GENES4/ANP3, Sep. -9, 3, Kyoto, JAPAN Paper 8 Full MOX Core Design in ABWR Toshiteru Ihara *, Takaaki Mochida, Sadayuki Izutsu 3 and Shingo Fujimaki 3 Nuclear Power Department, Electric Power Development Co., Ltd., Tokyo, 4-86, Japan Nuclear Plant Engineering Department, Hitachi, Ltd., Hitachi, Ibaraki, 39-88, Japan 3 Core Design Group, Global Nuclear Fuel Japan Co., Ltd., Yokosuka, Kanagawa, 39-836, Japan Electric Power Development Co., Ltd. (EPDC) has been investigating an ABWR plant for construction at Oma-machi in Aomori Prefecture. The reactor, termed FULL MOX-ABWR will have its reactor core eventually loaded entirely with mixed-oxide (MOX) fuel. Extended use of in the plant is expected to play important roles in the country s nuclear fuel recycling policy. bundles will initially be loaded only to less than one-third of the reactor, but will be increased to cover its entire core eventually. The number of bundles in the core thus varies anywhere from to 64 for the initial cycle and, to 87 for equilibrium cycles. The safety design of the FULL MOX-ABWR briefly stated next considers any probable MOX loading combinations out of such MOX bundle usage scheme, starting from full UO to full MOX cores. KEYWORDS: Full MOX, ABWR, Core Design I. Introduction The core design from full UO to full MOX loaded of ABWR ) ) 3) has been performed. The is 8x8 bundle configuration with a large central water rod, with 4GWd/t maximum burnup, and it is compatible with 9x9 high burnup UO fuel. The shutdown and thermal margins of the MOX core, from partially- to fully- loaded, are comparable to that of full UO core. Safety analyses based on MOX loaded core characteristics and property have shown its conformity to the design criteria in Japan. This paper shows the safety design for full MOX cores of ABWR. II. Full MOX Core. Design Concept One of the ABWR core features is a wider fuel bundle pitch. Specifications of fuel bundle are the same as those of the current BWR lattice. As a result, the non-boiling water area outside the channel box (bypass flow area) increases in the ABWR core. The bypass flow area can thermalize neutrons more effectively before absorption by the fuel material. The wider fuel pitch of the ABWR core decreases the absolute value of void reactivity coefficient and increases the shutdown margin, which collectively makes ABWR well-adopted for loading its core fully with MOX fuel. The main specifications are not changed from those of the standard ABWR, such as 396MWt thermal power, 87 fuel bundles and control rods. The main design concepts of are the following, and the basic specifications of core are shown in Table and those of fuel are shown in Table respectively. () The MOX bundle uses the well-proven design of STEP- UO bundle (GWd/t maximum exposure) having much operational experience. The bundle features a large water rod in the center of 8x8 fuel rod configuration, as shown in Fig.. () The MOX bundle maximum exposures conservatively are selected to be 4 GWd/t based on the MOX irradiation experience. This maximum exposure is equal to that of STEP-I UO bundle. (3) The bundle consists of rods using UO -PuO and UO fuel rods using UO -Gd O 3 as a fuel material. The bundle average fissile material content is selected to be about.9wt% of fissile plutonium (Puf) content and about.wt% of U-3 enrichment for the conditions of 3-month cycle length and the reference plutonium composition (67 wt% Puf ratio). Core Type Table Basic specification of core design Items Specification Advanced Boiling Water Reactor (ABWR) Thermal power (MW) 3,96 Core flow (t/h).x 3 (%rated) Core pressure (MPa[abs]) 7.7 Number of fuel bundles 87 Number of control rods Fuel bundle pitch (cm). * Corresponding author, Tel. +8-3-346-, Fax. +8-3-346-8, E-mail: Toshiteru_Ihara@jpower.co.jp
Table Basic specification of fuel design Items Specification Fuel assembly Lattice configuration Average 3 U content * (wt%) Average Puf content * (wt%) Maximum exposure (MWd/t) Batch averaged exposure (MWd/t) Number of fuel rods Pellet material Cladding outside diameter (mm) Cladding wall thickness * (mm) Cladding Material Number of water rods Water rod outside diameter (mm) * For reload fuel 8x8 9x9. 3.8.9-4,, 33, 4, 6 74 * 3 UO -PuO (MOX rods) UO -Gd O 3 (UO rods) UO,UO -Gd O 3.3..86.7 Zirc-(with Zr liner) Zirc-(with Zr liner) 34. 4.9 * Including Zirconium liner thickness of about.mm * 3 Including 8 partial length rods. Fuel and Core Design Rod arrangement of the bundle is shown in Fig.. Of the 6 fuel rods, 48 contain MOX and the rest UO bearing Gd O 3. Four kinds of MOX rods with different plutonium contents in the bundle are employed in order to reduce the local peaking. Although tends to become higher pellet temperature owing to lower thermal conductivity and increased FP gas release compared with UO fuel, the pellet center temperature remains sufficiently lower against the fuel melting temperature through the lifetime. Also, though tends to become higher rod internal pressure owing to increased FP gas and He gas release, the rod internal pressure remains still equivalent to 9x9 UO fuel rod at the end of lifetime due to increased gas plenum volume for MOX rod. The rod internal pressure of the is shown in comparison with 9x9 UO fuel in Fig.. The neutron multiplication factor (k-infinity) of MOX fuel is shown in Fig.3 in comparison with 9x9 UO fuel. The change in k-infinity with exposure is smaller for MOX fuel than that for UO fuel. This, in turn, leads to a smaller bundle peaking of full MOX core than full UO core and /3 MOX core (i.e. consist of 9x9 UO fuels and 36 MOX fuels). This brings about the decrease of the maximum control rod worth and the mitigation of the shutdown margin decrease for full MOX core. Figure 4 shows the fuel loading pattern of full MOX equilibrium core. Shutdown margin, Maximum Linear Heat Generation Rate (MLHGR), and Minimum Critical Power Ratio (MCPR) for the equilibrium cycle are shown in Fig. the comparison among full UO core, /3 MOX core and full MOX core. The shutdown and thermal margins of the MOX core, from partially- to fully- loaded, are comparable to that of full UO core. These results satisfy the design targets and the operational limits. W Control rod 4 3 3 4 3 3 w 3 3 4 3 3 4 Water rod 4 rod (Pu content 4<3<<) UO fuel rod containing Gd O 3 Fig. Rod arrangement of bundle
Fuel rod internal pressure (MPa[abs]) 8. 6. 4.. Coolant pressure 9X9 UO fuel 8 6 4 (kg/cm a) Shutdown margin (% k) 6 4 3 Design target Full core 9X9 UO fuel core /3 core 3 4 6 7 8 9 Cycle exposure (GWd/t) (a) Shutdown margin Neutron multiplication factor.3....9.8.7. 4 6 8 Peak pellet exposure (GWd/t) Fig. Fuel rod internal pressure Hot operation bundle 9X9 UO fuel bundle 3 4 Exposure (GWd/t) Fig.3 Neutron multiplication factor MLHGR (kw/m) MCPR 4 4 3 3.9.8.7.6..4.3.. Operating limit 3 4 6 7 8 9 Cycle exposure (GWd/t) Operating limit (b) MLHGR Full core 9X9 UO fuel core /3 core Full core 9X9 UO fuel core /3 core Full MOX /3 MOX 3 4 6 7 8 9 Cycle exposure (GWd/t) (c) MCPR Fig. Comparison of core performance in equilibrium Fresh reload bundle Bundle in nd cycle of operation Bundle in 3rd cycle of operation Bundle in 4th cycle of operation Fig.4 Fuel loading pattern of full MOX equilibrium cycle 3. Characteristics of MOX core 3. Reactivity coefficient and dynamic parameter The MOX core is characterized by an increase in the absolute void coefficient due to large neutron absorption cross section of plutonium relative to uranium. Void coefficient, Doppler coefficient, delayed neutron fraction and prompt neutron lifetime change depending on the MOX fuel loading fraction (Fig.6). While the void coefficient of the full MOX core is about % larger in absolute value than the full UO core, the delayed neutron fraction of the full MOX core is about % smaller than the full UO core.
The absolute values of Doppler coefficient of MOX loaded core is equivalent to that of 9x9 UO core. Relative value Void reactivity coefficient Doppler reactivity coefficient Delayed neutron fraction neutron absorption cross section, in the plutonium isotopes becomes small as Puf ratio becomes large. The MOX bundle can be used regardless of its plutonium vector to adjust the plutonium content of MOX bundle according to the reactivity compensation design. Shutdown margin, MLHGR, and MCPR of full MOX core with reference plutonium vector in equilibrium cycle are shown in Fig.9 compared with other vector (Puf ratio: 7 and 6 wt%). These results satisfy the design target or operational limit. (9X9 UO core) /3 (/3 MOX core) loading fraction 3/3 (Full MOX core) Fig.6 Safety parameter vs. MOX loading fraction 3. Influence of plutonium isotopic composition 4) () Reactivity compensation design of Since the plutonium mixed with the uranium come from the reprocessing spent fuels, the isotope composition of plutonium (plutonium vector) varies depending on initial enrichment, burn-up, cooling period, etc. of the reprocessed UO fuels. The main isotopes of plutonium obtained by the reprocess are Pu-39 and Pu-4 which are fissile, Pu-38, Pu-4 and Pu-4 which are fertile. The plutonium vector slightly changes also after reprocessing, since Pu-4 among these isotopes decays in short period (half life of about 4.4 years) and becomes Am-4 of a non-fissile. As the uranium of MOX base matrix is the tail uranium of enrichment to achieve effective use of plutonium, U-3 content of MOX pellet has the range from about.wt% to about.4wt% depending on the enrichment process of uranium. In the fabrication of, the plutonium content of MOX is adjusted depending on the plutonium vector, so that the MOX bundle is able to achieve same design discharge exposure. This is called the reactivity compensation design. The following are considered in the reactivity compensation design. (a) The plutonium average content of MOX bundle is adjusted to have the target k-infinity at GWd/t corresponding to the core average exposure at the end of cycle (EOC). The target k-infinity is equal to the k- infinity of UO bundle at GWd/t. (b) All MOX rods have uniform axial plutonium contents to simplify the fabrication. A design policy of having only four different plutonium contents is maintained irrespective of a change in plutonium vector from reference. Figure 7 shows the k-infinity changes with exposure depending on the plutonium vector of MOX. Figure 8 shows the average plutonium and Puf content of MOX bundle with the Puf ratio (range is about 8-77wt%). The plutonium content decreases as the Puf ratio becomes large. This reason is that the ratio of Pu-4, that has large k-infinity......9.9.8 3 4 exposure (GWd/t) Fig.7 Neutron multiplication factor vs. exposure Bandle average Pu and Puf content (wt%) 8 7 6 4 3 Hot operation MOX fule (Puf ratio 67wt%) MOX fule (Puf ratio 77wt%) MOX fule (Puf ratio 8wt%) UO fuel (batch averaged exposure 33GWd/t) GWd/t (EOC) Pu content (matrix U-3:.wt%) Puf content (matrix U-3:.wt%) 6 6 7 7 8 8 Puf ratio (wt%) Fig.8 Pu and Puf content vs. Puf ratio
Shutdown margin (% Δk) MLHGR (kw/m) MCPR 4 3-4 3 operation limit 44.kW/m (a) Shutdown margin 3 4 6 7 8 9.9.8.7.6..4.3.. design target % Δk 3 4 6 7 8 9 cycle exposure (GWd/t) cycle exposure (GWd/t) (b) MLHGR (c) MCPR Puf ratio 67wt% Puf ratio 77wt% Puf ratio 8wt% Puf ratio 67wt% Puf ratio 7wt% Puf ratio 6wt% 3 4 6 7 8 9 cycle exposure (GWd/t) Puf ratio 67wt% Puf ratio 7wt% Puf ratio 6wt% operation limit.3(from EOC-GWd/t to EOC).(other period) Fig.9 Comparison of equilibrium core performance () Influence on reactivity coefficients The influence of plutonium vector on dynamic void coefficient and dynamic Doppler coefficient (coefficient divided by delayed neutron fraction) are shown in Fig.(a) and Fig.(b) respectively. In Fig., the Y-axis indicates relative difference in dynamic coefficient expressed by (A-B)/B where A is a dynamic coefficient for plutonium vector X and B is the same coefficient for the reference plutonium vector. The influence of plutonium vector on delayed neutron fraction is shown in Fig.. Although the influence of plutonium vector on dynamic coefficient is relatively small as shown in Fig., the influence is taken into account in the safety evaluation. In the safety analysis of full MOX core, the dynamic void coefficient in the reference plutonium vector is multiplied by a factor.4, and dynamic Doppler coefficient by.96. As apparent from Fig., these factors are conservatively determined. The influence of plutonium vector on the scram reactivity worth is shown in Fig.. The difference of static control rod worth (in % k unit) between MOX core and UO core is small because the control rod is inserted in a water gap away from the fuel bundle in BWR. The dynamic control rod worth (in dollar unit) of MOX core is larger than that of UO core because the delayed neutron fraction of MOX core is smaller than that of UO core. The design scram curve has enough margin in either case. Relative value of (void coefficient / delayed neutron fraction) (%) - - ABWR,EOC, core average void fraction 4% -4% 6 6 7 7 8 8 Puf ratio (wt%) more nagative Fig.(a) Dynamic void coefficient vs. Puf ratio Relative value of (Doppler coefficient / delayed neutron fraction) (%) - - ABWR,EOC,cold +4% more nagative 6 6 7 7 8 8 Puf ratio (wt%) Fig.(b) Dynamic Doppler coefficient vs. Puf ratio
Relative value of delayed neutron fraction (%) Scram reactivity (dollar unit) - - 6 6 7 7 8 8 Puf ratio (wt%) Fig. Delayed neutron fraction vs. Puf ratio - -4-3 - - ABWR,EOC, core average void fraction 4% ABWR,EOC Puf ratio 67wt% Puf ratio 7wt% Puf ratio 6wt% UO core design curve...4.6.8. Control rod insert fraction Fig. Scram reactivity worth vs. Puf ratio III. Safety evaluation. Stability The thermal-hydraulic characteristic of MOX bundle is comparable to that of 9x9 UO bundle. The core loaded with both kinds of bundle has intermediate void coefficient between those of the core loaded with single kind of bundle respectively. Therefore the core stability and regional stability of the core loaded with both kinds of bundle is intermediate between the cores loaded with single kind of bundle. The core stability and regional stability about full MOX loaded core and full 9x9 UO loaded core are representatively evaluated assuming the conservative power distribution and void coefficient. The power distribution is determined to provide conservative assumptions taking account of power distributions for cores covering both full 9x9 UO core and full MOX core, and additionally, those falling between these cores. The largest negative void coefficient, observed throughout the lifetime of the core for each for full 9x9 UO core and full MOX core, is selected. Table 3.shows the decay ratios of core stability and regional stability at the condition of minimum pump speed and maximum reactor power. Those are calculated by the frequency domain code. Although the decay ratios of full MOX core are larger than those of 9x9 UO core, these are enough smaller than criterion.. Table 3 Decay ratio for core stability and regional stability Core stability Regional stability Full MOX core.68. UO core.37.6. Abnormal transient during plant operation One of the features of MOX loaded core is that the absolute value of void coefficient become larger according to the increase of MOX loading fraction. The larger void coefficient in absolute value brings the rapider decrease of reactor power in the case that the reactor power rises following the void fraction in the core increases by some causes. The loading brings higher self-control ability of reactor power. In the event in which void fraction increases such as partial loss of coolant flow, the large void coefficient in absolute value accelerates the decrease of power and mitigates the increase of MCPR (index of thermal margin). On the other hand, in the event in which void fraction decreases such as generator load rejection without turbine bypass, the large void coefficient in absolute value accelerates the increase of power and MCPR. The MCPR in loss of feed water heating and generator load rejection without turbine bypass that are void fraction decreasing event and the severest event in MCPR are shown in Fig.3. Loss of feed water heating is the event in which reactor power gradually rises according to the slow decrease of void fraction in the core due to the slow decrease of inlet coolant temperature. Since the reactor power slowly rises and the maximum power is limited by the scram of neutron flux high signal, the MCPR hardly depends on the MOX loading fraction, that is, the magnitude of void coefficient and the difference between scram curves for early stage and end stage in cycle shown in Fig.4. Generator load rejection without turbine bypass is the event in which reactor power rapidly rises according to the fast decrease of void fraction by the turbine control valve fast closure. The reactor power of MOX loaded core rapidly rises compared with 9x9 UO loaded core because of large negative void coefficient. The MCPR depends on the MOX loading fraction, that is, the magnitude of void coefficient and the difference between scram curves for early stage and end stage in cycle shown in Fig.4.
In the core applied to the scram curve of early stage in cycle, the MCPR of generator load rejection without turbine bypass is smaller than that of Loss of feed water heating regardless of MOX loading fraction, therefore the MCPR determining event is Loss of feed water heating. In the core applied to the scram curve of end stage of cycle, the MCPR determining event changes from Loss of feed water heating to generator load rejection without turbine bypass at about /3 MOX loading fraction (36 loading). The different two operational limit MCPR (OLMCPR) are selected to secure adequate operational thermal margin. One is used for the core applied to the scram curve of early stage in cycle regardless of MOX loading fraction, and the core applied to the scram curve of end stage in cycle below 36 loading. Another is used for the core applied to the scram curve of end stage in cycle above 36 MOX fuel loading. The void coefficients of 36 MOX loaded core and of full MOX loaded core in equilibrium cycle end are used in the MCPR evaluation of the core below 36 MOX loading and the core above 36 MOX loading respectively as the void reactivity feedback is conservatively estimated. % for 36 MOX loading core and 4% for full MOX loaded core are added to the void coefficient in the reference vector to appropriately consider the influence of plutonium vector. MCPR.4.3.. Generator load rejection (BOC) Generator load rejection (EOC) Loss of feed water heating 3. Accident analysis Loss of the coolant accident (LOCA), All pump trip accident (APTA) and Main steam line break accident (MSLBA) are the events in where the void fraction increases. The full MOX core with negative void coefficient larger than that of full 9x9 UO core tends to reduce the reactor power further during event. has the tendency to make the fuel cladding temperature higher, because the gap conductance and the pellet thermal conductivity are small compared with the UO fuel. For MOX loaded core, the time dependent reactor power identical to the UO core is adopted conservatively. The fuel cladding temperature of LOCA is evaluated for High Pressure Core Flooder system (HPCF) pipe break as the severest case for the drop of reactor water level. The cladding surface temperature rises due to boiling transition, but the core is always covered with two phase coolant mixture over the whole transient period. The peak cladding temperature (PCT) is about 6 o C, satisfying the criterion of o C, although PCT of MOX core becomes a little higher than UO core. The fuel enthalpy response of Control Rod Drop Accident (CRDA) is shown in Fig. as a representative event of transient and accident for control rod system, where the dropping control rod worth is assumed to be.3% k. The maximum fuel enthalpy in the full MOX core is almost equivalent to that of the UO core, because Doppler reactivity effect in dollar unit of full MOX core is more negative. Full core 9x9 UO fuel core...4.6.8 MOX loading fraction Fig.3 MCPR vs. MOX loading fraction - Fuel enthalpy Scram reactivity (dollar unit) -4-3 - - For early stage in cycle For end stage in cycle...4.6.8. Control rod insert fraction Fig.4 Design scram curve Time (sec) Fig. Fuel enthalpy response on CRDA IV. Others An increase of loading fraction causes the decrease in the boric acid water reactivity worth of the standby liquid control (SLC) system. This decrease has been compensated by the increase in the SLC tank capacity from 9m 3 to 36m 3.
Since the control rod is inserted in a water gap, which is out of the fuel bundle in BWR, the influence of loading on the control rod worth is not so large. This influence become smaller in the ABWR because the ABWR has a wider bundle pitch than BWRs. The total rods worth of a current designed full MOX core is almost comparable to that of full UO core in the ABWR as shown in Fig.6. To provide for the decrease of reactivity due to decay of Pu-4 associated with unpredictable delays of loading, the maximum core flow rate in full power operation has been increased from % in the existing ABWRs to % in the FULL MOX-ABWR. Control rod reactivity worth (% k) cold : total control rod reactivity worth : maximum control rod reactivity worth 9x9 UO core /3MOX core Full MOX core Fig.6 Total control rods reactivity worth V. Conclusion As the result of thermal-mechanical rod analysis and MOX loaded core analysis, it is confirmed that design criteria are satisfied in the case of MOX loaded to full MOX core as well as in the case of UO core. The parameters used for safety analysis are selected appropriately considering and MOX loaded core characteristics. As the result of such safety analysis, it is confirmed that criteria concerning to safety evaluation are satisfied in the case of MOX loaded to full MOX core as well as in the case of UO core. Nomenclature Puf ratio: (Pu-39+Pu4)/(total Pu+Am-4)*wt% Pu content: (total Pu+Am-4)/(total Pu+Am-4+total U)*wt% Puf content: (Pu-39+Pu4)/(total Pu+Am-4+total U)*wt% References ) Y.Kinoshita, et al., Design of Full MOX Core in ABWR, Karyoku-Genshiryoku-Hatsuden (J. Thermal and Nuclear Power Engineering Soc.), Vol. No., 6 (999) [in Japanese] ) S.Izutsu, et al., PROGRESS OF FULL MOX CORE DESIGN IN ABWR, International Symposium on MOX Fuel Cycle Technologies, IAEA-SM-38/7, Vienna (999) 3) S.Izutsu, Core design of Full MOX-ABWR, 3-th Seminar of reactor physics in summer, Yufuin, Japan, July 3- August,, p4-6 () [in Japanese] 4) Hitachi Ltd., BWR About Pu isotope change of MOX loaded core in the Full MOX-ABWR, HLR-67 (999) [in Japanese]