to 700 roughly 4. 0 roughly 20 to 100

1.

Return.)e.riods are estimated u.~e1~age recurrence inLr*>lls.-a,* ~.1r,:i~q*~,:{.:S of given size or greater over extrcm1?ly long time intcrv.:11s (,::.::ny tir;:cs -

• ,e length of the r~turn_p~riod).

In no w:.y are*thcy r.:~~nt to predict tile actua 1 occurrence of* an earthquake in a given year b~t r~ i:.h::i" the u.vcru!Je ch;;i*lce of its happ::::d ng.

Si1:1il arly th~ 1 imits of *;_:i:sc ri~turn periods -arc not meant to de 1:ote strict bour.dari es in \\*:hi ch a 11 pres ::nt or future estimates will be contained.

They arc simply the bread band of return periods determined from the interpolation and extrapolation of those studies ~xamined.

It should be emphasized thJt these return periods refer to 1;arthq11akc occurr~nce alone ;;nd do not refer to <"iS~ 1::... ::d rr::sponse of the

,i:Jing or sti**.1ct*.Jrcs.

They pro*1ide lo~*u:,* bounds for th.:?. r:;t*.;rn ~~rices i"ar response spectra ~hich could c~use calcul.:1tcd stresses in th2 piping to ~pproach these Villucs calculated in the original seismic analys~s. Considcraticn of parameters such as the increase in allow.:1ble dar;;pings for structures and piping, Jnd soil/structure. interaction (where.ap~licablc) ~ould tend.to provide an increase in this return period r~lative to th~ pipe st::ss c~lc~latcd for dcs~:n.

Extensive.:in,1lyses \\':ould be rcG1Jire:t1 to.pro*1idc ;;n,1CC'.~i".::tc ;;stii:1ai:c of this increase for all of these plants~

!fo*.-:c*,cr, for :k.'.!*1er* '.'iill~y t1nd St;lTJ, 0,1.se;d upon the *CDi:'lpariscn of the soi 1/structure interacticn r,:;:;-.~l.:;scs using c:.;rr;:;nt and the licensed spectra and da~pings~ the return peric~s for earth~uakes.with

.........*.. _*~:---:--

__ -----;:-----------~----

e e spectra for which cal~ulatcd stresses in the piping would ~p~ra~ch th~ir values calculated in their reanalyses would tend to be rcushly in t~c ranse of these given far the peak ground acceleration. rather than these given for the response spectra.

2.

PGA refers to peak ground acceleration and RS refers to response spectrum *

~ --......

I

~-*

-:.:.. *-................ - - - ~--**-*-*- - *--~.

QUESTION 7b:

e e

Based on the reanalyses using acceptable procedures, what are the recurrence frequency and magnitude of the earthquake that would have resulted in stresses above the allowable limit prior to any plant modifications.

ANSWER:

The review effort for the safety related piping systems on the 5 plants was focused at determining the adequacy of the systems to resist the specified earthquake design criteria and to implement any required modifications. It was not directed at determining the earthquake level at which the systems as built would reach their allowable stress limits prior* to modification.

The determinat1on t,,at ti,.: :itresses in a piping syst.em are within allowable limits for the specified design criteria requires not only an evaluation of the stresses within the pipe itself, but also support and nozzle loads and their resulting stresses. Additionally, the seismic load is considered in combination with other loads which also produce stresses in piping supports and nozzles. This further complicates the estimation.of the earthquake level at which the allowable stresses would have been reached in the unmodified con-dition.

In addition, the seismic analyses of piping systems do not predict the exact stress le~els in the piping under seismic levels. They merely pro-vide stress magnitudes for design purposes. It is impossible to uniquely charac-terize the nature of the ground motion at a site as a function of earthquake magnitude and to predict exactly the seismic responses of piping systems.

Using SSE design parameters and acceptance criteria {spectra, damping, allo-wable stress limits, etc.), the earthquake peak ground acceleration level at which allowable stress limits would have been reached in the as-built piping systems may be estimated from the infonnation we have to date for Beaver Valley Unit l.* For Surry Units land 2, Fitzpatrick and Maine Yankee, we do not pos-sess sufficient information regarding the stress levels predicted in the un-modified piping systems to make such an estimate.

For Beaver Valley Unit l, given the new response spectra based on soil/struc-ture interaction considerations, reanalysis results to date indicate that six pipe supports require modification, three snubbers must be added and at least one branch connection reinforced in order to bring all pipe stresses, support, and nozzle loads within their respective SSE allowable limits. However, many other supports could not be found acceptable until the SSE seismic anchor movement load, originally included, was removed in accordance with current ASME Code criteria. Several snubbers also could not meet original design

-... ~

e e

'II, t

I I' criteria and have been found acceptable after reevaluation of their capacity.

Without.knowing specific magnitudes of overstress or overload conditions, and given the acceptability of removing the SSE seismic anchor movement load and the one time snubber loadings, about 95% of all calculated stresses, support and nozzle loads would remain within their allowable SSE criteria for a ground acceleration of 0.125g. Utilizing the extrapolations and in-terpolations discussed in the response to 7a, this would roughly correspond to an earthquake of magnitude 5.0 and would have a recurrence interval on the order of thousands of years. The other 5% would require a substantial reduction, to possibly as low as 0.05g, before they could meet their allowable limits. Utilizing the extrapolations and interpolations dis-cussed in the response to 7a, this would roughly correspond to an earthquake of magriitude*4.6 and would have a recurrence interval 9n the order of hundreds of years or more.

The same caveats discussed in 7a would also apply to these rough estimates.

e e

QUESTION 8: What are the estimated costs of the shutdowns of the 5 plants in tenns of dollars and barrels of oil? The underlying as-sumptions should be stated.

ANSWER:

URC did not estimate a cost for the shutdown of Surry II, because it was al-ready shut down at the time the error in seismic analysis was found.

For the other plants,. the operating utilities were contacted to deteniline cost impacts. These costs agree with NRC calculations when the same assumptions are used.

Beaver Valley The replacement power for the 852 MWe unit is supplied by burning coal.

• Assum-ing a capacity factor of 74%, the monthly costs are $5.l million for coal (11.23 mills/kwh), $0.8 million for purchase of power (1.76 mills/kwh) and

$0.5 million (1.1 mills/kwh) for increased cost of non-fuel operation and main-tenance.

A savings of nuclear fuel cost is abbut $1.7 million per month (3.8 mills per kwh) leaving a net cost* of $4.7 million per month or $160,000 per day.

Maine Yankee Oil is burned for replacement power at $16 per barrel (27 mills per kwh) com-pared to the nuclear cost of 3.3 mills per kwh.

At a net capacity rating of 830 megawatts and a monthly capacity factor of 95%, 28,000 barrels of oil per day would be required and the net cost* of replacement power would be about*

$450,000 per day.

Surry l The replacement power is supplied by burning oil at $18 per barrel (30 mills/

kwh) at a net capacity rating of 822 megawatts and a monthly capacity factor of 75%, 23,000 barrels of oil per day would be reqµired and the net cost* of replacement power would be about $340,000 per day.

Fitzpatrick The replacement power is provided by burning oil at about $16 per barrel (27 mills/kwh).

At a net capacity rating of 821 megawatts and at a 75% capacity factor, the net cost* of replacement ppwer is about $330,000 per day and re-quires about 24,000 barrels of oil per day.

• The net cost is the cost of oil or coal minus the cost of nuclear fuel not consumed.

e f

e QUESTION 9:

ANSWER:

In the March 16 hearing. Mr. Denton remarked that much credit for bringing the computer error to his attention goes to the diligence of an NRC inspector who pursued the discrepancy in the results of the old and*new-codes.

Please provide the particulars in a chronology of the surfacing of the discrepancy and an asses-ment of the reasons for any delays *.

As stated previously by Mr. Denton and reaffirmed herein, the NRC Inspector deserves much credit for actively pursuing with Duquesne Light Company and Stone and Webster Engineering the problem in pipe stress analysis.

1"\\s the enclosed chronology indicates. there was persistent NRC staff effort to ob-t~in information that would accuretely define the safety issues so that appropriate actions could be taker,.

An assessment of the potential safety significance of the problem was considered throughout the fact finding pro-cess. The staff moved in a manner consistent with the safety significance perceived at the time based on the information provided to the NRC.

When the

~1use of the _discrepancy in the results of stress analyses was identified to

_.-,e NRC staff. prompt action was ta ken that led to the issuance of the Show Cause Orders.

Attachment:

Chronology Table

)

I

~

'~

I

'I

• \\..

10/25/78 10/27/78 10/27/7S 11/9/78

.. --. -----~..

e e

.. Prompt report L£R 78-053/DlP to -HRC Region I via teleccn fran Duquesne Light Company~ Reported infonnation received frtm Stone and Webster*th!t hand calculation errors-*resulted-i*n--stress lewels

• above ANS1B 31. l. 1957 but cnly in one case c'f six flow paths.

Daily Report by Region I to I&E headquarters included as a reportable occurrence - inadequate piping supports during review of safety inje=ion pipe stress analysis by the A/:. csiw) *. several pofats on the 6-inch and smaller piping were found tD be inadequate1y suppcrte:~ lr. th~ event of safety inje::~ion system oper!tion during a~. 5 points could exceed the code a11ow!b1e stress. A design change for safety injection piping supports wi1 l be accomp1 ished prior to uni! stal'""tup in mid-Rove:nber.

Written inte~im l£R submitted by Duquesne lig:t Company.

OlC characterized the errors repor-~ by Stene and Webster as resulting frcrn a hand c:a1cu1aticn method cf.ana1ysis.

IE Inspection 50-334/78 Region I* fo11 owp on 2~ hour report.

Inspector r!ised a mmi!::>er OT quest ions including: \\.'hat usurance can be given to show that the ca1 cu1 aticna1 error a;,pl ies cn1y to the si.x points in question? To cnly the Safety Injection system? To only the Beaver Valle-y faci1 ity?.

Second interim LER subQitted by DuqUo..sne Ligh't Ccmpany indicates that the original report*Wi?S etM)neQUs *.. Tne line stresses were thought to have been hand calculated

• only, when in fact they were subsequently ccmputer ca1cu1 ated and found acceptable-.. ou: al so indicated that *a fu11 report on the situation was in pTeparat*ion.

by Stone and Webster~

II r

I 11 11/16/78 11/30/7S 12/01/7S

• 1Z/04/7S 12/05/78

.. ~,1osna 3 -

e

• -IE Inspection S0-334ns-*33 - Region I f.'.nspectors fo11owp but no information availible cmsite.

Region I Daily Report indicated a r~re\\C'"iew by A/E found that the previously reported. condition W!S erroneous and that no inadequately sup~rted piping

.existed, a fu11 report of the situatiocr.is being prepared by the A/E and a fol 1 owup to the LER ~, 1 be submitted by the Licensee to HRC.

Fo1 lowup calls to site by the IE inspi:a-.aot attemi,ting tr, seek additiona1 infonnation.*

Fo11owup calls to site by the IE inspeC'tor attempting to se!i!k !dditicnal infcnnaticn.

Fo11owup calls to site by the IE inspe:c::tor attempting to seek additional information.

Fc11owup calls to site by the IE ins~or attempting to seek additional information.

LER 78-53/0lT-D: wes submitted to h'RC b~ licensee

• Conclusion was that "corrective action has been

. reviewed, -approved and satisfactorily ccxnpl eted*.

The rep:,r-t based on info*nnation supplied by Stene and Webster *attributes the pipe overst:-ess. to diff-erences between stresses analyzed by ?STRESS code and these done by the chart method.

I~ mentions differences between PS'i?.ESS and NUPIPE codes in fon:e sur.:nat ion but does not el aborat!! on them.

It. concludes that PS~ESS used met.ho~ acceptable for Beaver Valley Unit. 1 genei:-ation pl!znts. It st.ates that Reg. Guide 1.92 issued in iDec~ber 1974 established for facilities doc"ket:ed after..

. April 1975 mere conservative techniqu!:.S for intra-modal co.TJ)inations of generalized 1 ca::lings. The report st.ates that ana1ysi s showed th~t *only one safety injection system pipe required modification-*

the addition cf one snubber and the redesign of one support. The attach,-nerrt to this l.ER p~ovide::!

additional historical inforn.aticn as ~cl lows:

e

- *4 ~*

Duquesne Light Cocpany.reported in an attachment

. to the December 6. 197S I.ER 78-53/0lT-O that tc generate dat! needed fer inst.a11aticn cf a net positive suc:tion head modification.~ the Se!ver Valley Unit l safety injection system, they (Stone and Webster) decided to *code in" the six inch Sl 1 ines into a currently used ccrnputer pro9ram

{HUPIPE).

DLC indicated original design used the PS"TRESS code. Ho resu1ts cf an ana1,sis at this stage were reported by DLC to NRC.

Subsequent tc'the above activity the attachment states the Beaver Valley Power Station was notified by a vendor that check Y!lves in SI system were actua11y he!"ier _than used in design at construc:tion stage.

This increased weight was used as inp::t* to the l.bove NUPIPE model and found not to *affect* the piping design.

The Architect Engineer (Stone and Webster) also conc1uded that the hanger designs need not be changed as ! resu1 t cf using the co:-r-eet

{heavier) weight for these valves. However errors were said to have been discovered in the hand c:al cu-1 ation method. It was determined that piping anal,YZ-iS shewed 1 ocal overstress at several anchors but

• no overstress in *the pipe" al one.

Per attachment to LER 78-53/0lT-O, a mere thorough evaluation was initiated tD determine if "any other annulus piping* originally designed by the chart

{hand c:alc:ulaticn) method was overstressed.

Per attachment to LER iB-53/0lT-O, licensee foun~ that SI lines had been "as-built*. reviewed in 1974 an~ that two of the six lines had been (at that time) ceded into PSTRESS {not just hand calculation method).**

The PSTRESS c:ode was re-run using the correct valve weights and resulted in acceptable pipe st~esses.

A1so per attachment to LER-78-53/0lT-O, 1ic:ensee states "The models run in PSTRESS and NU?I?E ~re geometriea1ly similar; ho~ever, the mass distribtr:icn and *support stiffness are different. Further, the

• h ""

(

• d l )

• d *~z rne.. c... er, crce sur.ima.. ion, n.. ra-mo a., s

,,, ere....

h'U?!?£-i:ti1i:es more ccnser-vative tectlniques fer intra-r.:oda1 c~bi n!tions cf general i:zed 1 c.adings.

I**

• t 12/11/7~

12/12/78

. 12/14/78

........., **, JJo... --*

e

- 5 ~

e These newer techniques arose fol 1 ow; ng estab 11 shrnent cf

Beaver Valley Unit Ne. l design criteria.

In De::er..!)er.

1974. the USNP.C published Regu1ator-y.Guice 1 *.92. applicable to fac:i1 ities docketed after April. 1975. which required the use of the mere conservative eanbina.ticns. 'The PSTR£SS methods used 'ltere accepted dynamic analysis

• techniques for Be1ver Valley Unit 1 genei-ation plants, and is the.basis far all cccputerized Ca.tegory l pipe stress ana1_ysis perlcnned".

(It is HRC understanding that resu1ts were unsatisfactcey..

on two of three 1 ines, but snubber and sc;,pcrt modifications en one 1 ine reduced the. overstress en the second line such that no modifis:a.ticns on that. line we~G ne:essary.)

The pre Oece::!)er 6, 197S review of annuics seismic piping wa.s 1iQited to lines that had be~ previously analy:zed t:sing the hand cal cu1 ation methc:i (2-1/Z inch to 6 inch lines). 103 lines were id:entified, 55 were reviewed ind found acceptable.

Licensee noted that PSiRESS resu1ts were sti11 av~i1 able for 48 of the.103 lines free the 1974 as built review and were *acceptab 1 e*.

Licensee notes its Engineering Department is *continuing a review of the,rchite:t-engineer findlr:ogs 11 Follow-.2p ca1 ls to site by* the IE inspec.tcr to see~

additional infor:-..ation.

Region l IE inspector tel e;:hcned tii{R Licensing Prcje:t

~.anager tD cttain a contact for informal discussion of technical questions.

Region I Daily Report -

Furt.her review cf in-ccntain.-nen:

• SI system piping supports *identified one 1 ine *requiring support modification, attributed to an error in original design calculations.

Regional inspector wis telephoned by ~rtR i ndividua1 who wts desi or;ated ~s cor.:ar:t.

Piel i!:ii narv technica1 discussion wis he1 d at,c~ potelitil.1 prob1ems~

J: *

• 1 i211s-2ons 12/ZZ/78 1/18/19 1/23/79 About 2/2/79 2/2/79 2/5/79 3/1/79 * *

- -- -* - - ~ ~ -

~ -

~ ---.. ---

~ -

e e I£ Inspection 50-334/78 Region I fc11owu;, en 12/6 LER.

During this inspection, the insl)!tc:tor reviewed the detailed report submitted to the licensee by A/E and discussed the results of *that review with representatives of the licensee and A/E.

Region I inspector discussed with NilR i.ndividua1 s via te1ephone question$ he had as a resul~ ~f discussions he had with S&W on 12/18-20/78. The JaC individua1 s involved deter:uined that there was a pcssib1e problem.

Region r mai1 ed to I£ Headquarters*. a r:remorandum requesting that information be for.iaraed to NiiR for review.-The memo defined concerns~~ include:

1. Reconciliation of the differing ~~1ysis results to us1:re that the design methods :.ised are

. neither incorrect nor unconservative.

z. The need fer further 1 icensee review cf piping potentia11y affected by any incor.-ect or nonconservative r:a1 cu1 ation.

The IE Inspe*ctor provided copy of the 01/18/79 memorandum to ticensing Project Manage:-.

Discussion between IE inspector and Ji?.~ project manager determined' that a forrna1 transfer cf le-ad responsibility between I&E and h'RR had net been made* cf the Ol/lS/79 mesncrandurn *t:, N?.R.

A formal request fer DOR' s Engineerir:; Branch suppor-t (TAC fonn) was prepared by the proje::: manager.

IE inspector W!S informed by IE:HQ that telepbone.

discussion had established that NRR was wrking en the problem and that a fc~al transfe:- of 1 ead to tt'RR wou1d be made.

During.a conference cal1 to DU: and ~'.l, !. ccxnputer run w-as recuested for DOR review. Since S&W cci-porate policy W!S net to provide s:JJ::h proprietary*

data, a meeting w?s se~ up for S&W to bring in a ccmput~r run for DOR review at Bethesda

• 3/8/79

-:. 1:110.

~, -

J./10/79

. *- -~,..,...

e e A technical meeting W!S held between OLC. S&~. and the.NRC staff to discuss and review the PIPES'ra.£:SS and NUPIP£ cedes. The NRC approach~ the review with the be1ief that the two cedes were acceptable and that some mode1 ing or input pro~1em created the.

rl!Su1ts in question. It was re,*ealed th~t--the -------

PlPESiRESS code used an algebr!ic su:::natian af seismic lc!ds which in the absence cf a detailed time history analysis, gave unccnservative results in the seismic stresses.

Managemen: was i1m1edi ately informed and.a m!n!gement level Q!e:ing arranged with DLC and S&W.

A mana9erent 1 eve1 meeting was held with OLC and S&W tc air'an9~ for ~r..rnedi ate review *of the B~aver Va11ey pipe stress !na.lyses.

Commitments were requerted of S&l.' to identify the system$ and pl ants involved, the inadequacies expected and the reanalysis to confirm safe operation.

Ne definitive infor::ation W!S avail!b1e at that time.

DLC w!s re:uestcd to b!ve 1ts plant safety can::iittee review the si'tuation.

Numerous staff meetings were held at Bethesda t:,

scope the problem with res~ct t.o the effects if a seismic event were to occur. Tel econs were made to S&:W

.. h h d 1 f

J:.

~ *h on *,e sc e u e o comm, c.r.ien~s, or, L'i i. er,n, crma.,en on Seaver Valley.

The other utilities identified by S&W as h!ving pl ants with the S!:ite problem were.

notified. These plants and utilities were:

Fitzpatriek, Power Authority of the State of New Yon:.; Maine Yankee, ~~ine Yankee Atomic Power Cc~pany; Su:-r.y 1 and 2, Virginia Electric and PoW;r Cc::pany

• The Chairman was advised. Thre_e staff rtenbers were sent to Beston to provide ir.:nediate review and analysis of resu1 ts. DLC sent eight *pecpl e to Boston to a.ssist in expediting the review.

In view of the problems and with t.he Offsite Safe~y Review Ccrn:nittee c:on:u:-rence, the Beavei \\'al 1 ey t:tnit 1 w!s*

placed in hot standby for the ~eeke~~ by DLC to !W!it further analyses frcr:i SaW.

S d

.z t.?.i, rnee.,ngs con.,nue

!S ;,,e:es c.,r:.orna-ion were fed batk from Boston.

The !c~ ~uty Office~s we:-e advised cf actions. The t,SS.S \\'entiors*fo~ the plan~s were conta:ted to ass~re r.t ~:her co~es fc~

Jt12n9 e pipe stress during that period used 'the same a19e.braic approach. A DOR Assistant Directer W!S sent to Beston to provide management review*~nd coordination.

S&W 1 s canputer was dedicated fu11 -tiz:z to these stress calculations and extended wort hours for data reduction was instituted for S&W staff. HRc*cpticns were exp1 ored and draft materials de-tel oped to support a,:ipropriate action based o~ the technical results bect=ing available on Beave~ Va11ey.

Etrly Sal.' reanalysis results.on Beaver Valley runs indicated problems with pipes a.s well (originally thought only supports). L ir:ensees 1 * 't:)p management was*c:ontacted to assure action underw!y by all plants to idr::rtify inadequacies and c!:rtain r-eanal,YSes of stresses in a1l affected safety*sys:ems.

Additional infonnation fran DOR staff in Boston confirmed pipe stresses above allgwaf>1e and unaccept-able.

Arrangements were made to brief the. ~mrnission on this matter. A11 the licensees were notified cf a pending decision.

In view cf t~e safety significance of this matter ~s discussed above, the Director cf the Office cf Nuclear Reactor Regulation proposed tn the Co::1.Tiission that the public heal th and safety r-~i res that

• the present suspension of operation of the 'facility should be co:rtinued:

(1) until such time as the piping syste::is for all safety systems have been reanalyzed for e!rthquake events to demonstrate confonnance with General Desicn Criterion Ro. 2.

using ! piping analysis ccr:ipufer cede whic:h c5oes

• not contain the error discussed above, and {2Y if such reanalysis indicates that there are ccrnponents which deviate from appl.icab 1 e ASH£ Code requi rsnents, until such devi!ticns !re re~tified. The Co::cission concurred *;n the NRR Director's decision **

Prior to the h"RC fina1 de:ision to c:-::ier the piar,~s shutdown, the Beaver Va11 ey Offsite Safety Review Cc:::::i:t~e re:~-::.-:,ende~ the faci1ity be p1aced in cc1d shi.r:do~~ based en the data and anaiysis re:ieve:

frcc:i Sl:~.

ihe D!.C ordered -:he p1 ir.~ shir.dow.i.

.I e

- g -

. The 1 icensees confirmed by tel econ that the Dn:fers were received and provided times when each faci1 ity* wou1 d be in co1 d shutdown.

A11 fac:11 ities will be at or below 2oo*r by 10:40 p.=. on March 15,

• 1979 in ccnfcrniance with the Order.

Subsequently 111 affected 1 icensees 1e:-e no~ ifi ed

• by telephone that the Orders were executed ~d that a copy wou1 d be transmitted by facsimile.

Meetings wei-e he1 d with Stone and Webster w~:.h the Utilities to discuss acceptable Qethods cf c!!ma1ysis fer interim and long term fixes of the pipit:9 and supports.

• - ~-..... -

e e

QUESTION 10: Please provide available infonnation on the recent earthquake that occurred in the vicinity of the Maine Yankee plant.

How does it compare with the operating basis earthquake?

ANSWER:

A small earthquake (magnitude 4.0) occurred on April 17, 1979 at 9:34 p.m **

local time, near Brunswick, Maine and about 10 kilometers west of the Maine Yankee Plant site. The earthquake was felt over a broad area of New England and was recorded at many of the NRC-supported seismograph stations of the Northeastern U~S. Seismic Network.

An intensity investigation conducted by MIT suggested that the highest intensities were MM {Modified Mercall i) V.

No damageQ resulted anywhere from the earthquake. At Maine Yankee the licensee in fanned the NRC that the ear-thqu,ake wa-; fel.t..;in_the control_ room but not in the contail)lllent. According to the li~ensee, there was one operating s~rong_mot,on recorder at the s~te (trigger set at O.Olg in the vertical d1rect1on) and the earthquake did not trigger this device.

About 30 aftershocks were recorded in the first 24' hours following the earth-quake.

The largest of these aftershocks was magnitude 2.8. A plot of the apicenter locations describes a cluster of events centered approximately 10 kilometers west-northwest of the Maine Yankee site. There are no known structures in the vicinity of the earthquakes w~ich the NRC staff believes to be localizers of seismicity *

. Approximately one day after the magnitude 4.0 earthquake, personnel from the Massachusetts Institute of Technology and the Lamont Observatory installed networks of portable seismographs in the epicentral.area to record after-shocks *. They monitored for several days and recorded only one aftershock.

Two additional aftershocks occurred on May 11 and 13, 1979. Their magnitudes were measured at 2.3 and 2.7, respect,vely.

The NRC provided instrumentation for a sensitive portable sei smog*raph network in the epicentral area to attempt to detect and accurately locate any additional aftershock activity.

Weston Observatory of Boston College installed these stations about June l, 1979.

The Maine Geological Survey maintains these stations and perfonns preliminary analysis of the records, and ~eston Observa-tory performs detailed evaluations of the data.

Very-small earthquakes (about magnitude 1) were detected on June 6, 1979 {two events} ang.June 18, 1979 in the vicinity of the magnitude 4.0 earthquake. This portable network will remain in operation through July, 1979 and will operate after July only if there is additiona 1 activity.

- The NRC staff concludes that the Maine Yankee site did not experience ground motion exceeding the Operating Basis Earthquake (OBE) from the April 17, 1979 earthqu~ke because:

1) The maximum. intensities observed at the site are associated with.

ground motions less than those associated with the QBE.

2)

The earthquake did not trigger the strong motion recorder-at the site.

As noted in the answer to question 7a, the response spectra for the QBE using current NRC regulatory guide spectra approximates a magnitude 4.0 earthquake.

Given the wide range of expected acceleration levels, another earthquake of the same size as the April 17, 1979 event located near the site might equal

-,

,.. ex(-~ei:i the OBE.

I f

QUESTION 11: One of the plants ordered shut do\\lffl is the. Surry Plant which served as the. model PWR for the Reactor Safety Study (RSS).

The RSS included an extensive design adequacy study.

ANSWER TO. 11A:

(A)

What was the finding of the study team with respect to the seismic design of Surry?

(B)

What are the ramifications with respect to future quantitative ris~ assessments?

The Design Adequacy task of the Reactor Safety Study is reported in Appendix X to WASH-1400.

The met~od of combining modal inertial forces is discussed in 0 Section A6.3.3.1 of Appendix X (pp. X-47 * - X-49). *This discussion states that "The method of combining modal inertial forces.in the principal. directions to determine seismic stresses is not correct. i*

However, it was the understanding cf the Reactor Safety Study Group that the absolute value of the model forces~

were comined, rather than the algebraic values. Thus, the Reactor Safety Study concluded that the method used.

11* *** leads to conservative results.-~

11 We now know that this understanding was incorrect. The general findings of the Reactor Safety Study regarding the seismic design of the PWR analyzed (Surry) are presented on p. X-3 of Appendix X and are quoted below:

11 30 P'*!R items were examined \\*1ith regard to seismic design. Oft th~~e 2s' were found to be adequate {83~).

Design adequacy \\*J_as no d~mon~trated for*f.ive items (17%) (reactor coolant PUi:lP nozz~es, l~w hcild safety injection system instrumentation, rec, rcu! at~on 5 ra ump outside containment, the diesel generate~ day i.:~n,

a~d ~h~ AC and DC switchgea.r), because sufficient inform~t~on \\*:~~r not available to permit an assessment of adequacy t~ be t~!a_e.tion three! items ( the containment c~ane, the 1 ow. hca.d s~ i et~ l~J ~c the P1*ml)S and the reactor protect1an systc:n), 1 t \\*;as I ound 1.. a.

• c f *1 t

vpcc***ed un er se1sm1 l-i*:sisn \\*.'ilS i.idt::ic:;uate 111 :chat *a~ ure,s_no ;*:

~

d.

be lass than excitation.

Hm*1cvcr, tne mara1~ to fa1~ure \\*,as ou~ d-coualification thJt noriaa11y c;<pcctc:d considering appl1cable code QJ~

q t'

cquirGn~

;nts because either:

(l)(er)ror~ w:re fo~i~~i~~t~~~u~~s~~n!ere used in calculating stresses; or 2. se1sm1c qua, not sufficiently comprahensive or were not pcrforir.cd.

(Note:

No seismic modifications to the Surry Power Station were made as a result of the RSS conclusions.)

- - - - - - ~

e ANSWER TO 11B:

The impact of seismic design deficiencies recently identified has been es-timated to increase the risk and overall core melt probability by a factor of 3 to 4 over that. estimated in the Reactor Safety Study. *With respect to future quantitative risk assessments, this deficiency, plus analyses per-fonned by others on seismic risk potential ~uggest the following:

1. A comprehensive design adequacy review is necessary when considering the response to loadings not included in the available data base, e.g., severe seismic events.

2. A definitive need exists for improved modeling of* the seismic contri-bution to risk.

IR this regard, NRC has a large ongoing seismic research program which is in-tended to provide the infonnation needed to define the seismic risk contribution more precisely.

(J I

e e

QUESTION 12:

Please list all nuclear power plants that have been exported from the United States that were designed with the aid of the erroneous computer code involved in the five plant shutdowns.

ANSWER:

We are unaware of any nuclear*poweY:.plants. exported from the United States that were designed with the aid of *the computer code involved in the five plant shutdowns.

However, we are aware of a number of foreign organizations which have entered into Royalty Agreements with A. D. Little, Inc. ~f

• Cambridge, Massachusetts for the right to use ADLPIPE, a computer code which utilizes the algebraic sUII111ation technique. A list of these organi-zations is included in the attachment to this response.

e April 19, 1979 Mr. Vincent S. Noonan, Chief Engineering Brnnch Division of.Operating Reactors U.S. Nuclear R~gul~tory Commission Washington, J. c.

20.:: ~

Dear Hr. ~oonan:

98705.

I a.n enclosing a memorandum which confins the information furnished at a meeting with you and other members of the NRC Sta.ff Monday afternoon, April.16, 1979.

I a~ sending a copy of this letter (and its attachments) to John G. Davis, Acting Director, Office o: l""c;?P.~tfon and I::nforc*er.icnt, under cover of transmittal, a copy of 1o1hich is attached for your information.

A copy of this letter and its at:tac::1n1ents are being sent to the organizations listed in Appendix II, who are ADLPIPE licensees.

As discussed at our April 16 meeting, ~e will verify the five bench m.irk problem solutions (after l"l:~cdpt or the probler::s fro::,

NRC) published in ENL-NUREG 21241-RS and BNL-NUREG-23645 utilizing the present version of ADLPIPE, February 1977, Version 3C.

If you desire any further informatio~. cc net hesitate to call.

Very truly yours, c:J'W~1~

I. ~-:. Din;l':ell sp Enc.:osur::s

~lc::::.*randum Letter to John G. Dnvis fro~ I.~. Ding~ell of 4/19/79

'I e

e A 3rief Historv of ADLPIPE (see t.able l)

Ar:hur D. Little, Inc ** first prepared a prosram in 1952 to compute the flexibility and thermal deformations of piping systems for a private firm.

An ASME paper was delivered in Apr;l 1956, "The 6X6 Matrix Mechod of Piping System Stres~ Analysis".

Lnter durins the li~uid oxygen fueled ballistic missile program, Ar~hur D. Little, Inc ** adapted ~his program to make dynamic analyse~ of missile fueling systems.

A nev program, ADLPIPE,, was* developed _in the period 1967-1968, first for the static (deadweight, thermal, external force, applied displace-ment) analysis of elastic piping systems.

The program was written in FORTP.AN and designed to be independent qf the particular computer system used.

The second develo~ment~also'in 1968--(modificatinn one) was for the dynamic (modal) analysis of lumped mass ?iping systems.

The transient loading was described as a response spectra.

Fo:lowing-a prototype development period, a version w~s delivered in August* 1970 which enabled the user to implement ANSI B31. 7 "Nuclear Po~:er Piping".

This version* could not produce a iull stress report bu~ gave strcsse~ for particular loadings.

In 1972 a version ~as released which enabled the user to produce a partial" st,ess re~1ort to meet the rec;,uirements of ASME Section III.

In 19i2 -this vcrsicm \\>'.:lS relc.:iscd to Control Data Corpor.:itl,1n Cyburrtt.:l.

n 197J tlw cm::j)t;t;_t ion of :".:icl.i;ue usage factors was completed.

In 1974 a version was released for the utilization of AS!IE Section III, ~lass 2.

In 1975 n force time history analysis was included for the calcul.:it.ion oi hydraulic transients.

At the same time a one-dimensional thermal transient analysis was developed for the requirements of ASME Class l.

In 1976 the automatic computation c,f ;;,:;,i.s::ik c1nr1lyses in a,::corc:l;mco:? t,*i th Re;ulatory Guide l. 92 w.:is developed.:~C: c:1cckcc!.

The comj>lctc r.1.:it:-ix analysis portion of the program was rewritten based generally on the

• echniques of SAP IV with some improvements in the ~atrix storage methods.

-.1 addition, a.post-processor was developed which.:illm,eci the user tCI make lcc1d se: cc:nbinaci~ns for use ir..:;,;"1.i:.::ti~::-:s ~:hei::*r :!;.:i:i Rcb~l.:.:or::

Guide l.92. This version was released !.n February 19ii and upg:'aded in December 1977 and SE:_ptember 1978.:

In the period 1968 to 1973, ADL?IPE \\,:is the only cc,=.?uter program (\\-:hich

':..'as.:iv.:iilablc to tht! public.:) for cnmputir.:; piping rc,,;ponsc to v.:irious st:::ic.:lnd transient loads.

lltht'r r1**,*-~ra'.'.!s w,*n* ln t!i-;c.*. hut to nur kn.:-:-11.!dgc, the~c were: propricL:1ry.ind nut.1v;lilab] C! :,,r ;;cnci-;11 USC!,

Frc= its inception, ADLPIPE could be util~zed for a *:~::-icty of stress c~:.::ul.:1ricms not involving nuclear pn**<!r ;,ir:n;;.

In 1975.:iprlicntions

\\."C-!'c: c::-:tcndcd en ::iC'et the rc*<111ircmc:ns 0f chL*ndcal r::,:.t,mu re:iinery pi~.:.:.;;.:ind pctroll!u::i crazv;;,,r::ation jdpin:;. *

\\nh::: i >!.1ttk-.ln~*

. r 1967 1968 1969 TAIH.E I Dl:."VELOl'HENT OF Al>LP.ii'£ Development of static load version Delivery of static version Delivery of prototype dyn;1mic version e

1970 Delivery of stntic dyn.imic UJl.7 version 1971 1972 Inc;lusi'bn of ASHE Section III Class 1 Inclusion of clos~ly spaced modes' 1973 Inclusion of ASHE Section III Class l us:.;;e factors Inclusion of Metric ur.its 1974 Inclusion of ASME Section III Class 2 a:ic. 3, B3l.l 19i5 Revised input organization 1976 Force time history analysis 1977 1978 Transient thermal analysi~ (nne-dimensj,r.al)

Inclusion of l.92 modal summation (groc? ~ethoci)

Inclusion of post-processor for new l.~2 ::u~tion Revis'ed matrix storage and solution

\\ r : ! 1: ir I ) l 111 k-h

e

~fou:il An:i]_,:sis lw,\\DU'iP!*: Dur in~ th1.: Per :ud 1968 Th!"c-uch l <:*76 In this period of time, ADLPIPE was licenseJ to se\\*u*al clients and released beginning i~ 1972 to sever:il nationwide cc=puter service bure.ius.

A listing of ADLPIPE versions and docum~ntation is given in AppcnJi:< I.

Tiu.: names of ADLPl !'I~ licunsees and the cff cctivc d.it\\.'S of Lhc licL!nse.1;.;rcc1uc11ts.irL! givun in Ap1>endi~; II.

The dcvelop:nent of the scimnic nn:ilysis method was guided by av:iilable 1 itcr.aturc.inti the! Jc.:si;~n :-cquircnu.:nt-s oi our clients.

A method of

.i~lysis.....is dc\\'c]opcJ which w:is cxpl:iined by two documents published in 1969.

Th~e a-re enclosed.is Appendix III, 11Modi!ication One-Response to Ground Shock Spectra" and Appendix I\\', "Development of Modal Particip.:ition Hatrh: for Ceneral Three-Dimension Shock Input to Lumped Dyn:icic System".

In Appendi:l-; III on page_ _~r-;, I state "the modal amplitude, q

• is thus evalu.1ted as a scalar suoi:ation of the products of the ntfl vector of the mod~l p&rticipaticn-~tri.~ and the spectra :tmplitude (D,.) ".

The "spectra a!':'lplitude 11 =ie.ins the spectra

displacement component~ in the principal coordinates of the piping I system.

Fro~ these mod.il amplitudes, a set of dis?!ace~ents for each 1 mode of response is cor.:puted.

At e.:ich point in the piping system,

/ three mod.:il moment co~poncnts are then cocputed, one of each principal

'.axis. *. Each component "-'.ls then squared and then the square root of the 1sum of squ3res Yas taken. to combine the effect of all r.::,des.

This concept "-'3S used to modi?l earth motion nlon:; a vect=-r,-:hich was not necessarily aligned with a principal axis but was s:S.e,, and was decom-posed to three components.

The reason for this develop::ient is sho"'"'l'l in Figure l (;,age 4)-'1.*here a structurt: is not alignaci ;dti1 a gloo.-1 coordinate system.

An earthquake is assumed to act pe:-?endicular* to one wall of the structure. Mathematically, the ske~ axis of the earthquake is decomposed into two horizontal components in the global 3xes.

A user could calculate earthquake response with a ~ertical componen:

and a. single horizont.:il,:.:,::iponent i." ~:,.a tt,':J.:.x<:s i.:::::ri: i.:1.:ouple:d c.;:::-

bining several such analyses to create a worst case effect.

A user could make three or more different analy~es, one fo= each principal axis again combining the results.

Users who made a single analysis using a tri-directional earthquake would have printed out 3 single -set of modal moments.

If one isolated e.:ich response spc.:ctra component by a separate analysis and computed three sets of individual moment components, the resultant fro~ the single tri-

~

directional analysis would be the algebraic sum of each individual co:::?onent

-.;reach ea:-thquake dirc:ctional co~ponent.

T!"lc upper lev.::l would be the absolute st.::-: of the in:r:i-mod.Jl co:-:q,onl!nts.

Th<:: lc*.. :e= level could be zero within a mod=.

Ho~,*ever, it is my vie,-: that th: inter-::od.al su:.:::-.ation using the square root s:.1:r. of the squares... *~uld not ~e zerc and, in :.'.le:,

woulc! not v::.=y ;re~1tly (::33 percent) fro:::.i square :-cct su::: of squares

( SRSS) in tra-::iod.:il su::::::~1 t ion.

. nurneric

ll. e,:ample i.s ;:. *:e~ in Appen::i::,::

\\*/ -

e UNIDIRECTIONAL EARTHQUAKE WITH SKE;~ co:-1ro:-:~"TS 1 !\\(

I I

,,~. l

~

.. l'~ I' I

~-~~

f (-l- ;< )

X C.0W111P 0 ~

/

/ 1 --,

.If/

/

J "',f~

~

... -. Lt"-**' ~-

J u

I.."...

I r

I I

l

• 1 s f.!.-""-)

e

\\', "Dyn.1mic. An;i:lys i~ by AUI.I' 11'1~" wli i,. ii l dist r i iHJ t v~ : n

0:

. * ~.:-,- :* 1 ~ i 4.

Prior to 1971 any combination of loads or earthqu.1:...c:: h:l..: :.:- be ::-.:ide by hand or by another program.

In 1972 I released" s~-=-:~:icn ~=~ce-dure which enabled users to com~ine loads in accord:l::e ~:.:h B31.7 and Section III criteria. In 1973 the computation of f:l:igu= ~s~se factors was released, which included the cyclic effects due :o "*ar:.ous ea:-th-quake compone~ts.

If these summation techniques were used, the user could input several transient (earthquake) loadings,md c.o:::bine these loadings, one by one, with a sustained loading (dead-.:ei;::.:) to ac:lieve a "worst c2se" stress calculation.

Modal Analysis bv ADLPIPE During the Period 1977 to the Pres-ant A new option was made available in ADLPIPE in *rebru.;:-y 1~77 for t:1.e computation of earthquake response in accordance t..'it:: Re;..:.:atory.:;uide 1.92, Revision 1, l'!arch 1976.

In :iddition, 3 post-processor has. hc,in dc.:v~*lopcd,,iih<1,.-:~ ::*.._._-; ::11-:,

user to make a number of combinations of directional -a"r:~;uakes affects not included in-Regulatory Guide 1.92.

Verification of ADLPIPE Verification of ADLPIPE was undert:ik!'n in a series o: iu::*:**.:-:ie::1t::il -:hecks.

In important mo<iifications a supporting document was ?re;a:.-c.: as an ADLPIPE reference.

The verification procedure was as follows.

The thermal and deadweight loadings were checked by a Ho*,;.:..ard Be::.d and hand calculated systems given in "Design of Piping S:,-ste::s", :-1. i-:. Kellogg, Second Ei.lit:ion 9 1956, and "Formulas of Stress and St=ain". R.J. F.:iark,

?-lcGrat..'-Hill.

• The dyn:imic analyses.... ere checked by "Response of St:-.icti;r.:1 Syst:::ms to Ground Shock 11

, Shock and Structural Response, ASHE, 1960, :.:, "ADL?IPE Results of Model Given by Young (ADLPH'E Rdere.nce 4). a:::: ":)::n.:ir::.:.i:

BehavioT' of a Founc!ation-T H.e Structur.::.", t*h!.clwr.ic;:iJ !ncl,--.;::,c~ncc- ~-1cthods, AS:*lE, 1958, in "Experimental Verif ic.:ition of ADLPIPE :-:od i ** (.wLF:?E Reference 3).

TI1e time history anal~*sis was checked by a scpar.ite -.:l::D.l::::.::;:;: sC':ution of the p:oblem given in "Analytical Huthods of Vihr,1: :,_1:..:;, ** ;*-.1:.:::: ;95, Leonard !-ieinovitch, "ADLPI1'C Time Hlst,,.*:* Response C-::::;-:-,-:.*-=~ ~::'::1 :.

1r,t-.-n Solution :c ~-

.:i Hcnv*ily Dnmp...... J Sy:;ct:n (,\\I1L1'I!'~ :.~: ~:-;:-*::~;; *.. ~ ~

1

~\\

second check wzis made usin~ "Pr!!ssur1.: \\"css~l,m'u Pi;,:.::~

~ - C",-:::;:*.;ter

?:-ogress \\'erificat:ion", ASHE, 1972 (J*rnblci:l SJ.

,<' ti*,~r::::il tr:Jnsjc~t an.:1lyi::is t,*:u; \\*.::ritiv*! 1-y., s,:;*.:-.,:,:-.~:*... ::.-="-~. "T:-:m-

~il!nt Ti1~r::..il Gra.!1.:.:nt St::1::-.scs". I~. i:. !\\i-:11H"i1, lk.i~:::~.

.;.i. r

• _ * : : :: : r i ) l It! k* In:..

e CL,ntll:iun.ing, ":ul:.im~.:.J. 1971, Aruilysis" (Reference 15).

e The comput.. tion of -intra and inter modal =:i:::ent: CC'l..1;:~n"nt summation has been verified by a separate computer pros:-.i::: for that ;,ur*pose.

A report "ADLPIPE Modal Response Combination for Closely Spaced Hodes", is available as ADLPIPE reference 24.

Vnrious c:alculnti.nn procedures required by ASME Section III were verified in ADLPIPE references 10, ll, and 18 entitled "~~!.PU'£ Computation of Bending Stress in Tees and Branch Connections, AS~!E Section III, Class l Piping", "ADLPIPE Cc-::iput.:ition of Rasultnnt Mor::.cr. :s for S,*ctinn III Class 2 and 3 Stresses", and "ADLPIPE Stress Cor:puta:ion of Piping Compo-nents:

A Comparison with Hand Calculations for A~SI B3l :ind ASHE Section III."

In 1978 an independent third party revie't.' of ADL?I?E (Section III, Class

1) was performed "Verificacion of A.DLPIP!:. :\\~!E Se.:::.0-:1 !II, Class 1 Piping Stress Program", Teledyne Ensineeri::ig Ser:ices, Report No. TR-2864-l, August ll, 1978.

ADL?IPE Development Foli.:v The following policies have been in effect d:.irin; the devclo~ment of ADL?IPE:

1. The details of c.il~ulation processes are ~v... i1aole to the

?....

public by fre:1.: dj~:tributi:,:! *.:f *op,..:n1~i1..; ::-:.,~~.:.l., :ind ref _rences.

These are tabul.ited in Appcndi>: I.

F..,c:-: m.ij er new feature of ADLPIPE is documented for ur.~r rcvic~.

Program listings are !lklde available to !.ice~sees.

are noc restricted from m.,king pr,:,gram changes.

Licenseer:

3.

ADLPIPE is perio<i i.:ally irr,;'r<"*~ed... nd *J!*:'.:1 t2c ~-.,c: l icen~ec'.*~

are notified cf :he modifications.it th~ :i::::e u~ the:! reli:ase:

of the modified version.

4.

ADLPIPE is hand checked wherever ~ossib:e.

~~.en this is not possible, ADL?!:?:!: i:; che::!~.:;! ;;::.::-:j,.::-i::.~~-. :a: :i.".:.::;...:. ts or ;,.;,=

results of other calculation ;,roc:::dures.

E*:ery ::iodificati'1n, large or small, is checked.

5.

Special version:-- of ADT.l'll'E \\d]l :*,* ~,:ri~t,*:1 t;,, a lit:1.*nsc.*1.*'s spcc.:ificntjun.

llowc.*vL'r, Llil.' vc.:rsin:: of.\\::!.:*::*:: r 1:ll::1s(;d t**

con;'lllter service bureaus t!Cncra!.1::

• d,..,e:= ~.~: ::;ive: ~uch sreci;il aJdicions.

6.

Old \\'CI"sions of :\\ULl'Il'E ari.: nut r~:.:;1im.-~ :-:: :'.!*t:.:1r !J. Little:,

,In,.

Instcnd, :,c~innin~ in 19i:...:11 n..::-: *.*,.,:*:;i.:*ns,,f AD!.l':Z:?E i...*cre back..,,..:ird i:iccgratc:d.

Tiu.* ;,r..-sc:lt -.-~:-;.:i:*:-: <>:,\\DJ.PIP:

e e

rn.iint.:iins all p;:st f unturc:.; which h.:ive: :,E:;:~ =~-=.:: ;.*:.:.il.:iblr.:

to the: users during the period 19il to :9iS-.

)t{J)w~~

I. lJ. Din,*.-1ell Arthur D. 1.ittlc, Inc.

C.imbridge, MA O:?lt.O April 19, 1979

\\'ersion April 1968 April 1968 August 1970 January 19il July 1911 September 1971 November 1971 December 1971 June 1972.

July 1972 December 1972

.*..__.,_-.~.....,_. x*-**....---,~.... *#"*-**,..*,-~~,._...,.+*-*,.,,-*.*~***,,,.-,...

e e

,\\l'l'l*:~:111 X 1 ADL:'I!'E \\'ERS lO~S A:-:n noc:~*:-::::::-:. ::- :

Document~tion and Fe~turcs ADLPIFEThermal, Static, Dyn::=:.:.= ?ipe Stress Analysis Operacins Milnual, undated.

ADLPIPE Modification One:

Th~:-:-:a:. s~~:ic, Dynamic Pipe Scress Analysis:

Operacing :-:~:i:.:a:! first. version dated March 26, 1969.

'Features:

'!;:;.:-::;.:..:, deadweii:;ht, external, acceleration and shock lo~ll.l.'"i; :.. ~ :1:.::~ load sl rc..*.s.s :in:ilys.is; code - BJl.l (1955).

ADLPIPE **** Static-Thermal-Dyn~:.:.c ?ipe Stress Analysis dated August 15, 1970 Ne~ Features:

Code - BJl.l (:~:i:; equations 9-13, B31.7 ADLP!?E **** s~atic-Thermal-D~-n~:::.:.= ?ipe Stress Analysis dated January 15, 1971 New Features:

Four mod.11 St-7.':.:.z: ic:1 techniques:

maximum, ma~:imu: ;:me sc;~.s.=e :-oot si.;::: '.:If squares of remainin; ~ocies. s~~are root sum of squares, absolute; squa== rc:t su= of squares for str*ess c.... lcula:.:.: ns ADL?IPE **** Static-Then:lol-Jrn=:::~: ?ipe Stress Analysis dated A~ril 1, 19'1 New Feacures:

Stress su.~"::ar:: ::?C =t, B3L 7 for multiple loads ADLPIPE ***** Static, Ther1:1al. ::.*:-.3::-.:c Pipe Stress Analysis Input Preparation dated Ap~il :. :?72 New Features:

ASHE SectiC'n r::. C:ass: l (1971). summ:iry stress rej'.'C:'t ::: ::*.::ci;:,le loads; closely sp~ced mod~l s~:-~::on

References:

l.

ADLPIP!: :-:ath.::::.a:i:al..\\n.:i.l::sis and Logical Procedure

2.

Section I:! ~=~?:a ?ro~le~

3.

f:;,.pcrir*.,:m:::,~ *:.. :-::c~:.: kn llf ADJ;l'IPE

4.

ADLPil'E f-!:?s~.:.:~ ; : :-:oc::ll Given by D. Young

5.

ADLP lPE !-:~c:..::.:.a::.cn :, Response to Grounc:

sr~~*::r:i

6.

Dcvc>lop.:h.*nt.~:: :-::;al r.,rtic'.r.:ition M:itrix for <~cncral T:in.:i.: : :.::-.,;;nsion Si1ock Input to Lm:-q,~d ;Jyn:::-::i: 5::=:::-

ADL?IP!: *.*.* St.-:ttic, Till.!:*:::::.!..::*:*.-::-.:: !'ijh.! s:.:rC!SS Anillysis ln?~t Preparatinn d.:it~~ ~~::..:.:7-:

~:a*... : F 1:~ t.u res:

En*~ 1 is::.:::-:.:: :*:-::: :- ::. :.::: : : := : S:J?:::::~:-y s :rC?ss n*;,ort. :-.***:*,*:*. ::: ~:::!-,." * (197]); f.::t:i~ue

111::l*:si!-

.*:*... -:.;,*:.! rn1Lpt1C:

jso::ict:* ;c:...; fr*:- ::... _:.<.v.*;:.:.::::, u ir:wnsic,ncd i.,.;,,mctri.*,-. :-:.*.:**..

• • ,:,*fc,r::1..:J pipint;

.-\\1 tlmr Ii I.1i:k.ln-.:

• ~ *.-** -

,..,..,.~

~ -

r

,1 Hay 1974

• April 1975 July 1975 April 1976 e

Ref erencei.:

1.

.\\lll.P!i'f. :_l.~;1th'-*:=."lt L::ll :.::.:: :*~*-; ~.-,:;d Li,~i,*.11 l1roi.:1.*d11r1.:

2.

S1.*c t i.. *n I J I S:1r.:p I... Pr,*:*:,::*.

3.

Experimcmt.11 \\*eri.:ic:.i:i.:-~ oJi *.:.!J:..rIPE :*IOD l

4.

ADLPIPE Results of Mace: Give= by D.

Young

5.

Generalized Piping Syste:: Res?onse to Ground Shock Spectra 6.* A Method of Computing Stress ?.ange. and Fatigue Damage in a Nuclear Pipin; System by W. B. t~right and E. C. Rociabau;h.

ADLPIPE ***** St.itic, Tberm:il, D:*n.1rnic Pipe.Stress Analysis Input Preparation dated ~1.iy 1974 New Features:

Codes - B31.l (19i3): Section 1:1, Class 1, 2, 3 New

References:

7.

Section III s~mple Preble=, Class 2, 3

8.

A:-:SI B31.1 (1973) Sa::i;;le ?robletn ADLPIPE *** Static :m~ Dyn.:imic Pipe t-esi;n a=d S:ress Analysis:

Input Preparation ?-:anual ca tee :anuary 1975 New Features:

Revised input orsanizat!~= (;ec=etry and execution decks)

New

Reference:

9.

ADLPIPE April l9i5 Release ADLPIPE *** Static and D:-namic Pipe Desi;n a~= S:ress Anal'.',*sis:

Input Pr!'pr.:-atfon ?-:~nual d.:i:~~-:: :.. ~rn~'"Y 19i6 Ne"'* Features:

Section III Cl.:iss l, 1,.3 (:9i.:.); force time history dyn~:::ic an:11::s!s New

References:

j. Documentation uf ~D~?!?E :or Static and D:-*n:imic Lc,.:1,:s and S:ress ~-;aluation, Sej'te:nbcr 19iJ.

6.

z\\ HcthC'd C'lf Ci'.'::iruti::..:. 5tr,*;;s Ran~e

ind Fatigue D.l~~;e in a ~~~l~a~ Pipin;

~y~c~~. ~- b. ~ri;ht ~~~ :. C. koJab~u~h, Nuclear Engineering.ind !)csi;;::, 22 (19i2).

7.

Sample Stress.\\n::il:,-si:= of AS:*1E Section II! ~ucle~r Class l a~i C:3ss :, 3 Coi:lbined Pi;*:n; S:-*ste~ and,;~:s: ::.3:.1 (~.973) Piping Systci:l C,~putcd b:: ;:Jtl'::>::.

8.

A:>LPIPE S}:e*.. : Card Test Ru::. July 1975.

C}.

A:lT.PT1'F. Apr1l 197r, R.:-l,,:1~<"*.

111.,\\Ill.I' 11'1*: C:11111p11l:1l i1111,,;* B,*1;.I i 111~ SL l"l*s:; in T1*1*S :md l\\r;111,*i1 c:1111111*,*L :,,;;*;. :.:*:=-m Sl*t:L ion lLl, C.:l&.!.5.S 1 I'ijii&°&:,:., J*;~y.:..;-~5.

ii.,\\:H.i';;*r: C:,*::1;i11t.;1tii':1.~:-

i~,*,::-.1t:mt :foments

(,,r S,*c:- i,,11 I J 1 C:L1ss., :.:~.!

-itrl~sscs

. J,i : :: ! 'J ~I ~,

• 1:.,4:.JL?1P;": :!~t.... c::\\,n

.i~:.:i~ ::,.,,in,.:-~"'~f ::r:--.,r *. : - !,

!i:i::::: ion c: :\\urne~-

s~ ~~~;s ~n~ Stiff

I 1 1 f

Rov:il t *:.'\\r.r~\\.*r.1cn ts

. co:-!P k"r"Y Black &

Veatch Blaw-Knox Bro"-u. & 'Root Burns & Roe 10/04/74 10/04/67 11/07/75 7/22/77 P.O.

8/l9/i7 l

PENDING Comision Federal de Electricidad Framatone Gibbs & Hill H. w. Kellogg Company*

l l

2 1

Charles T.

Main, Inc.

P.O.

Montreal Engineering Company &

Monenc:o Comput-.

ing Service Ltd.

1 Northc:tst Utilities Service Comp:iny Po"-"er Pipine; Co.

7/ /74 2/23/76 11/29/72 11/16/75 7 /20/76 2/20/70 7 /06/i:.

11/01/78 5/12/70 11/19/i:;

10/26/":::

4/18/72 6/25/75

?

= l/Ol/77 5/12/70 I; -,

-*""'J.-

per?e~:Jal per;:etual per;.etual per;:-ctu.il auto:::&1:ic e:-:ten:5ion *

,erp

:ua.1.

au:o~a:ic exccnsion perpe::ual 11/1~ 'i6 ilU :c::m :ic t.>:*: :.c:1s ion

.\\id1;i: ! \\ l 1tth: In.

e Roy~lty A;re~mcn:s (ccn:)

Sencr Ingeniera y

• siste::i:as 9 S.A.

1 2

3 4

United

£ngincars

& Constructors westinshouse Electric:

Corpor::ition EF'FECT1VF. o:~ iE 2/01/72 10/01/73 6/16/75 5/01/i7 ll/01/78 5/20/70 ll/27/67 j ! - *.

' I

\\

1

--**-***-~* *-... *-

e e

Al'l"J:~::> I:*: J 1 J nae b;asic ;arprnnC'h to hr usc.u in r.ump111 it1r,.t!1c* l"***:;*,)~:.r t1f ririns~

Kystcm.c; to J;round shock inruts in trrm~ nf 1li!:i*t.:*u-,*::**r:r (,*r v*.*luc-ity nr

icc:clcr.:,tion) spl'r.trn com;isu, nf 1~1*1u*r:1lini~ 1111' dv1:.1:: i,* pr,11u*rti1.m nf the system and apply inJ: a mod11l suJh:rro!i i Lil*n n:..:thnc= (:ir 1wr::i.il r.iode mclhnd) to define the strur:tur:il rr*:;1mns,* lo tiu: :-.lull"l~ i:11'11L<;.

Tlw fnr-mul.ition in terms of nur:m:il modes fnllnws r.,*1wr.:11Jy :ii~.. :on:: disc:us~ed by Young.Cl)

As !ormul:itcd in this rc-fcr,~n~,... the C'C':1t:-i:rnt i,,ns from lhc individu:il normal.modes nrc dcf inccJ in l*.:nm; pf., :,c,::il ;,:art icipn-tion factor which.depends urnn the nc,rm:il c;l;,1p-.! ("*l:::~-~-.-... _*rn:-).i'nc! the distribution of the l....::..l ovc*r the slructun*.

• ,.. i:s

,* :-:1:;.-:t.iun is appU-c:able, however, to systems excited by one-tli::i1.:n!-ii.*:1.-::,.;;,_,,*.~ onJy. i.e.,

with the inertial elC':ncnts restrictl"'d to rnnt.io:1c: i:1:: ;-~:H~*-**

For the general three-dimensional shock input and rc~por:!::l' c:;.::,*. t:1c contribu-tions of the nor:rnal modes c,:m be shown. to :.>c? cl.:f inac: i:: ~ ~o.:.:il jlartici-pation matrix.

A description of tile.: steps lcc.1JJ.,1g :;,, t:,c.: ~.:L::::-:::::::i:..:.un of tue re-sponse due to ground shock is given in the f0ll.i*... *in; p.:;r.:-.~ro;,hs.

A.

Calculation of "Reduced" Stiffnci;s ~lntri:-:

In order to define the nonnal moJ'1s Cli :.;1~~ ;, i_;-1:::: :::=' :..:::.s, a f lc:-:i-bili ty or stiffness m&t1 b: rcl:itin;~ :*.,:-cl's :1::,:.* i:,* :. ;.

points in the system.must be gcner.:it,:t.1.

Feil io';,,"i:~.: t...:: 7"':"i'S:L*,:un! in ADLPIPE, a network stiffness matrix is fir!--t IL*r:-.~,,;,,;:; :i:: :: ;,y ~ array for a system of~ nctwnr~ pninls.

II

'. T'."'.'-.

6 x 6 subsets.)

The numbering of the nct*.,.or;: poir::::; i.;: -.1:-r:C'c out in LIil' f11J)c,wi11;~ 1>ri11rity:

fi1*!:I.,

Ll1,* :11;1!::: l'"i!:f*;; *,*,*. :1,.* r:,,. i111,*rior h r:1111:li pu In L:,; *;111t.l I j 11;al I y, 1 l1l' ;111d1ro, I'" i :i I *.*

I.

"1'1111111:

• ii;111;1 "Hc*~;I'""*;,*,,f

mJ !

:L n11*t nr:d H,*;.puns,*,

!i. r., J9ou.

!:r ru,*I,., ** i :>:* !.*r*,*. :

1.,

Ii 1-1 I

11

....., I :~:::* i lll'l" rs, Ard1~r D. Lttlc Inc.

I I o

e e

Llms furmt:J will be.: urcJc.:rccJ a!, incli.-.11,'..i lwl,,:::

A B,.D, E m.,ss poinLi=i suh-1n.al rix interior ri,,inLs suh-,:i.-it ricl'S C, F, C, II, J :mchor 0

poinL::. s11l,-maLric:r*s

\\s shown, the m.

itrix is pnrticinnecl inln tlil! Li1rl*** ~*:1lc!,~nri,*s nf rwtwnrk points.

The formation of the complete.: matrix is c~1rrit*d,111L L,y,\\l>LPil'E.

The rows and columns correspu~1din~ to anchor_ points.ire no,... deleted from the _stiffness matrix, lc.iving a m:itrix ch:ir.*1ctc.!rizin~ r."::iss j1oints.ind interior branch pqints only.

u O represents deflections at the m.iss point~* *.,nu

• , 1 r<.*r,i*, ~--:i~i;.11.*flcc-tions at the interior brnnc:li points.

imil

1rl*.-. i" rc*;,r,**.,

:*; 111.!*:o:.; :it*

I/

the mass points, and F1 lo.ids.it the interior hr:.mch po.i.x;c~.

ln

• the case of free vibration, the.lu.ids r0.ire inertial l11;1di; du~ to l:11* ::::iss points and the loads F1 are zero si:1.:e intcriur nct\\-:or:: ;-:oints,,r*.* ::.:.: l,J;.Jcd.

The equations then become

,~..

+ l~.

I(>

• CJ I~.

I II I-r*:.

II From thl: scc:onJ cqu~tion, -.I\\ 1 = -F.-I I'

'. 0 *

11

0::t itut:,11: l:1:P ~i1,* firi.t crpi:1tion I I 1-:~

Arthur Q Utt le. inc

,1 (A -

H *

- I 1-;

l, J II F II

.J,..Cl.:ctiuns :it mm*Ui,,oinL!i.

llais u1.alrix is ;an II x n ;irr:ay wlu.. n* n i.s the number of m:iss points.

For ilwrti;a Jn;icl::. llai:, m.aLri_."! l'*111;1Linn mny be written ~

2 1i1.* M.**, ()

. t, 0

where (A

ll 1( I lJ)

B.

Calcuh~ of ~rr:i:rd Modes Tiie eigenvectors. !J. 0,.ind the l*i1:~*nv:il111::;, ;.

, fCJr t*.,ch of the

1 n

normal modes arc computed by solvin,~ Lhc m;1t ri:-: c..*.111.,t ion

• 1

~* 11 0="

1-1,0 for c.:ich of iL.s n ch:ir:ic:tcristic snlut,ic.m:;~

Tl1i:~,*,,11;1ti1111 rn:1y In*

solved by iterative procedure when puc. into ti1c fc1r.:i

?.

AV

= w--v n

r:i n

This.transformation is pcrfnnncd by d~*finin,~

and M ""Ml/2 nl/.2 V n. = Ml/ 2 :\\ 0 11 ci1u~ dcfinln~ Lhc matrix A a~

It.,ssurt*s tl1:1L LIii' i tcrnt i,,n 1,,1i 11

,*1111*:**r;*,** ;111.f 11.,*1,* n*:* I :111d :,,,*; i I h*C' (2) 1.' lJ!l'nv;il111.*s

2.

~.,rt~.

B. Stiffn,*~s M.-1trj:-: :;tru,1111.11,\\n-1_!;.:_-_;_*.,,Jl*L Pr,,r,,1J:... i,1i1 I..,b,,r;itnry. T,*d111ic-,il l,,*pnrt...,.

1:1--i'.',. n,-1.. 1>,*r *i1, l'H,*,.

I I I - -~

Arth.ur D lJttleln~

e e

Wilh LIii' m;1lrix c.*r1w1Li1111 in tlai!, lunu, ti,,* ii, 1.il 1*.-,

1'1""11**.*. will

,*nnvcrt;,* musL rL*.uli ly un Llui c.*i,:,*11,..;al111* la;1vi111: 1111*

l;11"J:l*;:1 111;11~ni luu,*.

Fur s:rnund shoc-k rc.?spnmu: ;1p11lic.:1tions 1 il i:; rnura* &11*!;ir.1hh* rur th,*

process to c:onvcrJ;c: most rcildi 1 y to tlw i,.m,, I I,*st,. i s~,*nvn J 11,*.

Conse-quently. the m.Jtrix c:quat.ions are put in Llw inv,*rtLicl r,,rm where C

• CV n

V n

• n A-l. an,f.\\ n

• 11.. :*

n and.ilpplic.ition or an i tcrnti Vli me L hm.l, sud1 :as l Ill* SL P<l11 I :1 me: L lu,<l, will produce the SUCC'r.:ssivc mo<lal fn*qui!nc-ll*S (c.*i;:,.*nvalu~s) ;incl mnd.il columns (eigenvectors) oi a system in :tscc.*ntlia,r. nrJcr.

An altern.1tivc solution tcchniq,11.*, which h;as 1:,..cn uLiliz,*d in

'"LPIP~ MOD 1

• 1 J.

b'

• , du) 1 11 r

1

,=

1.s t1e aco 1.mct10 n t1111; pr11r<*u11n!,

1 tic eigenvalues and eigenvectors are prnt.fuced !dmul r.,n**nusly,,*i tii C'}u,,I.

accuracy.

This method may, thcrcfon*, cmpJ.11:,: tlu. r.wtrb:,*qu.,ti,111.in either form (i.e., with eigcnv.Jluc.-8 J /,..-2 n~

• J.

In !*W:J J, tiat! st:cond n

1 fonn, in terms of A, has been used..

l'hc m,,d.al rn!qm:nch*s arc stored n

in a "frequency vector", and the mod:il columns.ire stored in modified

£om as columns in a "modal m.Jtrix*".

111e mnu.il colu::ms an.* modified by first converting the V back to modnl deflect iom, '\\ 0 and ti1<m by nor-n malizin~ the column.

E.Jch-of the s,~t w£thin :1 mnrl:al col1:*~:1, -:., now l 11 rcprl!sents a nurmalized,fof l oct:i_on o i mjss* i in much: n.

c.

Calcul.1tion of.Eguivalcnt St.itic Ucfkr-t im1s As indicated in the appendix, gjvcn by the expression Ll1c mod.al m:;1,JiLuJc 'I i" shu,-:n Lt' D<:"

11

).

'I

i.

'j' (II

)

II V.

II ;*

V.

11 Grcc:nsL.1clt 1 J. "The: U1*t.,*nni11;1Li,,i,,,f th,* 1:J1.1r:1('1,*ri.,:i,* Rnqt:; nf.1 11.itri;. hy tlll' J.ir.nbi M,*Ll11*tl", 1:i1;rpt1*r 1 "' ::l.:.~_i,.'.'!..'.'.:~.i_:_.,_I_J1.!.:.1 iorul:;

fnr J)j1*it:il-C11m11utc.*rs.,.lnhn '1.'il,*::.

~ *.,.: Y.. rk, l'/'1 11.

111--',

Arthur D lJttle. lnc.

e wlu.*rc.* 'I' is Lhc mocJ.il 1,:irLic:ip.1L iun nwlri;-;.-and (IJ. J i!;

nr.

• n I !11* :,huc*k in-put displacement !or each courdinat_c ilnd fur,u1ch mm.le.

lhis ~\\!n~r.al

• three-dimensional form reduc.-L*s to tlw i; imr h*r fnrna;il.:it ion q *.,

I) n n

n i.n the c-.i:.c: that the.? jnpuL shn,*k mot i 1111.al 11... h.1:H' i :; tli,* :::1111,* i 11 every cl,nrdin.at"*

lt* ii; this lntt1..*r r,,nn whirh is d,*,*,*l,.,p,*d h:)*

Young (l).

For thii:. onc-dimc~nsinn.11 r.:mc ;1f;,I is1"11s:;,*,I in R,*f<'r,.,u*,* l, th(? mod.il participation factor.is J..:I inL"d ll*r,*;1d1 r.1t,rJc, l.'i&ilc fur Lhe I

general* three.-cii,:i~nsi.-;~* ~~- cai:.e, the moJal p:11*t i c*. pat inn is dcf'in<'d for each ffl.ltiti for oacia moJe,.it1J thus j,-. in., s11u.1n* ;irr.1y fnr::: r:llhl*r than in a linear arr~y form.

Th~ amplitudes (D 1 )

0.ire obtainl.'d from llu* dvc*n inr*st shoc:k spec-tra (e.f?. 1 Housn~r spectra for earthquake lonc.lin;s).

In ti11.ise spectra,.

the amplitudes arc defined by the moJ.il frl:

0 CJtJL'nC'y :mJ h,* the coordin.ite a.r.:is.

Fur" L':1ch v:i-luc of 1,., th~rcf1t1*,,; and f11r 1*:wl1 c-unrdi:1:1lr.* ;1:,;is in

. II which there is a prescribed input Sf7'Ctr:i. w,* h:1VL',"I v.ilu1..* c,f (i>.) *

'* 11 The modal amplitude q is then evalu.:1tc.!d as th..: scnlar summation of the n

products of the nth vector of the mncJ:il* i,:irl i,. i p,"'lt iun matrix :ind the spectra.amplitude CD1 )n' or, as given pruvi,,uslr, q

= i:

n R.

I n...

The modal.implitudcs are now converted to mnpl.iluuL'S in Lhl' ori1..:inal co-ordinate system by the relation

11.

'I

.1

• ju II

l.

Y IJ "I{

f **

I **

1*.

. '*1

• • 1 111111:~.

111:1 t*sp1111:;1! r,

.,tr11,*1111:*

..,v:.l**1w: l",1*1111f1*:.*,111,*;:, !!..!!.'.!:.!::.

111d Str1n*tur:il R<.*snons..:, Ameril:."111 Suc*i.*t::,,r :*%t*l*11:inic:1l J::m:inccr.s,

~

• y

• I l %0

• J T I - -,

. Arthur D Llttle Inc:

e 1l&is now 11r11v i cl,*:. a :a* t II I cl i =*I' I ;1n*11w11l :,

• II *

• I e

lttf **.1,*ii 11/

1111* II moJ,*s.

TI1csc individu.al sets of displ.ic,*m,*nls,*;an nnw h'"* ;1ppli..,I to th.:

system ~i:; cttuiv:ilcnt 1it.itiC' dr.flt>et.ions.

Th** rnrr,:::p,,nr!in:~ &U'lwnrk fnrci.:s nrc ubt:1i1111.~d hy thi.: mnml pru,*i.:dur,*:; 111*,\\IJLPll'I~.

It slm11ld he.*

noted Lh:it th'-' sLlffncss na;1Lrix to In* 11:u.-d fnr llai:: 11rrw,*1l11r,; must ht*

that resultins: when the rnwi:;_.:md cc,Jum.m; c-,*rr,*sronJinr. to ;mdaor p<'int!'-

arc drl~tcd, i.e.,

A I.I I>

,~

n1e reduced stiffness m.itrix (A-BE-JU) canrwt h(* us..:cl~ s;ncl! intt.:rior points (branch points) must be considered in tlic proc~ss of transferring interior loa*ds _and deflections from priint to Jwint.

ADLI'IPI::: ut*i] izcs the nclwurk forr.c. si.*ts I II l~l'llcr:1t.: sl rPSSl'S for each mode.

D1e upper bound for the StJ"C:SS JL*vc.ds.Jt aI~? pO int jn the system is given by the ~bsol.utc swnm;1tion <,f Li,.. ~ sl n*ss.::-: l!.. *n~ratL'U for each mode.

Such a summation assumes tlrnt Liu* r:nntribuUnns from l!.Jch mode reach their m.1ximum value at till' point in qu.. *stinn ~t the i:.amu tirr.e.

• Other methods of summ.ition m,,y he us'-'u,,,f C".Ourl-c~ r!,*:wnclini:: on tile degree uf conservatism u,1sirc.:d in UH: ~11.11::~;i;;,

,; s1;.:::**:,L**d,1!Lcr-native, for ex.Implc, might be the sur:1,,f thC" contribution of the funda-mental mode and the rms summntion of th\\! hii.:iicr mridc.

11 J -f, Arthur l) Lltt le Inc.

e e

(1o1l11*rl' *:*.

1r** Liu* u1111la I ru I 11111m; 11 I 111** 111,ul.il ru.il I i *. J ;111.(

I la**

I **I I **:,1u111*

Ill uiuJ; Lran~.Curm.JL.iuns ln:lwcc.:11 u 1

. mad "n' ii. ;111,I 'I *..

I 11 j ;uul I' * ;11111 l'

,md II

. 'j I>.

We further &!cfine the ~ener.ili.zc-u.incrti:i hy n

U,*c:iusu ur ortho,:unal ily, the* 1:c*nt*r:al i:.wcl i1..-r1 i.1 m:11 rix i:: :1 di:11~1111:il m:itrix, and hence ~y.bc writtt'n :JH where oki a Kronecker delta.

Because of the symmetry of tba inertia mnLrix, rn.*,

lJ thL' bilinear form ij m.. u. s.

l.J J

l.

bccomas i;ke. mij,;.ik. qk *i*j2. Pe.

I.

and bilinear the term l:

m ** u. s. *also ij l.J l

~-

Similarly*,* the quadratic forms J. Iii.. u.

lJ l

M p

=,. M k u

=*.

k i*

becomes I. mk qk pk k

u. anJ '

r.: ** s. s.

J 1.1 l

.l

!he kinetic energy may, therefore, be wl"ittC'n ns T

,: [ M 2 k k '11,.

I 01 '1

• ' I*. 'I k 1 'i-I,* I 10i:

L become

.i

( :....

I J -

,. k ;:: i i ) : i : u Frm:1 tl,c dciin.itiun of nurrn,,J modes, \\*.'l' h:n*,*

wile re

• ,j is :m cir.cnvcctor (mod,:11 c-olu:::n) :111d I\\'-.'

i:; till' c:r.. rrri=:pnnuing L

Anhur D ljttJe. lnc

J I I 1

.'r Now e

e For the nth mud.il column cJnd the nLh,*i,~1*nv:1l111*. this bccnmL*s *

r.

k.. i:-.

1.J Jn ij the potencial energy

>: l k 1 * -

(~, )- m.'. J *i*.

• 0 j

.J 11 l.J Jll

,: k. ;

j I I

,ii. **

lS

• {,:, ) 2 >: m **

11 I I j

(It.I,- i.

m.*.

• j II n

1.I ij

• jn

• ,*. I,...

given by V 1 I k..

is

= - r iJ

• 2 i.i l 1 J

l 1*:

• 2 i.i ns

.1.i 1..

C:

k.

2

. i i J

• ns

., - ~-l n

= I>!

11 II n.s

u.

.I qn..

qs in l.S

.i

• in CJ a )

'js n

• i:;

l. 'l"

= 2,.

,,,-_ m 5

q 11 '1.~

n n

ns

~

ns l

1

=

Q 2

n..

n n

l"hc ilppropriate form of Lagr:in~c 's cquntion j i,;

1o.*liid1 givl:S Lia! cqu:iLiu11s ol muLion, ;*1, +

I 111 1:

tj ~*

TIil:.solul ion cl Lliis cqu,aLit111 for Llw m,ul:11

,1111pi i r 11,lt- 'I,.

1:;

1

'I,

"( t) =

K l

I (r.-T),lT I'.'- I Arthur D. Little [nc.

• • ----* ---*-** **-*--*--*~*~- --**---*

e e

I j'

• t f

1-"rom the tr:111sfurm.*1L1un c.'qu;&l ion, tlw vc*rlur jik i:; r,*lal**rl 111 1111* v,*,*Lur

!:k which describes the.iccclcr:itions of th<"' b:isr of* thL' systL*m.

Thi"s

-1

-1.

I r&.*l.iLlun iN I'

.,. >: *I*

wlu*n* >: *I*

• I*

11... i,lc*nlily m.llrix.

k t

kt 1

1 k~

lm*

km" t

Hence, qk (t) * ~ i.£ ~*;! s1* (T) sin "'k -(l~T)<fr.

1 1 If [R (t)]k * -

S (t) sin ""k (t-1"),IT :nul the.? shuck inruL C

Wk l

0 displ.icemcnt {D )

• f (R (t) Jkf

• the in1ml sp,*ctr:i nrc.- cli.'f lnc*,I ns

f. k C

_ JD.'1X the m.iximum mudul us of the response v:i I uc RI', ( t).

The mu,l:11 amp l i l uclc.!'l now become qk ~ r *Jl~i (Dt)k, or.qk :;

.

: *j*k~. '{Jf_.)k ~?,ere \\-:I' is.in 1..'lemcnt

~

~

of the modal participation mntrix for mode k, m.iss point L*

In the special. case in which al 1 of the c*h*1:-u*nL!-. of Y...ire Lhc s.imC' and all of the (Di)k are the same, this exrrcMsion reduce~ to This is an alternative expression'for Youns's re~ult since it can be shown th.it tile modal r-'.lrticip,ition f.ictor, *,k, is C'rp1al to :: 'i"kc."

The modal particip.itiori factor is.tpproprfotc, however, only for the ~pcci.il case! wlwn nil nr LIii'.. 1,*111,*nl:: nf rlu* h:1s,* ;1,-,... l,*r:llinn Vl't.:Lurs s.Jrc Lhl! same.

Thi:; i:; 1i11I 1111* 1*;1:;1*,.. 1 1*11ur:;1*,

i11 1 !11 ****-

uina.:nsion.ll s_hoc..:k motjons with diffl*r,*nL sii..,c..:l; inpul:i (spl*,:Lr:i) in the

~arious _coordinate axes.

1...

I Arthur D Uttle Inc,,

I I JI e

i\\l'l'I*:: *** I:*; \\'

l>YNAZ*'lC S'rRE~S A::AJ.YSIS IS'i.\\Ul.1'11'1*:

by I. W. Dingwcll Arthur D. Little, Inc.

e On page 5 of the reference an expression for a set of displacements is developed for each masa degree of freedom and for each. mode:

• These displa.cementa are developed from che normaJ.j.zed aet of displace-ments q, as transformed by the modal matrix,:,in for mass direction i and mode n.

The displacements, Xi, represent the zero to peak Jisplacement of each mass degree of freedom when subjected to.:i shock lo.ii.ling. which is described by a (displacement/velocity/acceleration vs. frequency) res;;on::;e spectra.

The disp~acements h.ive a consist~nt set* of algebraic signs which define the mode shape of the deflected piping syst~m.

Reversing the signs of the.displacements gives the opposite peak modal deflec-rions of the piping system.

From this sec of modal displacements, X., the displa:cements of the

. 1 non-mass points are calculated. There are t1Jo types of non-mass points:

a) non-mass network points, and b) interior points within a pipe section.

Since ADLPIPE uses a transfer matri: technique for combining several pipe elements in series to formulatE: Lhe stiffoess of the section (a section is a series of connected elements), the non-mass network points

. are calculated f:l.rsc.

• then those deflections are utili.zeJ to calculate reactions at the network points. Fir.ally, internal forces,* moments, and deflections are calculated b)* transferring the initial boundary conditions across each member in' a section.

Thus, for each moJe, a set* of mumcnti:; j :; cal cul~i.'-'-d:

~!k. Jn

.. Genernli.z1.~d Piping System Respom::l' co i.round ShC\\ck SpC'c:tl*;1 h::,

Irving t-.'. Uin~ell. Arthur n. 1.jLtl,*. lnc., L1m!,rid;.;,', : 1:1!;s:1C'l111sett~.

\\'-I Arthur I) I 1ttlc l1ll..

\\.

e Al'Pl::N1>lX I\\'

IJl-.:VJ::J.OrHI~,* Of' HOUAL PART JC: I PAT J0:~.!!!\\_1]tE:__i,,~2!~ !=f::lE!f~~

TliRJ::E-UlMl*:NSIUN SIIClGK*rnru*r 10 l..UMl'l*:IJ uv:;,\\:-.ll: !iYSHM*

nu~ dcvc.*ln11mcnt of the modal p.'lrt i ci1ml i,an f:1rlor in l ht* :m.ilysis.

of tbe rcsponsu of :i lumped dyn.unic syst~'ffl Ln.1 nn1.-Jim.:nsinmil :;chock input is c:irri.:cl 011L by Yuun1:* i.n RL*f,*rt*ncr* I hy ;11*plir-:1li1111,,r l,:1~r:tn~<*'i.

eqm1tion with thu sys.tL'DI kin1.'tic :mJ pnt,*nti,al,*1ll*rs:i,*s,*:iq,n*s~;,,,1 in tar,m; of norm.:il cnuruin:ites.

In this :1pp1.*11di:*:. this cl,*v,*l,*111111.*111 is 1.*x-tc:ndcd to include.* th<" Jtl'ner:il ln:1di111: c*:u*H-in wlaic:la iliff,*r**11l ~.li,wk in-puts are allo~*=:3 :'..n each of Lhc syst**m cour,din:1t** :i:,,*s.

Tiu.* h,nninolc,gy uti lizccl by Young h.ii:. been ro llowcu LIi Llw *.*:-: I 1*11 t po~;s i I, I**.

• For the lumped syscum ucfined by til1., sy::1:m,.Lri,. itwT't. i,1 m:it:ri.x m..

lJ and by the stiffness matrix k.*, let u. be th,* cl:istic db,pl/lccmcnt in lJ 1

I

• tu d

• d 1

+

I I

I

• 1 I

11 t1c l.

coor 1natc, an ct u.

s.

l\\: tic :11*;n utr 1.1!-:p ;1n*ml*nt.

1c l

l elements u1 and s 1 arc, in the ~cner:11 cnse. :-;ix,.,J,*:i,*nt \\'l"C'tC'lri-- f,,r e:i\\*!1 m/l!:iS point.

The kinetlc energy T of Lhc sys Ll.'111 is s~i v1*n by

• r

• L t 2 ij

....!. i.

2 :.

l. 1

m.. (u. + s. ) (is. + s. )

lJ J

J I

l (m.. u. u. + m.. u. s. + m.. u. s.

JJ 1

l.J I

1 1.J I

.J The potential cncrlZY V is given 1.,y V = l

~

k 2 "

• .. u. u.

ij l.J

l.

J

+ m.. s.

I j l ;.. )

.I W,* iutru,hu;,* llal* uorm;il,*mini i11;1l 1*:: 'I (1) ;111cl. I' (I l !,,. I.Ii,* I in,*:IJ"

• 11 II l r ;1n:; f urma L j ui1:;

ll.

J.

( L) = i'.

n

'1 i l1 II

( l )

s 1.

(t) = f. ~-.

p (t)

J n n

n

• September 30, 1974 I\\"- I Arthur D Little Inc

,

• I e

where

.e k = orthogonal axi!; (X. Y, Z cnurdin.:1tc) j

• earthquake direction (X, Y, Z axis response spect~a) n

• mode -

With a normal mode analysis, all coup] ing and phaJ1;1.* relationships I

between modes are unknown. -* However,* since theRe momenu have algebraic signs and refer to a consistent position on the piping surface, the question of how to sum the modal moments arises.

The present version of ADLPIPE assumes that ear~h motion is oriented along n single vector and i~ composed by i1 spectra with cnmponents in the three orthogonal axes. Therefore, in.:1 single mode, the piping responds "in phase" 3

M nk (Equation 1) and the algebraic: sum is t.:iken of the motion which results from the single earthquake~

The response is in<lapen<lcnt of a.~is orie~tation.

Since there is no phase relationship between moJes, a mean summa-tion must be taken.

The present-version uses the square root sum of squares.

3

( !

(Equ:1tion 2)

There is an alternative. technique '-hich impl ~ es that closely spaced modes are coupled and are taken. to 'he in phase.

Therefore, when that occurs, the square root sum of s ~uares is taken of tbe 3hsolutc sum of the closely spaced mc,<lal nn-ient'.!..

For instance, modes 1 and 2 ~re c:lo5cly spaced J

3 3

((,.

I 'f I

I I

<\\

I:)

(Equn ti C\\n 3)

=

• kj 1 I + -

I~!.. I

) * +

.. k j=l j=l r'.J -

j=].. j -~'

V-2 ArtJmr I) I 1ttk*_ h'lL

e e

The test for the closely spaced moJcs is:

(f2 - fl) if---..---.

fl

< k, then the bandwidth factor (k~ for these modes

~l cause the program co form an absolute sum.

(This type summation must be requested of the progt'am by the analyst.

At present, the factor IC in percent is entered in the Z2 field on the SllOCK card. If Z2

• 0.,

then equation 2 is utilized.)

ALTERNATIVE SOLUTIONS A conservative assumption is that the vibratory energy in an earthquake is random and the component moments along each axis are independent of one another. Realistically, the earthqu.:ike acts as three different earthquakes, with the axis orientation a variable.

Therefore, since phase relationships are unknown, a mean solution is taken independently-for each shock direction.

In mode n, 3

Mnk * ( t jc:l 1/2 (M

)2) kjn (Equation 4)

Following the square root summation for the modes" ton max (Equation 5)

Since the absolute sum is overly conservµtive, an alternative is to take the maximum modal response plus the square root sum of the square cf the remaining moments.

l'1

-1 max 3

1/2 IM.j I>

+ct. c r 1-t. >:2>

""k m max 1

. 1 ltJm n*

J*

(Equation n)

Each of r.hese al.ternative solution sunmation scheru~s or var:i..:l Lions there-on can be insert:ed, upon request, into the Al>LPlPI:: proi:r;1.""::.

The resultins stress analysis ii:: dependent on the su::=.:i tiori of the modal moments.

The example siven here is not a statistic~l m~~n but certainly indicates that the present versio~ cf..;DLPIPE is unconscrvative but re_µist:ic.

As a consensus is r£'.:1chcd 1 oLhcr su::::.at.lt.*:~. r~"'c!miques will be introduced.

\\"-)

Arthur 1) I.ittlc. Inc

,111,ll' I ll',,I I

&*. :-.,I ;.. p l,~

I Al i,uint zero in,\\SHE *section 111 Sample r l*>blcm (C:las~ 1. Class 2)

(

Hx

• 1y Hz 111\\ldl*,,,... l Shock Dir.

X

-47467

-1297 67221 y

-153624

-4199 217557 z

-27185

-743 38498 olgl..'bralc sum (Equation 1)

-228276

-6239 323276 I

SRSS (Equation 4) 163071 4~57 230936 mo,lt~ n =2 Shock e

Dir.

X

-27343

-5128 2882 159117 29843

-16774 y

I

'\\

z

-851446

-159690 89758 1,1.whraic sum (Equation* 1)

.. 719612

-134975 75866 SRSS (Equation 4) 866617 162535 91357 r-:

I l:-

lllllJC 3

Shock ll =

Dir.

X 101890

... 3195 1426756-i y

-29914 *

-938

-418883 z

-8862

-278

-124098 I

j

,\\ l 1*,,,.1, r a I l: ~;um (t-:,111atlon l) 63114 1979 IHL.1725 e :

~;I: 5 5 (Equation 4) 106559 3341 1492144

--~-~---*..==~-*-

--c:r *--*-, * * * * * * *~; 2* ~ ~ *:r;; I>112

• • =....

r -

TOT.\\I.. mom~nt computed by R,\\'fll.J X

y Z

SRSS of a)g. sum (Equat Lon 2) 757641 135133 944098 1218031 1.0 SRSS l1 f SRSS (Equation 4) 828241 162630 1512670 1161701

. 1.44 C

Absolute sum of a lg. sum (Equation t) 1011062 143193 1282867 1639664 1.34

,\\b~o]utc sumo( SRSS (Equation 3) 1136247 170333 1814437 2147615 1.~6

'-...... r.

Nax. + SRSS of nl~. sum (Equntiun 1) 956512 141520 1215783 1553406 1....

K

-:ax, + SHSS of snss (Equation 6) 1061416 168)05 17110499 2045~11

-. __, *- -- *------*---~---**-

~----.-----------*--:-.-*---::....:. - -.. -

1..

QUESTION 13:

ANSWER:

e We understand that certain shortcomings might have been experienced with respect to informing all members of the Corrmission*of the development leading to orders for the shutdown.

Please clarify this matter and indicate any stops taken to improve conmunications to the entire Comnissfon. ______________ _

The problems encountered regarding infonning the Co11111issioners in a timely manner of important matters has been addressed and rectified. The Director of each NRC Office has bee~ instructed that each Commissioner is to be promptly and individually notified in such situations. It is our intention that this occurrence will not be repeated.

Regarding the particular instance at hand, the five plant shutdown, the fo1 *1ow-ing infonnation is supplied for clarification.

On Friday, March 9, the staff was aware that an area of concern existed regarding the seismic design of cer-tain nuclear power plants. Because of the preliminary nature of our infonna-tion, the potential severity of the problem was not identified at that time.

However, the infonnation available to the staff had been communicated to Chainnan Hendrie.

On the following weekend, NRC staff members went to Stone and Webster offices in Boston for further information. It was at this time that the potential magnitude of the problem was fully recognized and initial steps toward issuing the show cause orders were taken. Because of the short time period between our recognizing the need for the orders and their issuance and because much of what precipitated the final actions occurred over the weekend, the full ColllTlission was not kept properly informed.

The response.

to Question 9 provides a detailed chronoloqy of these events.

ED~UNO s! ~

MAJ.HE ROBERT T. STAFFORD, VT.

A

/,

J°'"t,."U.JIGS R.AHOOL.PH, W. VA., CHAIRMAN e

MIKE GRAVEL. AJ..J,..SKA..

HOWARD H. BAKER, JR., TENN.,

1..1..0'fO BENT~~

PETE V. OOMENICI, N. MEX.

QUENTii: NJ &IR::lOO<. H.. OAK.

JOHN H. CHAFEE, R.I.

JOHN C. CULVER.. tOWA ALAN K. SIMPSON, WYO.

=EL~,::..=....... H..Y. LARRY PRESSLER, s. DAK,

~ Cnifea..$£aJe$,$ena£e

.....= ~~.~~":;".. ~~R=:TOR COMMITTEE ON ENVIRONMENT AND PUBLIC WORKS Mr. Joseph M. Hendrie Chairman WASHINGTON, 0,C, 20510 June 26, 1979 U.S. Nuclear Regulatory Cormnission Washington, D.C.

20555

Dear 1'fr. Chainnan:

Please provide responses to the attached follow-up questions to the Subcommittee on Nuclear Regulation's Hearing o.n :March 16 regarding the shutdown of five nuclear power plants because of an error in the analyses of the seismic design. So.that the record may be completed, we would appreciate receiving your responses by July 20.

Sincerely,

~

ll y-1:,,uf:J" Gary Har-("_

Chairman, Subc mmittee on Nuclear Regul ENCL.

_____ J_

• L..........*.

l-***********

t~~~~~~~~:}pr

~------------ -

L:::::*::::::~:

[:;;;,:~~~~~

• --==---------

r-_._._._.,,,:::::

b.~JIIHHH h~~~:~Hfff fr

~:::*.--*-*.:::::.:

~::=.=.=.=.=.:::::~

• • • • • • • • • u****

r -:-:-:-:-:-:-: :::_=_= _:_

[:::::=.=.=::,:

I----------.--~

.... :::::::::1

'.:N~jji:f f N1

1

'f:~32{J

~~~-;-;-:-:-;-L:]

• • .*i

.... **-*** --~

.*.'::*.*:::::.J Ht!~}}~ r

~-. .I. i, *. L .... ~ i. e FIVE PLANT SHUTOOWN

1. When performing cost/benefit analyses of alternatives in NEPA reviews, how does NRC factor into those analyses costs such as those entailed in shutdowns (whether voluntary or by order or license conditions) of reactors because of safety concerns?

2.

How has NRC assured that the codes being used in the reanalyses of seismic design produce valid results?

3.

What steps have been taken to assure other computer codes currently being used for reactor designs do not contain errors?

4.

Please list each reactor which has been fotmd since March 13, including the five reactors which were the subject of the hearing, to have had an error in the s*eismic analyses of plant design. In your response, please include: (a) whether the reactor was shutdown because of the error; (b) whether the shutdown was voluntary or by order; (c) the systems involved; (d),~1ether the-systems are safety related or non-safety related, and -(e) the resulting corrective measures if any. 5.. (a) M1at technical standards/methods are being used to determine the adequacy of design for seismic events - those existing at the time the 5 plants were licensed or those in existance at this time? If the fonner, please describe: (b) the differences; (c) the rationale for not applying modern standards, and (d) a brief assessment of the relation between the existing seismic designs for the? plants and the existing standards.

6.

(a) How do the perc.eived risks associated with the error in the seismic design of the 5 plants compare with those associated with the Babcock and Wilcox plants during the first five weeks following the accident at Three Mile Island? (b)What factors led to the shutdown of all of the former within a few days of learning of the shortcomings, while some Babcock and l'vilcox plants never were shutdown? I

I-.*: - .i

• I-...

e Page 2.

7.

(a) 1~1atare the recurrence frequency and magnitude of the design basis and operating basis earthquakes at each of the 5 plants? (b) Based on the reanalyses using acceptable procedures, what are the recurrence frequency and magnitude of the earthquake that would have resulted in stresses above the allowable limit prior to any plant modifications.

8. What are the estimated costs of the shutdowns of the 5 plants in terms of dollars and barrels of oil? The underlying assumptions should be stated.

9.

In the March 16 hearing, Mr. Denton remarked that much credit for bringing the computer error to-his attention goes to the diligence of an NRC inspector who.pursued the discrepancy in the results of the old and new codes. Please provide the particulars in a chronology of the surfacing

• of the discrepancy and an assessment of the reasons for any delays.

10. Please provide available information on the recent earthquake that occurred in the vicinity of the Maine Yankee plant. How does it compare

,dth the operating basis earthquake.

11.

One of the plants ordered shut down is the Surry Plant which served as the model PWR for the Reactor Safety Study (RSS). The RSS included an extensive design adequacy study. (A) What was the finding of the study team with respect to seismic design of Surry? (b) What are the ramifications with respect to future quantitative risk assessments ?

12.

Please list all nuclear powerplants that have been exported from the United States that were designed with the aid of the erroneous computer code involved in the five plant shutdowns.. -

13.

W~ understand that* certain shortcomings might have been experienced iv1th_respect to informing all members of the Commission of the development leading to orders for the shutdm<JT1. Please clarify this matter and indicate any steps taken to improve communications to the entire Commission.}}