ML20207J793

Fort St Vrain Fuel Element Dynamic Response

ML20207J793
Person / Time
Site:	Fort Saint Vrain
Issue date:	07/09/1986
From:	Anderson C, Bennett J LOS ALAMOS NATIONAL LABORATORY
To:	Heitner K NRC
Shared Package
ML20207J773	List: ML20207J769 ML20207J779 ML20207J793
References
CON-FIN-A-7290 TAC-49055, NUDOCS 8701080597
Download: ML20207J793 (14)

Text

- . . . - . . . .- - -

'[-

Enclosure 2 O

't i

a i FORT ST. VRAIN FUEL ELEMENT DYNANIC RESPONSE l

NRC Fin No. A-7290 July 9, 1986 s

Los Alamos National Laboratory Joel G. Bennett Charles A. Anderson I

J Responsible NRC Individual j Kenneth Heitner Prepared for the i U. S. Nuclear Regulatory Comission Washington, DC 20555 4

l i

I I

i 1

l l

8701030597 861230 t

gDR ADOCK0500g7 m . . . , - - y,, -

---

w- - ,- -r - ,w ,--- .--r . ,-_m,- - ,-w- -- ,m--- --

l l

,/ ,

j i

/

i

/

/ .

l NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product or process disclosed in this report or represents that its use by such third party would

_. not infringe privately owned rights.

l l

l

TECHNICAL EVALUATION REPORT ON FORT ST. VRAIN FUEL ELEMENT DYNAMIC RESPONSE by Joel G. Bennett Charles A. Anderson BACKGROUND Following evaluation of the core segment 2, personnel of the Los Alamos National Laboratory, acting as technical consultants to the USNRC, raised the issue of the structural integrity of the cracked fuel elements under dynamic loading conditions such as during a seismic event. The licensee responded to this concern in a submittal to NRC (Docket No. 50-267, Ref. 1) dated August 13, 1984. The purpose of this Technical Evaluation Report (TER) is to report Los Alamos' evaluation of the submitted material. This task had four parts:

1. Review of the licensee's submittal and related material on the re-sponse of cracked fuel elements under dynamic loading situations.

2. Evaluate the licensee's justification of fuel element adequacy, based on the above documents and previous fuel element experience.

3. Determine if more extensive analytical and/or experimental research would be required to resolve the dynamic loading issue.

4. Submit a technical evaluation report to the NRC.

REVIEW OF THE SUBMITTED RESPONSE (REF. 1)

In the review of the licensee's response, several questions and comments have been generated that are covered in the following section of this TER.

1 l

The original issue of structural integrity of a cracked fuel block under dynamic loading conditions such as in a seismic event was primarily raised be-cause of the unusual construction of the FSV core. Because the method of con-struction requires stacking the graphite blocks into the core with sufficient clearance between fuel elements, gaps of varying size exist between the core elements and fuel columns. At one time for the FSV core, cumulative gaps on the order of several inches were possible. The latest value in the FSAR is

~

reported to be "slightly less than three inches" (FSAR 14.1-6 REV 3). Since installation of the top core region restraint devices, this cumulative clear-ance gap size has changed (i.e., it is more nearly uniform) at the top core plane.

It is difficult to see how the core midplane clearance gaps would be affected by the top region constraint devices, however.

Because of the presence of gaps between fuel element stacks, the dynamic l loads of concern are the interelement impact loads that will occur during a seismic event. The licensee indicates in the response that a nuximum load of 1500 lb has been determined to act on a fuel element during a seismic event.

It is not clear from the submittal how this number was determined. however, Ref. 2, "LHTGR Graphite Fuel Element Seismic Strength,' deals with this impact problem extensively.

Beginning with Newton's Second Law, the maximum impact force F or

_ max interelement impact (assuming a half-sine impact pulse) can be shown to be I T F =

2MAV m

max T 2M c i where M is the mass of each impacting element av is each element's change in velocity during impact, Tc is the contact time during impact.

Let F =

1500 lb (as given in the applicant's submittal)

W =

300 lb (approximate weight of fuel block)

T =

1.2 ms (from Ref. 2, p. 42 and supported by Ref. 3).

Then i

39 2 (1500 lb)(0.0012s) 2(300) lb (386.4 in./s )

or av = 1.16 in./s.

This small change in velocity during impact is implied by the licensee in 4

the submittal. On the other hand, Ref. 3, "HTGR Fuel Element Collision Dyanmics Program," which reports on one aspect of the extensive test program that considered HTGR-fuel element collision dynamics indicates *50 in./s repre-sents a maximum expected value of fuel element impact velocity for a design l basis earthquake (DBE)."

In light of the above calculation, the quote in Ref. 3, and the reviewers knowledge of extensive consideration given by Ger.eral Atomic (GA), NRC, Los Alamos, and others over the years to this problem, the following question is raised. Is the 1500 lb used by the applicant in his response a maximum cred-

~

ible load for an FSV cracked fuel element during a design basis earthquake?

References 4-6 are some examples of GA's analytical and numerical model-ing capability that was developed over the years to deal with the general problem of HTGR core response to a seismic event. There are several other reported efforts in the literature, including those of Los Alamos (Refs.

7-9). Apparently MC0C0 or C0CO or other computational methods have not been j

used by the licensee or its consultants to establish the impact spectra and 4

the interelement forces for a fuel element during an earthquake.

One of the outputs of such codes includes the dowel-socket forces that occur during the seismic event. If elements in adjacent vertical planes are moving with different velocities, a dynamic loading of the dowel-socket system i will occur. It should be noted that the concern is not failure of the dowel-l socket system, but rather what is the effect that loading a cracked element through the dowel-socket system will have on that element? It should also be noted that the licensee's "relatively simple stress cone" is a simplification

of the results of a complex state of stress (as opposed to uniaxial tension).

In order to get a handle on the interelement forces acting in the core of the FSV reactor, we have constructed a relatively simple lumped mass model of f

,.e. - - . ., - < - , -- , , - -- -. - , - - , , . . = - .

a graphite fuel / reflector stack sitting on the core support system as shown in Fig. 1. The column is pinned at the top, simulating the effect of the region constraint devices. Interelement springs, representing the dowel pin and socket arrangement, connect fuel elements together; a lateral spring connected from the core support block to ground stabilizes the column. The model was excited at the core support block with the horizontal component of the 1940 El '

Centro N-S earthquake adjusted to a peak acceleration of 0.2 g, which is twice the peak acceleration of the design basis earthquake for FSV. We used this multiplication factor to reflect the fact that forces will be amplified through the Prestressed Concrete Reactor Vessel (PCRV) (see the design basis earthquake response spectra Fig. 14.1 of the FSAR). l l

The velocity response of the midplane of the stack of blocks is shown in ,

Fig. 2--a peak velocity of about 4 in./s was calculated. In Fig. 3 is illus- ,

trated the time variation of the velocity of the bottom fuel block in the stack. l Thus, during our postulated seismic event we would expect on the order of 10 impacts with peak velocity of about 7-8 in./s.

Reference 2 concerns itself also with several aspects of seismic fuel l

element strength including the effect of multiple impact loadings as in a seismic event for flat-faced impacts and cumulative damage. For example, Ref. [

2 shows that 210 impacts at relative velocities of 58 in./s are required to initiate failure for a control element being hit by a reflector and af ter 376

[ impacts complete failure had occurred. These conditions of velocity and number of impacts for block failure are far in excess of what we calculate for the impact conditions for fuel blocks under a design basis earthquake, i IS MORE EXTENSIVE ANALYTICAL / EXPERIMENTAL RESEARCH NECESSARY TO RESOLVE THE DYNAMIC LOADING ISSUES?

In an effort to answer this question Los Alamos has performed a finite element parameter study under various loading conditions and looked at the effects on Weibull element failure probability estimates. This study was not meant to be either exhaustive or to represent bounding calculations, but its purpose was to give additional information on which a judgment regarding addi-tional analysis can be based.

I The finite element calculation in this study modelled a portion of the i dowel-socket area shown in Fig. 4. Two different studies were made using the ADINA-T and ADINA finite element codes.

In the first study, the model was used to evaluate the effects of flat-faced impact by first calculating the thermal stress field using a relatively cool bypass flow and a surface heat transfer coefficient of I W/in.2-F.

This combination shows that an interior web has a relatively significant prob-ability of failure (10-2). (Thermal stress calculations of this nature were discussed in a TER on the " Fort St. Vrain Segment 3 Restart," Ref. 11.)

Next, cracks were introduced in those elements that have the largest prob-abilities of failure. This calculation showed that the effect of crack initi-ation is to provide limited stress relief in that overall probability of fail-ure decreased slightly. However, in the region ahead of the crack, failure 3 probabilities increased by an order of magnitude (0.18 x 10-4 to 0.13 x 10-3) because of the increased stress level near the crack tip. Finally, for the cracked thermally loaded model, the "B"-face (Fig. 1) was additionally loaded with 20 and 40 psi pressures simulating maximum impact forces from velocity changes of 7 in./s and 14 in./s. These velocities were picked to be about the same as the expected PCRV velocities during a DBE. Flat-faced impact areas of 450 in.2 were assumed for these loadings. The results, which are shown in Table I, illustrate that a 15% increase in the probability of failure can be

, expected under a combined thermal loading and a 14 in./s flat-faced impact.

Higher impact velocities will, of course, increase this failure estimate.

The second condition studied simulated the effect of the dowel-socket system by imposing a loading from the interior of the # dement. For this study, the boundary conditions on the sides of the mesh were modified to reflect that a uniform elastic material constraint would be supplied by the adjacent mate-rial. This constraint was simulated by supplying a set of constant stiffness springs at each node to " ground" along both of the Y = a constant exterior surface of Fig. 4. The springs act in the Z direction. Then the mesh was loaded with a uniform pressure along the upper Y = 1.95 surface acting in the negative Y direction to simulate the stress condition resulting from dowel-socket impact. The load magnitudes must be estimated, because the dowel itself is not simulated in the model. A brief load-parameter study concluded this effort.

r I

i TABLE I &

FLAT FACED IMPACT STUDY r  !

r Loading and Impact Velocity Weibull Probability '

Case No. _ Geometry (in./s) of Failure i

1 Thermal only, ,

no crack 0.98 x 10-2 )

2 Thermal and web cracks initiated 0.96 x 10-2 l

[

3 Thermal and web 7 0.10 x 10-1 cracks  !

4 Thermal and web i

14 0.11 x 10-I cracks ,

The loads applied were estimated in the following manner. If an upper and lower dowel and socket pair are assumed to share equally in stopping a fuel element moving at 15 in./s, with d,wel-socket stiffness of 30,000 lb/in. '

(Ref. 12), the maximum force on the dowel-socket system is about 3,000 lb. If j

~ this force is distributed equally over a reduced (because of fuel and coolant I

' holes) area (say 225 in. ) the pressure would be on the order of 10 psi.  ;

Although this analysis is oversimplified it does give a method of generating l

the approximate stress that could arise from dowel-socket loadings during a seismic event.  !

Comparable results for the previous case of flat-faced impact i are shown in Table II with thermal only loading, thermal with crack initiation, [

and thermal with cracking and with 10 and 20 psi. For this parameter study a j

crack was introduced in element 230 (Fig. 4) and assumed to propagate in the

{

direction of the upper fuel hole, similar to the failures seen in the segment -

2 fuci elements. As can be seen from the table, less than 1/2 of 1% increase '

in failure probability occurred for a 20 psi loading on the upper face. Cau-tion must be used here in interpreting these results; however, because of the simple way the estimate has been made of how much of a dowel-socket interaction is necessary to produce a net 20 psi stress state pushing outward along this surface. In addition, the stress relief provided by the simulated material

constraining springs is probably too high. This study must be accepted as a parameter study to give a feel for the various effects. A true analysis of this effect will require a 3-D model.

TABLE II LOAD ASSUMED FROM DOWEL-SOCKET REGION Loading and Loading Weibull Probability Case No. Geometry (psi) of Failure 1 Thermal only, - 0.180 x 10-1 no crack 2 Thermal and web -

0.204 x 10-1 cracks initiated 3 Thermal and web 10 0.205 x 10-1 cracks 4 Thermal and web 20 0.206 x 10-1 cracks CONCLUSIONS AND RECOMMENDATIONS

l. We feel that the licensee did not do an adequate job of addressing the question of cracked fuel block behavior during a seismic event, consider-ing the analytical tools that have been developed over the years to treat this problem.

2. Our calculations show that there is a large nargin on uncracked fuel ele-ment strength during a design basis earthquake event at the Fort St.

Vrain reactor.

)

1

3. Our calculations show furthermore that the margin is not significantly reduced even when there is a crack present in the fuel block.

4. Therefore, we feel that more extensive analytical research is not needed to resolve this problem.

REFERENCES

1. O. R. Lee (PSC) Letter of E. H. Johnson (NRC) " Response to NRC/LANL Concerns on Cracked Fuel Elements," Docket No. 50-267, August 13, 1984.

2. L. Sevier, "LHTGR Graphite Fuel Element Seismic Strength," GA-A13920, April 30, 1976.

3. S.'M. Rodkin and B. E. Olsen, "HTGR Fuel Element Collision Dynamics Program," GA-A14728, September 1978.

4. H. D. Shatoff, " Approximation of Corner and Edge Loads from HTGR Core Seismic Analysis Codes," GA-A14247, April 1977.

5. R. W. Thompson, "MCOCO, A Computer Program for Seismic Analysis of the HTGR Core," GA-A14764, April 1978.

6. N. D. Richard, "C0CO, A Computer Program for Seismic Analysis of a Single Column of the HTGR Core," GA-A14600, Febrasary 1978.

7. J. G. Bennett, "A Physically Based Analytical Model for Predicting HTGR Seismic Response," Japan AEB/NRC Seminar on HTGR Safety Technology, September 15-16, 1977, Brookhaven, N.Y.

8. J. G. Bennett, R. C. Dove, and J. L. Merson, " Seismic Response of A Block-type Nuclear Reactor Core," Los Alamos Scientific Laboratory report, LA-NUREG-6377-MS, July 1976.

9. J. L. Merson and J. G. Bennett, "A Computer Method for Analyzing HTGR Core Block Response to Seismic Excitation," Los Alamos Scientific Laboratory report, LA-NUREG-6473-MS, September 1976.

10. J. G. Bennett and R. C. Dove. " Proposal for Analysis of HTGR Core Response to Seismic Input," Los Alamos Scientific Laboratory report, LA-5821-MS, January 1975.

11. C. A. Anderson and D. R. Bennett, " Technical Evaluation Letter Report on Fort St. Vrain Segment 3 Restart," Q-13:83:262, April 20, 1984.

12. D. D. Chiang, " Fatigue Tests of Dowel-Socket Systems," GA-A13861, June 15, 1981.

l l

l t

. 4 4

il ~

RB

-dREP-

. FB e -30 000 lb/in. EACH FB DOWEL AND SOCKET l m

FB 250in.

• l FB e RB - REFLECTOR BLOCK FB B - FUEL BLOCK l CSB - CORE SUPPORT BLOCK

• CSP - CORE SUPPORT POST FB 6

• e15 in.

RB

-N- LATERAL s

SPRING x

44 000 %

- s CSB <

s000 %ts n

7 a(t) 3 (( y g y D((V'~~ l

- ~ 8 in, t 48in. .

t CSP V . .

\

///////W/////// \

l Fig. 1. Fuel / reflector block column for seismic response calculation. I i

)

_9_

e ,

s

. a a - -

h -

l-

\

g - -

~

Fig. 2. Velocity response of middle fuel block in model from 0.2 g El Centro earthquake. Velocity in in./s.

I 1

l I

l I

l

+ 10 -

l l

ll l

- 10 ,

=

l ===>

1 fig. 3. Velocity response of bottom fuel block in model from 0.2 g El Centro earthquake. Velcotiy in in./s.

1

i l .

200sCLeQFt StX E T M ER , _

l Pressure loading ji I = 1 95 g

j 1

i g/ '"' '

a > a:-

N rr

! I! emelant $ Suel ble Crack location

[Ef I one i am Ij;< . mi 9(

Case 1

• • ) *4F '

1 4 I p m ,

l g

as, al < .s r

& t ,

$ Mb& & *'t

! A i 1 I '

, ec \ N , .

l aFM Crack location -

g; afe/afa Case 2 ,,,,3$, '

Springs to

" ground" on ?

'i -

3JMM

'*1 "I' 858'8 '8 N

y = constant

• 8 j i nis r J

' II' 41 eatsa i

Case 2 g 'PTT' /988 : sj a(

iL 79Vg i8VYN

,,, ,,a 8 CT- i 855; )1 8" r 4 ,

a y, l ] g:g j Springs to

,8! \ '0 8" 84 if T '1 ' '

i ~ " con [ tnt

• l Coolsat il hi a ll Lu b ,,3 mi, 58  ; Case 2

- I: Cnennet

's a a WF , F9 l

%]

', a m .

fm/

j

. a 8' ["I g

L. = n N

\s f-

• M 4:

a .

e ",9 TN m y y* /as as a i TTA. 1

. si t " ;; , . . = Y -

- 1 . e =, 4 = =

s "

ti Ogji a m # # # g g, #

c

, a , , ,

. e . . . s, A t' B-face pressure loading

Case 1

! Fig. 4. Finite element model used for parameter studies.

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