Conference Paper

0 downloads 8 Views 879KB Size Report
Jun 3, 2013 - Keys Building, 1 Plant Road, Chalk River, Ontario, Canada K0J 1J0. *E-mail: [email protected] KEYWORDS: heavy water, thorium, plutonium ...


Prepared by Rédigé par

Reviewed by Vérifié par

Approved by Approuvé par





Atomic Energy of Canada Limited

Énergie Atomique du Canada Limitée

Chalk River, Ontario Canada K0J 1J0

Chalk River (Ontario) Canada K0J 1J0

Manuscript Submission for ANS Nuclear Technology


HIGH UTILIZATION LATTICES FOR THORIUM-BASED FUELS IN HEAVY WATER REACTORS BLAIR P. BROMLEY Atomic Energy of Canada Limited – Chalk River Laboratories Keys Building, 1 Plant Road, Chalk River, Ontario, Canada K0J 1J0 *E-mail: [email protected] KEYWORDS: heavy water, thorium, plutonium, WIMS-AECL

Page 1 of 46

Manuscript Submission for ANS Nuclear Technology


ABSTRACT New fuel bundle and lattice concepts to implement thorium-based fuel cycles in pressuretube heavy water reactors (PT-HWR) have been explored to achieve maximum resource utilization. As an existing, operational technology, PT-HWRs are highly advantageous for implementing the use of thorium-based fuel cycles because of their high neutron economy and on-line re-fuelling capability. A PT-HWR is flexible in that it can use one, two, or more different types of fuels in either homogeneous or heterogeneous cores to optimize power production, fuel burnup, and new fissile fuel production. In a heterogeneous PT-HWR core, higher fissile content seed fuel will be optimized for power and excess neutron production, and lower fissile content blanket fuel will be optimized for production of U-233. Five different lattice concepts were investigated for potential use in a once-through thorium (OTT) cycle in a PT-HWR. The lattices involved 43, 35 and 21-element bundles with a central cluster of ThO2 pins, or a Zr-4 central displacer tube containing either stagnant D2O coolant or solid ZrO2, to help reduce coolant void reactivity (CVR). The fuel in the outer pins is a homogeneous mixture of Th and LEU (~5 wt% U-235/U) or reactor-grade Pu (~67 wt% fissile). The content of the LEU or Pu was varied to achieve different levels of burnup, and it is presumed that low-reactivity fuel would be used as blanket bundles. It was found that the various lattice concepts could achieve burnups ranging from ~10 MWd/kg to 80 MWd/kg, and that the fissile utilization could be up to 60% to 100% higher than what is currently achieved in a PT-HWR using natural uranium (NU) fuel. Burnup-averaged CVR ranges from ~+1 mk to +16 mk, depending on lattice type and fuel composition. Assuming a maximum linear element rating of ~50 kW/m, the maximum permissible bundle power ranges from ~520 kW to 800 kW.

Page 2 of 46

Manuscript Submission for ANS Nuclear Technology I.



The use of thorium-based fuels in pressure tube heavy-water reactors (PT-HWR) can improve the long-term sustainability of nuclear power, while making use of an existing, operational reactor technology. Thorium is approximately three times as abundant as uranium in the Earth’s crust, representing a huge long-term energy resource potential 1. However, as a fertile-only fuel, it must be mixed with fissile isotopes of uranium and/or plutonium to sustain criticality and to generate power. Fissile Pu (Pu-239, Pu-241) and/or U-235 can be obtained from recycled stockpiles of uranium-based fuels from light water reactors (LWRs), Magnox and Advanced Gas-cooled Reactors (AGR) and RBMK reactors. U-235, in the form of low enriched uranium (LEU) can be obtained directly from uranium enrichment facilities. U-233 can be obtained from recycled thorium-based fuels irradiated in thermal or fast converter or breeder reactors 2, 3, 4. Other potential sources of fissile Pu and U-233 are from sub-critical blankets of U-238 or Th-232 bombarded by fast neutrons from an accelerator-driven spallation neutron source 5, 6, or a fusion neutron source in a hybrid fusion-fission reactor (HFFR) 7, 8. Thorium-based fuels can be used in either homogeneous (one fuel type) or heterogeneous cores (two or more fuel types) in a PT-HWR

9, 10

. In the case of heterogeneous cores,

thorium-based fuels with a higher content of fissile fuel would serve as seed fuel, generating power and excess neutrons to sustain criticality. Thorium-based fuels with a lower fissile content would serve as blanket fuel, generating less power, and absorbing neutrons to breed U-233 from Th-232.

Page 3 of 46

Manuscript Submission for ANS Nuclear Technology


PT-HWRs 11, 12 are very attractive for implementing thorium-based fuel cycles because of their high neutron economy and on-line re-fuelling capability, allowing more flexibility in using different fuel types. Neutron-absorbing non-fuel materials are minimized as much as possible in PT-HWRs, and the relatively small and simple fuel bundles ease the fabrication and handling of radioactive fuels. The use of thorium-based fuels in PT-HWRs has been investigated in the past



more conventional 37-element bundle configurations (with  1.6 wt% fissile/HM). Studies have also examined the use of 43-element bundles with graded enrichments and burnable neutron absorbers (BNA) to help minimize coolant void reactivity (CVR) while achieving higher burnups ( 20 MWd/kg)

13, 14

. The CVR is the change in lattice

reactivity when the coolant density drops to near zero (~0.8 g/cm3  ~0.001 g/cm3). This work investigates five different lattice configurations and two different fuel types (LEU,Th)O2 and (Pu,Th)O2, which could potentially be implemented in once-throughthorium (OTT) cycles PT-HWR.

9, 10

in homogeneous or heterogeneous cores in a 700-MWe-class

Such lattice concepts may be able to achieve simultaneously high fuel

burnups with superior utilization of fissile fuel, and values of CVR that are lower than those found in conventional PT-HWRs running on NU fuel 11, 12, 15. The lattice concepts and associated computational models were modified versions based on earlier studies of PT-HWRs

13, 14

developed for using Pu/Th in an OTT cycle. The

lattice specifications are shown in Table I and related details can be found in earlier publications 14-17 (see also Section III.A).

Page 4 of 46

Manuscript Submission for ANS Nuclear Technology


It is intended that these lattice concepts could be implemented in a 700-MWe-class PT-HWR

11, 12, 15

, with 380 fuel channels (lattice pitch=28.6 cm), with 12 fuel bundles

(~50 cm long each) per channel, moderated and cooled with heavy water. It is anticipated that these concepts, based on modest modifications to conventional PT-HWR technology, will have higher probabilities of technological and economic feasibility and nearer-term implementation by vendors and utilities, than by using a completely different and untested reactor concept. An OTT cycle

9, 10

was chosen for this study because it was presumed to be more

economical and practical in the intermediate future, prior to the availability of commercial technology to recycle spent thorium-based fuels. For fuel manufacturing and operations, it is also simpler to use one type of fissile fuel at a fixed isotopic composition, rather than using graded enrichments within a given fuel bundle. The spent fuel (seed and/or blanket) from the OTT cycle will provide a convenient and valuable stockpile of U-233 to support a future generation of reactors 1, 2, 12, 16. The goals of the analyses were the following: 1.

Test a variety of lattice concepts using thorium-based oxide fuels mixed

homogeneously with either LEU (5 wt% U-235) or reactor-grade Pu (~67 wt% fissile). Depending on the composition of LEU or Pu, determine the maximum achievable burnup and the expected range of CVR for such concepts. It is desirable to have seed-type fuels that will achieve burnups  20 MWd/t, with burnup-averaged CVR  14 mk (1 mk = 100 pcm = 0.001 k/k).

Page 5 of 46

Manuscript Submission for ANS Nuclear Technology 2.


Determine the fissile utilization of various lattice concepts, and compare them

with the use of conventional PT-HWR using NU fuel (fissile utilization ~ 1,056 MWd/kg-fissile). 3.

Evaluate the expected bundle power limits for the different concepts with an

assumed linear element rating (LER) limit, and their impact on total reactor power, if they were implemented. 4.

Determine which concepts achieve the best compromise between good fissile

utilization and low CVR, and power level, and what fuels are best suited for seed and blanket fuels in a heterogeneous core in a PT-HWR. 5.

Based on analyses, identify opportunities for further development to improve

performance and feasibility of concepts.


Lattice physics calculations for the various bundle concepts were performed using WIMSAECL Version 3.1


, in combination with an 89-group nuclear data library, based on

ENDF/B-VII.0 19. WIMS-AECL was used to perform the detailed 89-group, 2-D collisionprobability neutron transport analysis of individual, single-cell lattice cells with burnup in a critical spectrum (keff=1.000).

The leakage model within WIMS-AECL solved the B1

equations with directional diffusion coefficients based on the Benoist formalism. Burnup calculations were performed using a constant cell-averaged thermal flux of 1.0e+14 n/cm2/s, as an approximation. WIMS-AECL was used to compute the infinite lattice reactivity (kinf, or k-inf) and the relative pin power densities in the bundle, and their variation with burnup.

Page 6 of 46

Manuscript Submission for ANS Nuclear Technology


A second set of WIMS-AECL calculations were performed for the various lattice concepts where the coolant density was set to zero (~0.001 g/cm3), but using the burnup-dependent fuel compositions determined from the previous cases with nominal D2O coolant density (~0.81 g/cm3). These data were required in order to estimate the infinite lattice CVR. Supplementary calculations of k eff (cooled and voided), based on a user input value of buckling (~0.86 m-2) representative of a 700-MWe PT-HWR, were also performed to get an approximate, qualitative estimate of the impact of neutron leakage on CVR.


III.A. Reference Concept

The lattice was based on a concept described in a previous study 14, which was a 43-element bundle with 42 fuel pins and a central non-fuel neutron absorbing pin. This reference model was modified to create five slightly different concepts, illustrated in Fig. 1 to Fig. 5. The calandria tube (CT) and pressure tube (PT) are made of Zr-2 and Zr-2.5Nb respectively, while the clad is made of Zr-4. Other details of the lattice specifications are shown in Table I, and are discussed further below.

III.B. Fuel Composition

Two types of thorium-based fuels were considered. The first was (LEU,Th)O2, and the second was (Pu,Th)O2. Both were homogeneous mixtures. The LEU was 5 wt% U-235/U,

Page 7 of 46

Manuscript Submission for ANS Nuclear Technology


which is readily available from current uranium enrichment facilities in the international community. The Pu was reactor-grade plutonium (2.75 wt% Pu-238 / 51.96 wt% Pu-239 / 22.96 wt% Pu-240 / 15.23 wt% Pu-241 / 7.1 wt% Pu-242), which is expected to be obtained from stockpiles of high-burnup LWR fuel. The Pu isotopic composition is based on that of sample SF97-4 taken from Reference 20. However, Pu from other sources could be used instead (as discussed earlier), and could also be obtained from fast breeder reactors operating on the Pu/U fuel cycle. The volume fraction (or weight fraction) of LEUO2 in (LEU,Th)O2 was varied from 0.10 to 0.70 for different types of blanket and seed fuel, the latter which could achieve different levels of burnup. Given that the LEU is 5 wt% U-235/U, higher volume fractions (e.g.,  0.40) may be required to achieve burnups  20 MWd/kg. One additional advantage of using homogeneously mixed (LEU,Th) is that the U-233 produced will be de-natured by the presence of U-238, which enhances the proliferation resistance of the fuel 21. In an analogous manner, the volume fraction (or weight fraction) of PuO 2 in (Pu,Th)O2 was varied from 0.01 to 0.07 for the different types of blanket and seed fuels. The Pu is ~67 wt% fissile, so only a small fraction of Pu is required to make the lattice critical. To achieve burnups of  20 MWd/kg,  3 wt% Pu content is required for the seed fuels. In one of the lattice concepts to be described below, a central region with 8 pins of pure ThO2 is used for production of U-233, with some supplementary power production from in situ fission of U-233.

Page 8 of 46

Manuscript Submission for ANS Nuclear Technology


TABLE I Specifications for Thorium-Based Fuels in PT-HWR Quantity Lattice pitch (square) Length of Bundle Calandria Tube (CT) CT Inner Radius CT Outer Radius Gap Material Pressure Tube (PT) PT Inner Radius PT Outer Radius Number of Fuel Pins Fuel in Outer Rings Fuel in Central 8 pins LEU enrichment Pu fissile content Cladding Material Fuel Element Radius Inner Zr-4 Tube Central Zr-4 tube filler Bundle HM fuel mass Moderator Temp., Density D2O Moderator Purity Coolant Temp., Density D2O Coolant Purity Fuel Temp., Density Cell-average thermal flux Burnup period Specific power in fuel

Value, Units 28.575 cm 49.5 cm Zr-2 6.4 cm 6.6 cm CO2 Zr-2.5Nb 5.2 cm 5.6 cm 43, 35, or 21 (LEU,Th)O2 or (Pu,Th)O2 ThO2 5 wt% U-235/U ~67 wt% Zr-4 0.57 cm 2.4 cm or 3.5 cm D2O coolant, or ZrO2,~4.3 g/cm3 43-pins: 16.043 kg 35-pins: 13.058 kg 21-pins: 7.835 kg ~69C, ~1.085 g/cm3 ~99.8 wt%D2O ~288C, ~0.81 g/cm3 ~99 at%D2O , ~586C, ~9.7 g/cm3 1.0e+14 n/cm2/s ~1,740 days ~10 to 70 kW/kg

Page 9 of 46

Manuscript Submission for ANS Nuclear Technology


III.C. Bundle Geometries

The 35-LEU/Th-8-Th (or 35-Pu/Th-8-Th) lattice concept is shown in Fig. 1. There are 35 identical outer fuel pins made with LEU/Th or Pu/Th, and 8 inner fertile fuel pins made of pure ThO2. The total bundle heavy metal fuel (HM) mass is ~16 kg. The inner pins serve as an inner blanket region within the bundle for breeding U-233, and to generate some power as U-233 undergoes fission in situ. The inner 8 pins will also help reduce CVR relative to the situation if there were 8 inner pins with the same fissile content as the outer 35 pins. The use of a central zone of ThO2 pins to reduce CVR has been proposed in previous studies, although LEU was used instead of LEU/Th in the outer fuel pins 12.

Fig. 1. 35-LEU/Th-8-Th (or 35-Pu/Th-8-Th) with 8 Central Fuel Pins of ThO2

Page 10 of 46

Manuscript Submission for ANS Nuclear Technology


The 35-LEU/Th-D2O-Rod (or 35-Pu/Th-D2O-Rod) lattice concept is shown in Fig. 2. It is similar to 35-LEU/Th-8-Th, except that the central 8 ThO2 rods are replaced by a central Zr-4 tube filled with stagnant D2O coolant. The HM fuel mass is ~13 kg. It is presumed that this tube would have Zr-4 plates on either end with a small orifice to permit the existence of a central region of essentially stagnant D2O coolant, which would drain in the event of a postulated loss-of-coolant accident (LOCA). It was anticipated that this concept would also help reduce CVR while permitting slightly higher neutron economy and burnup, due to the absence of a central neutron absorber. During a LOCA, the void region in the centre of the bundle will enhance axial leakage of neutrons.

Fig. 2. 35-LEU/Th-D2O-Rod (or 35-Pu/Th-D2O-Rod) - Zr-4 Tube Filled with Stagnant D2O Coolant

Page 11 of 46

Manuscript Submission for ANS Nuclear Technology


The 35-LEU/Th-ZrO2-Rod (or 35-Pu/Th-ZrO2-Rod) lattice concept is shown in Fig. 3. It is similar to 35-LEU/Th-D2O-Rod, except that stagnant D2O coolant is replaced with an inert matrix of ZrO2, at ~75% of its maximum density (5.68 g/cm3). Relatively little neutron absorption or scattering occurs in the ZrO2; thus, any changes in neutron flux and energy distributions and reactivity due to coolant voiding will be due only to the coolant surrounding the outer 35 fuel pins. This design was anticipated to be easier to operate, and to be able to achieve higher burnups with a reduction in CVR.

Fig. 3. 35-LEU/Th-ZrO2-Rod (or 35-Pu/Th-ZrO2-Rod) - Zr-4 Tube Filled with ZrO22

Page 12 of 46

Manuscript Submission for ANS Nuclear Technology


Lattice concept 21-LEU/Th-D2O-Rod (or 21-Pu/Th-D2O-Rod) is shown in Fig. 4. It is a more radical departure from the conventional 43-element or 35-element concepts. It is similar to the 35-LEU/Th-D2O-Rod, but the inner ring of 14 fuel elements have been removed and the central Zr-4 tube filled with stagnant D2O coolant has been enlarged, leaving a single outer ring of 21 fuel pins, giving a bundle HM fuel mass of ~7.8 kg. The large central stagnant coolant region in the 21-LEU/Th-D2O-Rod concept will contribute more to neutron moderation, and a LOCA event will result in enhanced axial leakage of neutrons, which will reduce CVR. The removal of the inner ring of 14 pins will also reduce CVR, given that there will be no inner pins to be depleted more slowly, or to experience resonance self-shielding by the outer pins. It is expected that such a bundle will have a reduced neutron economy due to the reduction in fuel volume, and that it will be over-moderated. It will also need to be operated at a lower bundle power to stay within LER limits. However, it will also simplify the bundle design and will permit the maximum utilization of the outer fuel pins, which determine the lifetime of the bundle. Inner fuel pins are always under-utilized relative to the outer fuel pins, which experience a larger thermal flux.

Fig. 4. 21-LEU/Th-D2O-Rod (or 21-Pu/Th-D2O-Rod) - Zr-4 Tube Filled with Stagnant D2O Coolant Page 13 of 46

Manuscript Submission for ANS Nuclear Technology


The final lattice concept is 21-LEU/Th-ZrO2-Rod (or 21-Pu/Th-ZrO2-Rod), shown in Fig. 5. It is similar to 21-LEU/Th-D2O-Rod, but the stagnant D2O coolant has been replaced with a solid rod of inert ZrO2. With a LOCA event, only very slight changes in the flux distribution and spectrum will occur; thus, reductions in CVR will occur relative to what is found in the 35-LEU/Th-ZrO2-Rod concept. The potential drawback is the presence of central ZrO2 rod will provide a pathway for axial neutron leakage, and so the neutron economy will be reduced somewhat. As mentioned previously for the 21-LEU/Th-D2O-Rod design, the peak bundle power will need to be reduced for the 21-LEU/Th-ZrO2-Rod concept.

Fig. 5. 21-LEU/Th-ZrO2-Rod (or 21-Pu/Th-ZrO2-Rod) - Zr-4 Tube Filled with ZrO2

Page 14 of 46

Manuscript Submission for ANS Nuclear Technology


III.D. Calculations and Post-Processing Analyses

For each of the 10 lattice concepts (5 geometries  2 fuel types) and for various compositions of LEU/Th (10 wt% to 70 wt% LEU) and Pu/Th (1 wt% to 7 wt% Pu), WIMS-AECL lattice physics calculations were performed.

Additional WIMS-AECL

calculations were performed where the coolant density was reduced to near-zero (~0.001 g/cm3) to approximate void conditions.

The values of cooled and voided infinite lattice

reactivity (kinf = k-infinity = k) were extracted for each burnup step. The maximum expected achievable burnup was determined when the burnup-averaged kinf dropped to ~1.03, as shown below: B  BU max

kinf BU max  



( B)

B 0

BU max

 1.03


The 3% of excess reactivity is used to account for neutron leakage in an actual core, although in some cores it may be less or greater (2% to 4%). With daily, on-line re-fuelling in a PT-HWR, with short fuel bundles (~50 cm), the re-fuelling process closely approximates continuous re-fuelling. According to the linear reactivity model 22, illustrated by the following equation: BU n  ~

2n BU 1 n 1


Page 15 of 46

Manuscript Submission for ANS Nuclear Technology


The maximum burnup for continuous re-fuelling (n>>1) is nearly double that of a single batch re-fuelling (when kinf(BUsingle) ~ 1.03). This doubling of the burnup is ~33% higher than what can be obtained in a 3-batch re-fuelling scheme (n=3, BU(3) ~1.5BU(1)), which is typically used for PWRs


. In reality, the reactivity variation with burnup is not exactly

linear, due to the creation of fissile fuels from neutron capture in the fertile fuel, and also due to the buildup of neutron-absorbing fission products. Thus, the maximum burnup based on the evaluation in Equation (1) may show that BUmax/BU(kinf~1.03)  2.0. It may be slightly larger, or smaller. To consistently compare the performance of different fuel types with different compositions of LEU/Th and Pu/Th in an OTT cycle, the performance parameter known as the fissile utilization (FU) is introduced:

FU 

BU max wt fraction fissile in bundle


The FU is the energy obtained per initial mass of fissile fuel in the fuel bundle. It is somewhat analogous to the uranium utilization, but can be used for any type of fuel containing any fissile isotope. Hence, for example, a PT-HWR running on NU fuel (~0.71 wt% U-235/U) has a typical maximum discharge burnup of ~7.5 MWd/kg


. Thus, the FU

~ 7.5/0.0071 ~ 1,056 MWd/kg-fiss. The infinite-lattice CVR was evaluated at each burnup step using:

CVRinf 

kinf void  kinf cooled kinf void  kinf cooled


Page 16 of 46

Manuscript Submission for ANS Nuclear Technology


The CVR will be different at each burnup step, due to changes in the fuel composition. Since a reactor core has fuel with various levels of burnup, it is useful to evaluate the burnup-averaged CVR to get an approximate estimate of the CVR for a large reactor core. The burnup-averaged CVR is given by: B  BU


 CVR ( B)

B 0



This formula is only applicable to a homogeneous reactor core with one fuel type. For heterogeneous cores with two or more fuel types, Equation 5 can be used to provide bounding estimates, based on the burnup-averaged CVR values for the different fuel types. The effect of leakage is neglected here, but it is estimated to reduce CVR by ~1 mk to 6 mk, depending on the lattice concept. To get a more accurate estimate, the CVR for a full reactor core would need to be determined from full-core reactor physics calculations. Another performance parameter of interest is the maximum bundle power, which will determine the maximum allowable power level in a reactor core. The maximum bundle power depends on an assumed upper limit for the linear element rating (LER) and the maximum relative pin power density. The maximum relative pin power, which usually occurs in the outer fuel element at or near zero burnup was extracted from the WIMS-AECL output file for use in determining the maximum bundle power using the following formula:

PowerBundle Max 

50 kW / m  0.495 m  # pins Max. Rel. Pin Power Density


Page 17 of 46

Manuscript Submission for ANS Nuclear Technology


It is assumed a priori that the maximum LER is ~50 kW/m. The bundle length is 0.495 m. An LER  50 kW/m will ensure fuel integrity


, although it may be possible that certain

thorium-based fuels could maintain their integrity at much higher LER values, perhaps higher than for uranium-based fuels. By comparison, the maximum bundle power in a PTHWR running with NU fuel ranges from approximately 800 kW to 900 kW, with an LER  57 kW/m 11, 12, 23.


IV.A. Reactivity and Burnup

The lattice reactivity (k-inf) and burnup characteristics for the various lattice concepts with different compositions of LEUO2 or PuO2 mixed with ThO2, are illustrated in Fig. 6 to Fig. 15. For each fuel type it is observed that for high initial fissile content (e.g., seed-type fuel), the reactivity starts high, and decreases approximately linearly, as is found generally for reactors using enriched fuel. However, for lower fissile concentrations (e.g., blanket-type fuel), the reactivity starts sub-critical, and increases to an asymptotic value. The near-asymptotic value of k-inf at high burnup ranges from ~0.82 to ~0.93. The fuel combinations that give a relatively constant reactivity over a large range of burnup (~ 0 to 40 MWd/kg) are most suitable for use as blanket fuel. The variation of reactivity with burnup is relatively flat for 20 wt% LEU to 30 wt% LEU fuels, and for 1 wt% to 2 wt% Pu fuels. For fuels with less than 20 wt% LEU or less than 1 wt% Pu, the blanket fuel Page 18 of 46

Manuscript Submission for ANS Nuclear Technology


becomes highly sub-critical (k-inf