ML19345G658
| ML19345G658 | |
| Person / Time | |
|---|---|
| Site: | University of Buffalo |
| Issue date: | 05/26/1966 |
| From: | Hall W, Lumb R, Macphee J NEW YORK, STATE UNIV. OF, BUFFALO, NY |
| To: | |
| Shared Package | |
| ML19345G655 | List: |
| References | |
| WNY-023, WNY-23, NUDOCS 8104080585 | |
| Download: ML19345G658 (39) | |
Text
.
EO lI lI I
i I
PULSTAR PULSE TESTS III AND IV COMPARISON OF NATURAL AND FORCED I
CONVECTION COOLING May 26,1966 Report No. WNY-023
,I lI I
l I
I i'.'
I I
I I
giov n os w I
I Sumary Report i
PULSTAR PULSE TESTS III AND IV COMPARISON OF NATURAL AND FORCED CONVECTION COOLING Compiled and Edited by John MacPhee*
I Ralph F. Lumb William F. Hall I
Western New York Nuclear Research center, Inc.
May 26, 1966 I
- Chief Process Engineer, Metal Flo Corporation. Formerly with AMF Atomics.
I b
,_,--._,,,.,,,._,.,-------,,,,,,,....,.,,,.,,,.,.,,-,.,,-,.,,,-.-,,w-,,
TABLE OF CONTENTS Page Introduction A
General 1
B Series I Pulses 1
C Series II Pulses 2
D Series III and IV Pulsas 3
E Refinement of Power Calibration 4
Discussion of Core Performance A
General 5
B Effect of Cooling on Core Performance 1
Natural Convection 5
2 Forced Convection 7
C Effect of Test Pins on Core Performance 9
D Effect of Pission Product Poison on Core Performance 10 E
Effect of Configuration on Core Performance 10 F
Significance of Total Energy and Peak Power on Core Performance 12 Summary and Recommendations 13 A
Conclusions 13 B
Recommendations l
1 Mode of Cooling 14 2
Limit on Pulse Performance 14 (a) Proposed Calculational Method 15 (b) Illustrative Safety Margin 16 3
Fuel Pin Inspection 17 l
Appendix Series III and IV Pulses A
introduction 19 B
Experimental Program 1
Series III Pulses (a) Test Pin Configuration 20 I
(b) Test Results 21 2
Series IV Pulses (a) Test Pin Configuration 22 l
(b) Test Results 23 l
3 Observations 23 l
C Analysis of Data 1
Cold Wall Effect 24 2
Initial Attempt to Corrt ste Data by Perimeter Ratio 25 3
Correlation Established 26 l
l
INTRODUCTION A GENERAL I
On September 27, 1963, the Western New York Nuclear Research Center, Incorporated (WNYNRCI) requested authority to install a PULSTAR oxide fuel reactor. Subsequently, the submittal dated May 28, 1964 requested authorization to (1) conduct a test program comprising steady state tests at low power and transient tests involving pulses of an energy I
release up to 90 Mw-sec., (2) operate routinely in the steady state mode to power levels up to 2 Mw, (3) operate routinely in the pulse l
l mode at energy release pulses up to 40 Mw-sec. On June 19, 1964 the WNYNRCI was authorized by the Atomic Energy Con: mission to carry out the l
l steady state program and to operate routinely at 2 Mw, while the I
routine pulse operation was authorized on May 12, 1965. In addition to routine operation, several pulse test programs have been authorized.
l The initial, or Series I, transient program with natural cooling was authorized on July 17, 1964. Two subsequent authorizations were l
issued permitting conduction of the Series II pulses with natural
(
l convection cooling, and Series III and IV pulses with forced convection cooling. The primary purpose of this report is to present recommen-l dations based on the results of these test programs regarding method of cooling, limits on pulse performance, and criteria for the selection and inspection of fuel pins in accordance with Section PS, paragraph (a)
(12) of the Technical Specifications.
B SERIES I PULSES The PULSTAR core was installed in June and achieved initial criticality
(
i l
l I L
on June 22, 1964. Fo11cwing completion of the steady state tests, I
the pulse program was initiated on July 31, 1964. On September 28, 1964, the reactor was pulsed on a period of 5.2 msee to a power level of 1290 Mw and a total energy release in the pui.se of 30.3 Mw-eec.
Maximum power densities achieved in this pulse were 367 watt-seconds per gram in the core fuel pins, and 513 watt-seconds per gram in the test pins-six fuel pins located in the water reflector region so as to lead the hottest core pin by a factor of approximately 1.8 in power per unit length, and by a factor of 1.4 in energy density.
(Flux gradients across test pins were relatively flat as compared to core pins; hence, the energy density ratio at the hot spots was lower than the ratios of power produced per unit length of y a.)
During the gradual escalation of energy release, these test pins pro-vided information on pin performance at higher energy releases, and constituted the basis for the " bootstrap" approach to power and energy increases.
Following nine repetative pulses at this power level and period making l
l g
a total of 100 pu' as for this first phase of the test program, the l
3 transient pulse 1
"- was suspended in accordance with the Technical Specifications. The Its of this program are reported in WNY-017(1).
C SERIES II PULSES During November and December,1964, the second series of pulse tests l
was conducted usin8 the same core and test pin configuration as was used in the initial series of PULSTAR tests. On December 9, 1964, a l
i 2
i l
I
I nominal pulse output of 50.6 Mw-sec and 4240 Mu was achieved with a reactor period of 2.83 maec. For this pulse the maximum power density achieved in the core pins was 623 watt-seconds per gram, which corre-sponds to a UO te perature of approximately 3675 F.
In the test pins, 2
the power density reached 872 watt-seconds per gram, which corresponds 0
to a UO temperature of at 1 cast 4500 F, and perhaps as high as 5000 F.
2 (The uncertainty reflects the fact that at this high energy release the is not well known.) Moreover, the equivalent energy heat capacity of UO2 release in the test pins, 72 Mw-sec, corresponds to a power of about 10,000 Mw.
The core was pulsed repetitively ten times at this level, following which an examination of the test pins was made in accordance with the test program procedure. This inspection disclosed that the test pins had become deformed, but that the integrity of the cladding had been maintained. Further escalation of pulse energy was temporarily post-poned, and a proc am of review and analysis was initiated (2).
D SERIES III AND IV PULSES Based on observations, it was concluded that departure from nucleate boiling (DNB) had occurred in all six test pins and in the hotter core pins ; i.e., core pins in which the UO2 temperature had exceeded approxi-0 mately 2000 F during the most energetic Series II pulses.
(The terms DNB and film boiling are used interchangeably throughout this report.)
Although metallographic examinations indicated that exceeding DNB did not produce any damage in the test pins which would jeopardize pin integrity, it was desired to eliminate this phenomenon in order to I
3 I
achieve higher pulse outputs.
As a result, the Series III and IV pulses using forced convection cooling were initiated. Because of their importance to the recommen-dations contained herein, the results of these series of pulses have been included in this report as an appendix.
I E REFINEMENT OF POWER CALIBRATION Subsequent to completion of the pulse testing with flow, it was observed that the core differential temperature, as read out from resistance bulbs by a new high precision electronic instrument, exceeded the differential temperature as measured with thermocouples and presented on the differential temperature meter.
An exhaustive, retrospective analysis of reactor operational data together with various test measurements indicated that the power as measured by the resistance bulbs is the more reliable. This measurement routinely exceeded the power indicated by the differential temperature meter by a nominal 14.77..
For this reason, all previous power and energy data were adjusted upward by 14.77..
Therefore, all data for all pulse series quoted in this report have been adjusted for this correction.
I I
I I
4
DISCUSSION OF CORE PERFORMANCE A GENERAL Prior to establishing criteria and limits governing pulse operation, the effects of variable factors upon basic core performance must be evaluated. In particular, it is important to know if pulse performance can be affected by such considerations as method of cooling the core during pulses, location and environment of test pins, increases in I
fission product inventory due to steady state operation, and physical changes in core configuration. In addition, the significance per se of gross performance parameters; i.e.,
total energy release and peak power, must be established.
B EFFECT OF COOLING ON CORE PERFOR&lNCE 1 Natural Convection It has been demonstrated that under conditions of natural convection, the limit of pulse performance for all practical purposes is dependent only upon the maximum enerav density in the core (or test pin associated with a core). No experin. ental evidence was found to indicate that pin environment is a significant factor in limiting pulse energy; that is, the environmental difference among the pins is not significant under conditions of natural convectior, cooling, and hence does not appear to influence the energy which can be released during a pulse without effect in a pin. This premise is illustrated by the following table which represents the results of an exhaustive inspection program of pins which had been present in the core during the 50.6 Mw-sec total energy pulse with the natural convection cooling.
5
2 Forced Convection Under conditions of forced convection cooling, the results of the Series III and IV pulses, as described in detail in the appendix, demonstrate that pulse performance is also limited by caxinum energy density in the core. However, performance with forced convection cooling is distinguished from performance with natural convection cooling in two very important aspects: First, forced con-vection cooling raises or improves the maximum permissible energy density in general; second, with forced convection cooling, pin environment also affects pulse performance and, therefore, cust be taken into account as a modifying factor.
The test results indicate that the heat flux at which DNB occurs with forced convection cooling is reduced by the presence of unheated surfaces or " cold walls" adjacent to the subject fuel pin. Thus film boiling would be expected to occur first in corner pins, next in edge pins, and last in centrally located fuel pins in a uniformly heated PULSTAR-type fuel element. However, the " cold wall effect" is not aufficient by itself to explain the observed behavior.
From the analysis in the appendix, it appears that differences in hydraulic diameter must be taken into account in any atte:npt to deter-mine the heat flux at DNB. This conclusion is also supported in the literature. Thus, the heat flux at DNB in corner pins and edge pins is reduced because of the smaller hydraulic diameter of the coolant channels formed by these pins and the fuel box. Bernath's correlation apparently describes this hydraulic effect adequately; however, that correlation does not take into account the " cold wall effect"(3).
l I
It is concluded that the data from the forced convection pulse tests on DNB threshold can be correlated by the method proposed by Bernath pro-vided a correction factor is used to account for the " cold wall effect."
This correction factor is equal to the ratio of the heated perimeter to the total perimeter as defined in Table II of the appendix. This correlation is illustrated in Figure 2 and 3, and predicts that film I
boiling will occur first in the corner pins adjacent to the reflector in core positions C-2 and C-5 in the test core. While Figure 2 dis-plays limiting fuel temperature as a function of the pin environment, Figure 3 shows the relationship between energy density at the thresh-hold of film boiling, and the correlation factor.
For example, the 4 x 5 core with no test pins will be limited by the energy density in the most energetic corner pin, or 587 watt-seconds per gram as obtained from Figure 3.
This most energetic corner pin has a hot spot factor for the 4 x 5 configuration of 3.55; hence, the maximum average energy release is 165 watt-seconds per gram. For a 289 Kg' UO core, this value represents a release of 47.5 Mw-sec. The 2
same core, under conditions of natural convection, can be evaluated by use of Table I values, which were derived from a 50.6 Mw-sec pulse j
(178 watt-seconds per gram average energy density). Bowing began to occur at a hot spot factor of 2.25; hence, an energy density of 400 watt-seconds per gram represents the threshold for that core under natural convection. Using the same maximum hot spot factor of 3.55, it can be seen that for natural convection the 4 x 5 core without test pins can be pulsed to an average energy of 113 watt-seconds per gram, or a total energy of only 32.7 Mw-sec. Thus forced convection cooling 1
8
provides a 467. increase in energy release without pin bowing.
C EFFECT OF TEST PINS ON CORE PERFORMANCE As originally conceived, the philosophy of test pins was intended to provide a " boot strap" approach to the gradual escalation of power and energy to maximum practical limits. For that purpose, three fuel pins in special fixtures were located in the water reflector at both sides of the core approximately 1 1/2" from the side of the core during natural convection pt.lsing. These test pino led the most energetic pin in the core by a factor of 1.8.
Under conditions of forced convection, test pins were moved to loca-tions in the outermost row of two peripheral fuel boxes. The fuel-reflector interface was diverted inward at this point by alternating four water filled pins with the three test pins at each side of the core (Figure 1).
Thus, these test pins lead the most energetic core pin by a factor of 1.2.
From the standpoint of energy density as a factor influencing pulse performance, the test pins represent the limiting pins. Moreover, in the case of forced convection, the unusual environment of the test pins further limits core performance, as noted in the previous remarks regarding " cold wall effect." Figure '3 provides a striking contrast in the performances of test pins and core pins in like locations, but with different environments. Test pins, therefore, constitute an
" Achilles heel" in routine pulse performance under conditions of forced convection cooling. This limit is illustrated by the fact that with no test pins, total core performance can be increased by 457.,
I
based on the same limits.
D EFFECT OF FISSION PRODUCT POISON ON CORE PERFORMANCE h
Pulses have been performed with significant residual xenon conditions (approximately 0.357. delta K/K) and measurable burnup (6000 Mw-hrs) and associated reactivity change (approximately 0.157. delta K/K) with ne significant deviation from predicted values in the resulting power and energy. From these observations, it is expected that fission product poisoning and burnup will not constitute a factor in affecting pulse performance.
E EFFECT OF CONFIGURATION ON CORE PERFORMANCE Experience has been acquired with various control rod positions, different experiments, differences in fuel loading, and two different reflector materials. These variances in configuration are characterized by the fact that they influence pulse performance only to the extent that the magnitude of the maximum energy density is influenced through the hot spot factor and average energy density.
I The position of the control rods in the guide channels affects the hot spot factor in pins adjacent to the channel by compressing or expanding the skewed cosine distribution of flux in the axial direction. Experi-ence to date w2'.h a variety of control rod positions indicates this is l
the only effect on pulse performance.
lI l
The presence of experiments in a pulsing configuration of fuel will i
j also affect the hot spot factor.
In the limiting case, a central position in the core from which the fuel assembly has been removed, I
10 I
l l
u_
I leaving a water filled hole, the hot spot factor in pics immediately surrounding the hole will be increased by virtue of the flux gradient across the diameter of these pins resulting from flux peaking in the water hole. This condition is similar to the gradients occurring across peripheral core pins due to the flux peaking in the water reflector.
While the presence of such water holes has an unfavorable effect upon I
the bulk or pool water temperature coefficient of the core, there is no evidence that either fuel or moderator temperature coefficients are affected; hence, pulse performance is not affected. A separate communication to the Commission described this condition (8).
The installation of an experiment in the water hole will have the effect of perturbing the overall flux in the hole, and if the water is largely displaced by the experiment, the flux gradient across surrounding pins will be lowered, thus lowering the hot spot factor.
I Although installation of experiments in the core during pulsing has I
not yet been authorized, experience with the reactivity changes repre-sented by the test pins in the reflector region and by changes in core reflector conditions (aluminum slab at the side of the corrlj show no evidence of influencing parameters such as neutron lifetime, which would cause pulse performance to deviate from predicted values. The only significant effect is upon the hot spot factor and hence, energy density.
Core size, in itself, will not serve to limit pulse performance. The addition or subtraction of fuel assemblies will, however, modify the hot spot factor and will alter the total mass of fuel in the core. With 11 I
I
I a fixed maximum energy density limit, increasing the number of assemblies in a configuration could lower the maximum hot spot factor slightly, increase the total mass of fuel, and consequently, increase the total energy release and peak power permitted in a pulse.
F SIGNIFICANCE OF TOTAL ENERGY AND PEAK POWER ON CORE PERF010fANCE Experience to date indicates that neither total energy nor peak power are limits per se.
That is, within the range of peak power and total energy released during the pulse test program, factors which respond to total energy and power (such as radiation levels, pressure generation outside the core, Nitrogen 16, radiolytic gases) did not approach significant values. Thus, total energy release and peak power can vary widely with no ill effects, provided that the safe limits on maximum energy dencity are not exceeded.
I I
I I
I I
12 I
I SU.;fARY AND RECOMMENDATIONS The following section susciarizes the conclusions drawn from the pulse testing accomplished to date, as well as applicable steady state expe-rience, and presents recommendations regarding method of cooling, limits on pulse performance, and criteria for selection and inspection of fuel pins based upon those conclusions.
A CONCLUSIONS 1 Pulse performance is significantly enhanced by forced convection.
While the improvement is principally associated with higher maximum energy density limits, forced convection also eliminates release of radiolytic and radioactive gases at the pool surface with attendant reduction of high radiation levels.
2 The single performance parameter limiting pulse operation is the maximum energy density in the core (maximum energy release per unit mass of UO ).
The limit of this parameter, which is a function of 2
coolant velocity and is af fected by equivalent hydraulic diameter and pin environment, can be predicted. All other parameters such as maximum power, total energy, core configuration (control rod position, I
l fuel arrangement, reflector materials, experiments), burnup, and fission product poisoning have been found to influence pulse performance only to the extent that they influence the maximum energy density in the I
core.
3 The best method developed to date for determining whether or not the limit on maximum energy density has been exceeded is visual 13 I
inspection of the fuel pins. Thus fuel pin inspection should be I
continued as a means of monitoricg core integrity and confirming that limits on maximum energy density have not been exceeded.
B RECOMMENDATIONS 1 Mode of Cooling On the basis of the findings of the experimental programs using forced and natural convection cooling, operation under conditions of foceed convection is recommended due to the significant improvement in fuel performance which can be realized.
For comparative purposes, under natural convection the energy release which corresponds to the threshold of DNB in a 4 x 5 core without water I
filled pins is 31.0 Mw-sec, while under conditions of forced convection the same core could release 47.4 Mw-sec. The improvement afforded by forced convection is, therefore, substantial.
2 Limit on Pulse Performance Under conditions of forced convection I
cooling, energy density, when combined with the conditions of hydraulic diameter and' bold wall effect" characteristic of edge and corner pins in a fuel assembly, constitutes the upper limit cf energy release for a pin in a given fuel assembly. In particular, the total energy release for a core configuration during a pulse will be dictated by maximum i
energy release for the pin most susceptable to deformation; therefore, i
it is recommended that a limit be established for this maximum energy density at 757. of the predicted threshold for bowing. The predicted thresholds for bowing for a variety of pin environments are shown on Figure 3.
The figure of 757. represents allowances for accuracy of reactor power determination, hot spot location, and reliability of the 14
correlation between burnout factor and energy density, as described in this report.
I The following sections outline the proposed technique for implementation of this recommendation, and illustrate the safety margin inherent in the technique through a sample calculation.
I (a) Proposed Calculational Method To calculate the limiting total energy release for a fuel pin the following relationship will be utilized:
75
_135 + 970(cf) _
100 hsf where E,x = limiting total energy l
cf
= correlation factor l
hsf = hot spot factor g
= total UO 1 ading 2
When applied to a particular core configuration, the lowest value of 1
E so obtained then represents the limiting total energy for that m
configuratien. The correlation factor will be determined as described l
in the appendix and illustrated in Table IV.
The hot spot will be l
identified by means of a technique utilizing the results of a series of probes by a miniature fission chamber of the coolant channels and reflector region opposite peripheral pins. From these data the hot spot factor will then be calculated.
I It has been demonstrated that results of fission chamber probes of coolant channels for the 4 x 5 test core correlate with the average of copper wire activation values for the four pins adjacent to the probed 15
I hole with a correlation coefficient of 0.98.
For determination of hot spots adjacent to the reflector region, where partial coolant holes exist which cannot be probed with available fission chambers, data obtained from locating the fission chamber in the reflector imediately opposite to the subject pin resulted in a correlation with the average copper wire activation values for the subject pins. A similar technique I
was employed to correlate pin energy densities around a water hole in the core with good results.
Subsequent to determination of the limiting total energy release for a core configutation not previously pulsed (or for a loading which has been pulsed, but which contains new fuel), a measurement of reactivity defect will be made to insure that the sign remains negative and that the magnitude has not changed significantly. Further, routine pulsing with such a loading will be approached by a series of test pulses of increasing magnitude.
I (b) Illustrative Safety Marain To illustrate the safety margin in-herent in this technique, consider a 4 x 5 array of fuel elements containing 289 kg of UO. The hot spot factor for a corner pin is 2
3.55 and the correlation factor is 0.46.
The hot spot facter for en edge pin is 3.46 and the correlation factor is 0.62.
From the relationship described above:
I for corner pin f
75 135 + 970(.46)
E,,x=
~
00 3.55 for edge pin 75
.135 + 970(.62)_
(289) = 46.1 Emax" 100 3.46 16 l
I
m-I Therefore, the corner pin is limiting and the limiting total energy 1
release for a 4 x 5 array is 35.5 Mw-sec.
To illustrate the consequences of errors in these measurements and correlations, the following tabulation indicates the extent cf expected pin bowing for pulses having specific energy releases in excess of the routine pulse average value of 123 watt-seconds per gram in a 4 x 5 array:
Average Power Fraction Ratio Bowed Maximum Power Density Density of Routine Pins to Total of Corner Pin (Watt-sec/gm)
Pulse Core Pins (Watt-sec/gm) 123 1.0 0
442 166 1.33 0
590 183 1.46 0.024 650 200 1.60 0.072 710 216 1.73 0.108 767 Clearly, the margin associated with the routine pulse is conservative.
It was concluded in WNY-020 that bowing per se constituted no hazard l
since the bowed pins would be effectively pre-stressed, and subsequent 1
(
repetitive pulees at the same energy release would not exceed yield stress.
Moreover, during the natural convection portion of the program, test l
l pins were repetitively subjected to specific energy releases of 872 watt-seconds per gram at the hot spots.
3 Fuel Pin Inspection It is recommended that frequency of pin inspections be as follows:
I 17
(1) After the first pulse in any fuel configuration not previously pulsed.
(2) Af ter every five pulses subsequently, or a minia:um integrated energy of 2000 watt-seconds per gram for the pin with the highest haf/cf ratio, through the 26th pulse.
(3) After each group of a minimum of 20 and a maximum of 30 pulses subsequently.
I It is recommended that each inspection consist of examination of a minicum of four core pins, which most closely approach the limiting energy density; i.e., pins having the highest values for the ratio hsf/cf.
It is recommended that the limitations on dimensional inspection, and upon corrosion and rupture of cladding be as set forth in Section P5a(10) and (11) of Appendix A to License R77:
Pulse operation shall be terminated and a report submitted to the Division of Reactor Licensing if any fuel clad rupture occurs.
I Evidence of abnormal changes in test pins, including corrosion deposits on the fuel pins, diametral growth in excess of 1%, or bowing in excess of 0.050 inch in any 6 inches of fuel pin length, shall require a re-view by the Operating Committee prior to continuation of pulsing. The review shall also include the results of an inspection of selected core pins. The results of the Con:mittee review and actions taken shall be documented and maintained as part of the licensee's record, i
l l
l
1 I
APPENDIX SERIES III AND IV PULSES A INTRODUCTION To date, four series of pulse tests have been carried out on the proto-type PULSTAR core at WNYNRCI. Series I and II were carried out with natural convection cooling and are reported on in detail elsewhere(1,2).
In these tests, a maximum pulse energy of 50.6 Mw-see was reached with a maximum UO te5Perature of 3675 F in the hottest core pin and a max-2 imum temperature of at least 4500 F in the test pins. Based on observations, it was concluded that DNB or film boiling had occurred in all six test pins and in the hotter core pins; i.e., core pins in which the UO temperature had exceeded approximately 2250 F.
Although 2
metallographic examinations indicated that DNB did not produce any damage in the core pins which would jeopardize pin integrity, it was desired to eliminate this phenomenon in order to achieve higher pulse outputs.
Based on further analysis of the pulses run with natural convection cooling and Bernath's correlation (3), it was concludea that the thresh-old of DNB could be increased significantly through the use of forced convection cooling. Accordingly, Series III and IV of the pulse tests were initiated using forced convection cooling. The purpose of this appendix is to document the results of these latter tests with regard to fuel pin performance, and to present a correlation for predicting zircoloy clad fuel pins in the pulse safe operating limits for UO2 mode.
I 19
B EXPERIMENTAL PROGRAM 1 Series III Pulses (a) Test Pin Configuration Heretofore, in Series I and II pulses, the energy per unit length released in the test pins exceeded that in the maxinum core pin by a factor of 1.8.
To have continued testing under this condition would have sharply restricted the energy release which could be realized in the core, without damage to test pins.
Consequently, in discussions with the Commission preparatory to forced convection testing, it was decided that the test pins should be so positioned as to lead the hottest core pin by a nominal 1.2 factor.
Further, to reproduce the lateral support, coolant flow, hydraulic diameter, and flux gradient of the hottest core pins, six test pins were located within the zircaloy boxes of certain peripheral fuel assemblies, and were interspersed with water filled pie in such a way as to displace the fuel reflector interface by one row of fuel pins in the region so modified. Figure 1 depicts the test pin arrangement.
I With this arrangement, the test pin power per unit length of pin at the maximum axial flux peak exceeded the maxinum power per unit length of pin at the axial flux peak of the core by the following values (ref os l
of energy densities at hot spots are included, since these values ultimately were selected as a basis for the analysis of results):
Ratio Power per Assembly Pin Unit Pin Length Ratio Energy Location Location at Max. Axial Flux Density at Hot Spot C-5 A-1 1.21 1.14 I
C-5 C-1 1.26 1.23 C-5 E-1 1.20 1.12 l I 20 l I
I (continuation of table)
Ratio Power per Assembly Pin Unit Pin Length Ratio Energy Location Location at Max. Axial Flux Density at Hot Spot C-2 A-1 1.20 1.12 C-2 C-1 1.26 1.23 C-2 E-1 1.21 1.14 (b) Test Results The " boot : strap" approach used in previous testing under natural convection conditions was applied to the Series III pulses with forced convection. That is to say, initial pulses involved very low levels of power and energy, and subsequent pulses yielded increasingly greater releases. Values of peak power and energy to peak power were plotted as functions of reciprocal period, and "least-l squares" fits were computed to establish 957, confidence limits for successfve pulses.
In addition, power and energy data were compared to the plots generated during Series I and II pulses.
I Pulse energy releases were gradually escalated to the point of 28.5 Mw-sec, (Pulse No.139) a maxin:um hot spot power density of 436 watt-seconds per gram in the test pins, where an inspection of test pins was conducted. No evidence of test pin bowing was found at this time.
Again, after Pulse No. 142, with a total energy release of 45.6 Mw-sec
,I and a maximum hot spot power density of 710 watt-seconds per gram in l
the test pins, there was a pin inspection. Evidence of bowing was l
detected in the inspection of test pins.
lI l
To determine the threshold for pin bowing, three pulses of lower energy
,I 21 l
releases were performed.
Pulse energy release increments were kept sufficiently small to bracket closely the thresholds.
I The perfe.
. e of various fuel pins in terms of UO temperature and 2
whether or not film boiling occurred is summarized in Table 1.
The data on pulse energy presented in Table I were obtained from graphical integrations of the power trace on the Sanborn recorder which is con-sidered the most accurate w;asure of this pulse parameter. The occurrence of film boiling was determined by ti.e Cegree of coherent bowing observed in the fuel pin in accordance with the theory postulated in Reference 2.
The performance of fuel pins in various environments is summarized in chart form in Figure 2.
While phyntcal equipment was employed during this test program to measure th.: extent of bowing, it was found that a subjective visual inspection was more effective in determining the onset of bowing condi-tions, since as-received fuel pins can be expected to be bowed in the same order as the detection limit of the measuring device (1/32" over 12" of the pin). The observed small deflection in the pin was usually accompanied by barely perceptible decreases in the lustre of the pin cladding in the region of the hot spot.
I I
Nine repetitive pulses were run at an energy release level (35.0 Mw-sec) just below the threshold energy value for bowing of any test pin to establish if any cumulative effects occurred. None were observed.
l 2 Series IV Pulses j
(a) Test Pin Configuration The analysis of the threshold concept of lI I
I pin bowing dictated further investigation of the effect of the environ-ment of a fuci pin, and in particular, the unheated wall adjacent to the pin of interest. To provide information on this effect, adjacent to one corner and one edge test pin, which are normally bounded on three sides by unheated walls, was placed an additional water filled pin, as shown in Figure 1, to create " hot" test pins.
Fluxes in these " hot" test pins were measured with copper wires at the maximum axial flux peak, with the following results:
Ratio Power per Assembly Pin Unit Pin Length Ratio Energy Location Location at Max. Axial Flux Density at Hot Spot I
C-5 E-1 1.39 1.22 C-5 C-1 1.39 1.20 (b) Test Reeults In a manner similar to the determination of the threshold for bewing for the normal test pins, a series of ten pulses was r d for the " hot" test pins. Three of these pulses were of a repea't nature, since the testing sequence had to be interrupted during Series IV.
Total energy releases were gradually escalated to the points of 26.4 Mw-sec and 28.7 Mw-see to obtain threshold energy density values of 433 watt-seconds per gram and 471 watt-seconds per gram in the " hot" corner and " hot" edge pins, respectively. The performance of these pins is also included in Table I and Figures 2 and 3.
3 Observations The most striking thing about these data is that they show that the threshold for film boiling is not simply a function of I
I
I UO2 temperature as indicated by the following observations:
I (1) The DNB threshold in all pins under conditious of forced convection cooling is higher than the threshold of approximately 2000 F under conditions of natural ccovection cooling.
In some pin configurations this difference is significantly lar88.
(2) The DNB threshold in the corner test pin with one adjacent unhenteo rod is approximately 350 higher than in the hot corner test pin which has two adjacent unheated rods.
(3) The DNB threshold in the edge test pin with two adjacent unheated rods is approximately 750 higher than the threshold in the hot edge test pin which has three adjacent unheated rods.
(4) The DNB threshold in the edge test pin with two adjacent unheated rods is approximately 500 higher than the threshold in the corner test pin which has only one adjacent unheated rod.
From the above observations, a number of conclusions can be drawn.
First,these tests demonstrate quite clearly that pin performance during pulse operation is enhanced through rhe use of forced convection cooling in that the threshold for Dtr? or film boiling is increased significantly.
The second conclusion that can be drawn from these observations is that the performance of fuci pins relative to film boiling cannot be corre-lated simply by UO temperature. Furthermore, any attempt at correlating 2
DNB threshold must consider both " cold wall effect" and hydraulic diameter differences, C
NALYSIS OF DATA 1 cold Wall Effect Based on the above observations, it was concluded 24 I
TABLE '
Natural Convection EDGE PINS CORNER PINS I
Hot Spot Hot Spot Bow Factor Bow Factor I
No 1.41 No 1.83 No 1.79 No 1.91 No 2.00 No 1.91 I
No 2.11 No 2.16 Slight 2.22 No 2.23 Slight 2.25 Slight 2.27 Bow 2.36 Slight 2.43 Slight 2.46 Bow 2.50 Bow 2.48 Bow 2.72 Sli ht 2.70 8
Bow 2.77 Bow 2.92 Slight 2.90 Bow 2.95 Bow 3.13 Bow 3.04 I
Bow 3.28 Bow 3.33 Bow 3.46 Bow 3.46 Bow 3.46 Bow 3.56
" Slight" implies definite evidence of bowing.
From the evidence tabulated above, it is concluded that with the same hot spot energy release, a corner pin is no more susceptible to bowing than an edge pin. That is, bowing was observed in both corner and edge pins at a hot spot factor of about 2.25.
(No correlation could be established with pins interior to assemblies since hot spots were in all such cases much lower than for ed e or corner pins.)
8 Thus, effects of pin environment on DNB, including hydraulic radius for coolant flow and unheated walls of coolant channels (" cold wall effect" discussed in paragraphs following), are negligible with natural convec-tion cooling.
that the DNB heat flux in a pin is reduced by the presence of unheated l
surfaces adjacent to the subject pin.
The reason unheated surfaces reduce the DNB heat flux cannot be stated with any confidence since the steam quality present in the channel, when DNB occurs, is not known. For example, if the steam quality is low, the effect of an unheated surface could be explained by the fact that it provides greater drag than a heated surface, and thus reduces flow. On the other hand, if the steam quality is high, the theory proposed by Becker(4) could account for a lowering of the DNB heat fi tx by an unheated surface.
In any event, the conclusion that ad-jacent " cold walls" lower the DNB heat flux is supported by these data and the observations of others(4,5,6).
2 Initial Attempt to Correlate bata by Perimeter Ratio While the literature (4,5,6) acknowledges the adverse affects of " cold walls,"
a suitable correlation has not been proposed. Becker(4) suggests using the ratio of heated channel perimeter to total perimeter for the situation where the steam quality at DNB is high. Briefly, Becker's theory holds that the steam quality at burnout in a channel is propor-tional to the product of the perimeter ratio (ratio of heated surface to total surface) and the steam quality at burnout in a channel complete 1,y surrounded by heated surfaces. Although it is not cicar that this is the case in PULSTAR pina, Becker's theory formed the basis for the first attempt at correlating the data in Table 1.
I The UO energy density at DNB threshold was plotted versus perimeter 2
ratio as the first attempt to correlate the data. The perimeter ratio 25
I for each of the fuel pin configurations was calculate $ as in Table II.
However, it became apparent that the perimeter ratio us not suf ficient to correlate the observed behavior.
3 Correlation Established Aside from the " cold wall ef fect," the coolant channels in question also differ with regard to hydraulic diameter. An effect due to these differences has been recognized in the literaturc (5) and furthermore, increasing the hydraulic diarnecer has been shown to increase the DNB heat flux in nultirod configurations (7). Therefore, a correction factor, to account for these differences, was calculated using Bernath's theory (3) as shown in Table III. A correlation factor was then calculated by taking the product of the perimeter ratio and the correction factor for hydraulic diameter. This correlation factor is summarized in Table IV for the fuel pin configu-rations of interest.
I If we are to follow Becker's approach, this correlation factor would be defined as the ratio of the heat flux at DNB in tha subject pin to the heat flux at DNB in a central pin surrounded on all sides by heated pins. However, since we are attempting to correlate performance with UO energy density rather than heat flux, such a definition does not 2
l apply in this cese. The plausibility of this method of correlating I
the data is illustrated by Figure 3 which is a plot of UO2 *"*f87 density at the DNB threshold as a function of the correlation factor.
l While it might be expected that such a plot would be linear (again following Becker's theory) there are a number of possible explanations for the actual shape of the curve.
l 26 l
I To begin with, the cur e extrapolates to a U02 energy density of approximately 370 watt-seconds per gram for a correlatio) factor of zero. inis fact, c.<Jpled with the fact that the DNB threshold energy den:' *y with natural cor.veccion cooling was also about 3/0 watt-seconds per gram, strongly suggests that this energy density represents a threshold for pin bowing. Since pin bowing is used to detect DNB or film boiling, the curve would be expected to extrapolate to some finite energy density if there is indeed a threshold for pin bowing. A second factor which could account for the nonlinearity of the curve in Figure 3 is nonlinearity in the relationship between energy density and heat flux due to such things as the variation of UO thermal conduc-2 tivity with temperature. That is, if thermal conductivity were constant, the relationship between heat flux and UO energy density 2
would be expected to be a: ore linear. However, for the actual situation where thermal conductivity decreases with temperature, the curve would be expected to depart from a straight line in the same way as the curve does in Figure 3.
Therefore, it has be2n concluded that the correlation represented in Figure 3 is plausible and can be used to predict the performance of PULSTAR fttel pins with forced convection cooling.
Actually, it cust be recognized that this method correlates threshold for measurable pin damage resulting from film boiling, assuming that l
pin bowing and film boiling threshold are coincident, at least in the i
temperature range of interest.
I I
27 I
I I
I FIGURE I
g TEST PIN ARRANGEMENT S ERIES III SERIES III I
THERMAL COLUMN THERMAL COLUMN 2
3 4
5 2
3 4
5 1
+
l 1
i Ob 8 0 a
%?
A OO J
J 00 eOO W
Ooe.
e g
B B, D,El M,
i 4l
,O 1
l g jj
. i O
D 1
h h]
f 3
l f
E I
E ll li
)
h-.
1 3
j r
! l a
k l.
e I
,,P. P,N - - -,
O HEATED PINS (FUEL) g e UNHEATED PINS (WATER FILLED)
I
T A l'. I.1 ; I l>ll J.SI:S ltIJ N T O 1)l;TI;(;T Fl1.M ltull. int i Tillt I;Sliol a >S DE SCR IPT10l!
TYPE OF PtJLSE EVIDENCE Pill. ';E HOT AVLRAGE t4AX liii.M
' 1A/' lfiUli PI?] Afl0
@ SUBJECT Piti !!UtiBrR OF F ll.H ft:rRGY SPOT I !!I RfiY Fili RGY 11 0 LOC AT IO?!
OllEATED PIN B0ll IllG (ftw-FACTOR PER GRAli l'l R t;RAli TiltP QtgjEATED sec)
(w-sec /gns (w-sec/gm) RISE (or SERIES Ill PULSES 142 No 45.6 3.25 162 527 2830 c
0.,
..~.
HOT EDGE CORE I
142 No 45.6 3 56 162 577 3100 I
0-1 IN 8-5 g g,
m.
CORNER
/
CORE 142 N0 45.6 2.89
'62 468 2520 A-5 IN 0-5 j,
m I
151 No 33.3 4.38 118 517 2780 EDGE TEST 150 No 35.0 124 544 2925 144 SLIGHT 37.3 132 578 3100
$ 8"#
143 YES 38.8 138 605 3240 142 YES 45.6 162 710 3780 lI 145 NO 33.1 4.02 117 471 2550 COR!lER 150 N0 35.0 124 498 2680 l
TEST
/
144 YES 37.3 132 531 2860 l
143 YES 38.8 13 8 555 2980 142 YES 45.6 162 651 3475 SERIES IV PULSES HOT EDGE TEST 171 No 25.2 4.62 89 5 413 2250 i
172 NO 26.4 93.8 433 2350 c-1 IN c-5 g
g 173 YES 28.7 102.0 471 2550
~.+<
110T CORNEP 170 NO 21.8 4.62 77.5 358 1950 TEST 171 NO 25.2 89.5 413 2250 E-1 IN C-3 172 YES 26.4 93.8 433 2350
M M
M M
M M
M M
M TABLEff CALCUI.ATION OF PERIMETER RATIO Uf! HEATED PERIMETEP' T10:1 TOTAL P
HEATED H
TYPE
@ SUBJECT Pl!,
p g g,ETEP' PIN PERIMETER
(! !)
P S HEATED Pil!
pH (IN)
P ill 30X TOTAL T
O UNHEATED PIN
!!!i:ER CORE 9'7 /4 5.904 5.904 1
b 4.428 1.206 1.206 5.634 0.786 C
O'G-6 CORNER 1.686 1.686 5.007 0.663 CORE 3.321 j j.
2.952 1.476 1.206 2.682 5.634 0.524 g3 2'.5 83 0.738 1.686 2.424 5.007 0.516 g
H 3.69 0.73 8 1.206 1.944 5.634 0.655
'E
/
TEST 9=w ~.
HOT EDGE 2.214 2.214 1.206 3.420 5.63 4 0.3 93 TEST f
1.845 1.475 1.686 3.161 5.006 0.369 CO ::ER TEST V9
M M
M M
M M
M M
M.
TABLE III CALCULATION OF HYDRAULIC DIAMETER CORRECTION CHANNEL i
TYPE
@$ HEATED PIN AREA g'ATER PERIMETER De De/De V/V V
! h CCORECT SUBJECT PIN (lIl. )
(lil)
(lM)
(ft/sec-l 80 CHANT!EL FOR De OUNHEATED PIN
(
- je e l
/7 l
6782 l 1.0
!!!NER NG G 0.1408 1.476 0.3816 1.0 -
1.0 5
eoe i
hh g 0.0704 1.3 41 0.21 0.55 0.74 3.7 5379 0.793 EDGE I
f0.699 CCRIER h9 0.0352 0.931 0.1512 0.396 0.63 3.15 4738 ee i
- BASED Cri BE?.N0TH'S CORRELATION WHERE:
SO =
0,890 ( g_)
48V h
+
De Di De 0.6 velocity (ft/sec)
V De : hydraulic diameter (ft)
Di = heater diameter (ft)
I l
l
m W
W W
m W
m M
TABLE IV CALCULATION OF CORRELATION FACTOR CHAN!!El DE SCR IPT PERillETER CORRECTION CCRRELAT10:1 TYPE
@ SUBJECT P:N RATIO FOR FACTOR PIN g HEATED PIN (P /P )
NEC g
O UN:iEATED Pit m. m/p" "
INNER CORE O
4 1.0 1.0 1.0 Nw U EDGE 0.786 0.793 0.623 0.663 0.699 0.463 CORE w w EDGE TEST 0.524 0.793 0.416 94 COR?!ER v -
TEST 0.516 0.699 0.3 61 go HOT i
E GE
.9 0.655 0.793 0.519 CORE I
1 I
HOT E03E 0.3 93 0.793 0.312 i
TE;T 1
0.359 0.699 0.25' TE R
l FIGURE 2 I
CHART OF PIN PERFORMANCE vs. UO TEMPERATURE RISE 2
l DURING SERIES II[ 8 E PULSE TESTS WITH FORCED CONVECTION COOLING I
E FILM BOILING g
B ranesnoto 4000 -
C NO FILM BOILING I
I g
3000 -
I w
- I
+
g I
m 2000 -
g 2w I
F-I Oa k
1000 -
z M
Z E
l
?
n 5
y z
i M
e H
8 E
e I
E E
8 M
e w
m
~
5 i
E 8
I a
s t
I O
O I
W I
W i
O-1
I l
FIGURE 3 l
CORRELATION OF PIN PERFORMANCE FORCED CONVECTION COOLING l
r lI 800
/
700 FILM B0lLING
/
/
o" 600 l
-)
o I
2 (0.519) 4 x
416)NO FILM o
(o s23)
O 500 o
I
(
si) g g
W (0.463) 400
.23,3 O
cc
'l W
z W 300 ll 2a 2
l Q 200 E
E E
E
,W 8
=
=
g g
o iOO g
g 8
0 5
5 e
I i
a l
l E
le 8 lS 8
lE S
y lI 0
O O.1 0.2 0.3 0.4 0.5 0.6 0.7 CORRELATION FACTOR
REFERENCES 1.
R. F. Lumb and J. MacPhee, PULSTAR PULSE TESTS, Report No. WNY-017, WNYNRCI, October 9, 1964.
2.
R. F. Lumb and J. MacPhee, PULSTAR PULSE TESTS - II, Report lo.
WNY-020, WNYNRCI, February 1, 1965.
3.
L. Bernath, "A Theory of Local-Boiling Burnout and Its Application to Existing Data," Chemical Engineer Progress Symposium Series, Vol. 56, No. 30, 1960.
4.
K. M. Becker, " Burnout Conditions for Flow of Boiling Water in Vertical Rod Clusters," AI Chemical Journal, March,1963.
5.
E. Janssen and S. Levy, " Burnout Limit Curves for Boiling Water Reactors," APED-3892, April 1, 1962.
6.
I. T. Alad'Yev and L. D. Dodoncv, " Critical Heat Flows in the Boiling of Underheated Water in Channels of Complex Shape in Convection and Radiation Heat Transfer," Report FTD-TT-61-395, (Abstract of pertinent data appears in Vol. 7, No. 3 of over Reactor Technolog,.)
7.
E. D. Waters, et al, " Boiling Bitrnout Experiments with 19-Rod Bundles in Axial Flow," HW-77303, August 1963.
8.
Letter to R. L. Doan from W. F. Hall on the subject of temperature coefficient dated May 19, 1965.
I I
t
,