13
Review Use of coupled code technique for Best Estimate safety analysis of nuclear power plants Anis Bousbia-Salah * , Francesco D’Auria Dipartimento di Ingegneria Meccanica, Nucleari e della Produzione, Facolta ` di Ingegneria, Universita ` di Pisa, Via Diotisalvi, 2, 56126 Pisa, Italy Abstract Issues connected with the thermal-hydraulics and neutronics of nuclear plants still challenge the design, safety and the operation of Light Water Nuclear Reactors (LWR). The lack of full understanding of complex mechanisms related to the interaction between these issues imposed the adoption of conservative safety limits. Those safety margins put restrictions on the optimal exploitation of the plants and consequently reduced economic profit of the plant. In the light of the sustained development in computer technology, the possibilities of code capabilities have been enlarged substantially. Consequently, advanced safety evaluations and design optimizations that were not possible few years ago can now be performed. In fact, during the last decades Best Estimate (BE) neutronic and thermal-hydraulic calculations were so far carried out following rather parallel paths with only few interactions between them. Nowadays, it becomes possible to switch to new generation of computational tools, namely, coupled code technique. The application of such method is mandatory for the analysis of accident conditions where strong coupling between the core neutronics and the primary circuit thermal-hydraulics, and more especially when asymmetrical processes take place in the core leading to local space-dependent power generation. Through the current study, a demonstration of the maturity level achieved in the calculation of 3-D core performance during complex accident scenarios in NPPs is emphasized. Typical applications are outlined and discussed showing the main features and limitations of this technique. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: Coupled code technique; Best Estimate tools; Nuclear safety analysis 1. Introduction Computer codes are widely used for safety analysis within the framework of the licensing and safety improvement programs of existing NPPs, better utilization of nuclear fuel and higher operational flexibility, for justification of lifetime extensions, development of new emergency operating procedures, analysis of operational events and devel- opment of accident management programs. The recent availability of powerful computer and computational techniques has enlarged the capabilities of making realistic simulations of complex phenomena in NPPs and more precise consideration of multidimensional effects. * Corresponding author. Tel.: þ39 050 2210354; fax: þ39 050 2210384. E-mail address: [email protected] (A. Bousbia-Salah). 0149-1970/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.pnucene.2006.10.002 www.elsevier.com/locate/pnucene Progress in Nuclear Energy 49 (2007) 1e13

Use of coupled code technique for Best Estimate safety analysis of nuclear power plants

Embed Size (px)

Citation preview

www.elsevier.com/locate/pnuceneProgress in Nuclear Energy 49 (2007) 1e13

Review

Use of coupled code technique for Best Estimate safetyanalysis of nuclear power plants

Anis Bousbia-Salah*, Francesco D’Auria

Dipartimento di Ingegneria Meccanica, Nucleari e della Produzione, Facolta di Ingegneria, Universita di Pisa,

Via Diotisalvi, 2, 56126 Pisa, Italy

Abstract

Issues connected with the thermal-hydraulics and neutronics of nuclear plants still challenge the design, safety and the operationof Light Water Nuclear Reactors (LWR). The lack of full understanding of complex mechanisms related to the interaction betweenthese issues imposed the adoption of conservative safety limits. Those safety margins put restrictions on the optimal exploitation ofthe plants and consequently reduced economic profit of the plant.

In the light of the sustained development in computer technology, the possibilities of code capabilities have been enlargedsubstantially. Consequently, advanced safety evaluations and design optimizations that were not possible few years ago can nowbe performed. In fact, during the last decades Best Estimate (BE) neutronic and thermal-hydraulic calculations were so far carriedout following rather parallel paths with only few interactions between them. Nowadays, it becomes possible to switch to newgeneration of computational tools, namely, coupled code technique. The application of such method is mandatory for the analysisof accident conditions where strong coupling between the core neutronics and the primary circuit thermal-hydraulics, and moreespecially when asymmetrical processes take place in the core leading to local space-dependent power generation. Through thecurrent study, a demonstration of the maturity level achieved in the calculation of 3-D core performance during complex accidentscenarios in NPPs is emphasized. Typical applications are outlined and discussed showing the main features and limitations of thistechnique.� 2006 Elsevier Ltd. All rights reserved.

Keywords: Coupled code technique; Best Estimate tools; Nuclear safety analysis

1. Introduction

Computer codes are widely used for safety analysis within the framework of the licensing and safety improvementprograms of existing NPPs, better utilization of nuclear fuel and higher operational flexibility, for justification oflifetime extensions, development of new emergency operating procedures, analysis of operational events and devel-opment of accident management programs. The recent availability of powerful computer and computationaltechniques has enlarged the capabilities of making realistic simulations of complex phenomena in NPPs and moreprecise consideration of multidimensional effects.

* Corresponding author. Tel.: þ39 050 2210354; fax: þ39 050 2210384.

E-mail address: [email protected] (A. Bousbia-Salah).

0149-1970/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.pnucene.2006.10.002

2 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

This technique could be applied for different purposes. A typical example is the coupling of primary systemthermal-hydraulic codes with 3D Neutron Kinetics Codes. Other cases include coupling of primary systemthermal-hydraulics with structural mechanics, Computational Fluid Dynamics (CFD), nuclear fuel behavior andcontainment behavior.

The capabilities of the coupled code calculations to simulate, in a BE way, plant behavior under a wide variety oftransient and accident conditions have been largely investigated through several international programs. These activ-ities include the OECD/NEA Benchmarks as the PWR Main Steam Line Break (MSLB) in TMI-1 (Ivanov et al.,1999), the BWR Turbine Trip (TT) in Peach Bottom (Solis et al., 2001), and the VVER1000 coolant transient (Ivanovand Ivanov, 2002).

However, notwithstanding the complexity of these codes and the level of the present scientific knowledge, a com-puter code cannot be expected to accurately model phenomena that are not yet fully understood by the scientificcommunity.

In general, the results of code predictions, specifically when compared with experimental data reveal somedeviations. These discrepancies could be attributed to several reasons as model deficiencies, approximations in thenumerical solution, nodalization failure, imperfect knowledge of boundary and initial conditions. Reliability predic-tion of BE coupled code tools includes the need of a general code qualification process. This could be performedthrough the consideration of experimental data issued from operational NPP data, Integral Test Facilities (ITF) orSeparate Effects Test Facilities (SETF) validation matrices (Aksan et al., 1993). In addition, nodalization qualifica-tions, as well as qualitative and quantitative accuracy of the code results are also needed for the code qualificationprocess (D’Auria and Galassi, 1998). Therefore, it is necessary to investigate the uncertainty of the results and thesensitivity effect of the most effective parameters.

The lack of immediate industrial interest to the coupled code technique, owing to the natural caution and conser-vatism from the regulatory bodies in accepting innovations, prevented so far the exploitation of the consideredtechnique. An attempt is made herein to emphasize the state of the art and the main features of such computationaltools through typical applications for simulating transients in different NPPs.

2. Coupled computational tools

The evaluation of complex phenomena in NPPs is closely related to the ability of determining the time-space coreflux distribution as well as the flow field conditions and the associated effects from heat sources and heat sinksthroughout the reactor coolant system. Online measurements at different locations of the NPP can provide valuableinformation in this context but important details will not be revealed by this mean especially for transients wherestrong feedback exist between core neutronics and coolant loop, and asymmetric phenomena events in the core areinvolved. The need of coupled code for safety analyses calculations is enough cute; this technique is performed usingthe following codes (D’Auria, 2004):

� Code for deriving neutron kinetics cross sections (CSC or Cross Section Code).� Thermal-Hydraulic System Code (THSC).� Neutron Kinetics Codes (NKC).

The CSC can be used out of line since the outputs of such codes are used by coupled THSC and NKC which interactat each time step during a phenomenon simulation. However, two fundamental pre-conditions shall be fulfilled forthe correct application of such complex codes to the prediction of transient scenarios expected in NPPs:

- The code should be frozen to ensure that no unjustified modifications of the constitutive models would alter theresults. The code must be able to correctly simulate almost all the transient dominant phenomena using themodels of the adopted frozen version.

- The code should be properly qualified through wide, preferably international, assessment programs, includingboth:� The verification through the comparison with results obtained by similar codes and the validation.� Through experimental data obtained from plant and/or test facilities.

3A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

3. Application range of coupled code technique

Coupled code calculation approach constitutes the normal evolution of analytical simulation methods applied forperforming safety analysis of NPPs. Until recent years most of the safety analyses, at least for PWR, have been madewith codes which model the neutronics only with point kinetics. For BWR, transient analyses have been carried outtraditionally with axially one-dimensional models since the coupling between neutronics and thermal-hydraulics isvery strong.

The need of coupled 3D neutronics calculation is largest in cases where strong feedback between the core kineticand the coolant loop as well as in situations where power excursion is important and its distribution changes during thetransient (Bousbia Salah, 2004).

Furthermore a reactor core is never uniform (even if it was initially constructed with uniform repartition of fuelassemblies) as a consequence of non-uniform consumption. The accuracy of the analyses can be improved signifi-cantly by modeling directly the interaction of the neutron kinetics and the fluid dynamics using the coupled codes’calculations. This is particularly true for the simulation of

� Almost all Reactivity Initiated Accidents (RIA) and especially1. If the reactivity increases are induced by thermal-hydraulic effects: In such cases strong interactions between

core kinetics and coolant loop thermal-hydraulics exist. The Turbine Trip tests are examples of such cases.2. If the perturbation is asymmetric in the core: Typical examples of a Design Basis Accident (DBA), which

cannot be well simulated without a three-dimensional core model, are the Main Steam Line Break(MSLB) and Control Rod Ejection (CRE) accidents. In this later case considerable deformation of the radialand axial power distributions occur in the core.

3. If there are possibilities for re-criticality in the later phase of the transient during the cool-down phase. This isparticularly emphasized for high values of moderator temperature coefficient, for increased high burnup fuel,or for extended use of MOX fuel.

4. The local boron dilution accident in PWR and VVER. The boron concentration is often non-uniformly trans-ported into and through the reactor core causing a reactivity transient that could be severe.

5. All Anticipated Transients Without Scram (ATWS) and other Beyond Design Basis Accidents (BDBA) needmore sophisticated calculations to eliminate any uncertainty due to spatial effects. In fact, using fewer dimen-sions in the core modeling did in some cases change the whole accident scenario.

� The BWR stability issues in plant conditions and beyond the stability threshold (D’Auria, 1997).� Nuclear Power improvement programs that generate the demand for reducing uncertainties.

4. Coupling process

The availability of a qualified system input deck and of qualified core model does not imply a qualified coupledmodel. There are certain requirements to the coupling of Thermal-Hydraulic System Codes and Neutron KineticsCodes that ought to be considered to provide accurate solutions in a reasonable amount of CPU time. More thanone possibility exists to perform this coupling, depending on the adopted computational tool.

Two different approaches are generally utilized to couple THSC with 3D NKC: serial integration coupling andparallel processing coupling.

1. Serial integration requires modifications of the codes; it is usually performed by implementing a neutronicssubroutine into the THSC. This could be performed through:� Internal coupling: The 3D neutron kinetics model is integrated into the core thermal-hydraulic model of the

system code. Each neutron kinetic node is coupled directly to a thermal-hydraulic node of the system code andis solved within each thermal-hydraulic iteration. This method requires a significant amount of information tobe exchanged between the two codes, but on other hand allows for detailed and direct system calculations. Onemajor disadvantage of this method is that it involves significant modifications in both codes.� External coupling: In this approach, the core calculations include 3D neutronic and the fluid-dynamic models,

while the system code is used only to model the thermal-hydraulics in the primary circuit excluding the core

4 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

region. This method facilitates the coupling procedure because it requires little or no modification of theTHSC or the NKC codes to be performed. However, numerical instabilities and slow convergence areobserved.� Parallel coupling: In such approach both codes are run simultaneously by exchanging information through the

boundary conditions at the core edges, and the core power. The main drawback of such coupling is that itrequires higher CPU time.

2. Parallel processing allows the codes to be run separately and exchange data during the calculation. The dataexchange is usually performed using Parallel Virtual Machine (PVM) environment. The NKC receivesthermal-hydraulic data from the THSC, such as fuel temperatures, coolant void fraction, coolant densitiesand temperatures, and boron concentration, and returns back the fuel power to the THSC. The main advantagesof this methodology are that only minor modifications are needed and the codes are isolated so they can inde-pendently be updated and maintained. However, data exchange between the two codes should be provided incarefully and properly selected time sequences, thus much attention must be paid to the process of data transferand associated time control of the execution processes of the two codes.

5. Uncertainty evaluation

BE codes, as any prediction tool, are far from being perfect; their solutions are approximate, and the degree ofapproximation constitutes the goal of uncertainty studies. For instance origins of uncertainties in BE tools aresummarized hereafter (Bousbia Salah, 2004).

� Numerical solution is approximate. The approximate balance equations for thermal-hydraulics conservation andthe kinetic diffusion equations are approximately solved by special numerical methods and/or simplifyinghypothesis. For instance in 3D NKC use generally limited number of neutron energy groups and the ADFflux corrections. The saving of computer memory and timing is a goal for such methods.� Correlations’ implementation and range of validity. Interaction between the phases and between each phase and

the walls are simulated by constitutive terms. Almost all of these correlations come from experiments and, assuch, are characterized by ranges of validity; uncertainties in code results originated by correlations are- Ranges of validity may not be fully specified (pressure, fluid and wall temperature, velocities, void fractions,

walls’ material properties, etc.).- The correlations may be unavoidably used outside their range of validity.- The correlations may be approximately implemented into the code due to the needs to fit with other

correlations, with the selected unknowns or with the numerical solution scheme.- Inherent scatter and error in the experimental results on which correlation is based.� State and material properties are approximate. These data are often provided in tabular format hence

approximations are, again, unavoidable and produce indefinite effects upon results.� Code user effect. Code users may interact at different levels with code results, involving the plant nodalization

phase, the choice of boundary conditions, the user expertise and the code guidelines.� Imperfect knowledge of Boundary or Initial Conditions (BIC). Boundary and initial conditions affect the

transient evolution in the way fixed by the code equations. The problem occurs when the BIC values areunknown or are known with some error.� Computer/compiler effects. A code installed in any computer machine should produce the same results provided

a unique input deck is adopted. This is not the case due to a number of reasons connected with the precision ofthe machine and with the compiler design (Trambauer, 1997).� Nodalization effects which tend to homogenize the complex systems.� Code/model deficiencies cannot be excluded. Such deficiencies may have an importance only in special transient

situations. However, they constitute an additional specific source of uncertainty.� Fuel assembly homogenization when using the nodal method to solve the kinetic equations.� The use ADF. This approach is introduced to mitigate errors in determination of neutron flow among assemblies.

However, this approach is somewhat not accurate under transient conditions, especially in areas where the fluxchanges rapidly.

5A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

These aforementioned concepts are applicable to any of the CSC, NKC and THSC. In nuclear power plants,thousands of parameters can affect the results of an analysis. However, only the variation of a few parameters canbe handled. In practice, good engineering judgment is the most important asset in choosing the appropriate parame-ters. The approaches pursued for uncertainty evaluation can be distinguished into two main categories, i.e. propagationof code-input and of code-output uncertainties which are connected to the GRS and University of Pisa CIAU method(D’Auria and Giannotti, 2000), respectively. The CIAU allows the achievement of continuous uncertainty bandssimultaneously with the BE calculation, thus avoiding the ‘uncertainty-methodology-user’ effect and the need ofresources for uncertainty prediction. However, a suitable experimental ‘‘error-database’’ must be made available topursue the CIAU approach. In a joint effort between University of Pisa and PSU the CIAU method has been recentlyextended to the evaluation of uncertainty from coupled 3D neutronics thermal-hydraulics calculations (Petruzzi et al.,2005). On the other hand, the GRS method is applicable to any computational tool, without modifications of the codes.The analyses are fully based on the statistical variation of model input parameters and the error margins could beobtained for any output parameter of the code. Attempts to apply this method to coupled codes were performed byLangenbuch et al. (2005) and Bousbia Salah et al. (2006).

6. Procedure to perform coupled code calculation

In order to perform adequate coupled 3D NKCeTHSC calculations capable to gain reliable results, a consolidatedprocedure should be used. The steps reported below outline the proposed procedure. However, four fundamental pre-conditions shall be fulfilled for the correct application of coupled codes’ simulation of transient scenarios expected inNPP:

- The codes should be frozen.- The codes should be properly qualified through wide set of assessment studies.- The developer of the nodalization should be a qualified code user for the selected codes.- The developed nodalization (of the plant) should be properly qualified.

6.1. Step 1: Input information and background

A comprehensive knowledge of the nuclear plant general features related to their design and operation should take intoaccount the plant geometry, operating conditions, material properties, subsystems having a role during the consideredtransients, including logic for actuation, and neutron kinetic parameters of the core, with known burnup and poison.The information should consider also specific data for the system object of the simulation, as well as the capabilitiesand limitations of used computational tools. These requirements constitute prerequisites for starting the analysis.

6.2. Step 2: Kinetic check

The adopted kinetic core model (to be considered for the analysis) should be carefully checked as well as the cross-section libraries. This is done through the so-called Hot Full Power (HFP) and Hot Zero Power (HZP) separate effectsimulations. These analyses are carried out using the NKC alone considering various core power and various controlrod configuration patterns. The kinetics response is checked through the feedbacks calculated by the code for fixedvalues of thermal-hydraulic parameters (which are the independent variables of the used lookup tables). The resultsof such calculations are generally compared with other calculations performed by some reference computational tools.

6.3. Step 3: Thermal hydraulic check

The adopted plant nodalization for the THSC is checked through a series of SS and TR runs. No kinetics calcula-tions are performed, in order to check the thermal-hydraulic system response. The core power is imposed as input. Thecalculations are stopped when almost all of the thermal-hydraulic parameters exhibit stable trend. The obtained valuesare afterwards compared with the plant available data and if necessary adjusted to fulfill the acceptability criteria ofthe nodalization at the steady state and transient level for THSC.

6 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

6.4. Step 4: Mapping process

In this phase a mapping scheme for fuel assemblies and thermal-hydraulic channels is selected. This could be oneby one (all the fuel elements of the core) or by an optimal mapping through a judicious user choice.

6.5. Step 5: Coupled code calculation

The availability of a qualified system input deck and of qualified core model does not imply a qualified coupledmodel. The coupling requirements should be fulfilled as discussed before. The run is restarted in Coupled Steady State(CSS) mode until a new stationary based upon the new core power, which is derived from the thermal-hydraulic andkinetic feedback balance is reached. A short transient is anticipated to be predicted by the code at the restart time,since, the shift from the imposed power to the calculated power causes a small adjustment of the plant condition(this happens because the neutronic flux shape imposed for the THSC initialization is typically a tentative shape,not known before, until the restart with the 3D NKC is run). The calculations are performed to match as close as pos-sible keff¼ 1. The degree of deviation from the unity depends on the calculation capabilities and on the number of usedneutron energy groups. If the value of keff is far from unity the THSC as well as the NKC inputs are checked againespecially the calculated steady states’ operating conditions and the adopted cross-section lookup tables for theTHSC and NCK codes’ calculations, respectively. After that, the transient calculations are performed (restartedfrom the previous steady state mode) and the analysis is made.

6.6. Step 6: Uncertainties’ evaluation

Code calculations are affected by uncertainties arising from several causes as discussed before. These can be char-acterized by hundreds of parameters that are typically part of the input deck. This occurs notwithstanding the highcode performance level and the systematic qualification processes. Thus, to be valuable, results obtained by the cou-pled code calculations should be associated with known uncertainties’ bands capable to give a quantitative range of thesimulation errors.

7. Typical results

In order to emphasize the capabilities and also the limitations of the coupled code technique some results per-formed for typical NPP using coupled 3D neutron kinetics thermal-hydraulic calculations are presented.

7.1. PWR case

The transient is modeled using RELAP5-3D�/NESTLE. The MSLB (Ivanov et al., 1999) is supposed to beoriginated by a double-ended guillotine break of one SL. The fast depressurization of one SG causes the pri-mary water cooling, as soon as a plug of cold water reaches the core inlet, a positive insertion of reactivity dueto the moderator neutronic feedback produces core power release. The initial power excursion is terminated byscram at about 10 s after accident initiation. However, owing to one stuck withdrawn CR in a critical positionof the core, return to power occurs at about 60 s. This is not a re-criticality accident and is phenomenologicallycontrolled by delayed neutron groups. The second power peak damps down without any active system interven-tion when number of generated neutrons physically decays to zero. In the DC and LP regions a partial mixingbetween ‘‘cold’’ water coming from the loop affected by the broken SG and the ‘‘hot’’ water of the intact looptakes place. A non-uniform distribution of moderator temperature in the core takes place and localized positivereactivity insertion occurs leading to higher relative power peak. Figs. 1 and 2 show results for core power evo-lution during the transient.

7A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

0

5

10

15

20

25V

3

510

15

V1

510

15

V2

XY

3.73.43.12.82.52.21.91.61.310.70.4

Fig. 2. PWR e relative power distribution at the second power peak time occurrence.

Fig. 1. PWR e core power evolution.

8 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

7.2. BWR NPP

7.2.1. Fast reactivity insertionIn this example the Peach Bottom (GE) is taken to represent the BWR NPPs. For this purpose an asymmetric

transient related to single peripheral Control Rod Ejection is considered. Such case will emphasize the capabilitiesof coupled technique under regional and local effects’ transients.

In this case a single peripheral control rod (shown in Fig. 3), initially completely inserted, is withdrawn into a periodof 0.1 s. The transient was modeled using adequate mapping of the core channels. The obtained results could besummarized by Fig. 4, which emphasizes the asymmetric effect involved during such transient. The power rise isstopped by control rod drop.

7.2.2. BWR stability issueThe purpose of this case is to emphasize the coupled code technique for simulating BWR core behavior under

instabilities’ conditions. According to the Core Power to Flow Map (D’Auria, 1997), a BWR plant may operate ininstability conditions approximately at 30% of nominal flow and 50% of nominal power. For this purpose, the coupledcode RELAP5/PARCS is considered (Joo et al., 1998; Ransom et al., 1990). The assessment of the technique has beenperformed against Peach Bottom-2 Low-Flow Stability Test number 3 PT3 (Carmichael and Niemi, 1978). Theexperimental test was performed by disturbing the reactor power with two peaks’ pressure of 0.055 MPa. After theperturbation, the pressure oscillation decreases rapidly and, in few seconds, turns to the initial value. In the meantime,the pressure perturbation propagates through the steam lines at sonic velocity and reaches the core zone from thesteam separator and from the downcomer. Consequently, as reported in Figs. 5 and 6, the core power experiencesthe same trend due mainly to its inherent void feedback reactivity (Bousbia-Salah and D’Auria, 2006). Some in-phaseinstability characteristics could be recognized for instance, frequencies in all the oscillations obtained in the analysesvaried from 0.3 to 0.5 Hz, i.e., in the typical frequency range of this kind of instability events.

Fig. 3. BWR e location of the ejected Control Rod.

9A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

Fig. 5. BWR e average radial oscillating power distribution.

Fig. 4. Radial power distribution during the BWR CR withdrawal transient.

10 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

7.3. VVER1000 case

The Kozloduy pump trip transient is selected to represent the VVER1000 reactors. During pump trip transientsstrong interactions between core kinetics and coolant loop thermal-hydraulic characteristics are involved. In factduring the pump start-up the cooling flow raises involving asymmetric downcomer coolant temperature distribution.The global effect on the core is a reduction of the moderator temperature and consequently the core power risesexponentially (Bousbia Salah et al., 2006). At earlier transient asymmetric core power distribution is estimated asshown in Fig. 7. Later, a quasi-equilibrium state between the moderator and Doppler effects is reached and thecore power exhibits a self-limiting behavior as shown in Fig. 8.

8. Conclusions

A noticeable progress in the capabilities of computer codes has been observed in the past decades. From the designand safety engineering point of view, coupled thermal-hydraulic system 3D Neutron Kinetics Codes are considered tohave reached an acceptable level of maturity. However, due to unavoidable approximations in the code models, anycalculation from a Best Estimate code, to be meaningful, needs an uncertainty evaluation. This is valid for variousapplications of the codes ranging from licensing studies to training of plant operators.

Through the current study, a typical application of the coupled code technique and a procedure for correct use isalso outlined. The main findings could be summarized as follows:

� Detailed requirements for user qualification should be developed.� The two energy group diffusion equations’ accuracy is quite good for common transients. However, better

solutions should be obtained with more sophisticated techniques, including Monte Carlo and detailed neutrontransport or multigroup diffusion equations and multidimensional cross-section tables to get more realisticflux distribution.� Constitutive models used to determine the evolution of the two-phase mixture, being mostly developed under

steady state conditions, should be made more adapted to transient situation, especially those connected withthe feedback between thermal-hydraulic and kinetic.� 3-D nodalizations for the core or the vessel regions as needed for Best Estimate simulation of some phenomena

as pressure wave propagation and flow redistribution in the core lower plenum zone.� A general-purpose system code should be developed essentially including multifluid capability and ‘‘open’’

interfaces for an easy coupling with other codes in areas like neutronics (for implementing presently available3-D codes), CFD, structural mechanics (e.g. for pressurized thermal shock studies), and containment.

1.40

1.60

1.80

2.00

2.20

2.40

2.60

45 50 55 60 65 70

Time (s)

Po

wer (G

W)

Fig. 6. Reactor core power response.

11A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

8.20E+08

8.30E+08

8.40E+08

8.50E+08

8.60E+08

8.70E+08

8.80E+08

0 20 40 60 80 100 120 140

TIME (s)

PO

WE

R (W

)

Fig. 8. VVER1000 Transient core power history.

Fig. 7. VVER1000 2D mean relative core power distribution map.

12 A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

� The importance and the need of uncertainty evaluations for coupled codes’ predictions are imperative owing toa large number of reasons discussed in this work. Therefore, uncertainty must be connected with any prediction.� Current computer capabilities should be properly used.� Finally, the industry and the regulatory bodies should become fully aware about the capabilities and the limita-

tions of the coupled code techniques.

References

Aksan, S.N., D’Auria, F., Staedtke, H., 1993. User effects on the thermal-hydraulic transient system codes calculations. Nuclear Engineering and

Design 145, 1e2.

Bousbia Salah, A., 2004. Overview of coupled system thermal-hydraulic-3D neutron kinetics code applications. Ph.D Thesis (Ref: 17033),

Universita di Pisa.

Bousbia Salah, A., Kliem, S., Rohde, U., D’Auria, F., Petruzzi, A., 2006. Uncertainty and sensitivity analyses of the Kozloduy pump trip test using

coupled thermal-hydraulic 3D kinetics code. Nuclear Engineering and Design 236, 1240e1255.

Bousbia-Salah, A., D’Auria, F., 2006. Peach Bottom Cycle 2 low flow stability tests analysis using RELAP5/PARCS. Annals of Nuclear Energy

33, 646e652.

Bousbia Salah, A., Vedovi, J., D’Auria, F., Galassi, G.M., Ivanov, K., 2006. Analysis of the OECD/DOE/CEA VVER1000 CT-1 benchmark using

the coupled RELAP5/PARCS codes. Progress in Nuclear Energy 48 (8), 806e819.

Carmichael, L.A., Niemi, R.O., 1978. Transient and stability tests at Peach Bottom Atomic Power Station Unit 2 at end of cycle 2 (EPRI NP-564).

D’Auria, F., 1997. State of the art report on boiling water reactor stability (OECD/NEA/CSNI OECD/GD (97) 13).

D’Auria, F., Galassi, G.M., 1998. Code validation and uncertainties in system thermal-hydraulics. Progress in Nuclear Energy 33, 175.

D’Auria, F., Giannotti, W., 2000. Development of code with capability of internal assessment of uncertainty. Nuclear Technology 131 (1), 159.

D’Auria, F., (Lead Author) (2004). Neutronics/thermal-hydraulics coupling in LWR technology: state-of-the-art report (REAC-SOAR).

CRISSUE-SeWP2. (OCDE/NEA-5436).

Ivanov, K., Beam, T., Baratta, A., Irani, A., Trikouros, N., 1999. PWR MSLB Benchmark. Volume 1: Final Specifications (NEA/NSC/DOC(99)8).

Ivanov, B., Ivanov, K., 2002. VVER-1000 Coolant Transient Benchmark: Phase-1 (V1000CT-1). Volume I: Final Specifications (OECD Paris,

NEA/NSC/DOC).

Joo, H.G., Barber, D., Jiang, G., Downar, T.J., 1998. PARCS, a multi-dimensional two group reactor kinetics code based on the nonlinear analytic

nodal method (PU/NE-98-26).

Langenbuch, S., Krzykacz-Hausmann, B., Schmidt, K.D., Hegyi, G., Kereszturi, A., Kliem, S., Hadek, J., Danilin, S., Nikonov, S., Kuchin, A.,

Khalimanchuk, V., Hamalainen, A., 2005. Comprehensive uncertainty and sensitivity analysis for coupled code calculations of VVER plant

transients. Nuclear Engineering and Design 235, 521e540.

Petruzzi, A., D’Auria, F., Giannotti, W., Ivanov, K., 2005. Methodology of internal assessment of uncertainty and extension to neutron kinetics/

thermal-hydraulics coupled codes. Nuclear Science and Engineering 149 (2), 211e236.

Ransom, V.H., Wagner, R.J., Trapp, J.A., Johnsen, G.W., Miller, C.S., Kiser, D.M., Riemke, R.A., 1990. RELAP5/MOD3: Code manual: user

guide and input requirements (NUREG/CR-4312 EGG-2396).

Solis, J., Ivanov, K., Sarikaya, B., Olson, A., Hunt, K.W., 2001. Boiling water reactor Turbine Trip (TT) benchmark. Volume 1: final specifications

(NEA/NSC/DOC).

Trambauer, K., 1997. Computer and compiler effects on code results report. OECD/CSNI Report (OCDE/GD(97) 6).

Glossary

ATWS: Anticipated Transient Without Scram;

BDBA: Beyond Design Basis Accident;

BE: Best Estimate;

BEPU: Best Estimate Plus Uncertainty;

BWR: Boiling Water Reactor;

CFD: Computational Fluid Dynamics;

CIAU: Code with capability of Internal Assessment of Uncertainty;

CPU: Central Process Unit;

CSC: Cross Section Code;

CSNI: Committee on the Safety of Nuclear Installations;

DBA: Design Basis Accident;

EC: European Commission;

GRS: Gesellschaft fuer Anlagen- und Reaktorsicherheit;

IAEA: International Atomic Energy Agency;

ITF: Integral Test Facility;

13A. Bousbia-Salah, F. D’Auria / Progress in Nuclear Energy 49 (2007) 1e13

LOCA: Loss of Coolant Accident;

LBLOCA: Large Break Loss of Coolant Accident;

LWR: Light Water Reactor;

MOX: Mixed UePu oxide nuclear fuel;

NEA: Nuclear Energy Agency;

NKC: Neutron Kinetic Code;

NPP: Nuclear Power Plant;

OECD: Organization for Economic Co-operation and Development;

PWR: Pressurized Water Reactor;

RIA: Reactivity Initiated (or Induced) Accident;

TH: Thermal-Hydraulic;

THSC: Thermal-Hydraulic System Code;

VVER: Water-cooled Water-moderated Energy Reactor;

3D or 3-D: Three-dimensional.