R2CA 4TH NEWSLETTER

Last news from R2CA

Most of the last year was devoted to updating accident sequence calculations using improved calculation chains and simulation tools (benefiting of the implementation of some refined models developed during the project itself), and to assessing the gains obtained in terms of radiological consequences by comparing these latest calculation results with the initial calculation results. As expected, due to the reduction of some of the conservatisms used in the initial calculations for most of the re-assessed reactor accidental scenarios the radiological impact was reduced.

The synthesis of all this work, currently underway, is feeding the formulation of recommendations for the harmonization of radiological consequences evaluation methodologies also in progress.

A great part of work has also been carried out in the work-packages dedicated to innovation, with a particular emphasis on optimizing accident management procedures in SGTRs and finalizing the development of a prototype expert tool based on neural networks, dedicated to the early diagnosis of defective fuel rods to improve their management. All this work has been and continues to be carried out under the best possible conditions, thanks to the 4-month extension granted to the project by the European Commission.

Finally, this last period was also rich in terms of communication and knowledge dissemination, notably with a well-attended summer school on DBA & DEC-A for LWRs held in July 2023 in Bologna and the preparation of 18 dedicated open papers to be published on-line in a R2CA special issue of Annals of Nuclear Energy.

To close the project, its most important outcomes and main advances will be presented and discussed during an open workshop organized at the end of November in IRSN headquarters in Paris.

Nathalie Girault (IRSN)

TASK2.5: Main outcomes of second set of reactor calculations

The main outcomes from the first set of reactor calculations were used to identify the needs in terms of code/model improvements, to reduce some of the conservatisms in modelling assumptions and to upgrade calculation chains by including for example more detailed (mechanistic) computer codes. The results of the first set of the simulations served as a reference point for a second set of simulations of the same LOCA and SGTR transients performed after the improvements and was used to quantify the gains in terms of radiological consequences (RC) of the updated calculation methodologies.

The improvements made were of two kinds:

Modelling improvements. For LOCA transient calculations, partners were improved thermomechanical models for clad ballooning and burst, new clad burst criteria were built. Also core nodalization was improved (made in more details or reorganized based on best practice and results of parametric analyses). For SGTR transient calculations, partners dedicated improvements on fission product modelling: initial primary contamination and FP transient spiking releases, dilution in RCS, transport, scrubbing, partitioning, atomisation, speciation etc. Partners also improved their thermohydraulic model using refined model for the relief/safety valves of the steam generators (SG) or optimizing the Emergency Operating Procedures (EOPs).
Improvements in the calculation chains. Detailed (mechanistic) computer codes were also used in partners calculations, often as a support of less detailed codes. For LOCA calculations for instance, detailed fuel performance codes were used to reflect fuel thermal and thermomechanical processes. Also detailed codes, coupling thermal hydraulics and thermomechanics models have been used to better predict clad ballooning and burst at subchannel level. For SGTR calculations, more detailed codes were also used for modelling spiking FP releases from defective fuel rods at transient onset, primary FP transport and chemistry and phenomena occurring at water/steam jet location (flashing…). Some partners used different modelling approaches from those used in the first calculation set.

In total 34 accidental scenarios (both LOCA and SGTR) were calculated in the second set of calculations on different kinds of reactor designs (VVERs, PWRs, EPR and BWR), covering both DBA and DEC-A conditions.

Mentioned improvements mainly concerned, for LOCA, thermal mechanical phenomena in fuel rods (LOCA) and, for SGTR, FP releases, transport and behaviour in primary circuit and in the failed steam generator. Thus, in the second calculation set, there was no significant change in thermo-hydraulics results, for most partners. The major changes were made in clad thermomechanics and in FP releases and behaviour.

For LOCA scenarios, most of the partners achieved significant lower fuel rod burst percentage after provided improvements. This directly involved a significant reduction of the primary contamination and FP activity transport to containment and environment. Thus, when very conservative assumptions were used in the first calculation set, a reduction of almost 99% the environmental activity has been observed, while this reduction didn’t exceed 22% when more realistic assumptions and/or more detailed simulation tools have been used. - Also, the results of second calculation set were found to be less scattered, especially for PWR type reactors, where the environmental activities differed by one order of magnitude compared to two in the first calculation set.

For SGTR scenarios as explained earlier, partners improved their simulation scheme mainly on the FP inventory and transport aspects. Therefore, except for two partners who also improved their thermohydraulic part as well, there is no change on the calculated cumulative steam/liquid water released in environment. Thanks to more realistic models for FP transport and interaction in the RCS and SG, partners achieved a reduction of the activity releases in the environment between 17 to 97 %. Nonetheless, two partners calculated higher activity releases with the second set of calculations.

Radiological consequences were evaluated for the first and second sets of calculations. The results are consistent with what is observed with the activity releases to the environment: radiological consequences are reduced in the second set of calculations compared to the first one, except for two SGTR transients.

The calculated doses for all analyzed transients (both LOCA and SGTR) and conditions (DBA and DEC-A) always stayed under acceptance criteria.

T. Kaliatka (LEI), P. Foucaud (TE), N. Girault (IRSN)

TASK 2.6:Towards harmonization of the RC evaluation methodologies

Several “recommendations” were expressed in order to sustain the evolution of the methodologies through more realistic (less conservative) results in terms of RC of LOCA and SGTR. These recommendations are sorted considering each barrier based on the structure of the initial review of the existing RC evaluation methodologies performed within the project. The choice was made to promote recommendations based on their impact on quantitative results in terms of RC.

Without being exhaustive, here is an overview of the recommendations:

LOCA. Fuel source term: regarding fuel rod burst failure, using a 3D core model appears to be crucial for asymmetric transient like LOCA; refining the evaluation of the FP’s inside the gap inventory appears to be important; coupling Thermohydraulics and thermo-mechanics; expressing specific rods initial conditions (internal pressure, burnup…); using more realistic criteria of cladding bust as developed during the project. Transfer from containment to environment: as important information, noble gasses represent at the same time the most dominant contributor and their behavior is easier to treat than other FP’s; valuable to evaluate the time dependent speciation of iodine in the containment in order to quantify their amount in each phase; selecting adapted hypotheses about containment venting and filtration is important as well as considering FP retention on its surfaces.
SGTR. Fuel source term: regarding gap release, the project illustrated the importance of the initial RCS contamination and spiking model, a large choice of best-fitted spiking correlations exist from the most simple to the most advanced, supported by codes or not. Transfer to and in affected SG: using more realistic value for the part of break flow which is transformed into aerosols based on test facility results for uncovered break scenarios; valuable to evaluate iodine speciation in the primary to better estimate iodine flashing and partitioning. Transfer to environment: distinguishing phases with their own activities in the secondary appears valuable in order to not directly release flashed part of the break flow.

Parmentier Francois (BELV)

WP5 (Artificial Neural networks for detection of defective fuel rods).

During normal reactor operations, the fuel rod cladding acts as a barrier to prevent the fission products (FP) generated within the nuclear fuel to be released to the reactor primary circuit. Some part of fission gases and volatile FP may be released from the fuel matrix, but they are retained within the fuel-cladding gap as long as the cladding layer is intact. The cladding may, however, be damaged during reactor operations due to the influence of various phenomena, such as stress corrosion cracking (SCC), grid-to-rod fretting (GTRF), debris induced failure, among others. Subsequently, the high-pressure coolant water enters the fuel-to-clad gap through the defect site, causing the FP retained within the gap to escape into the primary coolant, and oxidation of the fuel which may further enhance FP release. The released radionuclides can interact with the coolant and may cause serious corrosion problems to the fuel cladding or other reactor structures. Further deterioration of the defective fuel elements can also occur with continued operation due to secondary hydriding of the Zircaloy cladding. Such deterioration and activity release may affect the power plant economically, such as from lost burnup due to early discharge of the fuel at power. The detection of defective fuel rod is thus of primary importance. Different approaches have been adopted for fuel failure detection: (i) using the operating limit of the specific activity in the coolant (ii) using the release to birth ratio (R/B) slope versus decay constant in a log–log figure; (iii) using the fitted escape rate coefficient, which directly represents the degree of fuel failure. Due to drawbacks of these approaches, data-driven methods such as Artificial Neural Networks (ANN) have gained momentum in the recent years. In the framework of the R2CA project, a physical model for FP release from defective fuel rod has been developed at IRSN and used in conjunction with an artificial neural network for diagnosis and characterization of defective fuel rods.

The physical model considers the release from fuel pellet by diffusion and recoil, a generalized diffusion and first-order kinetic model for the FP in the gap region, and a first-order kinetic model in the coolant region, as illustrated below.

The physical model was run for a large set of input parameters to generate a computation database, which was then used to train and test the ANN. This ANN was designed to be a meta-model of the physical model, predicting isotopes activities in the coolant from the inputs of the physical model. The results obtained with this ANN on never-seen-before data, represented below on Prediction (ANN) vs Reality (Physical model) curves for normalized activity of 4 isotopes of interest, are satisfactory. The normalized activity values are regrouped in two sets which represent situations with and without defect, resulting in high and low activity, respectively. Based on these encouraging results, further work is being carried out to design a new ANN for predicting the occurrence of a defect from the coolant activity levels, and/or the time elapsed since the defect occurrence, hence providing interesting tools for the detection of defective rod during reactor operations.

KREMER Francois (IRSN)

COMUNICATION AND DISSEMINATION

R2CA OPEN WORKSHOP

Final Open Workshop of the R2CA project discussed the main results and outcomes of the project dedicated to Design Basis and Design Extension of Loss-Of-Coolant and Steam Generator Tube Rupture Accidents, Accident Management and Procedures, Innovative Tools and Devices.

Main topics addressed:

Fuel/clad thermomechanics
Fission products behaviour in fuel
Fission product transport from primary to secondary circuit and behaviour Accident Tolerant Fuels
Accident prevention & management procedures
Methodologies for radiological consequence assessments

R2CA OPEN WORSHOP:

Nathalie Girault (IRSN), Fulvio Mascari (ENEA)

R2CA SUMMER SCHOOL

A Short Summer School has been organized by ENEEA and IRSN in Bologna from the 4 to the 6 of July, 2023.

The main target of the summer school is to disseminate the knowledge consolidated and gained along the R2CA project to Masters and PhD students, young researchers and engineers involved in nuclear energy and reactor safety analyses. Along the school the main safety aspects related to DBA and DEC-A of LOCA and SGTR accidents will has discussed focusing the attention on the phenomenology, experimental knowledge available and current numerical modeling. Main advancements within the R2CA project will serve as a background to show the current state of art and the new ideas. The school targetted both fundamental knowledge, current nuclear safety best practices and innovation.

The agenda is here reported:

BROCHURE OF R2CA SUMMER SCHOOL:

AGENDA R2CA SUMMER SCHOOL

In addition, a panel of topics of interest for the future of nuclear safety research whas been presented.

The project briefly presented at the school have been:

CONFIDENCE

McSafer project

MUSA

SASPAM-SA Horizon Euratom Project

OPERAHPC

APIS

To have more infomation about the R2CA SUMMER SCHOOL we suggest the following posts:

FIRST DAY:

SECOND DAY

THIRD DAY

A video that summarize all the event can be find at:

R2CA SUMMER SCHOOL VIDEO

Fulvio Mascari (ENEA)

TRANSURANUS * SCIANTIX / MFPR-F TRAINING

On June 26-30, 2023, we have organized another in person training on TRANSURANUS, but this time coupled with the new SCIANTIX 2.0 version from POLIMI and the MFPR-F code from IRSN, both developed within the R2CA project. The training course took again place in Karlsruhe for 8 participants. The training provided an introduction on the three codes, focusing on the SCIANTIX capabilities and its functioning as a fission gas behaviour module within the TRANSURANUS fuel performance code, complemented with an online introduction about the fission product chemistry capabilities of the MFPR-F code. The participants had the opportunity to employ the coupled code suite TRANSURANUS-SCIANTIX to analyse representative LWR fuel rods and investigate the code calculations in terms of integral fission gas release provided by SCIANTIX.

Paul Van Uffelen (JRC), Lelio Luzzi (POLIMI), Francois Kremer (IRSN)

R2CA SPECIAL ISSUE:

A dedicated special issue on the Annals of Nuclear Energy Journal has been organized in order to collect all the outcomes of R2CA project and enhance the visibility of the activity maximizing the impact of the project. The special issue is under finalization. Guest editor: nathalie girault , Fulvio Mascari , Lelio Luzzi

Fulvio Mascari (ENEA), Nathalie Girault (IRSN), Lelio Luzzi (POLIMI)

PREVIOUS TRAININGS:

DRACCAR TRAINING

TRANSURANUS TRAINING

On January 17-21, 2022, we held the TRANSURANUS training course in Karlsruhe in person, following the start of the COVID-19 crisis. The training course was organised and hosted by the JRC for 12 young academic staff from technical safety organisations, research centres, universities and industry that include 3 organisations from R2CA. The TRANSURANUS course gives a theoretical basis on nuclear fuel behaviour in the nuclear reactor. It shows how to prepare and use the TRANSURANUS code for various fuel performance analyses. During the course, the TRANSURANUS is used for analysing thermal and mechanical behaviour of the nuclear fuel in the reactor and for analysing behaviour of gaseous fission products. Furthermore, the trainees learned to implement new model parameters in the source code and then create a new executable version and to verify compliance with safety and design criteria provided by the IAEA.

Paul Van Uffelen (JRC)

SCIANTIX TRAINING

On October 16, 2020, we held the first online SCIANTIX Training Course. SCIANTIX is an opensource code devoted to the simulation of inert gas behaviour within nuclear fuel, designed for inclusion in fuel performance codes. In the frame of R2CA, SCIANTIX is being extended to also model the production and transport of fission products within the fuel pellet.

The training included a general introduction to physics-based modelling of inert gas behaviour and proposed hands-on case studies for the participants to directly use SCIANTIX. The 30+ participants to the training came from both institutions within and outside the consortium of R2CA. The material used in the Training (slides, case studies with related documentation) is publicly available, together with the source code of the SCIANTIX version used. The recording of the Training is also available, divided in six videos covering all the topics presented. We take the occasion to thank all the participants to this first online training and look forward to organizing other activities!