ML20249A220

SER Accepting Licensing TR BAW-10221P, NEMO-K,Kinetics Solution in Nemo

ML20249A220
Person / Time
Site:	07001201
Issue date:	06/10/1998
From:	NRC
To:
Shared Package
ML20249A216	List: ML20249A215 ML20249A220
References
NUDOCS 9806160191
Download: ML20249A220 (6)

Text

{{#Wiki_filter:__ __- _ _ _ _ _ _ _ _ _ - _ - _ - _ _ _ _ - _ _ - . _ _ . .

1. )

4. ..

ENCLOSURE' SAFETY EVALUATION REPORT BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATING TO TOPICAL REPORT BAW-10221P.

                                                                                      "NEMO-K. A KINETICS SOLUTION IN NEMO."

FRAMATOME TECHNOLOGIES

                           .i. BACKGROUND By letter dated September 30,1997, Framatome Technologies, in conjunction with Framatome

Cogema Fuels; submitted the topical report BAW-10221P, "NEMO-K, A Kinetics Solution in NEMO" (Reference 1), for the staff's review. Topical Report BAW-10221P gives a description of the NEMO computer code and its application for reactor analyses. BAW-10221P describes the incorporation of the time-dependent solutions for neutronics, fuel temperature, and coolant properties to the steady-state NEMO computer code. NEMO solves the nodal balance equation in three dimensions to determine the neutron flux, the source, the relative power density

                         = (including pin power reconstruction), and the reactor core reactivity. The kinetics equations are added to NEMO so that both the static and the kinetic solutions are available in the san's code.

All the history files, cross section files, input and outputs for NEMO are used by NEMOS, thus allowing state-of-the art steady-state computer modeling to be extended to time-dependent

solutions with a reasonable amourt of effort. BAW-10221P describes the analyses and presents the assumptions, conservatism, limitations, benchmarking, and results of the NEMO-K computer program.

NEMO-K is benchmarked against several industry standards to qualify its time-dependent ] features and solutions. The benchmarking provides a check of the three distinct time-

                         ~ dependent calculations and the overalllinking ofinputs and outputs of these calculations. The three kinetic features (or modules) that are added to NEMO are the time-dependent neutron flux calculations, the time-dependent fuel pin temperature calculations, and the time-dependent
                         . core coolant cal-culations. Each of the modules is benchmarked individually against its respective benchmark. The benchmark analyses (comparisons) showed that the individual modules are accurately calculating the reference solutions, with insignificant differences.

2. TECHNICAL EVALUATION-METHODOLOGY 2.1 Model Desenptions 2.1.1 The Neutronic Model The neutronic model evaluates the cross sections, computes the three-dimensional flux shape, and calculates reactor power. Typical inputs to the model are the control rod position, the fuel pin temperature, the water density,. the boron concentration, and the neutron flux time step.

Typical outputs are the power generated in the fuel and the power directly deposited in the

water (energy transfer from the center of tne fuel to the cladding surface).

I 1

  <1 9906160191 9906'10N eM                    . f                                                                            )                                    .!

4g m

The main function of the neutronic modelis to calculate the magnitude of the three-dimensional

          . flux within the core. The neutronic model evaluates the cross sections, computes the three-dimensional flux shape, and calculates reactor power. Cross sections are evaluated in the same manner as the steady-state version of NEMO.

Core transient calculations require the coupling of neutronic, fuel pin thermal, and core coolant (thermal-hydraulic) models. Required kinetic constants have been added to meet the additional data requirements for the transient calculations. The neutronic model incorporates the solutions to coupled neutron diffusion equations and . . delayed neutron precursor equations in order to determine the three-dimensional neutron and j precursor flux and density in the core. The neutron flux shape and density calculations are, in l turn, used to perform power calculations. l 4 2.1.2 Fuel Pin Temperature Model The NEMO-K methodology assumes a circular fuel rod with azimuthally symmetric heat generation with no axial heat conduction. This assumption enables the radially dependent _ (one-dimensional) conduction equation to be solved to yield the temperature distribution in the fuel pin. Solving this equation requires setting up boundary conditions that are imposed on the pellet's inner radius heat flux and the outer-radius temperature. Solutions to the equations are obtained by integrating over time and space for each mesh and using appropriate material data within each mesh. The mesh properties are evaluated at the mesh-centered values. Thermal expansion is accounted for as a function of the mesh temperature and the material content for each radial mesh. The thermal conductivity correlations are modeled as functions of the mesh temperature, the material content, the nodal burnup, and the initial plutonium content. Specific heat is also accounted for as a function of the mesh temperature and material content for each mesh region. The material properties are evaluated at the mesh temperature values of the current time step. Both the fuel pin temperature calculation and the material property calculation are solved iteratively to determine the temperature distribution for the current time step. NEMO-K treats the fuel gap between the pellet and the clad as a very thin layer. It is accounted for by using a thermal resistance of contact and by using a modified coupling term to preserve the flux of the actual geometry. The pellet-to-clad gaps are treated using a gap conductance model. The gap conductance data are usually obtained from interpolation tables  ; as a function of fuel type, burnup, pellet temperature, and clad temperature. The gap thickness  ! is obtained by computing the thermal expansion of the pellet and the cladding. 2.1.3 The Thermal Hydraulic Model The transient thermal-hydraulic model in NEMO-K is based on the steady- state thermal-hydraulic modelin NEMO. In NEMO, coolant is assumed to enter the bottom of a vertical 1 channel in the core. The coolant inlet enthalpy is known for each channel. It is also assumed that the system pressure is known and is constarit throughout the core. No mixing is assumed to occur between channels, and the coolant enthalpy is determined for each node by solving i the_ mass continuity and energy conservation equations.

                                                                                                                ]

ll 2 f- ___-___-_____-____ _~

The same basic assumptions are made in the transient model of NEMO-K for determining the

         . average enthalpy between nodes. The power (energy) is transmitted to the fluid and is the sum of the direct heat generation in the fluid and the heat transmitted across the fuel cladding. The finite difference calculation assumes the density to be constant during the time step, where the time step assumes a relation that is dependent on the size of the axial node and the velocity of the fluid.

3. BENCHMARKING The vendor provided several benchmarking cases to qualify NEMO-K as a time-dependent solver. NEMO-K utilizes the fully implicit theta modeling for all the time dependent calculational modules, thus ensuring unconditional stability. Benchmarking is carried out for all three distinct time-dependent modules, the neutron flux, the fuel pin temperature, and the core coolant calculations.

3.1 Neutron Flux Benchmarks The neutronic effects of NEMO K were compared against two well-established benchmarked problems, TWIGL-2D step and TWIGL-2D ramp transients (Reference 2). The step transient is symbolic of a rod ejection, while the ramp transient is designed to simulate a fast rod withdrawal. Data provided by the vendor indicative of core average power versus time, for both the step and ramp transient benchmark cases, show excellent agreement with the reference results. 3.2 FuelTemocrature Benchmarks The fuel temperature aspects of NEMO-K were benchmarked against RELAP5. Three cases were examined: a steady-state case, a full-power transient, and a zero-power transient (Reference 3). 3.2.1 Steady-State Comparison Two steady-state cases were run with both NEMO-K and RELAP5. The first comparison was made for a constant fuel thermal conductivity (K), and the second comparison was rrcie for the temperature-dependent thermal conductivity (K(T)). Results indicate that NEMO-K and RELAP5 are calculating very similar results. If anything, NEMO-K overpredicted the centerline i tempera-ture by less than 10 degrees. This degree of accuracy is sufficient for kinetic applications because the accuracy associated with fuel temperature models is roughly on the order of 40 degrees. 3.2.2 Companson of PowerTransients This comparison involved the simulation of a typical rod ejection at hot full power. Analysis of the results shows that the NEMO-K fuel temperatures are consistently about 5 degree Kelvin higher than the RELAPS values at all times. This finding is consistent with the steady-state results from the temperature-dependent thermal fuel properties. The vendor also conducted sensitivity studies with NEMO-K to ensure that a sufficient number of intervals (rings) were used to appropriately model the fuel temperature dependence. This

         - step was accomplished by doubling the number of iterative intervals from 10 to 20. However,
             ~

3

I l

           . analysis of the results showed that'10 intervals (rings) in the fuel will provide sufficient        l resolution to accurately determine the transient fuel temperature.

3.2.3 Zero-Power Transient A NEMO-K test was performed to simulate the rod ejection accident at hot zero power, j Analysis of the results again indicated that there is very good agreement between the RELAP5 q

           - results and NEMO-K and that it is apparent from the results that both NEMO-K and RELAPS are calculating very similar temperatures within the fuel pellet versus time. Review and analysis of the data submitted by the vendor indicate that NEMO-K edequately models time-dependent fuel temperature effects.

3.3 Core Thermal Model The transient core coolant model in NEMO-K was compared to steady-state and transient

           , examples using the existing core model in NEMO. 'Two NEMO cases were compared with two NEMO-K steady-state cases (large time steps). Analysis of the results indicates very good agreement, which is not surprising since NEMO-K with large time steps reduces to the same equations used by NEMO.

The transient case consisted of progressing the thermal model from the steady state condition to a transient condition by linearly changing the inlet temperature over a given time interval. The result we a smooth transition from a prompt behavior to an asymptotical steady-state core power value. 3.4 Total Core Model Benchmarking 3.4.1 NEACRP Reference Benchmarks The vendor performed several calculations to benchmark the adequacy of the NEMO-K model. The benchmarking was carried out in accordance with the Nuclear Energy Agency Committee for Reactor Physics (NEACRP) reference benchmarking specification document (Reference 4) in which such things as the core configurations, the inputs, and the conditions for each of the transient cases are defined. Typical plant data for pressurized water reactor (or boiling water reactcrs) such as boron concentrations, moderator temperature, moderator density, fuel temperature, and control rods are utilized as defined by the reference benchmarks. Six cases were benchmarked, simulating such scenarios as rod ejection from three dWerent rod configurations, one from zero power, and one from full power. These cases will provide a range of reactivity insertions and core conditions that can be used effectively to test the response of the neutronic, core cooling (thermal), and fuel temperature transient models. The vendor reviewed and referenced a summary of results from severalindustry codes (Reference 5) that were compiled for these cases. The calculations from NEMO-K were compared against the reference code PANTHER (Reference 6). Analysis of the steady-state results between NEMO-K and PANTHER served to verify that the

input data are correct and that steady-state calculations by the two codes using the same input data are very similar, thus confirming that the initialization of the codes is equhlent i

V

7 The kinetic (o,- time-dependent) results between NEMO-K and the referenced code for power

                 . peaking were within 1 percent for all of the six cases. Transient fuel temperatures were within 2.3 "C and power distribution magnitudes for unrodded and rodded assemblies were less than 0.004 and 0.016 respectively for all times analyzed. Analyses of the time-dependent results of comparing NEMO-K with PANTHER indicate that the NEMO-K time-dependent results for the neutronics, the fuel temperatures, and the core coolant properties appropriately model the time-  )

dependent behavior, j 3.4.2 Dropped Rod Test The vendor also conducted a rod drop test in which the NEMO-K results are compared to excore measurements. Analysis of the reaufts indicated that NEMO-K responded within 1 percent of the measured results of Detectors 1,3, and 4 and within 2 percent for Detector 2. Further analysis of the results illustrated that, in general, the NEMO-K model slightly overpredicts (<1%) the initial power decrease from the dropped rod and tends to overpredict (<0.5%) the power increase after about 9 seconds from Detectors 1 and 4. The vendor conducted dropped rod sensitivity studies to determine the effects of such parameters as rod worth, beta effective, and gap conductances since the effects of these parameters could affect the results. The study showed that the minimum power is affected by the rod drop worth and/or the beta effective. The difference between the measured and the predicted results turned out to be very small and was attributed to the differences in the fuel temperature models. The final result is that the power distributions calculated by NEMO-K are in texcellent agreement with the measured results and that the small differences are well within the accuracy of the input parameters. The staff concurs with the vendor's results. 4 CONCLUSION The staff has reviewed the analyses in Topical Report BAW-10221P,"NEMO-K, A Kinetics Solution in NEMO," and concludes that BAW-10221P is acceptable. BAW 10221P is acceptable because NEMO-K calculations compare very favorably with those calculated by RELAPS and the NEA referenced benchmarks. NEMO-K is thus well suited for performing three-dimensional time-dependent calculations in which accuracy is very important in determining such parameters as reactivity, peak and core power, fast (rod ejection) transients and slow (rod drop) transients, differential and integral rod worth. This approval is applicable to the subject matter as statcd in this topical, BAW-10221P, NEMO-K, A Kinetics Solution in NEMO."

5.

REFERENCES:

1. Letter from J.H. Taylor, Framatome Technologies, submitting Topical Report BAW-10221 (Proprietary), "NEMO-K, A Kinetics Solution in NEMO," to the U.S. Nuclear Regulatory Commission, September 30,1996.

2. L.A. Hageman and J.B. Yasinsky, " Comparison of Altemating Direction Time-Dependent Methods and Other implicit Methods for the Solution of the Neutron Group' Diffusion Equations," Nucl. Sci. Eng.,38,8 (1969) 5

________J-_---- - .,

3. _ BWNT Topical Report "RELAP5/ MOD 2-B&W-An Advanced Computer Program for

               .    . Li0ht Water Reactor LOCA and NON-LOCA Transient Analysis," BAW-10164, Rev. 3, October 1992.

4. H. Finnemann and A.G. Galati, "NEACRP 3-D LWR Core Transient Benchmark, " Final Specifications, NEACRP-L-335, (Rev.1), October 1991, (January,1992).

5. H. Finnemann and H. Baur, A.G. Galati, and R. Martinelli, "Results of LWR Core )

Transient Benchmarks," NEA/NSC/ DOC (93)25, October 1993. l i

6. P.K. Hutt and M.P. Knight, "The Development of a Transient Neutron Flux Solution in l the PANTHER Code," Trans. Am. Nuct. Soc. 61,348 (1990). l I

l l 6 I =___--__________-_}}