[en] The performance and use of PSA:s in Sweden goes back about two decades. During all of this time, the field of PSA has been developing intensively, both internationally and within Sweden. The latest years have been characterised by an increased use of PSA models and results, and by major extensions of existing PSA models. The aim of this document is to describe PSA in Sweden with respect to development, scope and maturity, as well as to the contents of the analyses and the use of results. PSA activities will be described from the point of view of both the authorities and the utilities. The report gives an overview of the development within the area of PSA in Sweden both its history and current trends. The aim has been to include a reasonable amount of detail, both on the methods and results in PSA:s performed and on the numerous supporting research programs dealing with various aspects of PSA. 39 refs 39 refs

[en] The project deals with the modelling of components in connection with an external event, especially the modelling of area dependencies. A number of type components have been defined and analysed with respect to their sensitivity to external events. The focus has been on the interaction of the component with power supply and control. A systematic method is presented for adapting the type components to specific components, and the information needed for the adaptation is specified. The analysis has resulted in a fairly complete mapping of the required level of detail in the modelling of area dependencies of components. However, strongly conservative assumptions regarding the way safety components are affected by an external event are used in present PSA:s. Therefore, there is a risk that the detailed component model will result in unreasonably conservative risk assessments, making the basis for the ranking and selection of safety enhancing measures unreliable. Therefore, in addition to the detailed mapping of area dependencies, a realistic scenario analysis must be performed, describing how typical external events are handled by the plant, manually and via automatic functions. Only after this has been done, a true picture can be obtained of the sensitivity of components to external events. The analysis indicates that there are major differences in this sensitivity, depending on the mode of operation of the components. In general, a component that is activated (open/close or start/stop) is considerably more sensitive to an external event than a component that is to remain in its operating state (remain in operation). This fact can be used to simplify the analysis in some cases. Thus, if safety functions are activated early -before the interaction of it''s active component with the plant has become affected by the external event, then the operating mode 'No change' or 'Operation' is applicable, with the result that the logic model is simplified and the risk contribution due to the area dependence of the component is reduced considerably

No abstract available

[en] Post-processing of PSA results is very much a question of efficient data handling. Large amounts of information are reviewed, quantified, modified and re-quantified. This work may be part of the PSA itself, of the review process, or of some other safety-related analysis concerning the plant. In all of these cases both handling of the information and analysis are required. The use of an integrated program package, including programs for both analysis and information handling, and with smooth interfaces between the programs included, will make the post-processing easier to perform. It will speed the work up and reduce the amount of errors inevitably introduced into analyses requiring much manual input and data handling. Furthermore, certain types of analysis difficult to perform using a more traditional approach will be facilitated. (orig.)

[en] This paper describes the results from a number of analyses performed within a joint Nordic research project dealing with the reevaluation of Technical Specifications for nuclear power plants, based on a probabilistic approach. The analyses focus on various aspects of evaluation of periodic testing and failure data. Several pitfalls have to be avoided whenever generic failure data is prepared or applied, e.g. in a Probabilistic Safety Analysis. The analyses described were aimed at shedding light on the nature of the simplifications introduced, and on the extent of their influence on quantitative results, as well as to suggest methods for evaluation of various aspects of testing and failure data preparation

[en] The outcome of a probabilistic safety assessment (PSA) for a nuclear power plant is a combination of qualitative and quantitative results. Quantitative results are typically presented as the Core Damage Frequency (CDF) and as the frequency of an unacceptable radioactive release. In order to judge the acceptability of PSA results, criteria for the interpretation of results and the assessment of their acceptability need to be defined. Ultimately, the goals are intended to define an acceptable level of risk from the operation of a nuclear facility. However, safety goals usually have a dual function, i.e., they define an acceptable safety level, but they also have a wider and more general use as decision criteria. The exact levels of the safety goals differ between organisations and between different countries. There are also differences in the definition of the safety goal, and in the formal status of the goals, i.e., whether they are mandatory or not. In this first phase of the project, the aim has been on providing a clear description of the issue of probabilistic safety goals for nuclear power plants, to define and describe important concepts related to the definition and application of safety goals, and to describe experiences in Finland and Sweden. Based on a series of interviews and on literature reviews as well as on a limited international over-view, the project has described the history and current status of safety goals in Sweden and Finland, and elaborated on a number of issues, including the following: 1) The status of the safety goals in view of the fact that they have been exceeded for much of the time they have been in use, as well as the possible implications of these exceedances. 2) Safety goals as informal or mandatory limits. 3) Strategies for handling violations of safety goals, including various graded approaches, such as ALARP (As Low As Reasonably Practicable). 4) Relation between safety goals defined on different levels, e.g., for core damage and for unacceptable release. A number of important issues have been identified for continued studies in the next project phase. The results from the project can be used as a platform for discussions at the utilities on how to define and use quantitative safety goals. The results can also be used by safety authorities as a reference for risk-informed regulation. The outcome can have an impact on the requirements on PSA, e.g., regarding quality, scope, level of detail, and documentation. Finally, the results can be expected to support ongoing activities concerning risk-informed applications. (au)

[en] The project is connected to the development of RASTEP, a computerized source term prediction tool aimed at providing a basis for improving off-site emergency management. RASTEP uses Bayesian belief networks (BBN) to model severe accident progression in a nuclear power plant in combination with pre-calculated source terms (i.e., amount, timing, and pathway of released radio-nuclides). The output is a set of possible source terms with associated probabilities. In the NKS project, a number of complex issues associated with the integration of probabilistic and deterministic analyses are addressed. This includes issues related to the method for estimating source terms, signal validation, and sensitivity analysis. One major task within Phase 1 of the project addressed the problem of how to make the source term module flexible enough to give reliable and valid output throughout the accident scenario. Of the alternatives evaluated, it is recommended that RASTEP is connected to a fast running source term prediction code, e.g., MARS, with a possibility of updating source terms based on real-time observations. (Author)

[en] The first phase of the project (2006) described the status, concepts and history of probabilistic safety goals for nuclear power plants. The second and third phases (2007-2008) have provided guidance related to the resolution of some of the problems identified, and resulted in a common understanding regarding the definition of safety goals. The basic aim of phase 3 (2009) has been to increase the scope and level of detail of the project, and to start preparations of a guidance document. Based on the conclusions from the previous project phases, the following issues have been covered: 1) Extension of international overview. Analysis of results from the questionnaire performed within the ongoing OECD/NEA WGRISK activity on probabilistic safety criteria, including participation in the preparation of the working report for OECD/NEA/WGRISK (to be finalised in phase 4). 2) Use of subsidiary criteria and relations between these (to be finalised in phase 4). 3) Numerical criteria when using probabilistic analyses in support of deterministic safety analysis (to be finalised in phase 4). 4) Guidance for the formulation, application and interpretation of probabilistic safety criteria (to be finalised in phase 4). (LN)

[en] In recent years there has been a trend in Swedish Probabilistic Safety Analysis (PSA) work towards coordination of the tools and methods used, in order to facilitate exchange of information and review. Thus, standardized methods for fault tree drawing and basic event coding have been developed as well as a number of computer codes for fault tree handling. The computer code used by Asea-Atom is called SUPER-TREE. As indicated by the name, the key feature is the concept of one super tree containing all the information necessary in the fault tree analysis, i.e. system fault trees, sequence fault trees and component data base. The code has proved to allow great flexibility in the choice of level of detail in the analysis

[en] Nuclear power plant probabilistic risk assessment (PRA) studies utilise many models, simplifications and assumptions. Also subjective judgement is widely applied due to lack of actual data. This results in significant uncertainties. Three general types of uncertainties have been identified: (1) parameter uncertainties, (2) modelling uncertainties, and (3) completeness uncertainties. The significance of some of the modelling assumptions and simplifications cannot be investigated by assignment and propagation of parameter uncertainties. In such cases the impact of different options may (and should) be studied by performing sensitivity analyses, which concentrate on the most critical elements. This paper describes several items suitable for close examination by means of application of sensitivity analysis, when performing a level 1 PRA. Sensitivity analyses are performed with respect to: (1) boundary conditions (success criteria, credit for non-safety systems, degree of detail in modelling of support functions), (2) operator actions, (3) treatment of common cause failures (CCFs). The items of main interest are continuously identified in the course of performing a PRA study, as well as by scrutinising the final results. The practical aspects of sensitivity analysis are illustrated by several applications from a recent PRA study. The critical importance of modelling assumptions is also demonstrated by implementation of some modelling features from another level 1 PRA into the reference model. It is concluded that sensitivity analysis leads to insights important for analysts, reviewers and decision makers. (author)