Time filter

Source Type

Hellstrom P.,Lloyds Register | Knochenhauer M.,Lloyds Register | Nyman R.,SSM Swedish Radiation Safety Authority
10th International Conference on Probabilistic Safety Assessment and Management 2010, PSAM 2010

One of the basic requirements for nuclear safety is to maintain and to develop the Defensein-Depth (DiD). The overall aim is to prevent deviations from normal operation from occurring and, if prevention fails, to detect and limit their consequences, and to prevent any escalation to more serious conditions. Swedish regulation in SSM FS 2008.1, requires that the Defense-in-Depth is analyzed with deterministic and probabilistic methods. PSA studies are to be performed for all operating modes, as realistic as possible. In principle, the PSA can be used to investigate the application of DiD. A PSA addresses the frequency of initiating events that challenge nuclear safety (DiD Levels 1 and 2), the conditional probability of failure of safety systems and the frequency of core damage (DiD Level 3), the conditional probability of bypass or failure of the containment and the frequency of a large (early) release (DiD Level 4) and effectiveness of the off-site emergency response measures and the frequency of social and economic consequences (DiD Level 5) DiD levels are evaluated by the licensees with PSA, however, the results are seldom referred to in terms of weaknesses and strengths of the DiD levels The SSM research project has made an inventory of the potential and possible methods for using PSA in evaluating and ranking the system structures and components (SSCs) being part of the different DiD levels. The project has identified and described current use of PSA results and the possibility for extended use of PSA results in evaluating a plants current DiD levels, the impact on DiD levels due to plant changes and the importance of plant events reported in the Licensee Event Reports (LERs) for the SSCs belonging to each DiD level. The different pieces making up a PSA (initiating events, sequences, consequences in level 1 and level 2 etc) and DiD (definitions of DiD levels etc) are elaborated and a new extended framework for the different elements of DiD and PSA and their relations is outlined. Modeling features, PSA results and their presentation in support of providing a deeper insight in the Defense-in-Depth characteristics are presented, and provide a further basis for improved use of PSA and risk results in the evaluation of a plant's Defense-in-Depth. Source

Hedberg P.,SSM Swedish Radiation Safety Authority | Hessling P.,SSM Swedish Radiation Safety Authority
International Topical Meeting on Nuclear Reactor Thermal Hydraulics 2015, NURETH 2015

In measurement technology it is good engineering practice to estimate the uncertainties in the measurements. Estimates of the uncertainties in results from numencal simulations are much leys common. This can partly be explained by the additional computational cost it involves. This cost is a function of which method is used to propagate the input and model uncertainty to the result. Some methods require a large number of repeated simulations. Other methods suffer from the "curse of dimensionality" and can realistically only treat a few uncertain parameters. A novel method, called deterministic sampling (DS). is proposed to quantify the uncertainty in a numerical flow simulation and docs not display the previously mentioned shortcomings. The new method is efficient since relatively few simulations arc required, and many parameters can be present. In this paper the method is exemplified by uncertainties originating from turbulence model constants. The method is based on the idea that a continuous probability density function can be replaced by an ensemble of discrete deterministic samples, provided the two representations have the same statistical moments. The method is first illustrated in a simple example, and later applied to the classic simulation of turbulent flow over a backward-facing step. The results obtained here arc similar to what other researchers have found using Latin Hypercube Sampling for the same case, but at a lower computational cost. Both methods show that much of the experimental results fit within the calculated uncertainties, but not all of them. This indicates the presence of systematic uncertainties in the turbulence model, and/or systematic problems in the experimental setup, such as lack of two-dimensionality in the flow. Uncertainties from other ongins. such as boundary conditions and sources, can be also included in the method if desired The DS method is also not limited to fluid flow simulations and can for example be applied to systems analyses codes in thermal hydraulics. Source

Discover hidden collaborations