PSAM12 - Probabilistic Safety Assessment and Management
Monday, June 23, 2014
KEY: -Paper; -Biography; -Presentation
Commissioner George Apostolakis
M01 Consequence Modeling and Management
Chair: Ludivine Pascucci-Cahen, IRSN
Nuclear Refugees After Large Early Radioactive Releases
Institut de Radioprotection et de Sûreté Nucléaire, Fontenay-aux-Roses, France
However improbable, large early radioactive releases from a nuclear power plant would entail major consequences for the surrounding population. In Fukushima, 80,000 people had to evacuate the most contaminated areas around the NPP for a prolonged period of time. Had they remained where they lived, they would have received doses dangerous for their health in the long run. These people have been called “nuclear refugees”. The paper first argues that the number of nuclear refugees is a better measure of the severity of radiological consequences than the number of fatalities, although the latter is widely used to assess other catastrophic events such as earthquakes or tsunami. It is a valuable partial indicator in the context of comprehensive studies of overall consequences. Section 2 makes a clear distinction between long-term relocation and emergency evacuation and proposes a method to estimate the number of refugees. Section 3 examines the distribution of nuclear refugees with respect to weather and release site. The distribution is asymmetric and fat-tailed: unfavorable weather can lead to the contamination of large areas of land; large cities have in turn a higher chance of being contaminated. Variability with respect to site is quite intuitive; however, results show that simulations are far superior to an approach based on population living within 20 or 30 km around the site.
Multidimensional Risk Evaluation: Assigning Priorities for Actions on a Natural Gas Pipeline
Mônica Frank Marsaro, Marcelo Hazin Alencar, Adiel Teixeira de Almeida, and Cristiano Alexandre Virgínio Cavalcante
Universidade Federal de Pernambuco, UFPE, Recife, Brazil
This paper presents a multicriteria decision model application to define actions with a view to mitigating the risks involved in this mode of transportation. Natural gas is a fossil fuel that is important for society and is transported through pipelines. It is used for different purposes in industrial and civil applications. Although pipelines are one of the safest transport systems, some accidents involving natural gas have occurred. The Multicriteria decision model described in this paper is put forward as a means to minimize such possibilities. It incorporates MAUT (Multi-attribute Utility Theory), which considers a decision maker’s preferences and some aspects of the Decision Theory approach. Three dimensions of risk, namely the human, financial and environmental ones -are targeted in the context of probabilistic consequences. As an important result, the information obtained from the model is shown to be important in order to define how resources should best be allocated and to establish maintenance policies for managing and mitigating risk.
Development of Accident Consequence Assessment Scheme using Accident Cost and Consideration of Decontamination Model
Kampanart Silva (a), Koji Okamoto (b), Yuki Ishiwatari (b,c) Shogo Takahara (d) and Jiraporn Promping (a)
a) Thailand Institute of Nuclear Technology, Nakhon Nayok, Thailand, b) The University of Tokyo, Tokyo, Japan, c) Hitachi-GE Nuclear Energy, Ltd., Ibaraki, Japan, d) Japan Atomic Energy Agency, Ibaraki, Japan
Severe accident at nuclear power plants, including the Fukushima accident in March 2011, wreak various kinds of consequences, including health effects, economic, social and environmental impacts. The authors developed the scheme of the accident consequence assessment using “accident cost”, aiming for it to be an index that is as comprehensive as possible. Normalized accident costs of all accident sequences along with their breakdowns, and the breakdown of the average accident cost are presented. The radiation effect cost, the decontamination cost and the relocation cost are the three major components that dominate the accident cost. The decontamination model was reconsidered since decontamination effects were taken into account by very simple assumptions and decontamination cost was estimated by a rough calculation scheme in the former model. 99 decontamination-related parameters were selected and the model is formed. A sensitivity analysis was performed to identify parameters with large influence on accident cost calculation and large extent of interactions with other parameters. Parameters with high importance tend to have large extent of interactions with other parameters. Parameters influential to accident cost, e.g., the dose of setting decontamination target area, a number of waste management-related parameters, are identified.
Safety of LPG Rail Transportation: Influence of Safety Barriers
V. Busini, M. Derudi, R. Rota
Politecnico di Milano, Department of Chemistry, Materials and Chemical Engineering “G. Natta”, Piazza Leonardo da Vinci 32, 20133 Milano, Italy
The risk due to the road and rail transportation of liquefied petroleum gas (LPG) is well known. Severe scenarios were caused by road or rail accidents involving LPG pressurized tank cars. Consolidated approaches exist for the analysis, the prevention and the mitigation of risk due to the transportation of hazardous materials (HazMat) by road or rail. In Europe a specific regulation applies to the equipment used for the transport of HazMat and specific regulations apply to the qualification of equipment used for LPG transportation. Nevertheless, on June 29th, 2009, an extremely severe transportation accident involving LPG took place in the station of Viareggio, in Italy. A train carrying 14 tank cars of LPG derailed and several railcars overturned on the shunts in the Viareggio station. A tank was punctured, releasing its entire content that ignited causing an extended and severe flash-fire. The present study focused on the study of the effect of different parameters on the heavy gas cloud dispersion resulting from the accident, such as meteorological parameters and height of safety walls. It was found that, to be effective, the mitigation barriers must be carefully designed, with particular reference to their height with respect to the height of the heavy gas cloud.
Determination of Target Reliability Levels Based on Value to the Customer and Warranty Budgets
Michael Bartholdt, Volker Schweizer and Bernd Bertsche
University of Stuttgart, Stuttgart, Germany
The method presented here serves to determine the system’s as well as subsystems’ target reliability levels combined. It centers the functions of the product to be developed and knows to weight requirements in line with the voice of the customer. Each subsystem’s target reliability level is defined in accordance with its quantitative contribution to fulfilling the functions as desired by the customer. Statistically inevitable failures before the targeted product lifetime are often compensated by warranty and good-will expenditures. These costs are methodologically broken down and allocated to each subsystem purposefully in order to achieve the utmost customer satisfaction. By bringing together such costs per subsystem and its importance to the customer, reliability goals are obtained aligned with the value to the customer. In contrast to most of the existing methods to define and allocate reliability goals (typically realized by two different methods), subsystem target reliability levels are defined at first. The system’s reliability goal is then calculated by means of Boole-theory. Arbitrarily complex systems can be analyzed.
M02 Digital I&C and Software Reliability I
Chair: Hervé Brunelière, AREVA NP SAS
How to Integrate Correctly Hardware Common Cause Failures in Frequency Calculations?
Hervé Brunelière, Monica Rath, and Wenjie Qin
AREVA NP SAS, Paris La Défense, France
Hardware common cause failures are generally the highest contributors in the I&C systems reliability and availability studies. Comparisons of results from calculations of frequency of spurious actuations by a safety system or frequency of failures of a control system with operation feedback of such failures show that the frequency calculations are often overestimated. This is due to the use of « classic » common cause failure parameters. This is mainly explained by the fact that, for these undesired events, failures are generally not hidden ones and are then detected within few hours. Then, for common cause failures that are not simultaneous, the first failure is often repaired before the second one appears. This over conservatism can lead to inappropriate design choices like addition of redundancies or interlocks to minimize the frequency of an undesired event based on a calculation that does not reflect the real situation. This is then a concern for a designer and for a utility to limit as far as possible the impact of this over conservatism. One solution is to consider only independent failures in frequency calculations. In this case, the result is underestimated as simultaneous common cause failures that are possible and credible are not considered in the result. Then, the risk is not to implement some necessary measures in the design due to over optimistic results. The paper will discuss possible solutions to handle these types of failures in calculations based on real cases. Illustrations will be based on a typical architecture of an I&C system based on Teleperm XS platform similar to the ones currently implemented in nuclear power plants. The paper will also integrate discussions on relevance of the different methodologies including no consideration of CCF at all, degraded CCF factors values and possibilities of extrapolation. These methodologies will be compared based on their impact on calculation results and the consistency with operational experience.
The Basic Idea of Quantitative Model of Reactor Protection System Considering Stochastic Process
Tokyo City University, Tokyo, Japan
In nuclear power plants such as ABWR and the latest PWR, digital instrumentation and control system have been installed increasingly to reactor emergency shutdown system which is one of the important safety functions. However, it has been found that it is difficult to model the digital equipment reliability in probabilistic risk assessment (PRA). And some of issues such as taxonomies of failure modes have been studied in the international framework, OECD/NEA/WGRisk task group called DIGREL. In this paper, the reactor trip actuation failure event logics and frequencies resulting from the multiple failures and the demand following the initiating event are analyzed qualitatively and quantitatively. This paper presents the example of the reliability analysis of the digital Reactor Protection System (RPS) considering stochastic process, the approach given by this paper will be applicable to establish the PRA model of digital RPS of the actual nuclear power plant.
A Quantitative Software Testing Method for Hardware and Software Integrated Systems in Safety Critical Applications
Hai Tang, Lixuan Lu
University of Ontario Institute of Technology, Oshawa, ON, Canada
Most of today’s Safety Instrumented Systems (SIS) are hardware and software integrated systems. In these systems, failures can occur in both hardware and software. Hardware failures and their effects have been studied extensively in the literature. However, the methods and results dealing with hardware failure are not directly applicable for software reliability modeling, due to the difference of nature between hardware and software. This is especially of concern when the SIS is used for safety critical applications. In this paper, a hardware and software integrated reliability model is proposed to model the reliability of the integrated system. The requirement on software reliability is then determined based on the hardware reliability and the requirement on the Safety Integrity Level (SIL) of the integrated system. Following this, a Bayesian stopping rule is used to determine the minimal number of successful software runs, in order to provide a certain level of confidence that the reliability requirement on the software is achieved.
Abdallah Amri (a), Stefan Authén (b), Herve Bruneliere (c), Gilles Deleuze (d), Gabriel Georgescu (e), Jan-Erik Holmberg (f), Man Cheol Kim (g), Keisuke Kondo (h), Ming Li (i), Ewgenij Piljugin (j), Wietske Postma (k), Jiri Sedlak (l), Carol Smidts (m), Jan Stiller (j), and Nguyen Thuy (d)
a) OECD/NEA, Paris, France, b) Risk Pilot AB, Stockholm, Sweden, c) AREVA, Paris, France, d) EDF R&D, Paris, France, e) Institut de Radioprotection et de Sûreté Nucléaire, Paris, France, f) Risk Pilot AB,Espoo Finland, g) Chung-Ang University, Seoul, Korea, h) Nuclear Regulation Authority, Japan, i) United States Nuclear Regulatory Commission, USA, j) Gesellschaft für Anlagen- und Reaktorsicherheit, Germany, k) Nuclear Research and consultancy Group, the Netherlands, l) ÚJV Řež, Husinec -Řež, Czech Republic, m) Ohio State University, USA
The OECD/NEA CSNI Working Group on Risk Assessment (WGRisk) has set up a task group called DIGREL to develop a taxonomy of failure modes of digital components for the purposes of probabilistic risk analysis (PRA). The failure modes taxonomy is based on a failure propagation model and a definition of five levels of abstraction: 1) system level, 2) division level, 3) I&C unit level, 4) I&C unit modules level, 5) basic components level. This structure corresponds to a typical reactor protection system architecture. The failure propagation model consists of the following elements: fault location, failure mode, uncovering situation, failure effect and the end effect. These concepts are applied to define the relationship between a fault in hardware or software modules (module level failure modes) and the effect on I&C units (I&C unit level failure modes). The purpose of the taxonomy is to support PRA, and therefore focuses on high level functional aspects rather than low level structural aspects. This focus allows handling of the variability of failure modes and mechanisms of I&C components. It reduces the difficulties associated with the complex structural aspects of software in redundant distributed systems.
A Component-based Approach for Assessing Reliability of Compound Software
Monica Lind Kristiansen (a), Bent Natvig (b), and Harald Holone (c)
a) Department of Informatics, Østfold University College, Halden, Norway, b) Department of Mathematics, University of Oslo, Oslo, Norway, c) Department of Informatics, Østfold University College, Halden,Norway
Predicting the reliability of software systems based on a component approach is inherently difficult, in particular due to failure dependencies between the software components. This paper describes a component-based approach for assessing reliability of compound software, where failure dependencies between software components are explicitly addressed. This is done by finding accepted upper bounds for probabilities that pairs of software components fail simultaneously and then by including these into the reliability models. To find these accepted upper bounds, the approach applies principles of Bayesian hypothesis testing on simultaneous failure probabilities. In addition, the restrictions imposed on the simultaneous reliabilities and failure probabilities by the marginal reliabilities and failure probabilities are taken into account. To illustrate the approach, we use an example based on mobile positioning systems for backtracking. This is for instance used to help people with dementia to find their way home if they get lost.
M03 Enterprise Risk Management
Chair: David Johnson, ABS Consulting
Automated Evolutionary Restructuring of Workflows to Minimise Errors ViaStochastic Model Checking
Luke Thomas Herbert (a), Zaza Nadja Lee Hansen and Peter Jacobsen (b)
a) DTU Compute, Lyngby, Denmark, b) DTU Management, Lyngby, Denmark
This paper presents a framework for the automated restructuring of workflows that allows one to minimise the impact of errors on a production workflow. The framework allows for the modelling of workflows by means of a formalised subset of the Business Process Modelling and Notation (BPMN) language, a well-established visual language for modelling workflows in a business context. The framework’s modelling language is extended to include the tracking of real-valued quantities associated with the process (such as time, cost, temperature). In addition, this language also allows for an intention preserving stochastic semantics able to model both probabilistic-or non-deterministic branching behaviour. We further extend this formalism to allow for the introduction of error states which allow for both fail-stop behaviour and continued system execution. We explore the practical utility of this approach by means of a case study from the food industry. Through this case study we explore the extent to which the risk of production faults can be reduced and the impact of these can be minimised, primarily through restructuring of the production workflows. This approach is fully automated and only the modelling of the production workflows and the expression of the goals require manual input.
Enterprise Risk and Opportunity Management for Nonprofit Organizations and Research Institutions
Allan Benjamin (a), Homayoon Dezfuli (b), Chris Everett (c), Julie Pollitt (d), Dev Sen (c)
a) Independent Consultant, Albuquerque, NM, USA, b) Office of Safety & Mission Assurance, NASA Headquarters, Washington, DC, USA, c) Information Systems Laboratories, Inc., Rockville, MD, USA, d)Independent Consultant, San Jose, CA, USA
Enterprise risk and opportunity management (EROM) concerns the means by which organizations develop and implement their strategic goals through a portfolio of programs, projects, institutional assets, and activities. The overall objective of EROM is to reach an optimal balance between minimizing the potential for loss (risk) while maximizing the potential for gain (opportunity). The focus of this paper is on the development of guiding principles and an overall approach that serves the interests of technically oriented nonprofit organizations and research institutions. These interests tend to place emphasis on performing services and achieving technical gains more than on achieving specific financial goals, which is the province of commercial enterprises. In addition, the objectives of nonprofit organizations may extend to institutional development and maintenance, financial health, legal and reputational protection, education and partnerships, and mandated milestone achievements. This paper discusses the philosophical underpinnings of EROM in the context of nonprofit organizations, the integration of EROM with existing management processes, and the nature of the activities that are performed to implement EROM within this context.
Programmatic Assessment of RG-MOX Utilization Following Participation in the DOE Surplus Plutonium Disposition Program
David H. Johnson, Andrew A. Dykes (a), Andrew G. Sowder, and Albert J. Machiels (b)
a) ABSG Consulting, Irvine, CA, USA, b) Electric Power Research Institute (EPRI), Charlotte, NC, USA
EPRI is building a suite of tools for assessing nuclear fuel cycle options based on a platform of software, simplified relationships, and explicit decision-making and evaluation guidelines. This paper summarizes an example of an assessment from a utility perspective regarding continuing MOX utilization with commercial reactor-grade mixed-oxide fuel (RG-MOX) following successful utilization participation in the DOE Surplus Plutonium Disposition Program. This assessment reflects potential opportunities and problems based on topic familiarity and the perspective embedded in the scenario definition, as follows: (1) economic considerations will represent a primary driver for utilities operating in the U.S. commercial environment, and (2) back-end management issues must be flagged due to the number and magnitude of constraints in used-fuel management at U.S. nuclear plants for both wet and dry storage (and the important interface between them). While economic considerations are seen as the primary utility decision drivers with respect to RG-MOX use under the stylized conditions defined here, this assessment also showed that technical waste management issues could be showstoppers if not adequately resolved.
A Jointly Optimization of Production, Delivery and Maintenance Planning for Multi-Warehouse/Muli-Delivery Problem
Hajej Zied, Turki Sadok, and Rezg Nidhal
LGIPM-University of Lorraine, Metz, France
This paper develops à jointly optimization problem in order to establish an optimal production, delivery and maintenance strategy for a manufacturing system subjected to a random failure. The problem consists on several warehouses allow to satisfy random demands during a finite horizon, under service level. In order to assure an economical objective, we have determined the optimal production/maintenance plan and the economically delivery quantities plan considering the delivery time for each warehouse. The aim of the proposed approach is to show a jointly production/maintenance/delivery optimization, with a constrained stochastic production-delivery-maintenance planning problem under hypotheses of service level, delivery time for each warehouse and failure rate, which minimizes the total production, inventory, delivery and maintenance costs. A numerical example confirms the analytical results.
Investigating the Role of Statistical Models in Water Distribution Asset Management: A Semi-structured Interview Approach
Vikram M. Rao, and Royce A. Francis
The George Washington University, Washington DC, USA
A robust asset management plan needs to be in place for water utilities to effectively manage their distribution systems. Of concern to utilities are broken pipes, which can lead to bacteria entering the water system and causing illness to consumers. Typically, water utilities allocate a portion of funds every year for renewal of pipes and valves. However, pipe renewal is largely based on replacing current broken pipes, and long-term asset management planning to replace pipes is not a priority for water utilities. Water utilities are beginning to use probabilistic break models and other statistical tools to predict pipe failures. These models incorporate variables such as pipe length, diameter, age, and material. These models are emerging in the water industry; however, their direct impact on long term asset planning remains to be seen. In addition, the effectiveness of these models is questionable, as there is currently little research done to evaluate the ability of these models to assist in asset management planning. This paper discusses the role of probabilistic pipe break models in structuring long-term asset management decisions. We determine that there are many factors that are needed to contribute to the feasibility of statistical models in a water asset management program, including data availability, funds, and shared information.
M04 Environmental Modeling
Chair: Stefan Hirschberg, Paul Scherrer Institut
Modeling of Pollutant Dispersion in Street Canyon by Means of CFD
Davide Meschinia, Valentina Busini (a), Sjoerd W. van Ratingen (b), Renato Rota (a)
a) Department of Chemistry, Materials and Chemical Engineering “G. Natta”, Politecnico di Milano, Italy, b) TNO, Utrecht, Netherlands
Nowadays, pollution from traffic remains one of the major sources for contamination in urban areas and it is widely known that substances emitted by vehicles represent a serious hazard to human health; some traffic-related pollutants, such as NO, NOx and CO are responsible for both acute and chronic effects on human health. This is often the case near busy traffic axis in city centers or street canyons. Purpose of this work is to validate the CFD model predictions against the field measurements of pollutants dispersion in an actual urban environment: Göttinger Strasse, Hanover, Germany. In the location, the population exposure to traffic-related pollution is expected to be high. Steady-state simulations have been performed for 18 different wind directions, with an increment of 20°, in order to cover the whole wind rose. A grid and a Schmidt number sensitivity analysis have been carried out in order to determine both the most suitable resolution of the computational geometry and the most suitable parameter to model the turbulence conditions in the street canyon. All CFD simulations have been performed for neutral atmospheric conditions and have been carried out with the CFD code FLUENT 12.1.
Consideration on the Assessment of the Environmental Consequences and Impacts During Transport of Radioactive Materials (RAM)-A Safety Case
AREN, Bucharest, ROMANIA
The transport of Dangerous Goods-Class #7 Radioactive Material (RAM), is an important part of the Romanian Radioactive Material Management. The overall aim of this activity is for enhancing operational safety and security measures during the transport of the radioactive materials, in order to ensure the protection of the people and the environment. The paper will present an overall of the safety and security measures recommended and implemented during transportation of RAM in Romania. Some aspects on the potential threat environment will be also approached with special referring to the low level radioactive material (waste) and NORM transportation either by road or by rail. A special attention is given to the assessment and evaluation of the possible radiological consequences due to RAM transportation. The paper is a part of the IAEA’s Vienna Scientific Research Contract on the State Management of Nuclear Security Regime (Framework) concluded with the Institute for Nuclear Research, Romania, where the author is the CSI (Chief Scientific Investigator).
Health Effects of Technologies for Power Generation: Contributions from Normal Operation, Severe Accidents and Terrorist Threat
S. Hirschberg, C. Bauer, P. Burgherr (a), E. Cazzoli (b), T. Heck, M. Spada and K. Treyer (a)
a) Paul Scherrer Institute, Laboratory for Energy Systems Analysis, Villigen, Switzerland, b) Cazzoli Consulting, Villigen, Switzerland
As a part of comprehensive analysis of current and future energy systems we carried out numerous analyses of health effects of a wide spectrum of electricity supply technologies including advanced ones, operating in various countries under different conditions. The scope of the analysis covers full energy chains, i.e. fossil, nuclear and renewable power plants and the various stages of fuel cycles. State-of-the-art methods are used for the estimation of health effects. This paper addresses health effects in terms of reduced life expectancy in the context of normal operation as well as fatalities resulting from severe accidents and potential terrorist attacks. Based on the numerical results and identified patterns a comparative perspective on health effects associated with various electricity generation technologies and fuel cycles is provided. In particular the estimates of health risks from normal operation can be compared with those resulting from severe accidents and hypothetical terrorist attacks. A novel approach to the analysis of terrorist threat against energy infrastructure was developed, implemented and applied to selected energy facilities in various locations. Finally, major limitations of the current approach are identified and recommendations for further work are given.
Metal Remediation of Acid Mine Drainage Using a Hybrid System of Microalgae Reactor
Young-Tae Park, Hongkyun Lee, Hyun-Shik Yun, Jaeyoung Choi
Korea Institute of Science and Technology-Gangneung Institute, Gangneung, South Korea
Acid mine drainage(AMD) contains high concentrations of heavy metals and has become a serious environmental problem. A pipes inserted microalgae reactor(PIMR) was constructed to cultivate microalgae and purify AMD. The effects of metal concentration, pH and sulfate after pretreatment on the removal of iron and microalgae growth were investigated. Batch studies showed that PIMR and microalgae can adsorb iron with an uptake of 63.21 ± 9.8 mg/L iron. Microalgae growth was measured by optical density (OD) and dry cell weight (DCW); OD and DCW were 3.96 and 1.54g/L respectively. Continuous studies also proved that PIMR can be used for metal remediation and microalgae cultivation.
M05 Fire Modeling and Simulation
Chair: Shahen Poghosyan, Nuclear and Radiation Safety Center, Armenia
Experiences from Developing and Implementing Shutdown Fire PRA at Forsmark NPP
Erik Cederhorn, Maria Frisk
Risk Pilot AB, Stockholm, Sweden
The cold shutdown mode has earlier been considered as a safe mode without a significant risk for a major accident. However during the last few decades knowledge has improved regarding risks during shutdown mode. Many activities are on-going during this period and the risk of fire occurrence may be affected. Due to an increased number of plant activities the integrity of the fire compartments may not be intact and this could lead to more extensive fire spreading. At the same time important barriers may be unavailable due to maintenance and a fire event could become critical. Time available for recoveries before fuel is exposed in the reactor pressure vessel after an initiating event i.e. fire event, which results in loss of residual heat removal, is in many cases significantly longer than 24 hours. Area event analyses for shutdown mode generally tend to produce quite conservative results, which is why efforts have been made to increase realism in the analyses by using of improved methods. In order to increase realism dependencies between plant risk and maintenance activities, i.e. different combinations of safety system alignments, during the shutdown period have been studied in detail. This has had an impact on the estimation of both fire ignition frequencies and probabilities for fire spreading between different compartments. This paper will discuss the methodology applied to the fire PRA at Forsmark NPP during the cold shutdown period, with focus on fire frequency analysis and fire scenario analysis. The implementation of fire analysis in the PRA and lessons learned from this will also be addressed.
Fire PSA and Insights
F. Nicoleau, F. Corenwinder, G. Georgescu
Institute for Radiological Protection and Nuclear Safety (IRSN), Fontenay-Aux-Roses, France
IRSN (TSO of French Nuclear Safety Authority) develops simplified Fire Level 1 probabilistic safety assessments (PSAs) for nuclear power plants (NPPs) in order to establish his own independent opinion on the assumptions and results of the licensee Fire PSAs (EDF). IRSN Fire PSAs are extensions of the IRSN in-house developed NPP Level 1 PSAs for internal events. The licensee and IRSN studies are similar in scope; however the objectives and some main assumptions may be different. The licensee objectives are to answer to the Safety Authority requests to perform complete PSA studies as a complementary approach of the deterministic studies of the fire risks. The IRSN study objectives are to provide an independent verification of the licensee study and also to allow further PSA applications in the framework of technical instruction of safety issues. In particular, IRSN main goal is to focus on the most critical equipment and compartments in terms of fire-related risks. The paper gives two examples of specifics insights obtained regarding the licensee PSAs in the field of Fire. The first example is related to the ongoing third periodic safety review of 1300MWe NPPs. The second example deals with IRSN review of the licensee Fire PSA for the commissioning of the French EPR reactor (at Flamanville 3).
Complex Investigation of Fire PSA Dominant Scenario Related to Direct Flame Contact with Safety Related Pipes
Shahen Poghosyan, Tsolak Malakyan, Gurgen Kanetsyan and Armen Amirjanyan
Nuclear and Radiation Safety Center, Yerevan, Armenia
Fire risk is one of the complex problems and potentially serious challenges to the safety of Nuclear Power Plants (NPPs). Fire PSA is a powerful and systematic tool which can reveal critical safety issues from the point of view of fire. A detailed fire PSA study performed for Unit 2 of the Armenian NPP (ANPP) shows that overall fire risk is driven by several fire scenarios. However before applying the results in a safety-related decision making process, it is important to verify the robustness of conclusions related to the identified risk contributing factors. Observation shows that the results received for the confinement oil fire scenario imply a need to implement substantial modernization activities. On the other hand, the approach used for oil fire modeling in confinement is considered conservative and the results obtained are considered to have considerable associated uncertainty. The aim of the current paper is to present a more accurate complex investigation of the oil fire scenario in the ANPP confinement building in order to create an adequate basis for further plant modernization decisions. The aim of the current paper is to present a more accurate complex investigation of the oil fire scenario in the ANPP confinement building in order to create an adequate basis for further plant modernization decisions.
Fire Risks of Loviisa NPP During Shutdown States
Sami Sirén, Ilkka Paavola, Kalle Jänkälä
Fortum Power And Heat Oy, Espoo, Finland
Fire PRA for all 15 shutdown states of Loviisa NPP has been performed. The fire PRA for power operation and the internal event PRA for shutdown have been used as a basis for the analysis, reducing the time needed for investigating cable routing and potential of fire-induced initiating events. The hot states are mostly modeled using applicable power operation fire scenarios. For the cold states, 342 fire scenarios have been created and integrated with the PRA model. Fire frequencies have been estimated with an empirical Bayesian method using both plant data and international data. The importance of moving from conservative modeling towards best estimate is underlined in the shutdown fire PRA. The real availability of systems instead of the minimum requirements in Technical Specifications has been taken into account to decrease the conservatism related to maintenance activities. Fires inside the control building during cold states dominate the risk. (The shutdown fire risk is relatively small,) but it would be hundredfold without the backup RHR system.
M06 Human Reliability Analysis I
Chair: Jeffrey C. Joe, Idaho National Laboratory
Modeling and Quantification of Team Performance in Human Reliability Analysis for Probabilistic Risk Assessment
Jeffrey C. Joe and Ronald L. Boring
Idaho National Laboratory, Idaho Falls, USA
Probabilistic Risk Assessment (PRA) and Human Reliability Analysis (HRA) are important technical contributions to the United States (U.S.) Nuclear Regulatory Commission’s (NRC) risk informed and performance based approach to regulating U.S. commercial nuclear activities. Furthermore, all currently operating commercial nuclear power plants (NPPs) in the U.S. are required by federal regulation to be staffed with crews of operators. Yet, aspects of team performance are underspecified in most HRA methods that are widely used in the nuclear industry. Furthermore, there are a variety of "emergent" team cognition and teamwork errors (e.g., communication errors) that are 1) distinct from individual human errors, and 2) important to understand from a PRA perspective. The lack of robust models or quantification of team performance is an issue that affects the accuracy and validity of HRA methods and models, leading to significant uncertainty in estimating human error probabilities (HEPs). This paper describes research that has the objective to model and quantify team dynamics and teamwork within NPP control room crews for risk informed applications, thereby improving the technical basis of HRA, which improves the risk-informed approach the NRC uses to regulate the U.S. commercial nuclear industry.
Comparison of Task Loads between Usages of Computer-based Procedures in an Advanced Control Room
Yochan Kim, Wondea Jung, and SeungHwan Kim
Korea Atomic Energy Research Institute, Daejeon, Republic of Korea
With the development of a computer-based control room in an APR1400, the behaviors of operators in the control room have changed. To investigate the effects of the computerized instrument and control systems on workloads, the workloads of operators in an APR1400 who employ three different usages of a computer-based procedure were compared. The COCOA framework, a taskloading approach of workload evaluation, was employed to evaluate the workloads, and some statistical analyses were conducted to compare them. We performed a total of 22 experiments in a full scope simulator of an APR1400 under LOCA and SGTR scenarios, and obtained workload scores in cognitive, communicative, and operative dimensions. The results showed that the SS-centric usage requires many activities to the SSs, and the other usages require fewer activities to the SSs than the SS-centric usage. Based on the findings, we discussed whether the workloads between operators in an MCR can be adjusted by the CPS usages.
Study on Analysis Method of Operator’s Errors of Situation Awareness in Digitized Main Control Rooms of Nuclear Power Plants
Pengcheng Li (a), Li Zhang (a,b), Licao Dai, Jianjun Jiang, and Difan Luo (a)
a) Human Factor Institute, University of south China, Hengyang, People’s Republic of China, b) Hunan Insitute of Technology, Hengyang, People’s Republic of China
Situation awareness (SA) is a key element that impacts operator’s decision-making and performance in nuclear power plants (NPPs). The subsequent complex cognitive activities can not be correctly completed due to errors of situation awareness (ESA), which will lead to disastrous consequences. In order to investigate and analyze operator’s situation awareness error in digitized main control room (DMCR) of the nuclear power plants, the model of ESA is established, the classification system of SAE is developed based on the built SAE model, and the method of ESA is also constructed on the basis of the observation of simulator and operator surveys. Finally, a case study is provided to illustrate the concrete application of the method. It provides a theoretical and practical support for the operator’s SAE analysis in the digitized main control room of nuclear power plants.
Study on Human Errors in DCS of a Nuclear Power Plant
Licao Dai (a), Li Zhang (b), Pengcheng LI (a), Hong Hu (b) Yanhua Zou (b)
a) Human Factor Institute, University of South China, Hengyang, P.R.China, b) Hunan Institute of Technology, Hengyang, P.R.China China
More and more main control rooms in advanced nuclear power plants (NPP) use computer-based displays and controls, which are called digital control systems (DCS). DCS changes some technological aspects in a NPP control room, including information display systems, alarm systems, controllers and components and computer-based procedure systems. These changes on man-machine interface (MMI) alter the ways of operators acquiring information and controlling the system and thus give rise to new human error issue. In order to investigate the impact of the new MMI on human reliability, the researchers conducted a study in a reference plant with DCS. The practical operation data as well as the experimental data were acquired to study the causes, effects and recovery factors of the new human errors. The research makes an effort on providing a foundation for human error prevention in a DCS and human reliability analysis.
Experience Feedback from Fukushima towards Human Reliability Analysis for Level 2 Probabilistic Safety Assessments
V. Fauchille, H. Bonneville, J.Y. Maguer
Institut de Radioprotection et de Sûreté Nucléaire, Fontenay-aux-Roses Cedex, FRANCE
In the years 2000, the IRSN developed its first level 2 Probabilistic Safety Assessment (PSA) for the 900 MWe French PWRs. It was an ambitious project and one of the important tasks was to build a Human Reliability Analysis (HRA) model able to model the human actions to be implemented after the core melted. These actions are performed by operators in the main control room or by field operators outside but most of the decisions are taken, on the basis of the Severe Accident Management Guide (SAMG), by the crisis organization. A Human and Organizational Reliability Analysis in Accident Management (HORAAM) model is born from this enterprise. It is based on the “Decision Tree method”. HORAAM has been developed from the observation of the nuclear crisis exercises that are regularly practiced in France. Several influence factors which particularly affect human and organizational reliability in such a situation were identified. Currently HORAAM is used at IRSN but it has never been compared to the experience feedback of a real accident. After the Fukushima accident, IRSN conducted a study to confront HORAAM with the difficulties encountered to implement actions after the core meltdown. The purpose of this article is to present the main conclusions drawn from this study.
M07 Industrial Safety and Accident Analysis I
Chair: Thor Myklebust, SINTEF
The Impacts of Supervisor Attributes and Supervision-Related Policies on Safety and Environmental Outcomes and Reporting Behavior
Christopher J. Jablonowski (a), John J. Tolle (b)
a) Shell Exploration and Production Company, Houston, TX, U.S.A. b) Value Discovery LLC, Houston, TX, U.S.A.
This paper specifies detection-controlled regression models to investigate the drivers of health, safety, and environmental (HSE) performance and reporting behavior. The analysis confirms some results from previous research and also tests new hypotheses, with emphasis on supervision-related practices and policies. Most of the results are general and thus applicable to other regions, to other operators, and very likely to other industrial sectors. The results can be used to drive decisions regarding operating practices and HSE management system policy.
Change Impact Analysis as Required by Safety Standards, What to Do?
Thor Myklebust (a), Tor Stålhane (b), Geir Kjetil Hanssen, and Børge Haugset (a)
a) SINTEF ICT, b) IDI NTNU
Change Impact Analysis related to safety of products and systems is used by companies in many industries and is required by several standards. The International Electrotechnical Commission (IEC) has issued several standards with requirements and guidelines for the establishment of analysis like FMECA (IEC 60812), FTA (IEC 61025), Design review (IEC 61160), HAZOP (IEC 61882), Markov (IEC 61165) and RBD (IEC 61078) but no standard for Change Impact Analysis. Based on the aforementioned standards, a literature study and experience from several projects, this paper proposes a Change Impact Analysis Report adapted to the specific characteristics of the Railway and Process industry domains. The purpose of this paper is to serve as a tool to aid manufacturers in performing a Change Impact Analysis at the appropriate level which will be approved by assessors and certification bodies. This is important since the Change Impact Analysis report is one of the main inputs to the assessor/certification body. The paper starts by presenting and clarifying relevant terms and definitions, as these differ from standard to standard. The main part of the paper structures and describes the relevant topics for a Change Impact analysis report. Using the described approach will save time and cost and reduce the risk of having to re-issue the Change Impact analysis, thus ending up with a product having hidden defects. Using the mindset from SafeScrum -a method that introduces elements from agile into safety-related software development, will result in further savings. This work is part of a series of Railway and IEC 61508 certification projects and the SUSS2 Research projects.
Bucket Wheel Excavators: Past to Present Experiences in Safety Operation
Wrocław University of Technology, Wrocław, Poland
In Polish open pit mining there were and exist now 124 machines of 46 different types including: excavators and spreaders. After World War II it was observed 235 breakdowns and 38 of which were classified as major catastrophes related to 95 machines. Undesired events had both design and operational causes. Total number of multiple failures caused by five structural units reaches 205. Another 30 failures had operational and environmental causes. Each of 235 catastrophes was followed by official penetrating inquiry looking for basic causes. Results of investigations were introduced in design process, modernizations and regulations regarding safety operation and maintenance. Cost analysis shows that modernization or rebuilding of failed machine could be even twice cheaper than building a new one. Objective of the paper is to present closed chain of taking knowledge and data from current operation and applying it in the design and modernization process. In the paper there are historical data shown and analysis about catastrophes. Examples of included conclusions and recommendations from catastrophes of open pit mining machines in design and operation show progress in that branch of industry.
Verification of Risk Assessment and Treatment model and Software tool in Chemical Establishments in Slovak Republic
Katarina Holla and Jozef Ristvej
University of Žilina, Žilina, Slovakia
The major industrial accidents prevention is one of the principle conditions for ensuring the security and safety of the employees and citizens living close to the industrial establishments. The most hazardous industrial establishments in the EU are called the “SEVESO establishments” and their number is about 10,000 in the framework of all EU countries. Until June 2015 the individual member states are to transpose the new directive SEVESO III to their legal environment. In the implementation framework there is space for the individual member states to identify the problem areas in the existing legal environment and to implement new approaches especially for the risk assessment and treatment. Regarding to the space that was created during the SEVESO III transposition, it is possible to suggest unified procedures and methodologies which could beused by the enterprises for these types of analyses. This paper summarizes research results and recommendations (of University of Žilina in Žilina) in area of industrial accidents prevention in Slovak republic based on ongoing verification of created quantitative risk assessment and treatment model and software tool in two chemical establishments. The advantage is especially utilizing the bow-tie diagrams which link the fault trees and event tress and in this way create a possibility to utilize the generic trees for carrying out an analysis. Also a whole range of other methods and techniques can be utilized in the individual steps of this systematic model. This approach should be considered after its validation to be used in all SEVESO establishments in Slovak republic in the future and therefore it will be possible to compare results of analysis between SEVESO establishments in Slovak republic.
A Preliminary Accident Investigation on a Norwegian Fish Farm Applying Two Different Accident Models
Siri Mariane Holen, Ingrid Bouwer Utne (a), and Ingunn Marie Holmen (b)
a) Department of Marine Technology, NTNU, Trondheim, Norway, b) SINTEF Fisheries and Aquaculture, Trondheim, Norway
The aquaculture industry is one of the most dangerous professions with respect to occupational hazards in Norway. Hazardous operations are carried out daily on fish farms and safe operations are crucial. This paper aims to apply two methods of accident analysis for an accident at a fish farm. Accident analysis is necessary for understanding why accidents happen, helps us understanding the system in which the accident happened, and can provide for improvements for a safer system. The two methods, namely STEP and CAST are based on different assumptions of accident causation, and highlights different mechanisms that contributed to the accident happening. STEP provides a systematic guidance to ask the right questions to get a full view on what happened during the accident sequence, and portrays the accident in an easy accessible flowchart. CAST is a more comprehensive method that models all levels in the sociotechnical system to evaluate if there is inadequate control in any of the feedback loops of the different levels. Using CAST for accident investigation is more resource demanding, but will also give more information on safety problems which can be used to improve the risk management system.
M11 Lifetime and Ageing
Chair: Tunc Aldemir, The Ohio State University
Life Analysis for the Main bearing of Aircraft Engines
Peng Qin, Xiaoling Zhang, Liping He, Liangliang Ding
School of Mechanics, Electronic, and Industrial Engineering, University of Electronic Science and Technology of China, Chengdu, China
Life of main bearings in a aircraft engine directly affects the reliability, safety and feasibility of aircraft engines. In order to optimize and improve the performance of the existing aircraft engines, as well meet the needs of a new generation of aircraft engines, this paper analyses and estimate of main bearing’s life of aircraft engines by taking the deep groove ball bearing as an example. Firstly, the 3D model of deep groove ball bearings is established by using Pro-E software and then converted into a finite element model. Secondly, such features as stiffness, strength and fatigue life of the deep groove ball bearing are investigated by ANSYS software, which result in some theoretical are discussion and some relevant measures that enable improving the service life of main bearings of the aircraft engine.
Development of a Dynamic, Plant Condition-Dependent Probabilistic Safety Assessment
Radoslaw Lewandowski, Richard Denning, Tunc Aldemir and Jinsuo Zhang
The Ohio State University, Columbus, Ohio, USA
Although each nuclear power plant has a plant-specific probabilistic risk assessment (PRA) that reflects design differences from other plants, the condition of each plant changes uniquely with time. A great deal of surveillance data are collected for the plant that reflect the changing condition of the plant. In some instances, plant staff use these data to guide the plant’s preventative maintenance and surveillance programs. In general, however, these data are not used to characterize the evolving risk of the plant. Our understanding of the underlying mechanisms for the degradation of systems, although far from perfect, is improving with time. The possibility of developing a condition-dependent PRA is explored that would take a first principles approach to modeling the progression of degradation mechanisms, periodically adapting the model to account for surveillance results, and using the model as a basis for a time-dependent characterization of plant specific risk. Because surveillance data would be used to periodically assess the consistency of the observed behavior with model predictions, it might be possible to provide early identification of unanticipated degradation mechanisms. A case study is described involving a potential bypass accident sequence involving the progression of flow-accelerated corrosion in secondary system piping and stress corrosion cracking of steam generator tubes.
Risk-Informed Safety Margin Characterization Case Study: Use of Prevention Analysis in the Selection of Electrical Equipment to Be Subjected to Environmental Qualification
D. P. Blanchard (a) and R. W. Youngblood (b)
a) Applied Reliability Engineering, Inc. (AREI), San Francisco, California USA, b) Idaho National Laboratory (INL), Idaho Falls, Idaho, USA
Age-related degradation of electrical equipment is cited in numerous discussions of extended nuclear power plant operation as an important issue. Which SSCs matter? For which SSCs do we need ongoing assurance of performance? Replacement of all components and cables is a daunting prospect. Being able to focus on a subset of SSCs from an environmental qualification (EQ) perspective, while still maintaining plant-level safety and efficiency even if the other components and cables degrade, would be worthwhile. This paper summarizes a case study that examines SSC aging for components within a PWR large dry containment. The case study illustrates how an understanding of SSC margin can be characterized given the overall integrated plant design, and was developed to demonstrate a method for deciding on which SSCs to focus, which SSCs are not so important from an environmental qualification margin standpoint. The method chosen for selection of SSCs important to aging and environmental challenges is known as Top Event Prevention (TEP) or Prevention Analysis. TEP is a Boolean method for optimal selection of SSCs (that is, those combinations of SSCs both necessary and sufficient to meet a predetermined selection criterion) and allows demonstration that plant-level safety can be maintained by the collection of selected SSCs alone.
Risk-informed Prioritization of Modernization Activities Using Ageing PSA Model
Shahen Poghosyan and Armen Amirjanyan
Nuclear and Radiation Safety Center, Yerevan, Armenia
Nuclear Power Plant modernization is a continuous process, which is aimed to reduce risk as low as reasonably achievable. Modernization process is especially important for old design NPPs to keep them in compliance with current safety standards. In addition, modernization process is important for plants where ageing is becoming more and more significant factor in regard with equipment reliability. Development of modernization program requires not only listing the issues to be addressed, but also to come up with common understanding of importance of proposed measures and their priority. Traditionally prioritization of modernizations is mainly done using deterministic considerations. Meanwhile parallel application of PSA models allow to come up with numerically justified and optimal solutions. Incorporation of ageing aspects in PSA model provide additional information for modernization prioritization in regard with plant’s components ageing perspective. This paper describes a feasibility study aimed to use integrated risk-informed decision-making principles for prioritization of modernizations. Paper discusses proposed approach for prioritization, which implies combination of probabilistic and deterministic indicators. In addition, paper discusses comparative analysis of results obtained using base case PSA model and ageing PSA model.
The Reliability Effects of Transient-Induced Degradation on the Performance of Large Power Transformers
Brittany L. Guyer (a), Carl R. Grantom (b), and Michael W. Golay (a)
a) Massachusetts Institute of Technology, Cambridge, MA, USA, b) CRG LLC, West Columbia, TX, USA
Increased knowledge of the effects of severe operational transients on component reliability, in combination with currently used mechanistic component degradation models, could augment the predictive capability of reliability modeling. A new component reliability model has been developed that considers the effects of both types of degradation. An application of the new model was sought in order to provide insight into both the sources and consequences of severe component transients and how these considerations can be formulated into a new framework for component aging management supporting component reliability programs. The large power transformer was selected for demonstration of this new reliability model. This component was selected as it is a component that has failed prematurely, has experienced strong transients during its operational lifetime, data are available about the important effects that the occurrence of strong transients have had on this component, and the transients experienced have resulted in effects that are not readily repairable (i.e., requiring component replacement). In this work, a strategy is proposed for the development of a physics-of-failure model of large power transformers that could be implemented in order to make more realistic performance predictions, supporting improved long-term plant asset management.
M12 Maintenance Modelling and Optimisation I
Chair: Cristiano Cavalcante, UFPE -Universidade Federal de Pernambuco, Brasil
Risk-Informed Simulation Optimization for Engineering Asset Management
Jérôme Lonchampt, William Lair
EDF R&D, Chatou, France
This paper present a general method coupling genetic algorithms and Monte-Carlo simulation to address simulation optimization issues in the field of engineering asset management. After a description of the method, parameters tuning issues are analyzed through a test-case.
A Usage-Informed Preventive Maintenance Policy to Optimize the Maintenance Free Operating Period for Multi-Component Systems
Romain Lesobre (a,b), Keomany Bouvard (a), Christophe Bérenguer (b), Anne Barros (c), Vincent Cocquempot (d)
a) Volvo Group Trucks Technology, Advanced Technology and Research, Saint Priest Cedex, France, b) Laboratoire Grenoble Image Parole Signal Automatique, Gipsa-Lab, Grenoble INP, UMR 5216 CNRS,Saint Martin d’Hères, France, c) Laboratoire de Modélisation et Sûreté des Systèmes, UTT, Institut Charles Delaunay, UMR 6279 CNRS, Troyes Cedex, France, d) Laboratoire d’Automatique, Génie Informatiqueet Signal, Université Lille1, UMR 8219 CNRS, Villeneuve d’Ascq Cedex, France
This paper deals with the concept of Maintenance Free Operating Period (MFOP). This MFOP is defined as a period of operation during which the system should be able with a given level of confidence to carry out all its assigned missions without system fault or performance limitation. Based on this concept, a dynamic maintenance policy for a multi-component system is implemented. The main objective of this paper is to propose a method to integrate the usage information of the system components in order to optimize the implemented policy. The method is evaluated considering the Total Maintenance Cost (TMC) value.
Cristiano Cavalcante, Marcelo Alencar, Adiel Almeida, Ana Paula Costa, Rodrigo Ferreira (a), Maxwell Luna, Rogério Sá, Alison Ferreira and Adilson Vieira (b)
a) UFPE -Universidade Federal de Pernambuco, Recife, Brazil, b) CELPE -Companhia Energética de Pernambuco, Recife, Brazil
We observe that in electrical substations, issues often arise that directly influence the requirements for maintenance actions to be adequate. Maintenance policies are sometimes inappropriate because the aging of assets has been incorrectly evaluated, because technological upgrades are not properly reflected in maintenance plans, or because the operational regime is not taken into account. Thus, once the need for adjustments because of the presence of one or more of the issues mentioned above has been identified, it is essential that a different systematic be implemented to achieve the expected performance of the affected substation. Accordingly, this article proposes a model for establishing adequate maintenance policies to produce more effective results, taking into account not only the possible consequences of failure to which the system under study is subject but also the various specific concerns associated with the performance indices of the electricity system. A real electrical substation is used as a pilot system.
A Stochastic Production Planning Optimization for Multi Parallel Machine under Leasing Contract
Medhioub Fatma, Hajej Zied, and Rezg Nidhal
LGIPM-University of Lorraine, Metz, France
In this paper, we aim at optimizing the production planning. The problem consists on a several identical machines mounted in parallel and which are leased depending on a fluctuating demand over a finite time horizon under given service level. The objective of the production plan is to determine the best combination of leased machines numbers, production time (or level) and inventory levels, by developing a mathematical model, that minimize the average total costs over the finite time horizon. The contribution and newness of this work is that it treats this approach under new constraints related especially to leasing techniques and consequently we assume that the number of workstations varies from a production period to another. This characteristic is due to the leasing principle as well as to the fluctuating demand that we have to take into account. A numerical example confirms the analytical study.
Review of the Preventive Maintenance Requirements for the Safety Systems of the Mochovce NPP
Zoltan Kovacs, Robert Spenlinger
RELKO Ltd., Bratislava, Slovak Republic
A requirement to optimize the Preventative Maintenance (PM) tasks assigned to specified safety systems has been identified at Mochovce Nuclear Power Plant (NPP). RELKO Ltd has been tasked with optimising the PM tasks via application of the Reliability-Centered Maintenance (RCM) and PSA methodology. This paper details the results of the RCM analysis performed on the Core Cooling Systems. It is concluded that the PM tasks assigned to the Core Cooling Systems were, in the main, based upon the original equipment manufacturers’ (OEM) recommendations. Following the accumulation of about ten years of operating and maintenance experience it was concluded that many of the current task types and task frequencies required major revision in order to maintain the optimum levels of both reliability and availability of the Core Cooling Systems. It is also concluded that in several cases, specific components within the Core Cooling Systems will benefit from a shift in maintenance strategy from fixed interval invasive routines to a predictive maintenance (PdM) based strategy. Such a strategy will ensure close monitoring of system and component performance without compromising nuclear safety or availability. It is recommended that the Mochovce NPP replaces the current maintenance catalogue assigned to the Core Cooling Systems with new PM tasks detailed in the paper. In addition, the paper presents the impact of changes on CDF and LERF after implementation of the new PM tasks.
M13 Occupational Safety and Management
Chair: Thomas Wold, Norwegian University of Science and Technology
An Integrated Management for Occupational Safety and Health throughout the Plant-Lifecycle
Yukiyasu Shimada (a), Teiji Kitajima (b), Tetsuo Fuchino (c), and Kazuhiro Takeda (d)
a) National Institute of Occupational Safety and Health, Japan, Kiyose, Tokyo, Japan, b) Tokyo University of Agriculture and Technology, Koganei, Tokyo, Japan, c) Tokyo Institute of Technology, Meguro, Tokyo,Japan, d) Shizuoka University, Hamamatsu, Shizuoka, Japan
The main purposes of occupational safety and health (OSH) management are to assure safe and healthful working conditions for working men and women and to prevent industrial accidents by the establishment of process safety management (PSM) system in the company level as well as the improvement of safety engineering techniques. Business process model has been developed to systematize the engineering activities and information flow throughout a plant-lifecycle (i.e. from research and development through process/plant design, construction and active manufacturing period, including production and maintenance) of chemical processes. This paper proposes an integrated approach for OSH management based on the business process model of engineering activities. This approach consists of three level hierarchical PSM; 1) PSM framework at enterprise-level, 2) HSE (Occupational Health, Process Safety, and Work Environment Protection) management activities at middle-management-level, and 3) SQDC-conscious tasks at manufacturing-site-level. Hierarchical integration of the PSM at each level makes it possible to realize the consistent and collaborative OSH management.
End User Involvement in the Development of Procedures and Safety Management Systems
Thomas Wold and Karin Laumann
Department of Psychology, Norwegian University of Science and Technology (NTNU). Trondheim, Norway.
IT-based Safety Management Systems contains procedures, safety standards, checklists and descriptions on how different tasks should be performed, and are usually designed at an executive level in the organization, and then communicated to the lower level in the organization where they are being applied. This paper presents data collected from qualitative interviews with executives and operators from two companies in the gas and petroleum industry. The executives generally regard Safety Management Systems as important tools for all work in hazardous environments, while the operators weren’t that enthusiastic. How can end user involvement in the development phase of procedures and Safety Management System improve use? A central argument is that Human Factors must be involved as early as possible in the development phase, and that operators need to understand the purpose of the management system in order to use it as intended. The informants that had been involved in the development of the procedures at least to some extent, felt an ownership to the management system, while the ones who hadn’t been involved at all felt no ownership to the management system, and did not see the purpose of it.
Identifying Requirements for Effective Human-Automation Teamwork
Jeffrey C. Joe (a), John O’Hara (b), Heather D. Medema and Johanna H. Oxstrand (a)
a) Idaho National Laboratory, Idaho Falls, ID, USA, b) Brookhaven National Laboratory, Upton, NY, USA
Previous studies have shown that poorly designed human-automation collaboration, such as poorly designed communication protocols, often leads to problems for the human operators, such as: lack of vigilance, complacency, and loss of skills. These problems often lead to suboptimal system performance. To address this situation, a considerable amount of research has been conducted to improve human-automation collaboration and to make automation function better as a “team player.” Much of this research is based on an understanding of what it means to be a good team player from the perspective of a human team. However, the research is often based on a simplified view of human teams and teamwork. In this study, we sought to better understand the capabilities and limitations of automation from the standpoint of human teams. We first examined human teams to identify the principles for effective teamwork. We next reviewed the research on integrating automation agents and human agents into mixed agent teams to identify the limitations of automation agents to conform to teamwork principles. This research resulted in insights that can lead to more effective humanautomation collaboration by enabling a more realistic set of requirements to be developed based on the strengths and limitations of all agents.
Characterization of Resilience in Nuclear Power Plants
Florah Kamanja (a), and Kim Jonghyun (b)
a) Kenya Electricity Generating Company, Nairobi, Kenya, b) KEPCO International Nuclear Graduate School, Ulsan, South Korea
An emergency operation system in a nuclear power plant consist of operators, human-machine interface, procedures, and the interactions among these elements working together to respond to incidents. The complexity of dynamic systems such as nuclear power plants poses a challenge for safety as it can be a source of deviations from normal behavior during system operation. NPP control rooms consist of many elements that result in complex interactions between them. Resilience is the ability of a system to recover from a disturbance, so that it can sustain required operations under both expected and unexpected conditions. Nuclear power plants must anticipate the operating risks caused by either the hardware, human, or organizational failures in order to be resilient. The ability of NPPs to monitor the current status of the system, anticipate possible problems, react appropriately to events, and learn from past incidents is a measure of success hence the resilience. Although the significance of resilience has been stressed in the literature, there is a lack of adequate literature attempting to analyze system resilience. To achieve a practical an insightful understanding of the EOS resilience complexity, this paper aims at characterizing resilience attributes based on the existing literature.
M14 Operational Experience and Data Analysis
Chair: Shawn St. Germain, Idaho National Laboratory
Recent Insights from the International Common Cause Failure Data Exchange (ICDE) Project
Albert Kreuser (a), Gunnar Johanson (b)
a) Gesellschaft für Anlagen-und Reaktorsicherheit(GRS) mbH, Cologne, GERMANY, b) ES konsult, Solna, SWEDEN
Common-cause failure (CCF) events can significantly impact the availability of safety systems of nuclear power plants. In recognition of this, the international CCF data exchange (ICDE) project was initiated in 1994. The objectives of the ICDE project are: to provide a framework for a multinational co-operation; to collect and analyze CCF events over the long term so as to better understand such events, their causes, and their prevention; to generate qualitative insights into the root causes of CCF events which can then be used to derive approaches or mechanisms for their prevention or for mitigating their consequences; to establish a mechanism for the efficient feedback of experience gained in connection with CCF phenomena, including the development of defenses against their occurrence, such as indicators for risk based inspections; and to record event attributes to facilitate quantification of CCF frequencies when so decided by the member countries of the Project. Until January 2014, 1346 ICDE events had been analyzed and reported in public OECD/NEA reports This paper presents recent activities and lessons learnt from data collection on Control Rod Drive Assemblies and Heat Exchangers and on cross-component analysis on events which were due to external factors.
Internal Flooding According to EPRI Guidelines – Detailed Electrical Mapping at Ringhals
Per Nyström, Carl Sunde (a), and Cilla Andersson (b)
a) Risk Pilot, Gothenburg, Sweden, b) Ringhals AB, Varberg, Sweden
Eleven different tasks should be executed according to the EPRI guidelines for performing internal flooding PSA. Task 2 deals with identification of flood sources/mechanisms as well as with Systems, Structures and Components (SSCs). In this task it is briefly mentioned that not only the main components such as pumps and valves can be affected by flooding but also associated components such as circuit breaker, junction boxes and instrumentation and control circuitry are affected. It is fairly easy to locate the main components as well as the impact of flooding on these components. However it is more difficult to make a detailed mapping of the cable routing and the electrical dependencies (at Ringhals called electrical mapping) for the main components. This paper describes how this type of work is being executed and documented at Ringhals NPP in Sweden.
NRC Reactor Operating Experience Data
Shawn Walter St. Germain
Idaho National Laboratory, Idaho Falls, USA
Idaho National Laboratory (INL) has been providing technical assistance to the U.S. Nuclear Regulatory Commission Division of Risk Analysis in the Office of Nuclear Regulatory Research in the areas of data collection and reliability and risk calculation. INL collects, codes, assures the quality of, and maintains all reactor operating experience data necessary to support the Industry Trends Program and various risk-associated NRC studies requiring reactor operating experience data. The types of data collected under this effort include initiating event data, system reliability data, loss of offsite power data, common cause failure data, fire event data, and shutdown initiating event data. The data sources for this effort primarily consists of Licensee Event Reports (LERS), Event Notifications, and equipment failure reports provided by the Institute for Nuclear Power Operations (INPO). This data is analyzed and results published annually on the NRC website. The data is primarily used to support the NRC's standardized plant analysis risk (SPAR) models but also provides generic industry average values for use by the industry in their individual PSA models. This paper characterizes the types of data collected, the various uses of this data, and the methods of collection, storage and retrieval.
Component Reliability in the T-Book – The New Approach
Anders Olsson, Erik Persson Sunde, and Magnus Gudmundsson
a) Lloyd's Register Consulting, Stockholm, Sweden, b) TUD Office, Vattenfall, Stockholm, Sweden
T-Book is a reliability data handbook for use in Nordic Nuclear PSAs (Probabilistic Safety Assessments). Due to its ambitious scope, high level of detail, and high QA standard, it has become world-famous, and is frequently used even outside the nuclear field. Since 2008, Lloyd's Register Consulting, on behalf of the Nordic PSA Group (NPSAG) and TUD (the editor of T-Book), has performed a series of projects to enhance and consolidate the process, right from the classification and sampling of data, through parameter assessment, PSA modeling, and up to the final interpretation of results. Two aspects have proven to be of particular interest. Firstly, providing more homogeneous groups of T-Book components, which will have positive impact on PSA in terms of less conservative and more precise parameters, as well as increased consistency in the entire modeling process. Secondly, the benefits of said homogenization need to be weighed against the use of the multi-parametric model for standby components, because these two aspects are not fully compatible. A comprehensive approach, addressing both these aspects, is presented for selected components: pumps, batteries, diesel generators, and motor operated control valves. In this paper, the background and motives for the proposed strategy will be outlined, as well as the "tool box" to put it into practice. The presentation will also include what has been accomplished during 2013, and what is going to be introduced in the new version of the T-Book.
Trend Analysis of Input Data to Nordic PSA
Ostrovskii Dimitri (a), Lindahl Pär (b)
a) ÅF consulting, Gothemburg, Sweden, b) OKG AB, Oskarshamn, Sweden
In Swedish and Finnish NPPs the ”T-book” is one common source for reliability parameters used in PSA. These parameters are calculated based on the assumption that component reliability does not change with time. This assumption is e.g. violated if components degrade, due to ageing effects, or improve, due to improvements in maintenance strategies. It is thus relevant to ask how PSA results may be affected by the time independence assumption. To approach an answer, a non parametric test method, the Wilcoxon rank sum test, was used to analyze how observations of malfunctions deviate from a Poisson-distributed set that represents the case when malfunctions are independent of each other and occur with a constant frequency. It was found that deviations from the Poisson distribution, trends, can be detected in the gathered data, and that corrections for reliability trends may affect the PSA results significantly.
M15 Phenomena Modeling
1:30 PM Waianae
Chair: Nat Heatwole, University of Southern California
Preparation of Implementation Standard Concerning Severe Accident Management in Nuclear Power Plants
Shinya Kamata (a), Koji Okamaoto (b), and Tomoyuki Sugiyama (c)
a) Japan Nuclear Safety Institute, Minato-ku, Tokyo, Japan, b) The University of Tokyo, Tokai-mura, Naka-gun, Ibaraki, Japan, c) Japan Atomic Energy Agency, Tokai-mura, Naka-gun, Ibaraki, Japan
The Great East Japan Earthquake with a magnitude of 9.0 (The 2011 off the Pacific coast of Tohoku Earthquake) occurred on March 11, 2011, and the beyond design-basis tsunami descended on the Fukushima Daiichi Nuclear Power Plant by the earthquake. Eventually, the core cooling systems of the units 1, 2 and 3 could not operate stably, they all suffered severe accident, and hydrogen explosions were triggered in the reactor buildings of units 1, 3 and 4. In the light of these circumstances, Atomic Energy Society of Japan (AESJ) decided to establish a standard that consolidates the concept of maintaining and improving severe accident management. The standard also provides technical requirements for renovation and addition of the equipment, the formulation of procedures, and strategies. All these items enable the minimization of risks so as to prevent severe accidents, or otherwise enable the mitigation of impacts of severe accidents once occurred.
EPRI Fukushima Technical Evaluation—Evaluation of Flammable Gas Leakage from Fukushima Daiichi Containments using the MAAP5 Computer Code
David L. Luxat, Donald A. Dube, Andrew S. Dercher (a), Richard Wachowiak, Rosa Yang (b), and Jeff R. Gabor (a)
a) ERIN Engineering and Research, Inc., West Chester, PA, USA, b) Electric Power Research Institute, Palo Alto, CA, USA
This paper presents initial results from the investigations of flammable gas transport from the Units 1, 2 and 3 containments into their respective reactor buildings. This study is being conducted as part of the Phase 2 effort of the EPRI Fukushima Technical Evaluation, which is an extension of Phase 1 evaluation (Reference ). It builds upon the existing event evaluations conducted by TEPCO (References  and ) and Sandia (Reference ). The analyses are conducted using EPRI’s Modular Accident Analysis Program (MAAP), version 5.01. The analyses identify the potential for high temperature conditions in the drywell head region of Units 2 and 3 to contribute to the onset of leakage from each drywell—at drywell pressures below twice design. It is not likely that high temperatures in the drywell head region developed at Unit 1 prior to the onset of leakage from the drywell head flange (at about twice design pressure). The leakage at all units through the drywell head flange has been found to enhance the build-up of flammable gases on the refuel floor. Unit 1 may have experienced flammable conditions on its refuel floor for 10 hours prior to the combustion event. Unit 2 likely did not develop flammable conditions on its refuel floor due to the open blowout panel. At Unit 3, leakage from the hard pipe vent into the Standby Gas Treatment System soft ducting may have allowed hydrogen to build-up at lower elevations—this could have contributed to more damage to the reactor building structure.
Prediction of Complex Thermal-Hydraulic Phenomena Supplemented by Uncertainty Analysis with Advanced Multiscale Approaches for the TALL -3D T01 Experiment
Angel Papukchiev (a), Marti Jeltsov (b), Clotaire Geffray (c), Kaspar Kööp, Pavel Kudinov (b), Rafael-Juan Macián (c) and Georg Lerchl (a)
a) Gesellschaft fuer Anlagen-und Reaktorsicherheit (GRS) mbH, Garching n. Munich, Germany, b) KTH Royal Institute of Technology, Stockholm, Sweden, c) Technische Universitaet Muenchen (TUM), Garchingn. Munich, Germany
The thermal-hydraulic (TH) system code ATHLET was coupled with the commercial 3D computational fluid dynamics (CFD) software package ANSYS CFX to improve ATHLET simulation capabilities for flows with pronounced 3D phenomena such as flow mixing and thermal stratification. Within the FP7 European project THINS (Thermal Hydraulics of Innovative Nuclear Systems), validation activities for coupled thermal-hydraulic codes are being carried out. The TALL-3D experimental facility, operated by KTH Royal Institute of Technology in Stockholm, is designed for thermal-hydraulic experiments with lead-bismuth eutectic (LBE) coolant at natural and forced circulation conditions. No tests have been performed up to now. GRS carried out pre-test simulations with ATHLET – ANSYS CFX for the TALL-3D experiment T01, while KTH scientists perform these analyses with the coupled code RELAP5/STAR CCM+. In the experiment T01 the main circulation pump is stopped, which leads to interesting thermal-hydraulic transient with local 3D phenomena. In this paper, the TALL-3D behavior during T01 is analyzed and the results of the coupled pre-test calculations, performed by GRS (ATHLET-ANSYS CFX) and KTH (RELAP5/STAR CCM+) are directly compared. Moreover, this work is supplemented by uncertainty and sensitivity analysis for the T01 experiment, carried out at the Technische Universitaet Muenchen.
Cost-Effectiveness of Vehicle Barriers and Setback Distance for Protecting Buildings from Vehicle Bomb Attack
University of Southern California, Los Angeles, USA
Decision-making regarding implementing measures to protect buildings from vehicle bomb attack is often undertaken using highly judgment-based risk processes. This paper presents a quantitative risk-cost model for using vehicle barriers to create setback distance around a new office building. The model explicitly considers both the attack probability, and the damages in the event of an attack (both target building and collateral), as well as how both of these might change as mitigation measures are implemented. The attack damages are assessed using a new empirical blast model, which adapts the estimation methods used by the U.S. Geological Survey for earthquake damages, and is based on data from three well-studied vehicle bomb attacks. Monte Carlo simulation is used to carry the uncertainty in the inputs through to the final results. The model outputs are the mitigation costs, the attack damages, the “breakeven” attack probability (at which the benefits of the mitigation justify its costs), and the cost per statistical life saved (assuming an attack). The results suggest that this mitigation option is cost-effective only when the attack probability (for the case without the mitigation measures present) is rather high.
M16 Policy Making and Legislative Issues
From Prescriptive Arrival Times to Performance Based Fire Service Delivery – Parallels of Fire Service Planning and Fire Engineering
Adrian Ridder, Uli Barth
University of Wuppertal, Wuppertal, Germany
The fire safety design process of buildings underwent a substantial shift in the last roughly two decades, switching from prescriptive building codes to performance-based, fire-engineered designs. A similar process can be observed with Strategic Fire Service Planning which defines “how much fire service” is necessary per municipality. The methods used there become more and more sophisticated as well. However, with increasing complexity it becomes harder to explain and interpret results to the decision-makers, which applies both to fire engineering and fire service planning. The need for further research is made clear as the major outcome of this paper.
Issues in Incorporating Probabilistic Safety Assessment (PSA) in the Design and Licensing Stages of Generation IV Reactors
Ibrahim A. Alrammah
School of Mechanical, Aerospace and Civil Engineering (MACE), University of Manchester, Manchester, United Kingdom
Probabilistic approaches has been used and are also highly recommended to be used from the very early stage of the reactor design process. So far, Probabilistic Safety Assessment (PSA) approach is increasingly being utilized in the demonstration of safety in combination with deterministic approaches (e.g. to justify the classification of situations, to determine the sequences of sophisticated failures) and used also to verify the systems and components reliability in order to satisfy safety targets. However, epistemic problems such as uncertainties due to lack of design information, unknown phenomena, plant-specific hazards, data, etc., are larger than that from existing reactors, and will impose a significant challenge to the decision makers. This paper will discuss some technical issues related to applying PSA in the design and licensing stages of Generation IV reactors. These aspects include: initiating events, passive systems modeling, reliability data, common cause failure (CCF), modeling of novel design features, modeling of preventive maintenance, technical specifications, human reliability analysis (HRA), systems interdependencies, modeling of instrumentation and control (I&C), external hazards, continuous design risk monitoring, supporting studies, interpretation of PSA results for new plants.
Need for PRA in the Oil and Gas Industry
Matt Johnson, Nicholas Lovelace (a), and Michael Lloyd (b)
a) Hughes Associates, Inc., Lincoln, NE, USA, b) Risk Informed Solutions Consulting Services, Ball Ground, GA, USA
Probabilistic Risk Assessment (PRA) is widely used in the nuclear industry to assess the risk from hazards to nuclear power plants. This paper discusses the application of PRA methods to the oil and gas industry, and, specifically, to assessing production platform safety and optimizing levels of hydrocarbon production. Oil and gas platform safety can be analyzed with a focus on potential loss of life to platform workers from internal hazards such as uncontained liquid or gas hydrocarbon releases with subsequent ignition. Additionally, platform production capabilities can be analyzed with a focus on reducing production downtimes. PRA methods can be effectively utilized to identify both safety and operating issues for typical platform alignments, maintenance and testing frequencies, and prioritization of enhancements to platform operation.
Learning how to Learn from Failures: The Case of Fukushima Nuclear Disaster
University of Portsmouth, Portsmouth, United Kingdom
In this work, it is argued that learning from failures and safety competence should be an important part of the curriculum of Engineering and Management students. The case of Fukushima will be used to illustrate how to learn about learning from failures using multi-models inspired by reliability and risk analysis in order to investigate disasters. This type of analysis can offer richness to our understanding of the root causes and provide insight into policy making and support decisions for resource allocations for prevention of such disasters. The analysis is based on a workshop related to learning from failures where students and practitioners were first given a brief about the related theory of reliability analysis and decision science, followed by introduction of the analytical techniques that can be used (such as FTA, RBD and AHP). They were then given a brief in the form of a narrative of the accident from investigation reports, and they were then divided into small groups with the task to perform an analysis of the disaster followed by presentation of recommendations in the form of a written report and an oral presentation. Finally, a set of generic lessons and recommendations are provided in order to prevent future system failure.
Toward Demonstrating the Monetary Value of Probabilistic Risk Assessment for Nuclear Power Plants
Marzieh Abolhelm, Justin Pence, Zahra Mohaghegh (a), and Ernie Kee (b)
a) University of Illinois at Urbana-Champaign, IL, USA, b) YK.risk, LLC, TX, USA
Inefficiencies in the operation and maintenance of Nuclear Power Plants (NPPs) have caused unnecessary shutdowns, decreases in production, and increases in system risk. Probabilistic Risk Assessment (PRA), which guides risk-informed decision-making, helps expand the operational envelope by allowing more flexibility, adding to the efficiency of preventive and corrective actions and, therefore, generates more profit. However, the financial bottom line of PRA has not yet been formally estimated. This paper reports on the current status of first-of-its-kind research for estimating the monetary value of PRA. The proposed steps for this research include: (1) developing a Generic Financial Model (GFM) to estimate the Return On Investment (ROI) that results from profit generation or cost reduction associated with a typical PRA activity in an NPP (2) implementing GFM for one of the PRA programs and validating GFM, (3) conducting uncertainty quantification for the estimated ROI from Step 2, (4) identifying existing PRA programs at an NPP (i.e., South Texas Project Nuclear Operating Company; STPNOC), (5) obtaining ROI for all the PRA activities of STPNOC, running uncertainty analysis for the total ROI, providing a probabilistic monetary value of PRA, and (6) applying importance measure and sensitivity analyses to propose improvement approaches for PRA activities.
M17 Low-power and Shutdown
Chair: Stefan Eriksson, Ringhals AB
A methodology for determining of Plant Operating States of Low Power Shutdown Probabilistic Safety Assessment for the Next-Generation Nuclear Power Plants
Jae Gab Kim (a), Kwang Nam Lee (b), Hak Kyu Lim (a)
a) KEPCO-ENC, Integrated Engineering Department, Korea, b) KEPCO-ENC, Power Engineering Research Institute, Korea
This paper outlines the Low Power Shutdown (LPSD) Probabilistic Safety Assessment (PSA) portion of a methodology for the determination of the Plant Operating States (POSs). This is to determine how best to characterize them for inclusion into the LPSD PSA. The characterization of POS will begin a review of available shutdown PSA studies for current generation plants. The next-generation Nuclear Power Plants (NPPs) provide useful references for POS development. Several sets of current and next-generation NPPs including NUREG/CR-6144 of Surry Unit 1 shutdown PSAs have been reviewed to identify potential POS. The POS defined for the next-generation NPP PSA must represent all conditions that can occur over the course of a fuel cycle. This paper considers all plant conditions except full power operation which is addressed with the internal events PSA. The development of POSs can lead to group plant states that require similar equipment, timing, and operator action to respond to an upset condition. POS Grouping is based on Technical Specifications (TS) requirement as well as key factors associated with the main shutdown risk contributors like RCS temperature, RCS pressure, RCS inventory, State of RCS pressure boundary, and Decay heat levels.
Shutdown PSA for Ringhals NPP Unit 1. Insights, Overview and Results
Stefan Eriksson, Marie Gryte (a), and Erik Cederhorn (b)
a) Ringhals AB, Väröbacka, SWEDEN, b) Risk Pilot, Stockholm, SWEDEN
During 2011, 2012 and 2013 a Shutdown PSA (SPSA) has been developed for Ringhals NPP unit 1. Ringhals 1 is a Boling Water Reactor (BWR) made by ASEA-Atom situated at the West coast of Sweden. The SPSA supplement the existing PSA Level 1 and 2 for Ringhals 1 and the final outcome will give a complete risk profile for the unit, providing support for verification of plant safety and upgrades. This paper gives an overview of the level 1 SPSA. A description is made of the basic conditions for identification of Plant Operating States (POS), analysis of initiating events, sequence analysis and system analysis. The result for level 1 SPSA of R1 is briefly discussed.
Developing a Low Power/Shutdown PRA for a Small Modular Reactor
NuScale Power, LLC, Corvallis, OR, USA
A growing area of interest in the field of nuclear risk analysis is the application of PRA techniques to low power and shutdown configurations when the availability of systems and components may differ significantly from normal operation. Many operating plants have performed (or are in the process of performing) a PRA for low power operations, and new reactor designs are required to complete one as part of the design certification process. NuScale Power is developing a natural-circulation small modular reactor, and certain features of the design require refueling and maintenance procedures different from any in the industry. This uniqueness eliminates some sources of risk traditionally addressed in a shutdown PRA, but also introduces entirely new areas of risk. One major challenge is that all modules in the plant share a common refueling area, so each module must be lifted and moved from its operating location with fuel in the core. The module is completely disconnected and most systems credited in the full power PRA are unavailable when the module is in transit. This paper will give an overview of NuScale’s design and refueling process and discuss some of the challenges involved with developing a shutdown PRA for a reactor that is designed to be moved with fuel assemblies in place. Special attention is paid to determining a failure probability for a singlefailure-proof crane with little directly applicable publicly available data.
Risk-Informed Design Changes of an Advanced Reactor in Low Power and Shutdown Operation
Ji-Yong Oh, Ho-Rim Moon, Han-Gon Kim and Myung-Ki Kim
Korea Hydro and Nuclear Power Co. Ltd, Central Research Institute, Deajeon, Korea
APR+ has been developed in Korea since 2007. APR+ adopts various advanced safety features including passive auxiliary feedwater system, four emergency diesel generators. Through the implementation of the advanced designs, APR+ increased the safety to the world best level of evolutionary reactors. The full power core damage frequency or containment failure frequency decreased significantly comparing to APR1400 that is base model of APR+. However, low-power shutdown risk has not been improved substantially. This paper suggests several design changes that optimize low-power shutdown risk. Based on the design alternatives, this paper discusses risk effectiveness of the proposed design including various factors, e.g. equipment reliability, human error, training, procedure and so on.
An Implementation Strategy of Low Power Shutdown PSA for KHNP NPPs
Jang-Hwan Na, Seok-Won Hwang, Ho-Jun Jeon
Central Research Institute of Korea Hydro & Nuclear Power Co.,Ltd, , Daejeon, Korea
Rightly after the Fukushima accidents, the Korean Regulatory Agency with the support from a group of academic and research experts evaluated the safety of Korean nuclear power plants including plants on construction. The expert group particularly focused on any possible design vulnerabilities in view of ultimate heat sinks and power sources considering external hazards such as seismic, flood or complex initiated events. They identified several common or plant-wise improvement factors and elicited 49 post-action items as near term Fukushima accident measures. One of the measures is to develop SAMG (Severe Accident Management Guideline) during LPSD (Low Power and Shutdown) operation in addition to the existing SAMG on the full power operation. At first, KHNP (Korea Hydro & Nuclear Power) decided to develop the LPSD PSA (Probabilistic Safety Assessment) models to increase the quality of LPSD SAMG. To get a technical adequacy, KHNP decided to revise the full spectrum of existing PSA models including full power, external or level 2 PSA incorporating up-to-date reliability data and methodologies. This paper presents an implementation strategy of developing LPSD PSA models including the status of upgrading full power PSA models at the end of 2013.
M21 Reliability Analysis and Risk Assessment Methods I
3:30 PM Honolulu
Chair: Kaushik Chatterjee, FM Global
Proof Testing of Safety-Instrumented Systems: New Testing Strategy Induced by Dangerous Detected Failures
Yiliu Liu, Marvin Rausand
Department of Production and Quality Engineering, Norwegian University of Science and Technology, Trondheim, Norway
Some dangerous failures of safety-instrumented systems (SISs) are detected almost immediately by diagnostic self-testing, whereas other dangerous failures can only be detected by proof-testing. The first type is called dangerous detected (DD) failures and the second type is called dangerous undetected (DU) failures. Proof tests are usually carried out at constant time intervals. DD-failures are repaired almost immediately whereas a DU-failure will persist until the item is proof-tested. Many items can have a DU-and a DD-failure at the same time. After the repair of a DD-failure is completed, the maintenance team has two options: to perform an “insert” proof test for DU-failure or not. If an insert proof test is performed, it is necessary to decide whether the next scheduled proof test should be postponed or performed at the scheduled time. This paper uses Petri nets to model the proof test strategies after DD-failures and to analyze the effects of the different strategies on the SIS performance. It is shown that insert proof tests reduce the unavailability of the system, whereas the adjustment (or not) of the test schedule does not have any significant long term effect.
A New Interfacing Approach between Level 1 and Level 2 PSA
Nicolas Duflot, Nadia Rahni, Thomas Durin, Yves Guigueno and Emmanuel Raimond
IRSN, Fontenay aux Roses, France
IRSN (TSO of the French Nuclear Safety Authority) has been developing L2 PSAs for many years, using its own probabilistic tool, KANT (probabilistic event trees software) associated to a very fast-running source term code (MER). Since the IRSN L1PSAs event trees are developed with one other dedicated software, the L1-L2 PSA interface methodology is a key and difficult point of the IRSN PSA methodology. In the previous versions of the IRSN PSAs, L1-L2 PSA interface was a mostly manual process, resulting in significant resources allocation. To cope with such a difficulty, a new interfacing approach, allowing computerized generation of plant damage states (PDSs), has been developed. This approach is based on the introduction of flag events (basic events with a probability of one) into the L1PSA minimal cut sets (MCSs) in order to transfer information related to front lines systems (needed for accident management) status and operators actions. Afterwards, the MCSs are filtered to identify automatically the different PDSs of the L1-L2 PSA interface using a new dedicated tool. The automatic PDS generation allows implementing a very detailed L1-L2PSA interface easy to update. Since this new IRSN interfacing approach is based on fault trees only, it can be implemented with most of the level 1 PSA tools.
An Approach to Ensure the Availability of Complex Systems
Kaushik Chatterjee, Kumar Bhimavarapu, Robert Kasiski, and William Doerr
FM Global, Norwood, MA, USA
Availability of a system depends on: (1) the components’ reliabilities; and (2) the Inspection, Testing and Maintenance (ITM) characteristics (i.e., inspection/testing frequency, repair/replacement duration, and maintenance restoration factor). Complex systems typically have several sub-systems and components with complicated interactions and dependencies. In order to ensure a desired availability of a safety-critical complex system such as a fire protection system throughout its lifetime, it is necessary to: (1) ensure the needed reliability of the critical components through carefully planned durability/life tests; and (2) perform ITM actions at appropriate intervals (or frequencies). This paper presents a comprehensive approach to: (1) establish the reliability targets and the ITM frequencies for the critical components based on the desired availability of the system; and (2) estimate durability/life test duration and sample size requirements based on the established reliability targets for these critical components. The steps of the comprehensive approach have been demonstrated using a typical foam-water sprinkler protection system. The comprehensive approach, when applied to a safety-critical complex system, would help achieve the desired availabilities of the critical components, which in turn would ensure the desired availability of the system throughout its lifetime.
Reliability/Availability Methods for Subsea Risers and Deepwater Systems Design and Optimization
Annamaria Di Padova (a), Fabio Castello b), Fabrizio Tallone (a), Michele Piccini (b)
a) Saipem S.p.A., Fano, Italy, b) RAMS&E S.R.L., Turin, Italy
The restriction of construction licenses for onshore oil/gas treatment plants and regasification units along with energy demand growth has increased the development of offshore installations. Furthermore the discover of new offshore deep water fields enhance the engineering efforts towards the development of engineering of submarine systems and plants. Due to the complexity of these submarine systems, the severe environment where they operate and the difficulty or the impossibility to repair a component, a high system availability is becoming a key requirement. In this framework, to have a system architecture verified also from the reliability and availability point of view, the RAM analysis are becoming an essential part of the design. This paper describes the application of reliability/availability methods (RBD, Montecarlo method, FMEA risk assessment) to support the design of subsea deep water systems. In particular, two case studies are presented, the first aiming at the definition of the optimum configuration of retrievable and permanent deep water modules, the second addressing the verification of design configurations and the suggestion of tests and inspection plans to guarantee system integrity along operating life. Moreover the paper summarizes also difficulties to find subsea equipment reliability data and proposes solutions for reliability components characterization.
M22 Dependent Failure Modeling I
3:30 PM Kahuku
Chair: Andrew O'Connor, Acuitas Reliability Pty Ltd
Statistical Analysis of Common Cause Failure Events Using ICDE Data
S. Yu, M.D. Pandey (a), S. Yalaoui and Y. Akl (b)
a) University of Waterloo, Waterloo, Canada, b) Canadian Nuclear Safety Commission, Ottawa, Canada
Analysis of Common Cause Failures (CCF) is an important element of the Probabilistic Safety Assessment (PSA) of systems important to safety in a nuclear power plant. Based on the conceptualization of the CCF event, many probabilistic models have been developed in the literature. This paper utilizes a modern method, called “General Multiple Failure Rate Model”, for the probabilistic modeling of CCF events. To estimate the parameters of the GMFR model, the Empirical Bayes (EB) method is adopted. A detailed case study is presented using CCF data for Motor Operated Valves (MOVs).
Extending the Alpha Factor Model for Cause Based Treatment of Common Cause Failure Events in PRA and Event Assessment
Andrew O’Connor, Ali Mosleh
Center for Risk and Reliability, University of Maryland, College Park, United States
Common Cause Failure modeling for Probability Safety Assessments has become standard practice in many industries. Of the numerous models proposed to include common cause, one of the most widely adopted has been the Alpha Factor Model, which is supported by the US Nuclear Regulatory Commission CCF database and software tools. The Alpha Factor Model (AFM) uses an empirical ratio between the independent failures and CCF failures to quantify the model parameters. While this has been advantageous in allowing the prediction of system reliability with little or no data, it has been limiting in other applications such as modeling the characteristics of a system design or including the characteristics of failure when assessing the risk significance of a failure or degraded performance event (known as an event assessment). This paper proposes a new CCF model called the Partial Alpha Factor Model (PAFM), which extends the AFM to allow the explicit modeling of coupling factors between components such as shared maintenance, or shared location. Using this more explicit modeling allows the model to be tailored depending on how far the system design defends against such dependencies. By using the principles of the AFM as the basis for this new model, its implementation may be feasible without modification to existing PRA software or significant changes in data collection requirements.
Estimating Common Cause Failure Probabilities for a PRA Taking into Account Different Detection Methods
Kalle E. Jänkälä
Fortum Power and Heat Oy, Espoo, Finland
The methodology to estimate residual parametric common cause failure (CCF) probabilities consists of the selection of the data source, source plants, source systems and component type, failure mode, assessment of the impact vectors, determination of equivalent observations, calculation of CCF rates of different multiplicities with uncertainties using an empirical Bayes estimation method and finally determining explicit CCF basic events and their probabilities to be used in the probabilistic safety assessment model. The CCF probabilities are obtained as the result of unavailability estimation accounting for different detection methods and corresponding outage times. Typically CCF events of safety system components are detected by tests during plant operation or during annual overhaul. In CCF quantification this is often regarded as the only way of detection. This leads into CCF unavailability quantification in which the CCF rate is based on all kinds of CCF events and the corresponding outage time is always determined by the test interval and testing scheme. This approach might be overly conservative or sometimes optimistic. This paper improves CCF unavailability estimation by taking into account monitoring and different kinds of tests and outage times and considering failure modes in the failure rate estimation.
Time Dependent Analysis with Common Cause Failure Events in RiskSpectrum
Pavel Krcal (a,b) and Ola Bäckström (a)
a) Lloyd's Register Consulting, Stockholm, Sweden, b) Uppsala University, Uppsala, Sweden
Testing of components with common cause failures presents a challenge to a realistic analysis of failure probabilities. In reality, the most commonly used testing scheme is staggered testing. Common Cause Failure (CCF) models in Probabilistic Safety Assessment (PSA) studies often assume a sequential testing scheme. This might be overly conservative if the actual testing scheme is staggered. Some software tools, e.g., RiskSpectrum, offer time dependent analysis where one can model testing of components in time explicitly. This paper deals with effects of different testing schemes on the quantification of CCF events in time dependent analysis. Determining which formulae shall be used by software tools in time dependent analysis requires an in-depth understanding of how to model effects of tests on the common cause parts of failures. We analyze assumptions which lie behind different ways of modeling tests of common cause failure events.
M23 Risk and Hazard Analyses I
3:30 PM Oahu
Chair: Ingrid Bouwer Utne, NTNU, Department of Marine Technology, Norway
A State of the Practice Investigation Guiding the Development of Visualizations for Minimal Cut Set Analysis
Yasmin I. Al-Zokari, Liliana Guzman (a), Barboros Can Conar (b), Dirk Zeckzer (c), Hans Hagen (a)
a) TU Kaiserslautern, Kaiserslautern, Germany, b) University of Applied Sciences, Kaiserslautern, Germany, c) Leipzig University, Leipzig, Germany
Minimal Cut Set (MCS) analysis is used for the qualitative and quantitative safety and reliability analysis of systems. While many studies concerning MCS computation in the safety domain are found, no study gives a complete and detailed description of the tasks performed by practitioners during MCS analysis. The goals of this study are (1) to elicit the context (including the tasks) of MCS analysis; (2) to obtain the requirements and needs of the safety analysts, and the tools used; (3) and to assess the quality of the tools from the point of view of the safety engineers regarding their (3a) representation, (3b) interaction, (3c) performance, and (3d) usability. We found that the main purpose is nding improvements to increase the hazard's safety. The main tasks are identifying critical basic events, the related system components, and single points of failure. The stakeholders are mainly decision makers and system engineers. The main requirements are nding single points of failure, determining MCS order, and nding basic events with high failure probability and related components. The results show that the usability of the tools is accepted but their information presentation can be improved by providing overviews and the missing interactions.
Risk Analysis and Decision Theory: An Extended Summary
E. Borgonovo, V. Cappelli, F. Maccheroni, M. Marinacci (a), and C. Smith (b)
a) Department of Decision Sciences and IGIER, Università Bocconi, Milan, Italy, b) Idaho National Laboratory, Idaho Falls, Idaho, USA.
We reconcile Kaplan and Garrick's seminal definition of risk with classical subjective expected utility, filling in the relevant gaps and providing a framework that is ready-to-use in applications. We show that Kaplan and Garrick's “frequency” format can be set in one-to-one correspondence with 's utility theory. Kaplan and Garrick's “probability” format corresponds to the framework of  in which epistemic uncertainty is captured by a subjective probability over uncertain events. Finally, Kaplan and Garrick's “probability of frequency” format, the most general one, corresponds to the recently proposed framework of , which distinguishes aleatory and epistemic uncertainty in a Bayesian perspective. The classic Kaplan and Garrick's risk triplets are then cast in the powerful setting of axiomatic Decision Theory, with its solid behavioral foundations, allowing one to make explicit the often implicit decisions of a Risk Analysis.
Maritime Oil Spill Risk Assessment for Hanhikivi Nuclear Power Plant
Fennovoima, Helsinki, Finland
Fennovoima is planning to build a new nuclear power plant unit, Hanhikivi 1, on a greenfield site in Pyhäjoki in Northern Finland. A nearby maritime oil spill accident is one of the external events analysed in the probabilistic risk assessment (PRA) of the plant. The oil effects on a nuclear power plant are not well-known, but in the worst case the oil could cause a loss of the ultimate heat sink by blocking the sea water intake screens. By considering the maritime traffic, oil transport and oil spill accident data in the Baltic Sea area, it is evaluated that a nearby medium oil spill (100 -1000 tonnes) occurs with a frequency of 1,0•10-2 /a and a large spill (> 1000 tonnes) with a frequency of 3,0•10-3 /a. The probability that the spill drifts to Hanhikivi and oil combat measures fail is assessed by using event tree analysis. The spill behaviour is considered, including oil spreading, dissolution, dispersion and movement due to wind and currents. In addition, oil combat measures including the use of oil booms and skimmers are evaluated. According to the results, significant amount of oil could enter the plant intake tunnel with a frequency of 4,2•10-5 /a.
Using Bond Graphs for Identifying and Analyzing Technical and Operational Hazards in Complex Systems
Ingrid Bouwer Utne, Eilif Pedersen and Ingrid Schjølberg
Department of Marine Technology, Norwegian University of Science and Technology, Trondheim, Norway
Oil and gas exploration and production is moving into harsher environments, such as the Arctic, which increases the complexity of operations. Technology development introduces more advanced functionality and operators may not have the full overview or knowledge to handle deviations that propagate in the systems. The increasing complexity and couplings in systems and operations means that a systems approach is necessary to ensure sufficient risk management for accident prevention. Current risk analysis methods, however, have limitations. This paper investigates the use of bond graphs as a systemic method for analyzing risk in dynamic systems, for example, as a supplement to hazard and operability analysis (HAZOP) and system theoretic process analysis (STPA). The article uses a remote operated vehicle (ROV) for subsea operation as a case study.
M24 Risk Governance and Societal Safety I
3:30 PM Waialua
Chair: Woody Epstein, Lloyd's Register Consulting
Estimating Farmer’s Risk Aversion
IRSN, Fontenay-aux-Roses, France
In the early days of safety, together with his famous diagram, Farmer introduced a form of risk aversion. The first objective of the paper is to propose a general formulation of risk aversion along Farmer’s thinking. This theoretical framework is particularly well suited when accident severity cannot be mitigated and prevention efforts are aimed at reducing probabilities, as is the case with nuclear safety. This is shown to go beyond the expected utility theory. The second part of the paper reports on an attempt at estimating Farmer’s risk aversion as perceived by a panel of nuclear safety professionals. This tends to confirm Farmer’s views.
Development of a Methodological Approach to Strategic Fire Service Planning Combining Concepts of Risk, Hazard and Scenario-based Design
Adrian Ridder, Uli Barth
University of Wuppertal, Wuppertal, Germany
Strategic Fire Service Planning is a quite new field of research because of which there is a need for fundamental research and methodological work. Existing approaches like risk management, the hazard-concept and scenario-based design have been found to not be fully applicable on its respective own to the research question of “how much fire service is necessary in a city”. Based on analytical work and the analysis of incident data it is shown that a combined approach of risk and scenario-based methods is a good starting point for further research.
Ambiguity in Risk Assessment
Inger Lise Johansen and Marvin Rausand
Norwegian University of Science and Technology, Trondheim, Norway
This paper aims to shed light on the concept of ambiguity in engineering risk assessment. The objectives are to 1) Clarify the meaning of ambiguity in risk assessment; 2) Describe sources and manifestations of ambiguity in preassessment, risk analysis, and risk evaluation/decision-making; and 3) Outline a procedure for approaching ambiguity in practice. To address these objectives, we first review existing definitions of ambiguity, which are argued to be of limited relevance to engineering risk assessment. We then propose a new overall definition of ambiguity as a challenge in risk-informed decision-making, and define linguistic, contextual, and normative ambiguity as distinct categories of ambiguity that have different implications for risk assessment. Based on this, we list concrete sources and manifestations of ambiguity in risk assessment in a set of tables that can be used as a checklist for identifying ambiguity in the assessment process. We finally outline a stepwise procedure for approaching ambiguity in risk assessment, in order to provide practical guidance and stimulate further research on ambiguity in risk-informed decision-making.
How is Capability Assessment Related to Risk Assessment? Evaluating Existing Research and Current Application from a Design Science Perspective
Hanna Palmqvist, Henrik Tehler, and Waleed Shoaib
Division of Risk Management and Societal Safety, Centre for Risk Assessment and Management, and Centre for Societal Resilience, Lund University, Lund, Sweden
Several countries use capability assessments as a part of their efforts to manage risk. However, it is unclear how such assessments are connected to other risk management activities (e.g. risk assessment). Therefore, the aim of the present paper is to present a study of how capability assessment is related to risk assessment. Capability assessment methods were identified through a scoping study and the Swedish capability assessment method was investigated through interviews with Swedish public actors and analysis of legislative documents. The data was analysed using a design science perspective. The results of the analysis show that the purposes presented for some capability assessment methods are the same or similar to purposes common to risk assessment methods, and the actual form of some of the methods is similar to existing risk assessment methods. Nevertheless, the relationship between capability assessment and risk assessment is unclear. We conclude that if capability assessments are going to continue to be an important part of risk management activities more research is needed to better establish the relationship between risk assessment and capability assessment.
M25 Risk Informed Applications I
3:30 PM Waianae
Chair: Hao Zheng, Lloyd's Register Consulting
Analyses of AP1000® Expanded Event Tree Sequences Based on Best-Estimate Calculations
J.Montero-Mayorga, C.Queral, J.Gonzalez-Cadelo and G. Jimenez
Universidad Politecnica de Madrid, Madrid, Spain
The Westinghouse AP1000® reactor is an advanced design whose safety systems are based on natural mechanisms such as gravity or natural circulation, namely, they are passive safety systems. Because of the passive nature of the safety related systems and its dependency on small changes on certain variables (e.g. pressure), it is necessary to confirm that when core cooling is achieved, uncertainties are bounded. The thermal-hydraulic (T/H) uncertainty evaluation process performed by Westinghouse Electric Company (WEC) identified a set of low T/H margin by expanding probabilistic risk assessment (PRA) event trees. Expanded event trees contain more branches than classic event trees, including all possibilities for system actuation. Then detailed conservative computer codes were applied in order to analyze the bounding sequences that were significant to the core damage frequency and demonstrating that the T/H uncertainty was bounded. The UPM group has analyzed the low-margin sequences obtained by WEC with the best estimate computer code TRACE in order to verify the previous results and also to study the phenomenology of such sequences through a best estimate code. This paper presents the results obtained for the DVI line break case confirming that it does not exist damage in the bounding sequence selected for that case.
Application of Web-based Risk Monitor in Tianwan Nuclear Power Plant
Hao Zheng (a), Wei Wang (b), Xiaohui Gu, Yong Qu, Zhenli Bao (c), Xuhong He (b)
a) Lloyd's Register Consulting, Beijing, China, b) Lloyd's Register Consulting, Stockholm, Sweden, c) Jiangsu Nuclear Power Co., Lianyungang, China
As one of the specific applications of Living PSA, Risk Monitor, which is a real-time analysis tool used to determine the instantaneous risk based on actual plant configuration, has been widely used in risk-informed decision-making process during plant operation. A web-based Risk Monitor application for Tianwan nuclear power plant (NPP) is currently being used onsite to improve the PSA applications, popularize the concept of risk-informed management, and then enhance the whole risk management level in Tianwan NPP. Compared with the traditional windows based tools, this web-based Risk Monitor application is a natural multi-user program with great advantages. It has good interface with plant’s existing information system and can automatically update the risk information upon changes of component status/configurations. This paper presents the overview of the Risk Monitor application used in Tianwan NPP, including challenges and experience of implementing Web-based Risk Monitor, and major feature improvements which facilitate the application of Risk Monitor. Example PSA applications implemented in Tianwan NPP will also be presented together with future plan and challenges.
Analyzing System Changes with Importance Measure Pairs: Risk Increase Factor and Fussell-Vesely Compared to Birnbaum and Failure Probability
Janne Laitonen, Ilkka Niemelä
Radiation and Nuclear Safety Authority (STUK), Helsinki, Finland
Importance measures are used to rank components of a system according to a selected criterion depending on the decision problem. Sometimes, more than one importance measure may be used. In risk-informed decision making, a component that is critical to safety is usually prioritized higher in allocating activities, e.g. maintenance or inspection. One desired effect of this prioritization is to improve the reliability of the critical components. Changes in the system or component reliability affect their importance measures. If these feedbacks are taken into account, new ranking for the components may be obtained. This paper examines the properties of risk importance measure pairs in analyzing system changes with fault tree analysis. A common approach is to use risk increase factor (or risk achievement worth) and Fussell-Vesely importance measure. This approach is compared to an alternative method which utilizes Birnbaum importance measure and the failure probability of a basic event. It is shown that the first approach may lead to difficulties in understanding the effect of system changes whereas the latter seems to provide simpler and more robust alternative. The paper includes examples to show and compare the differences between the two methods. The key advantages of the alternative method are that it reflects the absolute instead of relative change, the variables are independent, and that the interpretation of the importance measures is straightforward, reflecting risk in terms of safety margin and failure probability.
Energy Loss Optimization in Basic T-Shaped Water Supply Piping Networks for Probabilistic Demands
KW Mui, LT Wong, and CT Cheung
Department of Building Service Engineering, The Hong Kong Polytechnic University, Hong Kong, China
Minimization of energy loss of water supply networks is a major concern of pump power reduction for sustainable water systems in buildings. This paper presents a mathematical model for energy loss optimization in a common basic T-shaped water supply piping network that serves infinite probabilistic demands. Optimized designs based on proper network pipe sizes are analyzed. Optimal pipe radius ratios (2 1/7 to 2 3/7) and their corresponding energy implications in the network are also discussed. The results show that existing piping designs are not optimized for probabilistic demands and there is potential for energy loss reduction.
Insights and Improvements Based on Updates to Low Power and Shutdown PRAs
J. F. Grobbelaar, J. A. Julius, K. D. Kohlhepp, and M. D. Quilici
Scientech, a Curtiss-Wright Flow Control Company, Tukwila, WA, U.S.A.
In several countries, the requirements for probabilistic risk assessments have increased beyond a Level 1 internal events PRA to add or address spatial and external hazards. In a growing number of countries the requirements have further increased to address all Level 1 hazards in all plant operating modes. Scientech developed its first shutdown probabilistic risk assessment (PRA) in the early 1990s for a European nuclear power plant. Since then several additional low power and shutdown PRA models were developed in the United States following the same approach. The original shutdown PRA model was expanded to evaluate hazards challenging fuel in the reactor vessel and fuel in the spent fuel pool; modeling Level 1 core damage for all hazards in all plant operating modes, with corresponding Level 2 (release) and Level 3 (consequence) models. This complete PRA of all hazards and all modes was incorporated into the European plant’s licensing basis, and in 2010 a peer review was conducted. In the last three years, the shutdown PRA model was updated and a follow-on peer review conducted. Plant operational state definitions were revised to better agree with technical specifications governing the plant operating modes. Additional initiating events were modeled for the fuel pool plant operational states as well as the refueling plant operational states. Initiating event frequencies have been updated to reflect recent operating experience. Success criteria and accident sequence development were revised based on insights from new thermal-hydraulic analyses. New shutdown procedures and "FLEX" strategies were considered in the accident sequence development. New operator actions were credited and human reliability analyses were performed. During the same period, additional model changes and refinements were developed on the USA shutdown PRA models. This paper presents the insights and improvements made in the PRA modeling of low power and shutdown states, and also presents a summary of insights and benefits that the plant obtained during the development and updates of the underlying shutdown PRA models.
M26 Risk Informed Licensing and Regulation I
3:30 PM Ewa
Chair: Dennis Damon, U. S. Nuclear Regulatory Commission
When Is It Justified to Delay the Implementation of Safety Improvements After They Have Been Approved?
IRSN, Fontenay-aux-Roses, France
After safety improvements have been approved, actual implementation can often be delayed significantly in the nuclear sector. This situation can be unsatisfactory for safety experts conscious of the safety benefits foregone during such a delay. They rightly assimilate this to a cost in terms of safety. The present paper proposes a cost-benefit analysis of this question. Two different types of delay benefits are distinguished: the time value of delaying the implementation on the one hand; and possible reductions in implementation costs. These two benefits are first approached separately after which a general formula is proposed and discussed. Delays generally appear difficult to justify, except when cost reductions are substantial and delays are limited.
The Underlying Principles and Quantitative Values of Risk Limits
Dennis R. Damon
U. S. Nuclear Regulatory Commission, Washington DC
The purpose of this paper is to clarify the principle that risks imposed on individuals should be limited in their magnitude. The paper also discusses implementation of this risk limitation principle by regulation, including quantitative risk limits. This principle of limits arises when an activity, although beneficial to society, nevertheless imposes some risk of harm on any individual without their consent and not for their benefit. In this case, fairness requires that the magnitude of the risk be limited. If the harm is fatality, then compensation is no remedy, and regulation should limit the risk a priori. This principle of limitation of imposed risk is not the same as the idea that it is desirable to reduce risk in general. The principle to be applied for general reduction of risk is optimization. Optimization considers all impacts, both beneficial and adverse, on all persons to improve net benefits to society collectively. The appropriate magnitude of a quantitative risk limit may vary if the individual has some degree of consent, or benefits from the activity having risk. Thus values recommended for limits on risk have typically differed between workers who benefit from the activity, and members of the general public who do not.
Development of A Framework for Establishment of Risk-informed Safety Goals for Nuclear Power Plants Operation in the UAE
Jun Su Ha (a), Sung-yeop Kim (b), Jamila Khamis Al Suwaidi (c), and Philip Beeley(a)
a) Khalifa Univ. of Science, Technology and Research, Abu Dhabi, UAE, b) Korea Advanced Institute of Science and Technology (KAIST), Republic of Korea, c) Federal Authority for Nuclear Regulation (FANR),Abu Dhabi, UAE
A framework for establishment of risk-informed safety goals for nuclear power plants (NPPs) operations in the UAE was developed in this study. The current regulatory circumstance to the safety goals in the UAE was addressed as well. Representative parameters related to the core integrity (Level 1 PSA) and containment integrity (Level 2 PSA) are used as surrogate measures, Core Damage Frequency (CDF) for cancer (latent) fatality and Large Early Release Frequency (LERF) for early (prompt) fatality, for risk-informed safety goals. Under this framework a conservative evaluation of risk-informed safety goals was performed on the basis of conservative assumptions and data which were obtained and/or derived from the PSA results of APR-1400, the same type of the Barakah NPPs which are under construction in the UAE, and public health risk assessments. The current safety targets specified in the regulatory guideline (FANR-RG-004) in the UAE were examined to be appropriately determined with sufficient conservatism from the evaluation results. Limitations of the study and recommendations for appropriate applications of the risk-informed safety goals were provided as well.
Insights from PSA Comparison in Evaluation of EPR Designs
Ari Julin, Matti Lehto (a), Patricia Dupuy, Gabriel Georgescu, Jeanne-Marie Lanore (b), Shane Turner, Paula Calle-Vives (c), Anne-Marie Grady, Hanh Phan (d)
a) Radiation and Nuclear Safety Authority (STUK), Finland, b) Institute of Radiological Protection and Nuclear Safety (IRSN) , France, c) Office for Nuclear Regulation (ONR), United Kingdom, d) NuclearRegulatory Commission (USNRC), United States of America
The paper describes the outcome of a limited probabilistic safety assessment (PSA) comparison on the following EPR designs: Olkiluoto 3 Nuclear Power Plant (NPP) in Finland, Flamanville 3 NPP in France, UK EPR design, and U.S. EPR design. The objective of this PSA comparison was to identify differences in the modeling aspects and results of EPR PSAs, as well as to assess the rationale for these differences. The comparison covered various types of initiators challenging a broad scope of safety functions. Insights from the EPR PSA comparison and rationale for the differences originated from modeling assumptions, applied reliability data, designs, and operational aspects. The EPR designs chosen for comparison represents various design and licensing stages, as well as level of detail, which gives the main rationale for the identified differences. The outcomes and lessons learned from the EPR PSA comparison have been used to facilitate the regulatory reviews and assessment work of various EPR designs and to enhance the scope, level of detail, and quality of EPR PSA models and documentation.
OECD WGRISK – Challenges and Recent Tasks
Marina Roewekamp (a), Jeanne-Marie Lanore (b), Kevin Coyne (c), Milan Patrik (d), Abdallah Amri, Neil Blundell (e)
a) Gesellschaft für Anlagen-und Reaktorsicherheit (GRS) mbH, Köln, Germany, b) Institut de Radioprotection et de Sûreté Nucléaire (IRSN), Fontenay-aux-Roses, France, c) U.S. Nuclear RegulatoryCommission, Washington, DC USA, d) UJV Rez, Rez, Czech Republic, e) OECD Nuclear Energy Agency (NEA), Issy-les-Moulineaux, France
The overall objective of the Working Group on Risk Assessment (WGRISK) of the OECD Nuclear Energy Agency (NEA) Committee on the Safety of Nuclear Installations (CSNI) is to advance the understanding of Probabilistic Safety Assessment (PSA) and to facilitate its utilization for enhancing the safety of nuclear installations. To accomplish this mission, WGRISK continuously performs a variety of activities to exchange information on PSA between member countries. This paper presents a brief overview on the actually on-going WGRISK activities and perspectives. In addition to on-going tasks covering more traditional PSA challenges (e.g. tasks relating to human reliability analysis (HRA) and digital instrumentation and control (I&C)), new challenges for PSA have arisen from the recent nuclear power plant operating experiences and the insights from the post-Fukushima stress tests†. In response to these new challenges, WGRISK conducted an international workshop on “PSA of Natural External Hazards Including Earthquakes” in June 2013. This workshop revealed valuable insights on challenges associated with external events such as scope consideration for PSA, the need to consider combinations of external hazards, and multi-unit impacts. Another ongoing WGRISK activity is the second follow-up workshop on “Fire PRA” to be held in April 2014. The Fire PRA workshop will address many of the technical challenges associated with including fire hazards, which typically provide a non-negligible contribution to the overall core or fuel damage frequency, in PSA. WGRISK recently initiated a task focused on obtaining insights from PSA related to the loss of electrical power sources. This task will collect examples of PSA insights related to a loss of electrical power sources, including those insights identified as a result follow-up activities to the Fukushima Dai-ichi reactor accidents. It is expected that this task will also highlight the capabilities of PSA as a tool for providing insights related to the potential consequences of the loss of a safety function, such as core damage frequencies or frequencies of radioactive releases. The use of PSA in this manner may provide a measure of defense-in-depth in case of loss of a safety function, which will augment more traditional analysis approaches that emphasize identification of failures that can lead to loss of system function.
M27 Automotive Engineering
3:30 PM Kona
Chair: Stefan Bracke, University of Wuppertal
RAPP: Method for Risk Prognosis on Complex Failure Behaviour in Automobile Fleets Within the Use Phase
Stefan Bracke and Sebastian Sochacki
University of Wuppertal, Chair of Safety Engineering and Risk Management, Wuppertal, Germany
The increasing complexity of product functionality and manufacturing process parameters often leads to complex failure modes during the product life cycle. These field information are the basis for risk analyses and damage case prognosis with the goal of an early risk detection and leads to the possibility of nearby interactions e.g. product and manufacturing optimisation or recall action. This paper outlines the essential procedure of the new developed method “Risk Analysis and Prognosis of complex Products (RAPP)”. The main focus of the RAPP method is the detection, visualisation and prognosis of risks and damage cases depending on their life span variables regarding to a product fleet -based on a risky production batch -in field. The RAPP method contains multiple steps: First steps include the mapping/prognosis of the failure behaviour and the mapping of product field load profiles. Next step is focusing on the estimation of the critical area regarding the life span variable (e.g. critical kilometer range). Based on these steps, it is possible to perform the risk analysis and risk prognosis regarding the product fleet. Finally, the last step of the RAPP method is the verification of risk analysis and –prognosis. The theory and application of the RAPP method is explained within an automotive case study oil tube leakage.
Stress-Dependent Weibull Shape Parameter Based on Field Data
Jochen Juskowiak and Bernd Bertsche
University of Stuttgart, Stuttgart, Germany
The Weibull shape parameter is often assumed to be constant, with no dependency on stress. However, some cases exist, in which it is a function of stress. If the stress-dependency is not considered, vague assumptions of the Weibull shape parameter may lead to inaccurate results, e. g. for reliability prediction or demonstration testing purposes. Drawbacks in choosing an adequate parameter are e.g. extensive testing at a specific stress level, or insufficiently established mathematical descriptions. This paper presents an approach which allows a stress-dependent derivation of the Weibull shape parameter based on field data. In order to do so, simulations of the customer behavior and additional information from the customers themselves are used. Linking the occurred failure with the corresponding stress-level is thus possible.
APTA Approach: Analysis of Accelerated Prototype Test Data Based on Small Data Volumes Within a Car Door System Case Study
Marcin Hinz, Philipp Temminghoff, and Stefan Bracke
University of Wuppertal, Chair of Safety Engineering and Risk Management,Wuppertal, Germany
Knowledge of failure behavior and failure modes regarding the component´s complete life cycle is fundamental within the early development phases of technical and complex products. Here, an overview of the design of prototype test procedures as well as the transformation of expected field failure behavior in prototype test characteristics is described. This provides the required knowledge for the understanding of accelerated testing and is the basis for understanding of the developed “Accelerated Prototype Test data Analysis” (APTA) approach. The APTA approach is demonstrated with the help of a case study with regard to a car door system. The analysis of the design principles, expected impacts in the usage phase and car door prototype test procedure is discussed. With the use of nonparametric as well as parametric statistical methods, the wearing and ageing of specific door mechanism characteristics (e.g. forces or displacements) in relation to life span variables are analyzed. Furthermore a method for the comparison of qualitative and quantitative characteristics and their impact on the door system is described. Finally the interpretation of the results and deduction of general issues and recommendations regarding to the design of prototype test procedures are presented.