EXPERIMENTAL STRATEGY FOR THE DETERMINATION OF HEAT TRANSFER COEFFICIENTS IN PEBBLE-BEDS COOLED BY FLUORIDE SALTS

Transcription

1 EXPERIMENTAL STRATEGY FOR THE DETERMINATION OF HEAT TRANSFER COEFFICIENTS IN PEBBLE-BEDS COOLED BY FLUORIDE SALTS L. Huddar, P. F. Peterson Deartment o Nuclear Engineering, University o Caliornia, Berkeley, CA 94720, USA lakshana.huddar@berkeley.edu, eterson@nuc.berkeley.edu R. Scarlat 931 Engineering Research Building, Madison, WI 53711, USA Deartment o Engineering Physics, University o Wisconsin Madison roscarlat@wisc.edu Z. Guo Deartment o Nuclear Engineering, University o Caliornia, Berkeley, CA 94720, USA Deartment o Nuclear Science and Technology, Xi an Jiaotong University, Xi an, , China zhangengguo@berkeley.edu ABSTRACT It is imortant to accurately model the heat transer coeicient between the uel ebbles and the libe coolant in order to correctly redict uel temeratures in the core o the luoride-salt-cooled, high temerature reactor (PB-FHR. We have been erorming exeriments using simulant oils that match key non-dimensional arameters exected in PB-FHRs. A 3.5 long test section is illed with ¼ coer ebbles, some o which are instrumented with thermocoules. Oil is circulated through the test section. The entering temerature o the oil is varied, and the time-varying exit temerature o the oil and temeratures o the instrumented ebbles are recorded. Using these temeratures the interacial heat transer coeicient can be extracted as a unction o osition and time. Correlations or interacial heat transer coeicients are available in the literature, but these are derived using exerimental data that do not entirely encomass PB-FHR oerating conditions. Generally or the PB-FHR the Reynolds and Prandtl numbers during normal oeration are higher than reorted exerimental results. Thus, it is imortant to erorm tests in the aroriate PB-FHR Reynolds and Prandtl number range. Preliminary results indicate that exerimentally measured heat transer coeicients may be at least 20% higher than redicted using Wakao and Funazkiri s correlation or Nusselt number in ebble beds. These reliminary results also reinorce the need to reduce uncertainties in the collected data. This aer will also resent exerimental techniques that aim to accomlish this. Seciically, we detail how requency resonse techniques can be used to extract heat transer coeicients rom a ebble-bed test section and how exeriments can be designed using simulant oils to achieve this. 1. INTRODUCTION AND BACKGROUND KEYWORDS FHR, ebble-bed, heat transer coeicient, simulant oils, scaled exeriments. Fluoride-salt-cooled high temerature reactors (FHR make u a class o advanced nuclear reactor designs. The University o Caliornia, Berkeley, is investigating a small modular ebble-bed reactor (PB- FHR. One o the key henomena o interest is the high Prandtl number coolant heat transer in the ebble bed core. The core is comosed o sherical uel elements that are randomly acked in an annular 1659

2 cylindrical geometry. The heat transer coeicient between the ebbles and the coolant needs to be well characterized in order to be able to redict the uel and libe temeratures in the reactor. Additionally, the heat transer coeicient needs to be known or a range o Reynolds numbers. We use simulant oils in lace o libe, the PB-FHR main coolant, while matching Prandtl and Reynolds numbers between the exeriment and the rototyical system [1]. 1.1 Literature Review Wakao s review aer lists all the exerimental data that was collected or heat transer coeicients in acked beds as a unction o Reynolds and Prandtl numbers [2]. The majority o the data was collected or air or water. The range o Reynolds or their roosed correlation was rom The exerimental data that this correlation was based on were all done with air or other gases, with Prandtl numbers ranging rom 0.7 to 1. The correlation roosed by Wakao and Funazkri is widely used in acked bed heat transer redictions. Carillo inds the heat transer coeicient between oil and a bed o steel sheres, and also studied the eect o orosity on the Nusselt number. They roose correlations or ive dierent orosities or ranges o Reynolds numbers that go rom 0.53 u to 412 [3]. Geb et al. low air through a randomly acked test section o steel ebbles heated via induction or a range o Reynolds numbers rom 20 to 500 [4]. This was done or a non-ininite bed so wall eects layed a role in the lower than exected measured Nusselt numbers. Handley and Heggs develo a correlation or heat transer in ixed acked beds or Reynolds numbers higher than 100 [5]. In the current work, heat transer coeicient is measured or higher Prandtl numbers, in the range imortant or heat transer to molten salts, and a range o Reynolds numbers in an ininite randomly acked ebble bed using oil as the heat transer luid. 1.2 Exerimental Aims o the Current Investigation There were three main aims to the exeriment: (1 To measure heat transer coeicients in ebble beds or a range o Reynolds and Prandtl numbers alicable to the ebble-bed FHR (PB-FHR cores, and understand any reasons or otential discreancies between measured values and values redicted using correlations rom the literature (2 To develo an exerimental basis or heat transer coeicients in ebble beds or a range o Reynolds and Prandtl numbers beyond those exected or FHRs during normal oeration (3 To generate data that could be used to conirm the use o simulant oils in redicting the heat transer behavior in ebble beds cooled by luoride salts. 1.3 Structure o the Paer The exerimental rocedure using a ste change in the inlet luid temerature is irst outlined. This includes the design o ebble bed test section and the use o simulant oils. The data reduction rocedure is shown and the sources o otential errors are discussed. The results obtained rom these exeriments are shown. Since all the aims outlined in Section 1.2 were not achieved in this irst set o exeriments, the design or another exerimental acility is described. This new exerimental acility design is otimized to measure heat transer coeicients in the ebble bed test section using requency resonse techniques. The aer ends with conclusions drawn rom the irst exerimental eort and rovides recommendations or how exerimental data can be best collected going orward. 2. EXPERIMENTAL PROCEDURE An exerimental loo was constructed or the urose o measuring heat transer coeicients in a ebblebed test section. A test section was illed with randomly acked coer ebbles, and was heated to an initial temerature. Then cold luid was assed through the test section and the temeratures o the 1660

3 ebbles and the luid inlet and outlet temeratures were recorded throughout the thermal transient using Tye T Omega manuactured thermocoules embedded inside selected ebbles. These temeratures were then used to extract the heat transer coeicient between the ebbles and the surrounding luid as a unction o time and axial location. The mass low rate throughout the transient was also recorded using a Coriolis low meter (Siemens MASSFLO MASS 2100 DI. The exerimental acility is shown in Figure Test Section Design The test section is cylindrical, with a length o 88.9 mm (3.5 inches and a diameter o 44.5 mm (1.75 inches and is illed with mm (1/4 inch diameter coer ebbles. The wall o the test section is dimled to break u ordered acking o the ebbles at the wall in order to better simulate an ininite bed. A icture o the test section is shown in Figure 2. Some o these ebbles were instrumented with thermocoules. Thermocoules were also used to measure the bulk luid temerature at the inlet and outlet o the test section. Figure 3 shows the location o thermocoules within the test section. Pebble temeratures in various axial and radial locations in the test section were recorded. The instrumented ebbles had holes drilled to their centers where thermocoules were cemented. Because o the high thermal conductivity o coer, the ebble surace temerature can be assumed to be the same as the center temerature. The Biot number ranges in an individual coer shere was rom to 0.024, which is smaller than 1. Thus, this aroximation is justiied. Figure 1: Test section illed with coer ebbles or heat transer coeicient measurement exeriments 1661

4 Figure 2: The test section is divided into 5 axial regions. Instrumentation locations or ebbles (labelled with s and bulk luid temeratures (labelled with are shown. 2.2 Use o Simulant Oils as the Heat Transer Fluid We use Dowtherm A and Drakesol 260AT to simulate the molten salt libe, which would be used to cool ebble bed FHRs. Table 1 shows tyical Reynolds and Prandtl numbers in a PB-FHR core during normal oeration (orced circulation [6] and Table 2 shows the obtainable ranges o non-dimensional numbers using the simulant oils. Currently, exeriments have been carried out using Drakesol 260AT only. The lowest achievable Prandtl number o the Drakesol 260AT is still higher than or Dowtherm A because o the higher viscosity. Since exerimental data in this range is generally lacking, it is helul to collect this data even i the Prandtl range does not overla with the PB-FHR oerating conditions. The hysics o the heat and momentum transer at Prandtl numbers higher than unity is very dierent to Prandtl numbers smaller than unity, so any data collected at higher Prandtl numbers is still valuable. Reynolds and Prandtl numbers are deined in Equations (1 and (2. Ud Re (1 (2 The suericial velocity U is deined as the velocity through an emty ebble-bed in m/s, the diameter d reers to the ebble diameter in m, and the luid roerties ν and α reer to the kinematic viscosity (in units o m 2 /s and thermal diusivity (in units o m 2 /s taken at ilm temeratures, which is deined in Section 5.2 in Equation (5. Drakesol 260AT matches the Prandtl number o libe at temeratures between 60 C and 140 C as shown in Figure 4. Table 1 shows the Reynolds and Prandtl number ranges or the 236 MWth PB-FHR design during normal oeration, which corresonds to orced circulation through the ebble bed core, and emergency shutdown cooling, which corresonds to natural circulation through the reactor core. These are aroximate based 1662

5 on the average inlet and outlet temerature in the reactor core, 600 C and 700 C resectively. We attemted to achieve the correct range o Reynolds and Prandtl numbers in the exeriment. Table 1: Tyical non-dimensional arameter values in the PB-FHR reactor core or orced circulation (ower oeration and natural circulation (decay heat removal Reynolds in the core Prandtl in the core Forced circulation Natural circulation Figure 3: Prandtl number comarison o the simulant oil Drakesol 260AT and Flibe The rocedure or heating the oil to the required temerature is outlined in Figure 5. In Phase I, low is directed simultaneously through the test section branch and the heater branch. Thus the test section temerature can be increased to the desired value. The maximum oil temerature that was obtained with the Drakesol 260AT was 60 C. In Phase II, the oil was raidly cooled down using the byass branch, at which time the other two branches were valved o. During this hase the loo was actively cooled by running chilled water through the heat exchanger. It is imortant to note that the test section had insulation wraed around it. Thus the temerature o the test section was maintained throughout the entirety o Phase II. The duration o Phase II is ~15 seconds. The aim o this hase is to decrease the oil temerature by a minimum o 5 C, so the temerature dierence between the ebbles and the luid will be suiciently large to erorm the data reduction with. In Phase III the colder luid is then re-routed through the test section branch, and then transient data collection begins. 1663

6 5a: Phase I 5b: Phase II 5c: Phase III Figure 4: Procedure or transient heat transer data collection or the ebble-bed test section divided into three hases. In all schematics, the um was run in the counter-clockwise direction. A summary o the exerimental runs is given in Table 2. The initial ΔT reorted is the temerature dierence that was achieved between the entrance control volume section in the test section and the initial inlet luid temerature as recorded by the thermocoules. Transient duration reers to the time it takes rom when the initial cold luid enters the test section to when the test section reaches equilibrium conditions. This number is much smaller than the luid residence time in the loo, which imlies that the luid inlet temerature is not aected during the transient. 1664

7 Table 2: Exerimental arameters or the 4 exerimental runs Exerimental run Reynolds range Prandtl range Initial T in section 1 ( C Transient duration (s 3. DATA REDUCTION PROCEDURE This section covers how the interacial heat transer coeicient was determined rom the raw exerimental data and when arameters went into the uncertainty analysis rocedure. There is also a discussion about the assuming quasi-steady state conditions during the exeriment. 3.1 Deriving Exerimental Interacial Heat Transer Coeicient The heat transer coeicient is extracted rom the temerature data by equating the rate o change o internal energy o the ebbles and the convective heat transer between the ebbles and the surrounding luid, as shown in Equation (3b. Fluid temeratures rom the middle sections were estimated using the luid energy conservation equation using inite dierences, as shown in Equation (4. Equation (3b and (4 can then be used or each control volume section. Thus the heat transer coeicient extracted rom the data is a unction o axial osition and time. There is not much variation o temerature in the radial direction in the test section. The variation that is resent alls within the error in reading o the Tye T thermocoules rom Omega Engineering, which is 1 C. It was ound that during equilibrium conditions, the ebble and the luid thermocoules were not reading the same temerature. This imlied that the calibrations o the thermocoules were not as exected. The luid temeratures generally read lower than the ebble temeratures by about 2.5 degrees at temeratures higher than 50 C. This was taken into account during the data reduction rocedure, by subtracting this discreancy rom the measured dierence between ebble and luid temerature. The uncertainty on the thermocoule readings was a contributing actor to the decision to build a new exerimental acility ocused on measuring heat transer coeicients in ebble-beds. (1 ( c ( c ( c T 1 t T 1 t Ts s t v section1 section1 section1 Ts (1 ( c s * t h ( t a ( T T ( t s * h ( t a section1 ( c ( c section1 v v z v ( T z T ( t s T z T z section1 section1tosection2 section1tosection2 h s h s ( T s1 ( T s1 T T 1 section1 1 section1 (3a (3b (4 The deinitions o the terms in Equations (3 and (4 are as ollows: ɛ is the orosity, (ρc s the volumetric heat caacity o the ebble (coer in the test section, T s the temerature o the ebble ( solid hase, t time, h* the heat transer coeicient, a v the seciic surace area o the test section, deined in Equation 1665

8 (10, T the oil temerature ( luid hase, (ρc the volumetric heat caacity o the oil, v z the axial suericial velocity o the luid in the test section, z the axial osition co-ordinate and h s the roduct o h* and a v. The subscrit section 1 reers to control volume section 1, shown in Figure 3. The subscrit 1 reers to roerties in control volume section 1. The subscrit section 1 to section 2 reers seciically to the advection term, and reers to the axial distance between the entrance to section 1 and the entrance to section 2. The orosity ɛ was assumed to be a constant and taken as 0.4 or a loosely-acked randomly acked ebble bed (re. The orosity will be measured in the next iteration o this exeriment. The Nusselt number was redicted at every time ste using the measured temeratures to evaluate the relevant nondimensional numbers. The correlations or Nusselt number along with the deinitions or Reynolds and Prandtl are given in Equations (6 and (7. The correlations are reerred to as Handley Heggs and Wakao [2], [5] resectively. The KTA correlation [7] was develoed or gas cooled ebble-bed reactors and was considered, but was not exected to accurately deict the interacial heat transer coeicient o the test section. This is because the Prandtl number range or helium, the coolant o the gas cooled reactors, is about 1, much lower than libe at the PB-FHR oerating temerature which is about 14. The ilm temeratures were used to evaluate the Reynolds and Prandtl numbers, which were in turn used to evaluate the Nusselt number. The exerimental h* was used to ind the exerimental Nu s so it could be comared to the correlations. This was done using Equation (9, which takes into account the article conductivity [8]. The ilm temerature is deined in Equation (5. The ilm temerature is used because it is the average temerature across the thermal boundary layer over each ebble. Ts T T ilm 2 (5 Nu s = Pr1/3 Re 2/3 (6 Nu s = Pr 1/3 Re 0.6 (7 h s = a v h * (8 1 h = d + d * Nu s k k s (9 a v = 6(1 d so a v = 6( = 567m1 in the test section (10 k is the thermal conductivity o the oil in Wm -2 K -1, k s is the thermal conductivity o the coer in Wm -2 K -1, Pr the Prandtl number (as deined in Equation (2, Re the Reynolds number (as deined in Equation (1, d the article diameter in m, β a geometrical constant deending on the shae o the article (or sheres it is 10. In the current study the heat transer coeicients rom the central three sections (sections 2, 3 and 4 in Figure 3 have not been evaluated given the uncertainties in the thermocoule readings o the luid bulk temeratures. These are the temeratures that would be used to estimate the luid temerature within the test section. The next iteration o this exeriment aims to ensure that heat transer coeicients rom the central control volume sections can be measured with reasonable accuracy. 3.2 The Assumtion o Quasi-Steady State Conditions 1666

9 Although this exeriment is a transient study, quasi-steady state conditions can erhas be assumed. The ratio o the volumetric heat caacity between the luid and solid hases is an indicator or whether quasisteady conditions may be assumed. Figure 6 shows the volumetric heat caacities o the coer ebbles and the Drakesol 260AT oil. I the ratio o (ρc / (ρc s is smaller than 1, this imlies that the temerature o the solid changes more slowly than the luid temerature. This is the case with the Drakesol 260AT oil and coer, with a ratio o This ratio is smaller than 1 but still close to unity. This means that there may be distortions due to transient eects. The measured heat transer coeicients are being comared to steady state correlations. The new exeriment discussed in Section 5 o this aer should be able to reduce any distortions due to transient eects as the temerature dierences are designed to be eriodic. Figure 5: Diagram o ebbles and oil (solid and luid hases 3.3 Uncertainty Analysis Procedure The biggest contributions to uncertainty in the exerimentally derived h* are rom dt s /dt, T s and T. The orosity ɛ also has an associated uncertainty with it since it was not measured rior to the tests. It is imortant to note that the uncertainty analysis is aroximate because the uncertainties associated with the material roerties o the oerating luid Drakesol 260AT are unknown. The thermal conductivity and seciic heat caacity o Drakesol 260AT were assumed to be constants, which may not be the case because these roerties are temerature deendent in other similar oils. It was assumed that the uncertainty in the Drakesol 260AT roerties were about 10%. The uncertainties associated with the roerties o coer were assumed to be small enough to neglect. Uncertainty is also associated with the redicted Nusselt numbers, as the non-dimensional numbers used to calculate them were based on luid temeratures. Thermocoule errors are associated with them. Table 3 shows the uncertainties associated with the instrumentation readings. Even though the error associated with the thermocoule temeratures was 1 C as reorted by Omega Engineering, we ound a larger discreancy between the readings at higher temeratures. Table 3: Uncertainties associated with instrumentation readings Instrumentation Error Tye T Thermocoule +/- 2.5 C Coriolis Flowmeter 1% o reading 4. RESULTS AND DISCUSSION FROM STEP RESPONSE TESTS The exerimentally derived Nusselt number is lotted as a unction o time. The redicted Nusselt number is also lotted or the corresonding exerimental runs. The redicted Nusselt number was calculated at each time ste using the measured Reynolds and Prandtl numbers. The range o Reynolds and Prandtl numbers are given or each run in Table 2. The error bars rom instrumentation readings and luid roerty uncertainties have also been lotted. The results rom control volume section 1 and 5, the entrance and exit sections resectively, are reorted here. The temerature transient or Run 2 is also lotted as an examle in Figure 7. Towards the end o the transient the temerature dierence between the ebbles and the oil decreases, and thus the calculated heat transer coeicient becomes arbitrarily 1667

10 large. Thereore these time stes are not taken into account in the calculation o the heat transer coeicient. Figure 6: Smoothed therocoule readings as a unction o time. T5-1, T5-2 and T1-1 are the luid temeratures, and the rest are ebble temeratures 1668

11 Figure 8: Nusselt number comarisons or Run 1 Figure 9: Nusselt number comarisons or Run

12 Figure 10: Nusselt number comarisons or Run 3 Figure 11: Nusselt number comarisons or Run 4 Figures 8 to 11 show the Nusselt numbers as a unction o time or the entrance and exit control volume sections. During these runs the achieved Reynolds and Prandtl numbers were lower than the oerating PB-FHR conditions, and this is artly because o the high viscosity o the Drakesol 260AT. In uture iterations o this exeriment, it would be better to use Dowtherm A as the oerating luid given that Prandtl numbers o 15 can be reached at a temerature o about 90 C. Generally, it was ound that the exerimental Nusselt numbers were higher than the redicted correlations by u to 360 %. More exerimental data or a lower Prandtl number and higher Reynolds number needs to be collected in order to draw any inal conclusions. This is esecially the case because a high temerature dierence between the ebbles and the oil was not achieved. 1670

13 5. EXPERIMENTAL DESIGN USING FREQUENCY RESPONSE TECHNIQUES The data collected in the exerimental acility described in Section 2 o this aer has signiicant uncertainty attached to it, and thereore the quality o the results can be substantially imroved using other exerimental techniques. For this reasons we are constructing a new exerimental acility that will have the caability to erorm tests that study the resonse o the test section temerature due to ste changes in the luid temerature (as beore and also due to sinusoidal changes in the luid temerature. Frequency resonse techniques have been used reviously to measure heat transer coeicients in luidized beds, in which the solid articles have an associated velocity [9]. One o the advantages o requency resonse techniques outlined in these studies is that the solid temerature within the test section need not be directly measured. These studies use a model to estimate the solid and luid temeratures within the test section. In UCB s test section, the temeratures o some o the coer sheres are measured. Thus we will have less uncertainty in the derived heat transer coeicient comared to revious studies. There are three main reasons we consider using requency resonse techniques to measure heat transer coeicients in the ebble-bed test section: (1 In the data reduction rocedure outlined in Section 3.1 o this aer, h is a unction o the derivative o the ebble temerature T s. With a ste change in the ebble temeratures, the sloe very stee and is thus subject to errors. I the ebble temerature was made to vary sinusoidally, the derivative could be easily obtained, and to a higher accuracy. (2 The data collection eriod can be much longer than it was in the revious exeriment as we would be oerating the exeriment under eriodic steady state conditions, thereby reducing any otential distortions due to transient eects. (3 With the new exerimental coniguration we would have more control over the minimum and maximum temeratures o the luid, and thus we could have larger ΔT than was reviously obtained. Thus more data can be collected. 5.1 Governing Equations in the Test Section and Analytical Solutions The governing equations or the solid and luid hases are given in Equations (11 and (12. The boundary conditions are also given (Equations (13 (15. These equations can be non-dimensionalized using the arameters in Equations (16 (22. The non-dimensional governing equations and boundary conditions are given in Equations (23 (27.These equations can be solved and used to redict the luid and solid temeratures in the test section as a unction o time and one satial derivative. In the rerediction models, h will be estimated rom the Wakao correlation [2]. T T hs vz ( Ts T t z ( c (11 T h s s ( T Ts t (1 ( c s (12 T ( t 0 T 0 (13 Ts ( t 0 Ts0 T 0 (14 T z 0 T T sin( (15 ( 0 0 t In equations (11 to (15, T 0 is the initial luid temerature in C, T s0 the initial solid temerature in C, α the amlitude o the luid temerature oscillation and ω the requency o the oscillation in s

14 T F T T Ts T 0 S (17 T z x L (18 hs (c t (19 z s hs L v (c (20 z s hs L v (c (21 c s s (22 h (16 F F ( S F x (23 S ( F S (24 F ( 0 0 (25 S ( 0 0 (26 F ( x 0 sin( (27 1 Here, π relaces to simliy the equations. (1 Analytical solutions or the entrance section (section 1, where x = 0, can be determined and are given in Equations (27 and (28. 2 sin cos e S ( x 0 (

15 The solid and luid non-dimensional temeratures as a unction o non-dimensional time are shown in Figure 12. The solid temerature lags behind the luid temerature, and this lag can be controlled by choosing the aroriate non-dimensional constants. Non-dimensional temerature Non-dimensional time S F Figure 12: Fluid and solid non-dimensional temeratures as a unction o time at the entrance (x = 0. The red line reresents the luid inlet temerature and the blue line reresents the solid temerature An aroriate requency o the inlet luid temerature needs to be determined. We can use the solutions at x = 0 to estimate an aroriate value or η, and thus ω. In order to do this, σ and φ do not need to be estimated, as they do not aear in the x = 0 solutions. The values or η and ɛ were used to lot the grah in Figure 12, where was taken as 1.67 and η arbitrarily chosen as 1. The value o η can be adjusted deending on how much lag between the luid and solid temerature is needed. It is imortant to note that during the exerimental run, the numerical value o the non-dimensional numbers in equations (16 to (22 will change i they are deendent on h s or luid roerties. These are a unction o temerature, and thus will be changing throughout the run. 5.2 Exerimental Design A schematic o the new exerimental loo is shown in Figure 13. The oerating luid will be Dowtherm A in order to match the Prandtl number o libe at lower temeratures than or Drakesol 260AT. The old thermocoules that were reviously installed on the test section will be relaced with new and calibrated tye T thermocoules. A tankless heater will be used to heat the oil. This tye o heater was chosen or its low thermal inertia. It is designed to instantaneously heat water to the required temerature. The ower into the heater will be controlled using a DC ower suly. LabView will be used to control the DC ower suly to the heater to allow or a sinusoidal heater ower inut. A late-and-rame heat exchanger will be used to cool the luid beore it enters a large reservoir o the oil. There are no exerimental constraints on choosing the requency o the temerature oscillation, and thus it can be chosen reely. A variable seed drive controls the um seed, thus allowing or easier control o the mass low rate through the loo comared to the revious exeriment. 1673

16 Figure 13: Diagram o the new exerimental design including imortant comonents 6. CONCLUSIONS AND FUTURE WORK The main aims o this investigation were to measure heat transer coeicients in ebble beds and comare the measurements with redictive correlations. This was done in suort o develoing a better understanding o heat transer in the PB-FHR reactor core. A test section illed with randomly acked coer sheres was heated to a certain temerature, ater which a ste change in the inlet luid temerature was established. The temeratures o the luid inlet and outlet were recorded throughout the transient, as well as the ebble temeratures. Exerimental heat transer coeicients were extracted rom this. It was ound that the correlations underestimated the exerimental values. This needs urther investigation, and the design o another exeriment to study heat transer coeicients using requency resonse techniques was roosed. 7. ACKNOWLEDGEMENTS This material is based uon work suorted by the National Science Foundation Graduate Research Fellowshi under Grant No DGE and suorted by the US Deartment o Energy Oice o Nuclear Energy University Programs Integrated Research Project. 1674

17 8. REFERENCES [1] P. M. Bardet and P. F. Peterson, Otions or Scaled Exeriments or High Temerature Liquid Salt and Helium Fluid Mechanics and Convective Heat Transer, Nucl. Technol., vol. 163, , [2] T. Wakao, N., Funazkri, EFFECT OF FLUID DISPERSION COEFFICIENTS PARTICLE-TO- FLUID HEAT TRANSFER COEFFICIENTS IN PACKED BEDS, Chem. Eng. Sci., [3] L. P. Carillo, Convective heat transer or viscous luid low through a metallic acked bed, Interciencia, vol. 30, no. 2,. 81, [4] D. Geb and I. Catton, Internal Heat Transer Coeicient Determination in a Packed Bed From the Transient Resonse Due to Solid Phase Induction Heating, vol. 134, no. Aril, [5] P. J. Handley, D., Heggs, Momentum and Heat Transer Mechanisms in Regular Shaed Packings, Trans. Inst. Chem. Eng., vol. 46, no. 9, [6] P. F. Andreades, C., Cisneros, A. T., Choi, J. K., Chong, A. Y. K., Fratoni, M., Hong, S., Huddar, L., Hu, K. D., Krumweide, D. L., Lauer, M. R., Munk, M., Scarlat, R. O., Zweibaum, N., Greensan, E., Peterson, Technical Descrition o the Mark 1 Pebble-Bed Fluoride-Salt-Cooled High Temerature Reactor (PB-FHR Power Plant, [7] N. S. S. C. (KTA, Nuclear Saety Standards Commission ( KTA Reactor Core Design o High- Temerature Gas-Cooled Reactors Part 2 : Heat Transer in Sherical Fuel Elements ( June 1983, [8] D. L. Dixon, A. G., Cresswell, Theoretical Prediction o Eective Heat Transer Parameters in Packed Beds, AIChE J., vol. 25, no. 4, , [9] A. P. Littman, H., Stone, Gas-Particle Heat Transer Coeicients in Fluidized Beds by Frequency Resonse Techniques, Fluid Part. Technol. - Chem. Eng. Prog. Sym. Ser., vol. 62,