CBBI-13 Page n° 1

Out-of-Pile Thermo-Mechanical Testing of Breeder Pebble Beds for

HCPB TBM for ITER

G. Dell’Orcoa, P. A. Di Maiob, R. Giammussob, A. Malavasia, L. Sansonea, A. Tincania, G. Vellab

(a) ENEA Brasimone, 40032 Camugnano (Bo), Italy (b) DIN-Dipartimento di Ingegneria Nucleare, Università di Palermo, Viale delle

Scienze, 90128, Palermo, ITALY

ABSTRACT The Helium Cooled Pebble Bed (HCPB) Blanket is one of the reference concept for the Test Blanket Module (TBM) to be tested in ITER. In the HCPB TBM module, alternate staked beds of Lithiated ceramics and Beryllium pebbles act respectively as Tritium breeder and neutron multiplier. The thermo-mechanical behaviours of the pebble beds and their nuclear performances in terms of Tritium production, are dependent from the reactor relevant boundary conditions (heat flux and neutron wall load), the pebble sizes, the breeder cell geometries (bed thickness, pebble packing factor) and the heat transfer parameters (temperatures, thermal conductivity, convective heat transfer coefficients etc.). ENEA, in the frame of the EU Fusion Technology Programme, has performed several experimental testing of small scale mock-ups. The “Dipartimento di Ingegneria Nucleare” (DIN) of the Palermo University has campaigns to determine some of these behaviours by out-of-pile adapted the thermo-mechanic constitutive models (non-linear elasticity and plasticity models), available on commercial FEM code, for the prediction of the thermal and mechanical performances of breeder pebble beds and for the comparison with the experimental results of the ENEA tests. More recently, among the EU Associations involved in the ceramic breeder qualification (ENEA-FZK-NRG), a benchmark exercise has been launched aiming at selecting the pebble bed thermal mechanical constitutive models to be implemented in a commercial FEM computer code. The paper presents the main experimental results of two test campaigns on HELICA I and II performed on HE-FUS 3 facility of ENEA Brasimone. The paper also presents the first comparisons with the theoretical calculations, carried out by commercial FEM code, in the frame of the benchmark exercise.

1. INTRODUCTION The prediction of thermo-mechanical behaviours of the breeder and multiplier pebble beds, in reactor relevant conditions, is one of the main concern of the design of the Helium Cooled Pebble Bed (HCPB) Blanket for DEMO and its Test Blanket Module (TBM) to be tested in ITER. In fact, the bulk thermal conductivity of these beds are dependent on both the bed local temperatures and strains. In the interface regions of the

CBBI-13 Page n° 2 steel containment walls, the heat transfer is dominated by a pressure-dependent thermal conductance. Furthermore, the thermal cycles due to the plasma pulses and the high design temperatures of both materials could also affected these dependencies due to pebble relocations during the pulses and creep effects at high temperatures. Therefore, several out-of-pile experimental test campaigns were launched at FZK (D) and ENEA (I) to determine these behaviours in small scale mock-ups with cylindrical [1] or prismatic geometries [2-3]. The trial to adapt a FEM commercial code to have a tool for the estimation of the pebble bed behaviour has given good results only for the first thermal-mechanical cycle [4]. Therefore, the EU Associations involved in the ceramic breeder qualification and modelling have agreed on the EFDA proposal to compare the experimental results of the out-of-pile experiments, carried out by ENEA on the mock-up HELICA with the theoretical calculations performed adapting the available constitutive models of the pebble beds in the non-linear elasticity, in the plasticity and in the creep regions. Two different cell orientations have been tested : i) vertical (polo-radial); ii) horizontal ( toro-radial).

2. HELICA MOCK-UP The HELICA mock-up, shown in Fig.1and Figs. 2.1-2.2 (in both two orientations), consists of a single welded cell made in ferritic-martensitic steel ASME SA 387 grade T91. The breeder cell, provided with two flat electrical heaters designed to reproduce the reference pebble temperature increase, is divided into three sub-cells 446 mm in width, 192 mm in depth and 4.6 mm in thickness. The total cell volume is 2012 cm3 (1197 cm3 in heated zone, 747 cm3 in the superior dead space and 68 cm3 in the level monitoring tube). The pebble bed temperature gradient is controlled by two external helium-cooled plates. The mechanical constraint on the lateral plates of the cell is obtained by using stack of SCHNORR springs set at variable preload. The test cell is closed by a flanged plug provided with the power supply and instrumentation leadthroughs. The pouring of the pebbles inside the cells is operated via a proper filling tube also used for the monitoring of the pebble level during the tests, Figs. 2.1 2.2. The bed level is also measured by a displacement gauge (LVDT) elastically pushed against the pebbles. The temperature distributions across the beds are measured by thermocouples located in the beds, on the heaters and on the CP surfaces. The cell lateral displacements are measured by 3 LVDT transducers for each side, as shown in Figs. 2.1-2.2.

3. TESTS RESULTS 3.1 HELICA I test campaign The first HELICA I test campaign has been performed on Li4SiO4 pebbles, 0.2-0.4 mm in diameter as qualified by FZK [5], with the cell oriented in vertical direction, Fig. 2.1. The breeder cell has been filled with 3.186 kg of pebbles, by pneumatic hammering at 5 kHz as in previous test campaigns [2-3], at a final packing factor of 65.6% and with an average density of 1.584 kg/dm3. HELICA has been helium cooled at a inlet flow rate of 8 g/s, 1.5-1.7 MPa and 250 °C. The breeder cell was purged by an helium flow of 2x10-

3 g/s, 0.05-0.1 MPa and 50 °C. The helium velocity among the pebbles has been kept lower than 1 cm/s. A total of 20 thermal ramps have been performed, the heat flux to the resistors being progressively increased from 0 up to an average value of 42 kW/m2 with

CBBI-13 Page n° 3 six increases of 7 kW/m2 and a duration of an hour per step, as shown in Fig. 3. The pebble bed maximum temperature close to the resistors was about 800 °C. During these tests, both the pebble temperatures and the lateral plates deformations have exhibited reproducible cyclical behaviours, as in Fig. 4 [4]. As in previous experiments carried out in vertical orientation [3], the pebble bed height has showed an irreversible overall bed height reduction per each thermal ramp, as illustrated in Fig. 5. The total measured overall bed reduction was about 12 mm, with a volumetric bed reduction of 4156.3 mm3 corresponding to a bed height reduction of 0.674 mm. The packing factor increase, estimated as +0.20%, is considered negligible. However, after 20 cycles, this irreversible height reduction has not yet reached an asymptote. The comparison between the experimental pebble average temperatures and the theoretical value obtained by 3D FEM model is showed in Fig. 6. The results are in very good agreement. 3.2 HELICA II test campaign The second HELICA II test campaign has been performed on Li4SiO4 pebbles, 0.2-0.4 mm in diameter of the same batch qualified by FZK [5] for HELICA I, with the cell oriented in horizontal direction. The pebble filling in the breeder cell and the thermal hydraulic boundary conditions were the same as in HELICA I test campaign. A total of 34 thermal ramps have been performed, the heat flux to the resistors being progressively increased from 0 up to an average value of 42 kW/m2 similarly as performed in HELICA I test. The pebble bed maximum temperature close to the resistors was about 860 °C, higher then obtained in HELICA I due to a better thermal insulation on the lateral sides of the mock-up. During these tests, both the pebble temperatures and the lateral plates deformations have exhibited reproducible cyclical behaviours but at a value higher than measured in HELICA I test carried with vertical orientation, Fig. 7, probably due to the overall bending of the containment cell and its frame at high temperatures. On the contrary, during this experiment the pebble bed height has not showed any irreversible overall bed height reduction due to the negligible effect due to the gravity on the pebble relocation in the breeder cells. The pebble bed thermal conductivity, showed in Fig. 8, was calculated from the measured heat fluxes and the temperature differences between the thermocouples located across the beds.

4. THERMO-MECHANICAL NUMERICAL ANALYSES The Department of Nuclear Engineering of the University of Palermo (DIN) has launched a theoretical research activity to assess the constitutive model previously developed to realistically describe the thermo-mechanical behaviour of fusion relevant pebble beds and intended to be implemented in a quoted Finite Element Method (FEM) code for design-oriented applications. The DIN model is based on the fundamental assumption that a pebble bed can be considered as a continuous, homogeneous, uniform and isotropic medium, characterized by effective properties whose values depend on temperature, pressure and or mechanical volumetric strain and may be determined by means of the typical tests for granular systems (oedometric, biaxial and triaxial tests, hot wire and current pulse techniques). The model is articulated into two coupled sub-models, namely the mechanical and the thermal one. The former is composed of a non-linear elasticity model that takes into account the dependence of the effective deformation modulus of the bed on its

CBBI-13 Page n° 4 equivalent pressure, a Drucker-Prager modified with Cap plasticity model that has been tuned to simulate the compaction induced irreversible behaviour of the beds and a Cap creep model. Recently the potential adoption of the Gurson plasticity model has been taken into account since its numerical implementation in the commercial code seems to be more stable. Further details can be found in [6, 7, 8]. The thermal model is based on the assumption that the effective thermal conductivity of a pebble bed can be represented as a second rank diagonal tensor, whose diagonal term depends linearly on both temperature and mechanical volumetric strain. At the interface region between the bed and the steel containing wall a thermal conductance model has been adopted that foresees a power law dependence on the contact pressure. A more detailed description can be found in [9, 10]. The model has been adopted to numerically simulate the thermo-mechanical behaviour of HELICA test section within the benchmark exercise promoted by ENEA-Brasimone labs. The six steady state configurations corresponding to each of the reference step increase of internal power densities assigned to the heater have been simulated by iteratively carrying out uncoupled steady state thermal and mechanical analyses. A 3D realistic FEM model of HELICA has been set-up (Figs. 9, 10) paying a particular attention to the simulation of the interface regions among the bed and both the box and the electric heater, where a contact model has been implemented allowing simulating both mechanical friction and gap thermal resistance. Attention has been focused also onto the simulation of helium flowing through the Cooling Plate (CP) channels to properly take into account the spatial distribution of the thermal power it extracts from the box. To that purpose forced convection elements have been adopted for helium and they have been thermally coupled with the cooling channels ones by means of a proper gap conductance condition. A proper set of loads and boundary conditions have been applied to the model as summarized in Fig. 11. Further details con be found in [9]. The spatial distributions of the thermal field at the end of each power step are shown in Figs. 12-17. In particular, together with the distributions within the whole box and the bed, that one relevant to the test section middle section has been reported too in order to highlight the temperature poloidal-radial profile within the bed. The temperatures calculated at the main position instrumented within the bed are reported in Figs. 18-19, together with the experimental ones. The results obtained seem to be encouraging especially as far as thermocouples 1, 2 and 4 are concerned. In fact, errors lower than 10 °C have been found between the numerical prediction and the experimental measurements. However it should be noticed that the gap conductance model at the interface between the bed and the heater should be improved since the temperatures predicted at thermocouple 3 seem to be overestimated in each one of the six power step considered. The theoretical-numerical investigations are still in progress to further asses the gap conductance model at the interface between bed and both bed and box. As far as the mechanical analysis the first trials have evidenced problems on the numerical convergence in coupled thermal and mechanical analyses. On the contrary, the model needs further modification to accelerate the convergences if decoupled analyses are performed. More detailed reproduction of the mechanical constraints and boundary conditions are also necessary to better reproduce the experimental results.

5. CONCLUSIONS The two HELICA test campaign carried out at ENEA Brasimone has given a further confirmation, already verified in previous test campaigns, that the Li4SiO4 pebble beds,

CBBI-13 Page n° 5 filled and packed in a prismatic breeder cell, also swept by helium purge flow, have exhibited repetitive thermal-mechanical behaviour when the bed temperatures are increased by stepping ramps. The thermal conductivity depends mainly on the bed average temperature with some not well estimated differences probably due to measuring errors. Although the pebble filling was performed by pneumatic hammering at high frequency, reaching a very high packing factor (65.6%), an overall bed height reduction was measured only in HEICA I tests with vertical orientation due to the effect of the gravity in the pebble bed relocations. However, in that case the packing factor increasing of +0.2% was considered negligible but, after 20 thermal cycles, the irreversible bed height reduction has not yet reached an asymptote. The measured lateral cell deformations are considerably different in the two orientations. In fact the values measured during the HELICA II test in horizontal direction are about the double than the vertical case. The preliminary decoupled thermal-mechanical 3-D analyses have been performed with an FEM code with the implementation of specific models to reproduce the non-linear elasticity and plastic regions of the mechanical strains. The results of the thermal analysis are encouraging because the errors in the estimation of the bed temperatures are within 10 °C. The results of mechanical analysis needs some modifications and simplification to accelerate the numerical convergence. More detailed reproduction of the mechanical constraints and boundary conditions are also necessary to better reproduce the experimental results. REFERENCES [1] J. Reimann, S. Hermsmeyer, Thermal conductivity of compressed ceramic breeder

pebble beds, Fus. Eng. and Des. 61-62 (2002) 345-351. [2] G. Dell'Orco, A. Ancona, P.A. Di Maio, L. Sansone, M. Simoncini, D. Zito, G.

Vella, Experimental Tests on Li-ceramic Breeders for the Helium Cooled Pebble Bed (HCPB) Blanket Design, Fus. Eng. and Des. 69 (2003) 233-240.

[3] G. Dell'Orco, A. Ancona, P. A. Di Maio, M. Simoncini, G. Vella, Thermo mechanical testing of Li-ceramic for the Helium Cooled Pebble Bed (HCPB) Breeding Blanket, Jour. of Nuc. Mat. 329-333 (2004) 1305-1308.

[4] A. Ancona, P. A. Di Maio, G. Vella, La sezione di prova HELICHETTA. Analisi termomeccaniche numeriche e sperimentali. Quaderni del DIN, 5/2003.

[5] R. Knitter et al., Fabrication of Lithium Orthosilicate by Melt-Spraying: Yield and Reproducibility, FZK Internal Report N° 231- December 2004.

[6] G. Vella, E. Oliveri, P. A. Di Maio, M. Dalle Donne, G. Piazza, F. Scaffidi-Argentina, Modelling of the Thermal Mechanical Behaviour of a Single Size Beryllium Pebble Bed. Fusion Engineering and Design, 58-59, pp. 635-640, 2001.

[7] G. Dell’Orco, A. Ancona, P. A. Di Maio, M. Simoncini, G. Vella, Thermo-mechanical testing of Li-Ceramic for the Helium Cooled Pebble Bed (HCPB) breeding blanket. Journal of Nuclear Materials, 329-333, pp. 1305-1308, 2004.

[8] G. Dell’Orco, P. A. Di Maio, M. Simoncini, G. Vella, D. Zito, TAZZA mock-up pebble beds. Experimental and theoretical investigations. Proceedings of the 10th International Workshop on Ceramic Breeder Blanket Interactions (CBBI), October 22-24, 2001, Karlsruhe, GERMANY, ed. L. V. Boccaccini, pp. 115-127.

[9] P. A. Di Maio, R. Giammusso, G. Vella, Sezione di prova HELICA. Analisi termomeccaniche. Rapporto per l’ENEA, Centro Ricerche Brasimone, Giugno 2005, Palermo.

CBBI-13 Page n° 6 [10] A. Ancona, G. Dell’Orco, P. A. Di Maio, L. Sansone, M. Simoncini, G. Vella,

Experimental test on Li-ceramic breeders for the Helium Cooled Pebble Bed (HCPB) blanket design. Fusion Engineering and Design, 69, n. 1-4, pp. 233-240, 2003.

FIGURE CAPTION Fig. 1 - HELICA mock-up scheme Fig. 2.1- HELICA I mock-up Fig. 2.2 -HELICA II mock-up Fig. 3 - Power supply and Heat Flux and during HELICA I-II tests Fig. 4 - LVDT displacements on lateral plates and Heat Flux during HELICA I test Fig. 5 - Overall LVDT displacement on top of the pebble bed and Heat Flux during

HELICA I test Fig. 6 - Comparison between the average pebble bed temperature in HELICA I test Fig. 7- Comparison of LVDT displacements on lateral plates and Heat Flux during

HELICA I - II tests Fig. 8 - Thermal conductivity calculated during HELICA II test Fig. 9 - The HELICA 3D FEM model. Fig. 10 - Middle section of the HELICA 3D FEM model Fig. 11 - Summary of loads and boundary conditions. Fig. 12 - Thermal field distributions at step 1 Fig. 13 - Thermal field distributions at step 2. Fig. 14 - Thermal field distributions at step 3. Fig. 15 - Thermal field distributions at step 4. Fig. 16 - Thermal field distributions at step 5. Fig. 17 - Thermal field distributions at step 6. Fig. 18 - Comparison between calculated and measured temperature at 100 mm. Fig. 19 - Comparison between calculated and measured temperature at 55 mm.

CBBI-13 Page n° 7

Fig. 1 – HELICA mock-up scheme

CBBI-13 Page n° 8 Electrical

Leadthroughs Top LVDT

Filling and monitoring tube

Lateral LVDTsFlanged plug SCHNORR cup springs

Fig. 2.1 HELICA I mock-up

Fig. 2.2 HELICA II mock-up

CBBI-13 Page n° 9

HEAT FLUX IN HELICA I - II

0

5

10

15

20

25

30

35

40

45

0 50 100 150 200 250 300 350 400 450 500

TIME [min.]

HEA

T FL

UX

[kW

/m2]

0

500

1000

1500

2000

2500

3000

3500

4000

POW

ER S

UPP

LY [W

]

Heat Flux (HELICA I)Heat Flux (HELICA II)Power (HELICA I)Power (HELICA II)

Fig. 3 – Power supply and Heat Flux and during HELICA I-II tests

LVDT 1-6

-35

-30

-25

-20

-15

-10

-5

0

5

10

15

1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000

Time [min]

LVD

T D

ispl

acem

ents

[1*E

-02

mm

]

-5

0

5

10

15

20

25

30

35

40

45H

eat F

lux

[kW

/m2] LVDT01

LVDT02LVDT03LVDT04LVDT05LVDT06Heat Flux

Fig. 4 – LVDT displacements on lateral plates and Heat Flux during HELICA I test

CBBI-13 Page n° 10

LVDT 0

20

21

22

23

24

25

26

27

28

29

30

4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000

Time [min]

LVD

T D

ispl

acem

ent [

mm

]

-5

0

5

10

15

20

25

30

35

40

45

Hea

t Flu

x [k

W/m

2]

Heat Flux

LVDT

Fig. 5 – Overall LVDT displacement on top of the pebble bed and Heat Flux during HELICA I test

AVERAGE BED TEMPERATURE [°C]

0

100

200

300

400

500

600

700

800

0 5 10 15 20 25 30 35 40 45 50Heat Flux [KW/m^2]

T [°

C]

Experimental Temperature (Polo-radial orientation)

Theoretical Temperature

Fig. 6 –Comparison between the average pebble bed temperature in HELICA I test

CBBI-13 Page n° 11

HELICA I - II tests - LVDT Displacements

-0.7

-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7

0 50 100 150 200 250 300 350 400 450 500

Time [min]

D [m

m]

0

5

10

15

20

25

30

35

40

45

Hea

t Flu

x [k

W/m

2]

HELICA I - LVDT lateral mmHELICA I - LVDT Axial mmHELICA I - LVDT lateral mmHELICA II - LVDT lateral mmHELICA II - LVDT axial mmHELICA II - LVDT lateral mmHeat Flux (HELICA II) kW/m2

Fig. 7– Comparison of LVDT displacements on lateral plates and Heat Flux during HELICA I - II tests

Li4SiO4 - Thermal conductivity

0.600

0.620

0.640

0.660

0.680

0.700

0.720

0.740

0.760

0.780

0.800

500 600 700 800 900

Average bed temperature [K]

Ther

mal

con

duct

ivity

[W/m

K]

Axial - 84 mm to FWLateral - 140 mm to FWLateral - 180 mm to FW

Fig. 8 – Thermal conductivity calculated during HELICA II test

CBBI-13 Page n° 12

Fig. 9 -The HELICA 3D FEM model.

Fig. 10 - Middle section of the HELICA 3D FEM model.

CBBI-13 Page n° 13

Natural convection with airT=20 °C

( ) ( ) ( )2q r,T T J rρ′′′ =

Mechanical and thermal symmetry

p=0.09 MPa Helium Tinlet=250 °C P = 1.5 MPa GTot=3.55 g/s

Heliump=1.5 MPa T fixed with the step GTot=0.44 g/s

Fig. 11 -Summary of loads and boundary conditions.

CBBI-13 Page n° 14

Fig. 12 -Thermal field distributions at step 1.

Fig. 13 -Thermal field distributions at step 2.

CBBI-13 Page n° 15

Fig. 14 - Thermal field distributions at step 3.

Fig. 15 - Thermal field distributions at step 4.

CBBI-13 Page n° 16

Fig. 16 - Thermal field distributions at step 5.

Fig. 17 - Thermal field distributions at step 6.

CBBI-13 Page n° 17

200

300

400

500

600

700

800

900

0 60 120 180 240 300 360

Time [min]

Tem

pera

ture

[°C

]

Thermocouple 4Thermocouple 3Thermocouple 2Thermocouple 1

Fig. 18 - Comparison between calculated and measured temperature at 100 mm.

200

300

400

500

600

700

800

900

0 60 120 180 240 300 360

Time [min]

Tem

pera

ture

[°C

]

Thermocouple 4Thermocouple 3Thermocouple 2Thermocouple 1

Fig. 19 -Comparison between calculated and measured temperature at 55 mm.