Search

We have developed a void reactivity evaluation method by using modified conversion ratio measurements in a light water reactor (LWR) critical lattice. Assembly-wise void reactivity is evaluated fro...

We have developed inexpensive and easy-handling measurement methods on intra-pellet neutron flux. A foil activation method with metallic foils, which were fabricated by punching out technique and etching technique to reduce fabrication error and positioning error, was used for the intra-pellet neutron flux distribution measurement. The developed method was applied to measure intra-pellet neutron flux distributions in a reduced–moderation light water reactor (LWR) lattices, and uncertainty of the distributions was estimated to be 1% to 2%. Measured values were analyzed with a continuous energy Monte Carlo code. Comparison of measurements and analyses revealed that the developed method is useful for the validation of an advanced fuel design method considering neutron behavior in fuel pellets.

Erbia-credit super high burnup (Er-SHB) fuel offers a means to introduce >5 wt% 235U enrichment fuel; small amounts of erbia added to all the high-enriched UO2 powder can reduce the initial reactiv...

A parallelization method that does not entail much data communication and allowseasy implementation was developed for the method of characteristics (MOC). In the parallelization method, azimuthal angles were grouped to compute, on the same processor, theangular flux before and after the reflection on the outer boundary. The method can be easily applied to existing MOC codes without the highly technical knowledge on efficientdata communication in parallel computing. It was implemented into the GALAXY code and numerical comparisons were performed. As a result, computation speedup by a factor of 9 to 10 and good parallel efficiencies of 70 to 80% were achieved by using twelve processors. The speedup benefits the practical reactor core design using MOC codes. Other thanthe parallel efficiency, the proposed method allows easy implementation because no drastic change in the conventional algorithm for existing sequential process codes is needed. This is carried out with the absence of data communication between ...

For the method of characteristic (MOC), a system with large-gradient neutron flux caused by a strong absorber or neutron leakage is reported to entail large errors in spatial mesh discretization and require very fine mesh spacing. To apply the MOC to such fine models, the ray-trace path width has to be fine in order to make many paths cross each region. Our new method intends to obtain good accuracy with a coarse path width. With a coarse path width on the MOC, some tiny regions have less ray-tracing paths. The reliability of flux calculation for the regions can be evaluated with the calculation volumes that are estimated in ray tracing. If the discrepancy between the calculation and true volumes becomes large, the accuracy cannot be expected. In this study, the discrepancies were numerically evaluated, and it is found that the discrepancies occur on a very tiny region. To make the flux calculation of such tiny regions more reliable, an approximation, in which the outgoing flux is equal to the incoming ne...

A new uncertainty reduction method is proposed to evaluate the prediction accuracy of neutronics characteristics by combining the generalized bias factor method and the cross section adjustment method. The present method is applied to a fuel fabrication process with erbia-bearing fuel. The cross section uncertainty of erbium is improved by means of cross section adjustment using experimental data of the erbia worths. For a blending machine (H/U = 0) used in the fuel fabrication process, the uncertainty reduction, which shows the rate of reduction of uncertainty, of the k eff is 0.604 for the present method and 0.555 for the conventional bias factor method. Thus, the prediction uncertainties are reduced by the present method compared with by the bias factor method.

ANCK/MIDAC is a three-dimensional (3D) code system that has been developed, which couples the 3D nodal kinetic code ANCK and the 3D drift flux thermal-and-hydraulics (T/H) code MIDAC. The adequacy of the ANCK feedback model was confirmed by analysis of the SPERT-III E-core reactivity accident tests. The chord weighting method was adopted in the effective fuel temperature (Teff) calculation model of ANCK and ANCK/MIDAC since it provides results that are in good agreement with those of a Monte Carlo calculation. The adequacy of the Teff model was also confirmed. The calculation results for hot startup and hot standby conditions using the ANCK code showed excellent agreement for the main parameters such as peak power, peak time, and behavior of reactor power. It was also confirmed that the ANCK feedback model with the chord weighting method was applicable to the reactivity-initiated accident (RIA) calculation for the case of 4.8 wt% UO2 core.

Two extended bias factor methods, the LC and PE methods, were applied to the prediction accuracy evaluation of neutronic characteristics of a breeding light water reactor, using data of FCA-XXII-1 critical experiments, in order to investigate the features and effectiveness of these methods on the basis of an actual core design and existing experimental results. The present study confirms the following features of these methods. Both the LC and PE methods can improve the prediction accuracy the most when all the experimental results are used. The prediction accuracy improvement is achieved mainly by reducing uncertainty due to errors in cross sections. This is done by realizing a profile of sensitivity coefficients closer to that of the target core and suppressing the influence of errors in experiments and experimental analysis methods. The PE method always improves the prediction accuracy with the use of any combination of experimental results. It is always superior to the LC method in the improvement of ...

The evaluation technique of prediction error reduction by the bias factor method is extended to the cases where the neutronic property of a target core is related to the properties in critical experiments. This extended procedure can utilize several experimental results, although the conventional procedure utilizes only one experimental result. Calculations were performed where the measured property is the reaction rate ratio of 238U capture to 239Pu fission (C28/F49) in FCA XXII-1 cores and the target core property is C28/F49 or breeding ratio in a breeding light water reactor (BLWR). The prediction error of the breeding ratio in BLWR is reduced from 0.57 to 0.43%, whereas the error of C28/F49 is reduced from 0.72 to 0.19%.

The three-dimensional (3D) neutronics and thermal-and-hydraulics (T/H) coupling code ANCK/MIDAC has been developed. ANCK/MIDAC consists of the 3D nodal kinetic code ANCK and the 3D drift flux T/H code MIDAC. In order to verify the adequacy of ANCK/MIDAC, the Phase II problem in the “OECD main steam line break benchmark (MSLB benchmark)” was analyzed. This MSLB benchmark has been defined in order to simulate the core response and the reactor coolant system response to a relatively severe steam line break accident condition. The Phase II problem has a conservative condition that the control rod with the maximum worth is stuck in a fully withdrawn position throughout the transient. The simulation was performed using the core inlet temperatures and flow rates for 18 different regions, which were provided by the PSU best-estimate TRAC-PF1/NEM calculations. The comparison of the ANCK/MIDAC results with other participants' results shows the excellent agreement on main core parameters. ANCK/MIDAC has good capabil...

Extended bias factor methods are proposed with two new concepts, the LC method and the PE method, in order to effectively use critical experiments and to enhance the applicability of the bias factor method for the improvement of the prediction accuracy of neutronic characteristics of a target core. Both methods utilize a number of critical experimental results and produce a semifictitious experimental value with them. The LC and PE methods define the semifictitious experimental values by a linear combination of experimental values and the product of exponentiated experimental values, respectively, and the corresponding semifictitious calculation values by those of calculation values. A bias factor is defined by the ratio of the semifictitious experimental value to the semifictitious calculation value in both methods. We formulate how to determine weights for the LC method and exponents for the PE method in order to minimize the variance of the design prediction value obtained by multiplying the design cal...

A new approach to calculate the burnup matrix using a simplified formula is proposed. We used the perturbation form of the burnup matrix and the Zassenhaus formula as the numerical procedure. The merit of the approach comes from the fact that it treats only the perturbed matrix whose eivenvalues are shifted so that the absolute values are much smaller than the original. The unperturbed matrix, which generally has large eigenvalues, is treated in advance and stored to be repeatedly used in each time step. The modified procedure leads to a drastic reduction of the computing task. A small test calculation was done to check the accuracy of the simplified formula. The accuracy of the simplified formula seems to be reasonable provided that the time step is appropriately set.

Three-dimensional (3D) neutronics and thermal-and-hydraulics (T/H) coupling code, ANCK/MIDAC, has been developed. ANCK/MIDAC consisted of the 3D nodal kinetic code ANCK and the 3D drift flux T/H code MIDAC. In order to verify the adequacy of ANCK that is a kinetics engine of this coupling code, several international benchmark problems have been performed. The calculation results of LMW (Langenbuch, Maurer and Werner) benchmark problems, PWR rod ejection benchmarks and PWR benchmarks on uncontrolled withdrawal of control rods at zero power are shown in this paper. The comparison of the results with the reference solutions shows very good agreements with the main core parameters.

A new methodology to produce subgroup parameters has been developed based on orthogonal factorization of weighting functions. In the existent methods the weighting functions do not appear explicitly, which causes the inconvenience in producing subgroup parameters in some situations. In the present method, the weighting functions are tailored to the required conditions by the use of orthonormal factorization method of which mathematical background is based on the Lanczos method. The obtained weighting functions can be commonly used for any physical quantities in order to produce corresponding subgroup parameters. The superiority of this approach becomes eminent especially when multiple conditions are specified at the same time. The numerical results of two simple examples have revealed the potential of the applicability of the present method to general problems. We have concluded that the present method may be developed into an efficient tool to deal with wide variety of problems using subgroup-method-rela...

A generalized bias factor method is proposed to improve the prediction accuracy of neutronics characteristics of a target core. The generalized bias factor method uses conventional bias factors calculated for several critical assemblies. The weighting factors for individual bias factors are determined to minimize the variance of neutronic characteristics of the target core. Numerical calculations are performed to investigate the uncertainty reductions of neutronics characteristics for a tight-lattice core. Though the uncertainty is not remarkably reduced for keff , that for the reaction rate ratio of 238U capture/239Pu fission is remarkably reduced: For example, the uncertainty reduction of the reaction rate ratio in the upper core is 0.871 for the present method, and 0.657 for the conventional bias factor method.

The Spatially Dependent Dancoff Method (SDDM) was recently developed to evaluate the power distribution within a fuel rod that has spatial variation of isotopic contents. The method was validated and verified by comparison to Monte Carlo calculations and measurements. However, those evaluations and comparisons were based on the assumption that the temperature distribution within a rod is flat. In this study, an equation used in the SDDM is enhanced in order to more accurately treat the temperature distribution. The enhancement was carried out with the knowledge that a Monte Carlo calculation shows no effect of temperature distribution on spatial flux within a rod. This leads to the cancellation of reaction rate changes due to temperature distribution between inner and outer regions within a rod. The knowledge gained from these evaluations was then applied to the equation used in the SDDM with the temperature distribution. The improved SDDM was validated and verified by comparison to MCNP4C calculations. P...

A new theory and methodology to optimize subgroup parameters were established by the use of Chebyshev approximation. The optimization of the fitting method was realized by the minimax approximation provided by the Remes algorithm. As for the moment method, a new definition for moments was proposed to cure the poor reproducibility of effective cross section by the original moment method. The new moment method uses transformed Chebyshev polynomials as orthogonal bases. This method is mathematically stable and the efficiency was verified by a numerical example. It was also shown that use of preconditioning can furthermore enhance the efficiency. The first dominant term of the new moments can be approximated by old moment definition σα with α= 0.1–0:6, including the case with α=−1/2. This fact can explain the efficiency of the reservation of half-integer moments in the reproducibility of effective cross section that was mentioned by Unesaki.

To estimate the uncertainty of the neutronic parameters it is required to calculate sensitivity coefficients to cross section changes. The sensitivity coefficients are usually calculated based on the diffusion theory. However, the accuracy of the sensitivity coefficients becomes doubtful because of the use of diffusion theory. The purpose of this paper is to derive the generalized perturbation theory using the transport theory, and to evaluate transport effect on the sensitivity coefficients. Sensitivity calculations were performed in 70 energy groups. The calculated sensitivity coefficients were compared with those obtained from direct transport calculations to check the accuracy of the present method, and those obtained from the generalized perturbation calculation based on the diffusion theory to investigate the transport effect on sensitivity coefficients. From the comparison, the difference was remarkably large for the neutronic parameters such as reaction rate distribution, reaction rate ratio, and ...

Most of the lattice physics codes, used for routine core design calculations, are based on the spatial flat-flux assumption representation to solve the neutron transport equation. Consequently for regions (like fuel or control rods) with strong flux gradient, a fine computational mesh becomes required for better accuracy of the predictions. In the case of a PWR assembly, this situation particularly occurs with Gadolinia fuel (UO2-Gd2O3) or Erbia fuel (UO2- Er2O3) rods. The aim of this study is to determine, for UO2-Gd2O3 and UO2-Er2O3 pins, the optimal number of computational fuel mesh rings that preserves good accuracy and at the same time is consuming minimum computational running time. This study will be carried out using PARAGON lattice physics code. PARAGON is a two-dimensional neutron/gamma transport code used mainly to generate the group constants for core simulator codes such as ANC. PARAGON (with its new module SDDM) can treat the multi-region resonance self-shielding effect for all resonant isot...

A method for calculating effective microscopic and macroscopic cross sections for media containing randomly dispersed particles, which was originally derived by Shmakov et al., is improved to extend its applicability to a wider range of criticality calculations. The newly modified method can be applied to media containing more than one particle type. This technique is incorporated into a continuous energy Monte Carlo code MCNP and can be applied to a wide variety of criticality problems. The cause of an inadequacy of the original method for larger particles is investigated, and the accuracy is found to be improved by using an adjustment parameter. This method originally underestimated fission neutrons’ tracks in particles, which degrades calculation results for larger particles. A simplified method to implement the effect of the fission neutron tracks into MCNP is developed. We demonstrate that the new technique is successfully applied to MOX fuel rods with plutonium spots, fuel solution containing absorber particles, etc. The newly improved method can treat media containing particles with unequal diameters. However, a sample calculation shows the accuracy of criticality calculations becomes worse with increasing variation in particle diameters. The new method can also successfully perform a criticality calculation for media containing different particle compositions with an equal diameter.

A new spatially dependent resonance self-shielding method (SDDM: Spatially Dependent Dancoff Method) was developed based on the generalization of the conventional Dancoff method to multi-regions in a fuel pellet based on the Stoker/Weiss technique. SDDM correctly accounts for radial power distribution within fuel rods in a fuel assembly. SDDM is fully consistent with the conventional method if the pellet is not sub-divided. It also has the advantage of being less computing time consuming when compared to more rigorous resonance shielding method such as sub-group and special fine energy mesh methods. Moreover, it can be installed easily into the lattice physics code widely used in commercial LWR design. To validate the method, spatial concentration of isotopes and burnup distribution within a rod are evaluated using SDDM and the results are compared to the destructive measurement data. From the comparison, it is concluded that the spatially dependent Dancoff method, SDDM, is appropriate for generating the ...

A series of critical experiments has been analyzed by the deterministic method code CHAPLET-3D in two- and three-dimensional core configurations in which explicit core structures are represented. The results show that the three-dimensional core calculation model employed in CHAPLET-3D code is valid and useful to obtain fine resolution results by the deterministic method. Moreover, the conventional two-dimensional axial buckling calculation for critical experiment analysis has also been validated, through the comparison between the results of two- and three-dimensional experimental core analyses by CHAPLET-3D code.

A three dimensional heterogeneous core transport analysis code CHAPLET-3D which is based on deterministic methods has been developed. In CHAPLET-3D code the non-linear iteration technique which is commonly used in advanced nodal diffusion codes is employed to perform three dimensional heterogeneous core calculation in a form of conventional finite difference method with the accuracy of the method of characteristics in radial two dimensional geometry. For an axial direction solver in addition to finite difference method and nodal expansion method in diffusion theory the method of characteristics has been incorporated in order to take account of transport effect. According to the verification tests compared with the results of multi-group Monte Carlo reference calculations it is found that the accuracy of CHAPLET-3D code for three dimensional heterogeneous core analysis is almost the same level as that of the reference calculation and also demonstrated that the three dimensional core analysis method utilizi...

The effect caused by the circular approximation of the geometry for cell calculations in light water reactors is studied using the continuous-energy Monte Carlo code MVP. It was found that the k{sub inf} values were underestimated with this approximation of the geometry, especially in the case of a mixed-oxide fuel cell. To treat the square geometry, including the resonance calculation, KRAM-B was developed based on the two-dimensional neutron transport code KRAM as a deterministic cell calculation code. KRAM-B solves the neutron transport equation using a combination of the subgroup method and the characteristics method. The subgroup method is able to perform the resonance calculation faster than the ultrafine energy group calculation and predict the resonance cross section more accurately than the Dancoff factor method. It was found that the k{sub inf} values and the effective microscopic resonance cross sections by KRAM-B agreed well with the reference MVP results.

When the continuous energy transport equation is integrated over an energy interval, the total cross section for the energy group becomes region and angular dependent, because the angular dependent neutron flux must be used as a weight. Usually the angular dependence is neglected. We investigate the approximation, and show that the error is proportional to the product of the neutron current and the difference between the flux weighted total cross section and the current weighted total cross section. The formulation of angular dependent cross section is derived in 1-dimensional slab geometry based on the multiband method. The numerical results are shown for UO2 and MOX fueled cells.

The effect of the moderator density distribution of annular flow on the fuel assembly neutronic characteristics in a boiling water nuclear reactor was investigated using the SRAC95 code system. For the investigation, a model of annular flow for fuel assembly calculation was utilized. The results of the assembly calculation with the model (Method 1) and those of the fuel assembly calculation with the uniform void fraction distribution (Method 2) were compared. It was found that Method 2 underestimates the infinite multiplication factor in the fuel assembly including the gadolinia rod (type 1 assembly). This phenomenon is explained by the fact that the capture rate in the thermal energy region in gadolinia fuel is estimated to be smaller when the liquid film of annular flow at the fuel rod surface is considered. A burnup calculation was performed under the condition of a void fraction of 65% and a volumetric fraction of the liquid film in liquid phase of 1. It is found that Method 2 underestimates the infin...

The effect of radial subchannel-wise void distribution in a fuel assembly on assembly neutronic characteristics has been investigated using the assembly calculation code SRAC95 and the subchannel analysis code THERMIT2. With the iterative calculation of assembly calculation and the subchannel analysis (Method 1), subchannel-wise void fraction distribution, pin-power distribution and the infinite multiplication factor of the assembly are calculated. The results are compared with the result of the assembly calculation using uniform void distribution as input (Method 2). The calculation is performed for two assembly configurations in the present study: one is a fuel assembly that does not include a water rod (Case 1) and the other is the assembly that includes a water rod (Case 2). The differences in the infinite multiplication factor and pin-power peaking factor between the two methods are small in both cases. In typical BWR fuel assemblies that are investigated in the present study, the method that does no...

A new 3-D nodal SNtransport method has been developed for Cartesian and hexagonal geometries. In the present method, the neutron angular distributions of intra-node fluxes and transverse-leakage are represented using the SNquadrature set, and the spatial distribution of neutron source is approximated by a quadratic polynomial expansion. Several considerations in numerical procedures to achieve stable convergence are also mentioned. Some numerical results are represented to show the validity of the method.

A space/time nodal diffusion code based on the nodal expansion method (NEM), EPISODE, was developed in order to evaluate transient neutron behavior in light water reactor cores. The present code employs the improved quasi-static (IQS) method for spatial neutron kinetics, and neutron flux distribution is numerically obtained by solving the neutron diffusion equation with the nonlinear iteration scheme to achieve fast computation. A predictor-corrector (PC) method developed in the present study enabled to apply a coarse time mesh to the transient spatial neutron calculation than that applicable in the conventional IQS model, which improved computational efficiency further. Its computational advantage was demonstrated by applying to the numerical benchmark problems that simulate reactivity-initiated events, showing reduction of computational times up to a factor of three than the conventional IQS. The thermohydraulics model was also incorporated in EPISODE, and the capability of realistic reactivity event an...

The effective technique to accelerate the convergence of a fission source distribution in a Monte Carlo simulation is proposed as an application of a fission matrix. It is found that this acceleration method is especially useful for large core analysis where the fission source distribution is slowly converged in a Monte Carlo simulation by the source iteration method. It is found that the number of inactive cycles can be automatically determined and reduced by the acceleration method in this investigation.

A three-dimensional time-domain core analysis code was applied to numerical simulations for an actual regional neutron flux oscillation observed in a commercial BWR core, in order to investigate potential nonlinear behavior in its coupled neutronic and thermohydraulic system. The present study shows existence of the nonlinear reactivity interaction between the fundamental and first azimuthal spatial harmonics modes of neutron flux distribution under the regional event. The spectrum analysis of the simulated data provides a unique result, that is, temporal harmonics peaks are excited at the even- and odd-order multiples of the characteristic resonance frequency in the fundamental and first spatial harmonics responses, respectively. The numerical simulation also shows that the strong nonlinearity of the coupled neutronic and thermohydraulic dynamics locally appears where the power unstably oscillates with large amplitudes, inducing the power shift and reactivity bias which are shown in the core-wide situati...

The present paper discusses the effect on accuracy of eigenvalue by the degree of discretization at the cell boundary within the framework of the multiband-CCCP, a combination of the multiband method and the CCCP method. A study on sensitivity of discretization has been done concerning the dependence of surface flux distribution in space and angular domain, or interface currents. It is found that appropriate discretization with more than five for both of segments and sectors is required to accurately calculate effective cross sections and eigenvalue in the multiband-CCCP method. However, in practice, discretization schemes with one segment and several sectors can be employed with some biases on eigenvalue. The distribution of surface flux at the boundary of a typical PWR-MOX cell has been studied. It is found that discontinuity and strong angular dependence in the distribution of surface flux on the off-centered segments have large influence in the multiband-CCCP method. An improved scheme of angular discretization has been examined for the feasibility to retain calculation accuracy with less degree of discretization in angular domain.

To determine rapidly the core-power ratio in a coupled-core system, a method is proposed on the basis of the control rod drop experiment. A formula of an asymmetrical two-point version was derived to deduce the core-power ratio and subcriticalities of the individual cores. It requires only a familiar measurement technique and tools for the conventional rod drop experiment to apply this formula for the purpose of obtaining these quantities. The present method was applied to the rod drop data measured in coupled-core systems, where the core-power ratio sensitively depended on the rod patterns. The validity of the proposed method was experimentally demonstrated through the comparison between the measured core-power ratios obtained by the present method and that derived from the flux distribution measurement.

Analysis of the benchmark problems on the void coefficient of mixed-oxide (MOX)-fueled tight-pitch cells has been performed using the Japanese SRAC code system with the JENDL-3.2 library and the Fr...

A new method was applied to calculation of reaction rates in blanket of LMFBR using the multiband Sn theory. This procedure leads to the use of direction dependent total micro cross sections, which advances the penetration of neutrons into the blanket. Test calculation with RZ model of a prototype fast reactor shows that reaction rates tend to rise with the penetration into blanket compared to the conventional multigroup calculation: the maximum difference was about 3% for 238U capture, 4% for 235U fission, 4% for 239Pu fission, and less than 1% for 238U fission. This tendency shows the same direction as the difference observed between the continuous energy Monte Carlo method and the conventional method.

The effect of hydrogen concentration and loaded mass of minor actinides (MAs) in the target on the core performance and MAs transmutation rate was analyzed in this paper. An optimum core was propos...

A formulation has been established to estimate the error propagation in Monte-Carlo burnup calculations. The uncertainties in cross sections and the statistical errors in Monte-Carlo calculations a...

The multiband method has been extended to treat the resonance interference effect between two nuclides based on the intermediate resonance approximation. The integral equation of the flux belonging to different bands of the two nuclides is derived for a heterogeneous cell system. In the equation, a new band parameter is introduced. The new parameter denotes the conditional probability that a nuclide takes a certain band under the condition that the other nuclide takes another band. The calculational procedure of band parameters is described in a homogeneous medium. This method has been applied to a homogeneous medium and a thermal reactor cell containing 235 U and 238 U. The effective cross sections calculated by this method and the conventional multiband method without considering the interference effect are compared with the results by a reference continuous-energy Monte Carlo method. It is seen that the conventional multiband method greatly overestimates the fission and capture cross sections of 235 U for energy groups where there are both resonances of 235 U and 238 U, and the present method remarkably improves the overestimation.

An analysis method for BWR regional stability is developed in frequency-domain. Conventionally, the energy spectrum of fundamental eigenfunction has been applied to the calculation of first-harmonic void reactivity instead of the spectrum of first-harmonic eigenfunction. In the present method, the fine energy spectra of first-harmonics are applied to the evaluation of void reactivity. The spectra are obtained by assembly calculation considering the geometric bucklings of spatial harmonics. The present method is applied to the Ringhals 1 benchmark test problem. Comparing the present method with the method without the spectrum of first-harmonic eigenfunction, we can estimate the effect of the spectrum on the frequency-domain modal analysis of regional stabilities in BWRs. The decay ratio (DR) of regional stability is also calculated by these two methods. It is found that the latter method tends to underestimate the absolute value of void reactivity by about 10%. By considering the spectra of first-harmonic ...

The first-harmonic eigenvalue separation, the difference between the fundamental and the first order eigen-values of the higher harmonic neutron transport equations, which were measured at the Kyoto University Critical Assembly (KUCA) has been analyzed. A method was proposed to calculate the first order eigenvalue based on the discrete ordinate method. The 3-D effect, energy group effect, mesh size effect, and transport effect were investigated. Among these effects, the transport effect was significant and when it was taken into account, the calculated eigenvalue separation approached the measured value on the KUCA coupled-core.

To provide accurate effective cross sections for core calculations, the multiband method was applied to light water reactor assembly calculations. The multiband method has been extended to arbitrary geometries by introducing band-dependent currents at the boundaries of a region. The transport of neutron is treated by the angular space-dependent current coupling collision probability method. A fuel assembly is divided into heterogeneous domains where the multiband method is applied directly by using collision probabilities. Several examples of numerical calculations for UO{sub 2} and mixed oxide fuel assemblies are shown. The space dependence of the effective cross section can be expressed accurately by this method, which leads to an accurate prediction of k{sub {infinity}} values.