The activities of NUBIKI serve to ensure safe operation of nuclear power plants. NUBIKI staff members perform deterministic, as well as probabilistic analyses to improve the safety level of nuclear power plants.
In the frame of deterministic safety analyses, NUBIKI performs computational simulations to analyse different event sequences which can occur in a nuclear power plant. One of the fields of our activities is the analysis of the containment response in case of design basis accidents using computer codes. The computed results are validated with post-test calculations of experiments. Important condition for the safe operation of a plant is the containment leak rate analysis. Our traditional field is the compilation and evaluation of the measured data of numerous leak test measurements performed since the commissioning of the power plant.
Thermal-hydraulic processes and activity release during severe accident sequences has been analysed applying computer models. In this field analyses in the AGNES project dealing with the safety analysis of the Paks Nuclear Power Plant, and during the project for the Level 2 probabilistic safety assessment of the plant have been done. Important field of the severe accident analyses is the evaluation of the hydrogen distribution and management in the containment. For this purpose CFD tools are used. Our computer simulations – among others – support the development of safety management guidelines of the plant. NUBIKI participates in several international experimental projects with post-test calculation of experiments. The objective of this work is the verification and validation of computational models.
probabilistic safety analyses to identify event sequences that lead to a severe accident and to determine the likelihood of such sequences. We develop detailed event sequence models for a wide range of accident scenarios. The responses of plant systems, system components and the interactions of plant personnel are all subject to in-depth analysis. We apply state-of-the-art analysis methods and computerised tools for model development and quantification. Over and above a quantitative description of nuclear safety, use can be made of our analysis results in support of risk-informed decision-making. The nuclear safety authorities as well as the utilities can take advantage of this decision-making process. We devote substantial efforts to developing appropriate methods and tools for risk-informed applications.
Nowadays we make uses of our experience and expertise from more than two decades of probabilistic safety assessment in providing support to plant time extension activities, design and implementation of plant changes, making decisions related to plant operation and maintenance and preparation for new nuclear builds in Hungary.
We regularly apply deterministic and probabilistic analyses in combination as complementary methods.
Deterministic safety analyses
The team has a long-term history in performing DBA calculations for the VVER-440 containment. The applications included the update of DBA analyses for the Final Safety Analysis Report (FSAR) of Paks NPP. Recently uncertainty analyses were performed to explore the margins of the containment parameters under DBA conditions.
Severe accident analysis
The behaviour of the VVER-440 plant under severe accident conditions was studied with code calculations in the AGNES project, and later in several PHARE projects.
Further studies involved accident progression and fission product transport analyses in the Level 2 PSA project for Paks NPP. Level 2 analyses also served as a basis for the development for a severe accident management strategy for the VVER-440 plant.
Severe accident management
Verification analyses of the severe accident management guidelines (SAMG) with code calculations were performed to support the development of the guidelines and to check the time frames available for the interventions. Hydrogen distribution in the containment and hydrogen management with passive catalytic recombiners was analysed with 3D containment calculations.
Probabilistic safety analyses (PSA)
We apply level 1 probabilistic safety assessment to analyse and evaluate accidents that may result in severe damage in the reactor core or in the fuel assemblies stored in close vicinity to the reactor. We have completed such analyses for a wide range of potential initiating events and operational states of the four Paks nuclear power plant units. We have studied internal events and failures (e.g. pipe ruptures, system malfunctions), internal hazards (on-site fires, internal flooding) as well as external events (earthquakes) as initiating events that divert the plant from normal operationally conditions. As to plant operational states, our analyses cover plant operation at full power and low power and shutdown states of the reactor too. The scope of our PSA for fuel damage accidents in the spent fuel storage pool within the reactor hall of the Paks plant is comparable to that of the reactor PSA. We apply internationally recognised computerised tools in support of model development, input data compilation and analysis, and risk quantification. Moreover, we also make use of specific analysis tools that are the result of our in-house developments.
We develop level 1 PSA further up to level 2 probabilistic safety assessment. We determine the potential large releases of radioactivity to the environment and the likelihood of such consequences by taking into account the severe accident processes within the containment. Similarly to the level 1 analyses, the level 2 PSA for the Paks nuclear power plant include a broad spectrum of potential severe accidents. The Risk Assessment Division and the Safety Analyses Division of our institute work in close co-operation to perform level 2 PSA.
We regularly review and update the existing probabilistic safety assessment studies for the Paks nuclear power plant including PSA model, reliability data, analysis results and documentation. Usually an update is made every year to ensure a credible picture about the safety level of the plant. During updates we give considerations to safety related plant changes, new experiences from plant operation and maintenance, extensions to analysis scope, and application of advanced analysis methods and new knowledge.
In our efforts towards method development we put much distinguished emphasis on the establishment of models and approaches that enable a probabilistic description of interactions made by the plant personnel. We have performed numerous experimental studies, observations at the full scope, replica training simulator of the Paks plant. We collected data on the characteristics of control room crew operations in emergency during the observations. Simulator data were subsequently used in combination with expert opinion in direct support to method development. We have made uses of these methods in human reliability analysis as part of the plant PSA.
From among the analysis tools we developed the so-called ADRIA computer code should be highlighted. ADRIA is a combined database, analysis as well as design tool we constructed originally for the purposes of fire and flood PSA. ADRIA includes a detailed inventory of plant components including cables and cable routing in particular. In addition to its uses in risk assessment, it has also important functions in designing cable routes and providing support to operational tasks and decisions related to cabling. In order to enable lifetime extension of the Paks plant ADRIA is applied in the ageing management program for cables. Also, it can be used to underpin decommissioning activities when the plant has come to the end of its lifetime.
Over and above the fulfilment of certain nuclear safety requirements, probabilistic safety assessment is utilized in actual PSA-applications within a risk-informed decision-making framework. Activities of both the nuclear safety authority and the licencsee can be supported by these applications.
If justified by safety analysis results, we develop proposals for plant changes to increase safety by removing vulnerabilities to accidents. We determine the expected improvement in plant safety if such changes are made. We have performed risk assessment in the design and in the implementation phases of several safety related modifications made to the Paks nuclear power plant. These assessments were of high importance during a large-scale safety improvement program that was implemented between 1996 and 2002 at Paks.
We developed a proposal for the Hungarian Atomic Energy Authority to develop and implement risk-based safety indicators that could be used to monitor and predict the safety performance of a nuclear power plant. As one element of the proposed indicator system we have adapted analysis and evaluation method used in the US to determine the plant level risk impact of events occurred at nuclear power plants. In order to make such analyses simpler we have created a computer programme that uses the level 1 PSA model of the Paks nuclear power plant. We make use of this programme and the adapted analysis method during the analysis of events reportable to the nuclear safety authority by the Paks plant. We determine the conditional core damage risk for each and every licensee event as a regular support to the authority.
We have developed a computer code that is capable of calculating core damage risk and displaying risk graphically in accordance with the actual configuration of safety related systems and components in a nuclear power plant. This programme is called Risk Supervisor and it is intended for regulatory use. Similarly to PSA-based event analysis, we analyse all the reported licensee events by the Risk Supervisor too.
We also develop a risk monitor for the four units of the Paks nuclear power plant. This risk monitor builds upon the RiskWatcher risk monitor software of Scandpower AB and the unit specific PSA models. It can calculate and evaluate configuration specific risk in a more detailed manner than the Risk Supervisor and it can also take into account more factors that could influence risk. It is envisaged that the risk monitor will be the basis of risk-informed decision-making in relation to a number of operational and maintenance tasks.
We have constructed a computerised tool that can be used to determine the likelihood of a severe accident during plant emergency. This tool is installed at the Centre for Emergency Response and Training Awareness of the Hungarian Atomic Energy Authority. If an event calls for the emergency operation of the centre, then this computer program can calculate the conditional probability for the development of a core damage accident. The precision of the prediction is driven by the scope and level of detail of the information available about the actual plant event.
We are engaged in method development to enable the use of risk information (derived from probabilistic safety assessment) besides the traditional, deterministic safety classification process for the systems and components of a nuclear power plant. At an early stage of thus development we determined the safety significance of systems and components modelled in the plant PSA. We seek for combined, mutually complementary uses of deterministic principles and quantitative risk information in a common framework. The key objective of this effort is to enable the uses of safety requirements that differentiate in accordance with an improved, more risk-informed safety classification.
We have reviewed and evaluated the potential areas of risk-informed decision-making that can be supported by probabilistic safety assessment. We have produced a series of technical reports in this subject within our consultants’ services to the Paks Nuclear Power Plant Ltd. and, also, as a technical support organisation to the Hungarian Atomic Energy Authority. In these reports we identified the potential applications the underlying objectives, as well as the technical and administrative conditions for their introduction. Use is to be made of these reports in the current co-ordinated efforts of the authority and the licensee towards risk-informed decision making in Hungary.
We have developed methods for risk-based review and modification of the surveillance test intervals and the allowed outage times of safety related systems and components laid down in the Technical Specifications of the Paks Nuclear Power Plant Ltd. We have made a number of trial applications to demonstrate the use of our proposed method for revising allowed outage times. As to surveillance test intervals, we have programmed our methods in a computer code and we have applied it the full range plant systems modelled in the PSA for the Paks plant.
We have studied and evaluated to what extent the training scenarios included in the continuing training programme for the control room crews of the Paks nuclear power plant cover the accident sequences found dominant in the plant PSA.
We utilise the methodology of probabilistic safety assessment and also the experience from such studies indirectly in support of strategic tasks aimed at ensuring long-term safe operation of the Paks nuclear power plant.
The summary description of probabilistic safety assessment is part of the Final Safety Analysis Report which is the most important document that demonstrates the safe operation of the Paks nuclear power plant by describing and evaluating the compliance with nuclear safety requirements. Moreover, we have performed fault tree based reliability analysis for a range of plant systems in order to enable an evaluation of the numerous nuclear safety requirements related to system reliability. Theses system level analyses form the basis of safety assessments presented separately for the different systems within the Final Safety Analysis Report for the Paks plant.
We have elaborated a guide on the methodology to monitor the effectiveness of maintenance by regularly keeping track of and evaluating performance indicators for safety related plant systems at the Paks nuclear power plant. We have identified the performance indicators for all the plant systems that are subject to monitoring and we have determined the acceptance levels of these indicators based on fault tree analysis. Also, we provide support to the plant in implementing the proposed monitoring programme which is a necessary condition for lifetime extension at Paks.
We utilise our PSA knowledge and skills to contribute to the ongoing scientific and technical preparatory work aimed at extending the capacity of nuclear power at Paks by building new reactor units. Further details on this subject can be found under References within this website.
International co-operation plays an important role in our analyses and also in the underlying method developments. An example of such co-operation is the risk assessment of the Proton Therapy Facility erected for cancer treatment at the Paul Sherrer Institute in Switzerland. We have contributed to that assessment by developing fault tree models for the electrical and control and instrumentation systems of the facility.
For further information on probabilistic safety assessment and its applications please contact Attila Bareith, Head of Risk Assessment Division, by e-mail or phone (+36 1 392 2716).