ML20205B008

Mixing Phenomena of Interest to Sblocas, to Be Presented on 850623-27 at Specialists Meeting on Small Break LOCA Analysis in Pisa,Italy

ML20205B008
Person / Time
Site:	Crystal River
Issue date:	04/30/1985
From:	Iyer K, Theofanous T PURDUE UNIV., WEST LAFAYETTE, IN
To:
Shared Package
ML19284E734	List: ML20204F041 ML20205B000 ML20205B008
References
NUDOCS 8506050228
Download: ML20205B008 (30)

Text

f s

6

~

,o'

,r' j"',

1 s'

^

MIXING PHEN 0 GENA 0F, INTEREST TO SBLOCAs l

by T.

G.

Theofanous and K.

Iyer School of Nuclear Engineering Purdue University West Lafayette, Indiana 47907 ASSTRACT

$=all break LOCAs may lead to flow stagnation with high primary system pressure. The thermal mixing of the high pressure safety injection under such conditions has important implications for reactor vessel integrity. Experimental an*d analytical work in this area during the past few yearsskave pro-duced a ecmprehensive understanding of the prob-lem. A summary account of these developments, and some new tests of predictive capability are presented.

Invited paper for the " Specialists Meeting on j

Small Break LOCA Analyses in LWRs," Pisa, Italy, 23-27 June 1985.

l l

l April 1985 l

6 l

e.

l g

r-

-l-

+ y' e

(, '.

I.

INTRODUCTION

~

refef' to e,, process leading to In a broad sense, mixing s

the annihilation of spatial gradients in any scate, thermal, or kinematic propbrty within a fluid volume. The opposite of

,_ mixing is called 's t r a t i f i c a t.i o n.. Buoyancy forces act to create,and preserve s t r a t i *: fe a t i o n. Inertial forces and associated instabilities act to promote mixing. The task of mixing analysis is to determine: (a) the length scale of inhomogeneity or flow patt'ern and (b) the degree of remain-ing stratification.

With prevailing low fluid velocities, S=all Break Loss of Coolant Accidents (S3LOCAs) exhibit strongly stratified behavior. One aspect of this behavior is associated with phase-change / separation phenomena and has long been recog-nized as the principal characteristic of SBLOCA events.

Accordingly, it has been studied in detail and its analyti-cal representation has served as the cornerstone of all SBLOCA simulation tools. The other aspect of this behavior is associated with thermal stratification / mixing phenomena.

These latter develop as a result of cold safety injection and, unlike phase stratification, have been left outside the

=ainstream of SBLOCA analyses. There are no indications at this time that this is inappropriate. However, the need to consider them has arisen in impostant, specialized contexts, and current indications point to further developments along such lines. Thermal mixing in the contIxt of Pressurized Thermal Shock (PTS) is a prominent example in this regard

~"

and is the topic to which this paper is devoted.

Our task is to quantify Etratification and associated cooldown effects owing to High-Pr'essdre Safety Injection (HP1). Such effects have been shown (1,2) to survive only in the absence of natural circulation loop flow. Among all potential PTS scenarios, therefore, the proble= is only relevant to 53LOCAs and, in fact, only to the small subclass of these events that exhibits stagnation at high pressure.

The mixing / stratification problem is, therefore, considered

.here in this narrow, but important practical context. Dura-tion of stagnation period (i.e.,

extent of the cooldown transient) and associated pressure levels are also important in putting the results of such studies into the proper risk perspective. These systems / scenario dependent aspects, how-ever, will not be considered further here.

After three fears of intense research activity we now consider the problem solved. Our purpose in this paper is to provide evidence that this is indeed the case and to present an overview of the developments along the path to resolu-tion.

o

'n'

~

~,-

2.

INITIAL DEVELOPMENTS Starting in the 1980-81 time (pame, both experimental and numerical simulation efforts were in - pl a c e. The numeri-cal simulations aimed for three-dimensional finite differ-ence models of the~ cold l'e g and downcomer and considered

_ turbulence generation and t rans por t-e f f ec t s. With various

)

degrees of involvement, t h e s'e' e f f o r t s included work with COMMIX-1A [3], TEMPEST [4], and the development of a new code, the SOLA-PTS [5]. The experimental simulations were 1/5-scale

• geometric model of the cold leg carried out in a and downcomer at CREARE. In this facility a few cooldown transients under Froude number scaling were reported early in 1952 [1]. although =ost data were obtained in the pres-ence of loop flow. Nearly perfect mixing and rapid cooldown were indicated. Code comparisons with these data were made with pre ference for the loop flow rune, which rapidly yielded a steady-state and showed generally good agreement.

To our knowledge, no comparisons with these or any subse-i quent transient cooldown runs have been documented. As a i

first attempt in this direction, a well-mixed control volume approach was proposed by Levy and Healzer [7]. This model I

was subsequently developed further [8] but remained empirl-

]

cal and of limited use. It served as the point of departure for the Regional Mixing-Model described in the next section.

t 1

In support of the three-dimensional code development efforts experiments w e r_e also c a r r ie d GEL a t SAI. The SAI

[

~ ~

facility also employed a cold leg /downcomer geometry although geometric similarity was abandoned in favor of a

~"

full height represen,ta, tion of a longitudinal slice of the cold leg. Mixing results in t he presence of loop flow [18]

as well as for transient 'ooldowns, [25] were presented.

c As it turned out, this cold leg /downcomer geometry was an inadequate representation of the practical situation.

This was pointed out by Theofanous and Nourbakhsh [6], who i

claimed that any experimental or analytical simulation must

?

also include the lower plenum, the pump, and the loop seal volumes, as illustrated in Figure

1. This is because at the

. low flow throughputs of interest, the system is elliptic, that is, it gives rise to recirculating flows extending to

}

all positions from which the injection location can be reached by any combination of horizontal and vertical upwards displacements. Any other truncation in the system i

gives rise to inappropriate boundary conditions, which in j

experimental simulations, as already pointed out [6], gives rise to ambiguities,.and which in numerical simulations may even prevent convergence. This ratter consequence was recently confirmed by Chexal [9). As a result, finite l

difference models become prohibitively expensive to run, and j

in fact, except for a single TEMPEST calculation [4], no j

other cooldown predictions have been documented.

i t

?*

-3*

/,P '.

. e W

A'nother difficulty with the large-scale finite. differ-ence models relates to the problem of numerical di.ffusion.

Given the i=portance of this issue T'or the proble= at hand, considerable efforts were made to' improve"pewfoy=ance in this regard. H ow e ye r, no comprehensive assessment in this area has been docu=ented as yet.

3. THE REGIONAL MIXING MODEL, REMIX The physical situation may be described qualitatively with the help of Figure 1.

Initially, this portion of the pri=ary syste= is entirely filled with coolant (water) at a te=perature near that of normal operation ( 550 K). The cooldown transie'nt is initiated by actuation of the HPI.

Typical injection rates are 10 to 20 kg/s although, depend-ing upon the system pressure, both lower and time-dependent injection rates are possible. Coolant exits at an equivalent rate through the reactor vessel, as indicated in Figure 1,

1 and eventually out the break. Dimensions and volumes typical of a 3,000 MW PWR are summarized in Table 1.

g The ensuing flow regime was first analytically esta-blished by Theofanous and Nourbakhsh [6], as schematically illustrated in Figure 1.

A " cold stream" originates with the HPI plume at the point of injection, continues towards both ends of the cold leg, and decays away as the resulting plumes fall into the d own c ome r and pumpittop-seal regions. A

" hot stream" flows counter to this " cold scream," as indi-

-r cated, supplying the flow necessary for mixing (entrainment) at each location. This-mixing is most intensive in certain locations identified as mixing' regions (MRs).'MRI' indicates the mixing associated with the highly-buoyant, nearly axisymmetric RPI plume. MR3 and,MR5 are regions where mixing occurs because of the transitions'(jumps) from horizontal layers into falling plume s. MR4 ' is t he region where the i

downce=er (planar) plume finally decays. Ter ec=pleteness MR2 is also shown, as the interface between the hot and cold streams, however, because of stable stratification, the mix-ing is negligible in this area. The cold' streams have spe-

'cial significance because they induce a global recirculating flow pattern with flow rates significantly higher (several-fold) than the net flow throughput (Qy,z). This keeps a major portion of the system volume in a well-mixed condition (at temperature T,in Figure 1), so that the uncle process

=ay be viewed as the quasi-static decay of the cold streams within a slowly varying " ambient" temperature.

~

.The quanticative aspects of this physical behavior may be found in the. Regional Mixing Model (RMM), [2,6,10]. This model integrates' local mixing behavior into an overall sys-tem response. Its basic ingredients may be summarized as follows:

- - _ ~,

m.-

+

-4

's, (a)

Plume mixing rates are consistent with data from ideal-

~

i:ed single plume geometries.

(b)

The dimensions of plumes and hot and co rd s t reams are not a rb i t r a r tl y spec *ified but rather are obtained as parts of the' solution.

o (c)

Countercurrent flow limitations bet ween the hot and cold streams within the cold leg ste observed.

The computation proceeds at two levels. The first is global and seeks to establish a "mean" system response, referred to above as the " ambient." The second is local and seeks to partition = ass and energy into the hot and cold streams con-sistent with mixing rates and countercurrent flow require-ments. The global computation also takes structural heat into account and proceeds from the initial conditions march-ing out in time. The local computation provides, at arbi-trarily selected times, snapshots of detail constructed upon the global results. In practice, it is convenient that these two levels of computation proceed in parallel. For constant properties and rough accounting of structural heat, even hand calculations are possible. For a full and detailed representation the REMIX code can be used [11,12]. The com-putational effort is trivial.

A nodal model with a flow regime sjm,ilar to that of Figure I was presented Tater by Oh et al. [133. The choice of control volumes, their sizes, and a number of empirical 7F parameters were adjusted to obtain agreement with one of the CREARE 1/ 5-s c. a l e t e s'e s '.' H o w e v e r,,

no other comparisons with experiments or any reactor calcula,tions have been reported.

To conclude this section, two refinements of the RMM,

~

which were made possible through the use of Purdue 1/2-scale data, will be mentioned.

(a) Mixing within the HPI line.

It was suggested [6] that for injection Froude number, Fr a pr, less than unity, back-

. flow of the hot stream fluid and mixing within the i n j e c t i,o n

-line sho'1d be expected. Purdue's initial 1/2-scale experi-ments confi med this prediction and provided the basis for taking this effect into account [ 14 ]. The approach is to define an effective HPI plume origin that moves into the injection line as the Fr decreases below 0.6, by a length un L y, the value of which, for 0.2 <

F r,:

< 0.6, is given by e

g the relation L,y /D 3-5Fr In the calculations, g,1 ge.

this additional length is used rn' con j unc t ion with the height of the. hot stream to obtain an estimation of the plume length in MRI and, from it, the entrainment. For Fr

> 0.4, this effect is negligible [11].

gy (b) Mixing in the downcomer.

A highly complicated three-

1

~ J

~.

,7^*

dimens'ional mixing pattern occurs in MR3. In the original formulation of the RMM, the approach was to conservatively neglect this contribution to the mining in the downcomer.

Rather, the cold stream exiting the c o l d '1'e g -was a s s ume d to form smoothly in t a ' a planar plume within the downcomer and to decay according" to t h e ci-t - G' t u r b u l e n c e model predic-

-~~

tions. Rapid decay was precitted nevertheless. The Purdue 1/2-scale tests indicated that within MR3, the cold stream entrains an equal volume of the ambient fluid [15]. Taking this into account, we modified the procedure as follows

'[ 11). The planar plume is taken to form within a distance of two cold leg diameters below the cold leg centerline and to be fed in equal volumetric flow rates by the cold stream and the surrounding hot fluid. The resulting temperature is termed initial p'lanar plume temperature, Tj. Below this point, the decay is approximated to that o.

a planar plume of initial width equal to the cold leg diameter, Det and 1.0 [6,11). It should be emphasized that since the cold Fr stream always adjusts itself to a Froude number value of

~0.5, as was indeed the case in the experiments, this empir-ically determined entrainment behavior should be generally valid.

4.

MORE ON FLOW REGIMES AND SCALING, NEWMIX Aside from the system truncaEion problems mentioned above, the choice of geometric s imilarity and Froude number as the principal scaling criteria is straightforward. In addition, of course, we have to be certain that as the size of the scale model decreases, the Froude number similarity does not force the experiment into laminar flow [6,11,143.

An even more interesting c on s id e r a-t io n relates to potential non-similarity owing to momentum flux effects [63.

Specifically, a concern was expressed that variations in the intensity of i= pact of the injected plume with the opposite cold leg boundary may lead to alterations in the resulting flow regimes. This point was addressed by the Pur-due 1/2-scale experiments'. For the vertically downward, low

'Froude number injections ( F r gpg

~ 1) of interest to Westing-j house and Combustion Engineering reactors, no discernible effect'vas noted. The reason is that the cold stream builds to a depth sufficient to prevent any wall-induced deflec-tions in the plume from entering the hot stream area [15].

This is consistent with the RMM bases and intentions. Furth-ermore and also in agreement with RMM..such highly buoyant plumes exhibit little, if any, in e r t i a, so that the angle of injection is immaterial to the flow pattern and amount of mixing observed ( 14 ).

On the other hand, for high Froude number injections,.

which are pertinent to the. Babcock and Wilcox reactors,I a

. E

.r* [^.

highly forceful jet, rather than a buoyancy-driven plume, is obtained. Under such conditions, independent of the'orienta-tion of injection, substantial agit.ation and local mixing a

owing co wall deflections and splattering..vould be expected.

The RMM is not directly applicable to this kihd*of situa-tion. However, sin'ce the' extent of mixing in RMM is also controlled by the' countercurrent flow limitations at the i

cold leg e x i t., it is reasonafie in such situations to con-sider applying the RMM with only this limitation to mixing.

that is, the mixing at MRI is considered to occur with the maximum entrainment allowed by the countercurrent flow pro-cess. Otherwise the computation proceeds as in REMIX. For detailed computations at this limit, the computer code NE2 MIX has been prepared. The results of such co=putations are in excellent. agreement with the few CREARE 1/5-scale and Purdue 1/2-scale data obtained at this regime [163 1

5.

THE CREARE 1/5-SCALE AND PURDUE 1/2-SCALE EXPERIMENTS 4

The first confirmation of the trends predicted by the 3

RMF came from the last round of the CREARE 1/5-scale tests.

These tests were conducted soon after the presentation of the RMM [17]. They were carried out in an original acrylic facility expanded to include the lower plenum, pump, and loop-seal components or several combinations of them. Both thermal and salt-induced buoyancy were utilized. Temperature transients at various locations in the JJstem were measured.

j

- ~

~

The essential aspects of the geometry are summarized in Table 2. The basic experimental conditions are compared with

}

those expected in a PRR, in Table'3. A visual appreciation of

~"

the scale-down in the facility..may be gained from Figure 2.

a Excellent agreement of these testJresults with RMM predic-tions has been documented [2,10].

The Purdue 1/2-scale experiments were aimed at explor-ing scale effects on the one hand and-at providing a more detailed data base, against which the internal consistency of the RMM could be checked, on the.other. They employed salt-induced buoyancy (allowing low pressure, acrylic con-

-struction). The expansion of geometric and other parameter ranges from those of the CREARE 1/5-scale tests'is depicted in Tables 3 and 4 The scale-down is illustrated in Figure

~

2. As shown, the lower portion of the downcomer and lower i

plenum.were distorted. This-was done to keep the height of the whole facility menageable and was demonstrated to convey no ill effects [2,15]. Measurements of instantaneous salt concentrations'and velocities were made by means of conduc-L tivity and hot wire probes. Fron~these, the mean values, as well.as.various quantities _ characteristic of the fluctuating components (i.e.,

root-mean-squares, etc.),icould be

~

deduced. The-probes were mounted on motor-driven traversing-mechanisms, and data were recorded at the race of 1,000 Hz t

n v,, - -

,,e-.

~

..s 7

J together'with the instantaneous probe locations. T h~i s pro-cedure provided an excellent resoluqion of the steep velo-city and concentration gradients in'the gold leg. In addi-tion, visualization studies were made through-dye injection and video recordings..In the absence of backflow in the injection line, (the case,Frun 0_. 4

'f a l l s in this category as per discussion in Se c t i. f'3) t h e experimental results were in' excellent agreement with the RMM (REMIX) predictions

[2,11,15]. For lower Froude numbers, a small correction accounting for this backflow, as mentioned above, was derived. Also as previously mentioned, a refinement of the initial planar plume conditions in the downcomer based on these data was derived.

6.

THE CREARE 1/2-SCALE EXPERIMENTS This facility was planned prior to the last round of the 1/5-scale tests mentioned in the previous section. It was intended to comple=ent them with regard to scale and wall heat transfer effects. It was also expanded to include all appropriate ec=ponents following the RMM and the last round of the.1/5-scale tests. The facility is depicted in its final configuration in Figure 2.

Its major geometrie characteristics are given in Table 5.

It employs thermally induced buoyancy and with its 200. psi pressure capability, a substantial af// range can be obtained without resorting to adding salt. The =ajor ~ operating parameters are compared with those of the other facilities and the reactor in Table

- ?"

3.

Temperatures, velocities, and wall heat fluxes were meas-ured at selected loc'ations t h r o u'gh o u t the system.

The first cooldown transients we're reported in late 1983. However, these data were considered ambiguous because of a spurious heat source in the lower plenum owing to the particular design of this component. The magnitude of this source was quant-fied in the experiments and, by taking it into account in the RMM, excellent post-test predictions were obtained. Nevertheless, because of this potential ambi-

.,quity these data, although in our opinion valuable, were not relea' sed. After the problem was corrected, the tests were repeated and reported one year later [19].. Comparisons of these test results for test MAY 105 with the REMIX predic-tions are documented in Figures 3(a) to 3(h). The agreement shown is similar for the other run, MAY 106, which was con-

-ducted at an injection Froude number of 1.0.

Furthermore, the measured velocit.i.es and wall heat fluxes imply a forced convection dominated heat t r a n s-f e r regime in the upper por-j tion of the downcomer, confirming the results of analysis (20].

7.

THE HDR REACTOR TESTS i

i

=~..

-8.

~

7"

,f'.

1 These tests which were conducted at Battelle-Frankfurt, provided the opportunity for double-blind pretest predic-j tions. This means that the first such tests in the facility were predicted from geometric data and experiment operating i

specifications. This assessment opportunity is.particularly important for two reasons. First, these are the first tests

_ conducted at full pressure.ang temp.e_rature and thus at fully' prototypic buoyancy and all o' t h e r thermal effects present.

Second, because of the much larger injection Troude number used in relation te U.S.

reactor designs, a much stronger stratification was obtained which provided a test of predic-tions much more stringent than tny previous experiment.

1 The relevant geometric data of HDR are summarized in i

Table 6.

The rel,arTon of this. reactor to a full scale U.S.

PWR is illustrated in Figure 2.

The key operating parameters t

are su==arized in Table 3.

Temperature transients at selected locations ~throughout the system are measured. It is i

hoped that

. eventually, velocities will also be deduced from correlation analyses of fast-response thermocouples. A sum-

=ary of all pretest predictions attempted in comparison with 1

the experimental data has just become available [213. Here j

only the REMIX results will be documented for tes T32.18.

l The comparisons for the other test available, T 3 2.15, at i

injection Fr 1.35 are similar.

)

The pretest predictions of the cold stream, hot stream, 3

~

initial planar plume, and ambient temperature transients are i

3 compared with the HDR data in Figures Cta$ t o (d). Excellent j

agreement is noted. The pretest predictions of the downcomer i

re plume temperatures below three cold leg diameters under the cold leg nozzle, unfortunately, contained an input error j

caused by the significantly laYger downcomer length to cold t

leg diameter ratio of the HDR comp'ared with U.S.

reactors.

The comparisons in Figure 5 were obtained with this error i

j corrected, so they cannot be labeled as " pretest." However,

)

to the extent that the calculations are based on well-defined planar plume decay [2,6,113, they still should be I

considered.as true predictions. Additional tests in the HDR are forthcomin~g.

3

6. THE IVO AND UPTF EXPERIMENTS Transient cooldown tests have also been conducted in a low pressure, transparent facility by Imatran Voima Oy (IVO) in Finland. This facility is a scale.model of the Loviisa i

' reactor. Its scale relative to the other experiments'is illustrated in Figure 2.

It employs bottom cold leg injec-tion (High Fr.oude numbers), and the lover' plenum is 4

separated into two compartments'by a perforated flow distri-

. E butor. plate. Its uniqueness is its representation of-one-j half of=the downcomer circumference and the associated i

l --..

3

+

,a.

+m

,n, nn.,.

+,

,,,. ~,

,n..,_,,-

.,. ~,.,, - ~ -. + - - -,

. = -.

1-s

,- /.g,,

number of cold legs. As in the CREARE 1/5-Scale tests, buoy-4 ancy is induced by salt and mixing is inferred from tempera-ture measurements. The RMM prediction of these' data [223 i

cannot be discussed at-this time because of proprietary con-7

- straints. However,' through

a. cooperative program with U.S.

4 NRC and Purdue, a new series'6f tests involving injection geometries typical of U.S.

reactors are currently in pro-gress.to explore possible plume. interaction effects in the downcomer. Results and interpretations are expected by the end of 1965.

i Finally, a full-scale test in the UFTF facility in Ger-many is being planned. The overall size is very similar to that of a PWR, a's shown in Figure 2.

However, the design pressure is only 20 bar, and the buoyancy cannot be fully represented. In addition to its scale, another unique and l

4 most attractive feature is that the facility lends itself to multiloop operation at full downcomer representation. Injec-tion geometry and run conditions are-presently in the plan-l i

~

ning stage with the help of REMIX and NEWMIX calculations.

i The actual tests will be conducted later in 1985.

?

i 9.

CONCLUSIONS Thermal mixing is_now fully recoggized as an important aspect of SBLOCA analysis in the narrow but important prac-tical context of pressurized thermal _ shock. In response to this issue a multifa_ceted, comprehensive, research program, on HPI mixing, has evolv,ed ovtg the past several years with 1

the support of USNRC, E P R'I, and U$, Industry. In addition,

~

international cooperative efforts.came into existence (US-FRG, US-Finland). In order to fully address the importat i

problem of scaling these programs included integral simula-i cions at several scales, both thermal and salt induced buoy-ancy, and-a variety of measurement approaches. This data.

base is culminating with the full scale tests at the HDR and I

UPTT. facilities.

U On the theoretical side three-dimensional finite i

difference codes were developed-in parallel.. However, the practicality of these tools in addressing this particular 4

problem seems to be limited. Rather,-a mechanistic, funda-mentally oriented approach (RMM, REMIX, NEWMIX) has emerged i

as most appropriate. Extensive comparisons with the f

comprehensive data base, including detailed' flow regime stu-l dies in visualizing low pressurt" facilities as well as L

double-blind pretest preictions of full-scale reactor vessel (HDR) tests indicate that both the scale of stratification and the i

_ degree of mixing can be predicted with high confi-dence. Sample-results for U.S.

reactor designs and parameter sensitivity studies have also been documented h1,20,23,243.

'l 4

/

-. /..o *

~

4 j

e As a result of this multipronged approach and after four years of effort this thermal mixing problem can be con-sidered solved.

10. REFERENCES 1.

Rothe, P.F.

and M.F.

Ackerson, " Fluid Thermal Mixing in a Model Cold Leg and Downcomer with Loop Flow," EPRI NP-2312, April 1982.

2.

Theofanous, T.G.,

H.P.

Nourbakhsh, P.

Gherson and K.

Iyer, " Decay of Buoyancy Driven Stratified Layers with Application *s to Pressurized Thermal Shock," NUREG/CR-3700.

~

3.

Marston, T.

et al.,

" Robinson 2 Reactor Vessel: Pres-surized Thermal Shock Analysis for a Small-Break LOCA,"

EPRI NP-3573-SR, Special Report, August 1984.

~

4

Eyler, L.L.

and D.S.

Trent, " Pressurized Ther=al Shock:

TEMPEST Computer Code Simulation of Thermal Mixing in the Cold Leg and Downcomer of a Pressurized Water Reac-tor," NUREG/CR-3564, PNL-4909, April 1984.

5.

Daly, B.J.

and M.D.

Torrey, "SOLA-PTS: A Transient, Three-Dimensional._ Algorithm for F Luid-Thermal Mixing and Wall Heat Transfer in Complex Geometries,"

{

NUREG/CR-3822, LA-10132-MS, July 1984 6.

Theofanous, T.d. an.d H.P.

Nourbakhsh, "PWR Downcomer Fluid Temperature Transients'.Due.to High Pressure Injection at Stagnated Loop Flow," Proceedings of Joint NRC/ANS Meeting on Basic Thermal Hydraulic Mechanisms in LWR Analysis,. September '14-15, 1982, Bethesda, MD, NUREG/CP-0043, pp. 583-613.

7.

. Levy, S.

and J.M.

Healzer, " Approximate Prediction of Heat Transfer During Pressurized-Thermal Shock with No i

. Loop Flow and with Heat Additi'on," SLI-8220,~Rev.

1, Prepared by S.

Levy, Inc., August 1982.

8.

Levy, S.

and J.M.

Healzer, " Simplified Predictions of Pressurized'Ther=al Shock,"_ Proceedings of Eleventh Water Reactor Safety Research Meeting, Vol.

2, NUREG/CP-0048, pp. 115-134.

.9.

Chexal, V.K. - Private c o m'mu n i c a t i o n, April 1985.

4 10.

-Nourbakhsh, H.P.

and T.G.

Theofanous, " Decay'of Buoy-ancy Driven Stratified Layers with Applications to PTS Part:I: The Regional Mixing Model,"-Nuclear Engineering and Design (in press);

s

~

+,., _ -

,.m

-m,

,y,

-ll-s' 7

t u

~

Iyer, K.

and T.G.

Theofanous, " Decay of Buoyancy Driven Stratified Layers with Applica.p. ions to PTS: Reactor Predictions," to be presented in the.1985 National Heat Transfer Conference, August 6-9, Denver 7" Colorado.

12.

Nourbakhsh, H.P.

and C.G.

Theofanous, " REMIX - A Com-puter Forgram for Te=$eYature Transients Due to HPI at Stagnated Loop Flow," NUREG/CR-3701 (in press).

l 13.

Oh, S.

et al.,

"A Mixing Model for Transient Cooldown in a Reactor Cold Leg and Downco=er Under Stagnant Loop Flow," AIChE/ASME PAPER No. 83/HT/13.

14.

Theofanous, T.G.,

P.

Gherson, H.P.

Nourbakhsh and K.

Iyer, " Decay of Suoyancy Driven Stratified Layers with Applications to PTS Part II: Purdue's 1/2-Scale Experi-

=ents," Nuclear Engineering and Design (in press).

15.

Iyer, K.,

P.

Gherson, and T.G.

Theofanous, "Purdue's i

One-Half Scale HPI Mixing Test Program," Second Proceedings of Nuclear Thermal-Hydraulics, Annual Meet-

)

ing of the American Nuclear Society, New O r l e a n s,., LA,

June 3-7, 1984, pp.. 165-171.

16.

ly ~ t,

E.

and T.G.

Theofanous, " Flooding Limited Thermal e

'tix in g - The Case o f H igh-F r-In j e c t i on," Submitted for 4

presentation in the Third Interna (ional Topical Meeting 4

on Reactor Thermal Hydraulis.co be held at Newport, Rhode Island, October 15-18, 1985.

4 17.

Roche, P.H.

and

'. W. Fanning, " Transient.Cooldown in a i

Model Cold Leg and'D'owncomerQ",E,PRI Report NP-3118, Project 2122-3, Interim Report May 1983.

18.

-Hashe=i, A.

and J.

Goodman, " Thermal Mixing in a-Rec-tangular Geometry Mo' del of a Cold Leg with High-Pressure Injection and a Downcomer," EPRI NP-2924, March 1983.

19..Valenzuela, J.A.

and F.X.

Dolan, " Thermal and Fluid i-Mixing in !/2-Scale Test Facility," Vol. 2 - Data Report, EPRI NP-3802, NUREG/CR-3426, Nov. 1984 20.

Theofanous, T.G.,

K.

Iyer,.P.'Gherson, and H.P..Nour-bakhsh, " Buoyancy Effects on Overcooling Transients Calculated for the NRC-PTS Study," NUREG/CR-3702 (in press).

21.

Draft Data Report for Three Preliminary Thermal Mixing Tests at the HDR-Facility HDR-TEMS ExperimentsLT32.15,-

T32.17, and T32.18, Project HDR, Kernforschungszentrum Karlsruhe GmbH,-FRG, March 1985.

i 1

y e

,.r,.,,

s

~,

is'

/

f'.

22.

" Fluid and Thermal Mixing Tests of the Loviisa. Pressure Vessel and Downcomer," Imatran Voima Oy, Propriatery Report, April 1984.

23.

Selby, D.L.

et al., " Pressurized Thermal Shock Evalua-tion of the Calvert Cliffs, Unit i Nuclear Power

~ ~ ~ '

~

Plant," NUREG/CR-4022' 01iNL /TM;9 4 0 8, 1985.

24.

Selby, D.L.

et al., " Pressurized Thermal Shock Evalua-tion of the H.B.

Robinson Unit 2 Nuclear. Power Plant,"

NUREG/CR-4183, ORNL/TM-956, 1985.

25.

Eashemi, A.,

et al.,

" Transient Thermal Mixing in a Full-Height PWR Cold Leg and.Downcomer," EPRI NP-3478,

~

May 1984.

l i

i 1

I 1

f e

h YU e

W 4

0 I

L

.m8'**

b

=T

t

.i I

l

'l 6

Tahle l l

Typical 3000 MW Power Planti Ceometric Configuration of a injector Diameter: 21.6 cm Cold Vessel /

. Lower 1.o o p Core Thermal I.e g Downcome.r Plenum Pump Seal Barrel Shield Inner Diameter 69.8 340.0 362.3 (cm) 69.8 394.9 Lengtli '(cm) 704.0 634.3 643.3 466.3 i

Clad Thickness (cm) 0.3 0.6 0.6 0.3 Insulatinn l

Thickhess (cm) 8.9

8. 9-8.9' 8.9 8.9 H.9 t

.,9 Wall llea t Tr.*

Area to W.i t e r l

10

l.5 2.6 1.I I.7 2.3 3.6**

(cm 2) x i

Fluid Volume

• 10

2.7 4.7 4.6 2.0 1.7 (cm 3) x

• Per cold leg.

• • BotI sieles included.

t3-loop Wes t inghouse Plant ( ll. R. Robinson).

%,g 5

e t ' 8.,

4 i

o

,i I

l

,k t

6 g

t l

l l

Table 2 Geometric Con f igeir a t ion of CREARE 1/5-Scale Test Facility.

Injector Diameter: 5.0 cm Cold Vesse1/

I. n w e r I.o o p Co r e-Thermal

1. c g Ilowncomer Plenum P ii m p Seal Barrel Shield Inner Diameter (cm) 14. 3 14.3 i

1.ength (cm) 130.5 114.3..

1-l 14. 3 84.5

~

Acrylic' Wal1 Thickness

',-1, (cm)

I.3

1. 9 l l.9 I.3 1.9 1.3 Wall lle a t Tr.*

l Area to Water i

(cm 2) x 10

5.9 5.8 7.9 4.3 5.5 5.8 II.4*

I i

Fluid V o l is m e (cm 3) x 10'3 25.3 26.1 42.3 26.5 19.5 I

L l

l

• noth sides included.

.i i

9 i

s l

i

.g

,1 6

s 8

T a li l e 3 Typical Operating Parames.crs tised in t li e Various integral Test Facilities l(E II P I Mass Flow Rate Ap /p,p R e,,p g F,p x

y Facility (Kg/s)

PWR 13.6 0.25 8.0 0.5 CREARE _l/5-Scale (Run 100) 0.45 0. 14

1. 0,

0.7 Purdue 1/2-Scale

()i-IC*)

l.41 0.09 1.6-0.45 CREARE 1/2-Scale (HAY 105) 5.18 0.12 5.8 1.4 I

IID R (T32.18) 1.49 0.28 3.8

2.0

• Typical We s t i n gliou s e reactor simulation run.

.(

13 l

I s

.i i

l

'l e

s

)

Table 4 Ge ome t. r i c Configuration of Purdine 1/2-Scale Test Facility Injection Diameter: 10.8 cm f.e g Downcomer Plenum Pump Seat Barrel Sliicld..

inner I) i a m e t e r (cm) 34.3 34.3 263 i.ength (cm) 366 180

+.

Acryl(c e

Wall Thickness

-(cm) 1.3

1. 3' '

l.3*

1.3 1.3 1.3 1,

D Flicid Volume 10' #

2.3 3.2 9.10 t 2.8 1.2 1.3 (cm 3) x t

lAlso simulates part of downcomer volume.

s'I 9

\\

• g s

G e

I

i I

'l

~

I

,e Table 5 Geometric Con f i gie r a t i on oI CRF.ARE I/2-Scale Test Facility injector Diameter: Il.4 cm Cold Vessel /

1.over 1.c o p Core Thermal

1. c g Downcomer Plenum Pump Seal Barrel Shield

• f Inner Diameter 36.3 (cm) 36.3 254 334.0 24 3.6 I.ength (cm) 376.9 334.0

% e Base Metal I

(cm)gickness Wall Ti 2.1 3.8

~

2.1 5.1..

0.6 Clad Thickness (c.)

l Insulation Thickness (cm) 5.1 5.1 5.I 5.1 5.1 5.1 5

Wal! IIc a t Tr.

Area to Water 10-4 4.3 5.4 3.7 2.9 5.4 7. 9 'I~

( c m ')

x Fluid Volume 10-#

3.9 5.8 5.8 2.7 2.8

'(cm 3) x N

'lnoth sides i n c l ueleil.

. ~ - i, i

s' E

. _ _ _. _ - -. = _.. ___ _ _.. _ _ _ _ _

e i

l l

l

'l t

i Tahle '6 i

,e e

Geometric Configuration of the IID R Reactor injector Diameter: 5.0 cm Cold Vessel /

I.ower I.o o p Core Thermal I.c g Downcomer Plenum Pump Seal Barrel Shield Inner Diameter (cm) 19.1 296 266 i.ength (cm) 600 685 I

757 7--

Base Metal Wall Thickness 2.0 10.5 21.0 2.3

- z, l

Clad { Thickness (cm)

, 0,,8 0.,8 Insulation Thickness (cm) 10.0 10.0 10.0 Wall lleat Tr.

Area to Water 10~4 3.6 63.7' 23.1 63.7 (cm 2) x Fluid Volume 10-#

1.7 77.7 110.4 (cm 3) x

t. k s

a 4

~

O MR2_

HPt i.

MR4 Q,M R 1

,Th MR3 MR5 g,y

/

x

-m-37;~~ -

, J, COLD LEG N T

'l' 2D N

Tg r

c T} -

1Y, /

, et I

DOWNCOMER Oo N

7 7

t

\\

[

'T e,p i

LOOP SEAL

,., Y D

. ~

~

v Tm LOWER PLENUM

^

Region RMM. Temperature Cold stream all along the bottom of cold leg TC Hot strea= all along the top of cold leg Th i

HPI plume ().RI)*

THPI,p(I} " '

7 HPI Pump (MRS)

T.

J Loop Seal T,

Downce=er plume at MR3 T.

J Downec=er ile=c (MR4)**

T

'I

~ 8 (7 d,p CL Outcide of devnec=cr plu=e T

m I

Lower plenum T,

Mixed =ean temperature is computed using entrainment as a function of y and FrHPI*

• • This temperature is the lowest for the given level.

Figure'l Conceptual definition of flow regime and the Regional Mixing Model (RMM).

4 2

\\

.i

.r'*

Y, '.

i i

HDR i

PWR or UPTF -

.- D

..n ff 1

f I

h 2

l 1

l l

-3 E-

_t. _ _ _ _ _t _ _.

I i

1 1

n 6

1 C RE AR E(1/2)

PURDUE(1/2)

., ~

/

i

//

d l

I~-

4 i

4 i

t i

J l

i

~

t 1

d

~d)

I 4

i d

CR E A RE( 1/ 5)

I IVO i

Figure 2 Relative sizes 'of-various integral test facilitier 1

in comparison to full scale commercial PWR.

i I

k

~~

o

9 190 i

i

~

m.

l

'h

'c COLD LEG o

,,j E

's

-w-g 110

'~.,,

L..,.

w p

/

/

..m

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

~

30 i

8 t

0 600 1200 1800 2400 TIME (S)

Figure 3(a)

Comparison of transient temperature response at the exit of cold leg for CREARE run May 105 with REMIX predictions (T-and T ).

h e

190.

i i

i PUMP o

_x n

w C.2 110 W

-1.

-r 30,-

0 600 1200 1800 2400 TIME (S)

~

Figure 3(b)

Comparison of tran'sient temperature response in pump for CREARE run MAY 105 with REMIX predictions (T.).

J 1SO i

LOOP SEAL s

n a

w C.

110 2

w l-

.. ~ ~ '

~ _.

"~

30 i

i 0

600 12 00~ ~

1800 2400 TIME (S)

Figure 3(c)

Comparison of transient temperature response in loop seal.for CREARE run MAY 105 with REMIX predictions (T ).

m

s s

_?

7 s.

o

190, a

.g

^

3' DDWNCOMER O

".Q CORE BARREL SIDE v

g-b"y I I 'd k;.g,,

2 110 24.

VESSEL SIDE W

y, %.g,?,*

p ws s

- ~

4._. h,.

i

- 2

-A %, -. ~__._.,__ -

.30 0

600 1200 1800 2400 TIME (S)

Figure 3(d)

Comparison of transient temperature response in

- downcocer at 1.I D below cold leg centerline for CREARE run CL MAY.105 with REMIX predictions (Td,p).

190 s

i i

pi CORE BARREL SIDE DOWNCOMER a q ~fr h'g m

O c.

110

%.4 'O n VESSEL SIDE

. m un j

" ' 1R;. y"J _

~

w

% 7.-

3 "N

30

' 7=,g-g y, 3

O 600 1200 1800 2400 TIME (S) '-

Figure 3(e)

Comparison of transient temperature response in downco=er at 2.I D below cold leg centerline for CREARE run MAY 105 with REMIX predictions (Td,p).

190 i

g

'- )%

DOWNCOMER N.Y,.. -

CORE BARREL SIDE

~

n O

7. P F

VESSEL SIDE c.

110

- n.

2

<' DN v-w p

7._

~ ~.-

~

30 i

f ~-

-.s. ;,3.

. c -- - y. --

0 600 1200 1,800 2400 TIME (S)

Figure 3(f)

Comparison of transient temperature response in downco=er at 3.1 D below cold leg centerline for CREARE run et May 105 with REMIX predictions (!d,p)*

e

1.0 g

i g

COLD LEG 0.8 e

e 0.6 0

Cc e

s 0.4 e

e 0.2 f

i 0.0 15 50 85 120 155 190 TEMPERATURE (C)

Figure 3(g)

Comparison of spatial temperature profiles at the

~

exit of cold leg at.192 e for CREARE run MAY 105 with REMIX I

predictions (T

,T ).

==

1.0 i

i

,i i

COLD LEG

~

0.8 e

~

0.6 o

$c e

5 0.4

~

e

~

0.2 p

I i

1 0,o l

15 50 85 120 155 190 TEMPERATURE (C)

Figure 3(h) Comparison of spatial temperature profiles at the exit of. cold leg at 462 s for CREARE run MAY IOS with REMIX r

predictions (Th'I c

~

D

~

7'

~

310 i

i COLD LEG O

270 w

W C

3 230 E i C

5 i

if.,I d.{

91b'iJ1

($, li

[" "h;,9 er..: + m--.7S.. M" b Ny it!"q;[Sp ;m: gg$fdMh;-

I

' ie t.

4 l

2 i

si - r g

H 190 f

1 5

9 0

360 720 1080 1440 1800 TI M E ('S )

Figure 4(a)

Comparison of transient temperature response of cold stream at the exit of cold leg for HDR run T32.18 with REMIX pretest predictions (T ).

~

~#

310

~.

O v

W C

270 COLD LEG

~

D l-i C

W c_

230 2

W H

190 e

e 0

360-720 1080 1440 1800 1

i TIME (S) i Figure 4(b)

Comparison of transient tem'perature response of i

hot stream at the exit of cold leg for HDR run T32.18 with REMIX predictions (Th)*

l h

i l

7' f'.

310 a

i

.. _. +

D'd W NC.QME R n

C W

270

Q__,

, ?.

- g j!

D f'g

?:

.4 '

w 1

r --

1-190

~

0 360 720 1080 1440 1800 TIME (S)

Figure 4(c)

Comparison of transient temperature response in downco=er at 2.6 D e

w cold leg centerline for EDR run CL T32.18 with REMIX predictions (T.).

J l

320

~

DOWNCOMER

~

275 v

W C

D H

< 230 C

W C.

2

'W 185 l--

l 140 O

450 900 1350 1800 TIME (S)

Figure 4(d)

Comparison of transient temperature rerponse in downcomer faraway from the. plume with REMIX predictions (T,).

I I

A

.- 'f

l 310 i

DOWNCOMER n

ll f

W 263 l

~

i C

pi!

iJ

<I 1

i: ' '

j 1-lc] & '-

"3 u..

.r i,

i H

T G

!;; i;l t;

.!! ! @ij :t t.

p l'M C

l l

N j!

{'

i 1

,If a, J i.3 l ht f"

r jjd L'

jl[ ['r!? d<

W 216

p

'L i f, i

fh 3

c.

i

}

i-h 5

h d

I' I

-j g

i I fji.',.l:LM i[.

1 u r

q' i,t a

f' :1 il. -I.'y y';l{

M w'!g I H

. 1w..

.n

..a

• I p1 r " lF..

UT 169 0

360 720 1080 1440 1800 TIME (S)

Figure 5(a)

Comparison of transient qmperature response in downcomer at 1.0 D belo'wcold leg centerline for HDR CL run T32.18 with REMIX predictions (Td,p).

_f

~~

~

310 DOWNCOMER n

3 270

~

,q 'l W

l

.C l.-

t g,g!g{;];,,,g j

230 j

c.

4 Tita F. e.t 9,g g..r., y.y.

7 3

W 1

190 0

360 720 1080 1440 1800 TIME (S) i Figure 5(b) co=pariso. of transient temperature response in downcomer at 2.6 D below cold leg centerline for HDR run T32.18 with REMIX $kedictions T d p)-

~

r

,..i'.s.

~

.~

31O i

DOWNCOMER i b ift ! F j

ii.I :'

SIf j

j

.i.

l j.

230 I

g W

1

~

2 m

I-190

~

i i

0 360 720 1080 1440 1800 TIME (S)

Figure 5(c)

Compa ris on-o f transient t ee.p e r a t u r e response in downcomer at 5.2 D below cold leg centerline for HDR CL run T32.18 with REMIX predictions (T d,p 310 DOWNCOMER p.

L 7

.g

i. g 4. ?! - n

.e

.;w 230 g

m C.

.2w F

190 i

0 360 720 1080 1440 1800 TIME (S)

Figure 5(d)

Comparison of transient temperature response in downcomer at 8.6 D below cold leg centerline for HDR

~

run T32.18 with REMIX redictions (T d, p) -