Original Article
Development of a Probabilistic Safety Assessment
Framework for an Interim Dry Storage Facility
Subjected to an Aircraft Crash Using Best-Estimate
Structural Analysis
Belal Almomani
a, Dongchan Jang
a, Sanghoon Lee
b,*,1, and
Hyun Gook Kang
c,*,1a
Department of Nuclear and Quantum Engineering, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea
bDepartment of Mechanical and Automotive Engineering, Keimyung University, Dalgubeol-daero 1095, Dalseo-gu, Daegu, Republic of Korea
cDepartment of Mechanical, Aerospace and Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY 12180, USA
a r t i c l e i n f o
Article history:Received 7 December 2016 Received in revised form 27 December 2016
Accepted 27 December 2016 Available online 23 January 2017 Keywords:
Interim Storage Facility Probabilistic Safety Assessment Radiological Consequence Structural Response Targeted Aircraft Crash
a b s t r a c t
Using a probabilistic safety assessment, a risk evaluation framework for an aircraft crash into an interim spent fuel storage facility is presented. Damage evaluation of a detailed generic cask model in a simplified building structure under an aircraft impact is discussed through a numerical structural analysis and an analytical fragility assessment. Sequences of the impact scenario are shown in a developed event tree, with uncertainties considered in the impact analysis and failure probabilities calculated. To evaluate the influence of pa-rameters relevant to design safety, risks are estimated for three specification levels of cask and storage facility structures. The proposed assessment procedure includes the determi-nation of the loading parameters, reference impact scenario, structural response analyses of facility walls, cask containment, and fuel assemblies, and a radiological consequence analysis with doseerisk estimation. The risk results for the proposed scenario in this study are expected to be small relative to those of design basis accidents for best-estimated con-servative values. The importance of this framework is seen in its flexibility to evaluate the capability of the facility to withstand an aircraft impact and in its ability to anticipate po-tential realistic risks; the framework also provides insight into epistemic uncertainty in the available data and into the sensitivity of the design parameters for future research. Copyright© 2017, Published by Elsevier Korea LLC on behalf of Korean Nuclear Society. This
is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).
* Corresponding authors.
E-mail addresses:[email protected](S. Lee),[email protected](H.G. Kang). 1These authors contributed equally to this work.
Available online at
ScienceDirect
Nuclear Engineering and Technology
journa l home page:www.elsevier.com /locate/ne t
http://dx.doi.org/10.1016/j.net.2016.12.013
1738-5733/Copyright© 2017, Published by Elsevier Korea LLC on behalf of Korean Nuclear Society. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
1.
Introduction
An intentional aircraft impact (AI) into hazardous facilities became a prominent issue after September 11, 2001. The questions then arose as to what extent an interim storage facility (ISF) can adequately protect against a large commer-cial AI and what momentous consequences an AI may pose to society and the environment. To answer these questions, possible damage scenarios with various AI loading conditions should be investigated to estimate the potential risks. Several safety assessment studies have considered various mechani-cal and thermal loadings that represent severe impact con-ditions in the frame of a probabilistic safety assessment with conservative assumptions, or just to evaluate the structural integrity of a cask assembly based on a deterministic approach[1e10]. However, no risk assessment with a scenario of a realistic aircraft crash into an ISF has yet been completed. Development of a comprehensive and systematic framework that comprises several elements of risk evaluation would
make it possible to determine the feasibility of using such a framework to predict the expected hazard posed to the public, in addition to enhancing the reliability of facility/cask designs under possible severe impact loadings from aircraft crash. Further, this approach will enable the development of a so-called risk-informed regulatory framework, which is more reasonable than conventional approaches that rely more on conservatism. Thus, under the frame of a probabilistic approach, this study attempts to integrate more elements into the development of a realistic aircraft crash scenario, including possible damage consequences and associated release of fission products to the environment.
The current study covers only the first assessment stage of the accident scenario, which is a direct mechanical impact from an intended AI. The second assessment stage in follow-up research will examine associated radiological re-leases from subsequent fire and multicask collisions that will supplement the radiological consequences of the first stage. The current analysis is divided into four major parts:
Part 1 sets up the reference impact scenario; Part 2 provides a structural response analysis; Part 3 outlines the radiolog-ical consequence analysis; and Part 4 summarizes the event tree analysis results and the accident doseerisk estimations. Illustrative examples are presented in this paper for high-,
median-, and low-resistance structures subjected to
concentrated impact loadings from aircraft crash. As this paper only touches on the methods used in each analysis stage with brief discussions and summarized results, readers are referred to the list of references for further
details on the methods, assumptions, and information used in this study.
2.
Methodology overview and reference
modal study
It is commonly known that ISFs are ruggedly built and storage casks are conservatively designed to comply with stringent regulatory requirements; only the local response[11], which
Table 1 e Reference specifications employed in the analysis.
Item Parameter Value
Engine model (CF6-80C2) (1) Length (L) 4.3 m (2) Average diameter of engine (d) 1.4 m (3) Dry weight (W) 4.4 tons
(4) Missile shape factor (N) 1 (average bullet nosedspherical end) Storage facility wall (1) Concrete strength class (fc): Normal reinforced concrete
a. Low-resistance structure (L) C16 b. Median-resistance structure (M) C20 c. High-resistance structure (H) C25 (2) Weight density of concrete (rc) 2,300 N/m3
(3) Thickness (tw)
a. Low-resistance structure (L) 70 cm b. Median-resistance structure (M) 95 cm c. High-resistance structure (H) 120 cm Metal storage cask (1) Cask diameter (D) 2.1 m
(2) Lid closure thickness 20.1 cm (3) Number of bolts 24 (4) Number of fuel assembly 21 (5) Total weight 97 tons (6) Cask height (H) 5.4 m
(7) Seal type Metal seal
a. Low-resistance performance (L) (Elastic recovery distance (er): 0.03 mm)
b. Median-resistance performance (M) (Elastic recovery distance (er): 0.15 mm)
c. High-resistance performance (H) (Elastic recovery distance (er): 0.25 mm)
Fuel assembly (1) Manufacture, array CE 16 16 PWR (2) Fuel enrichment 4.5%
(3) Number of fuel rods 236 (4) Total length of a fuel rod 4.094 m (5) Number of grid spacers 12 (6) U-235 burnup rate
a. High burnup rate (L) 60 GWd/MTU b. Intermediate burnup rate (M) 45 GWd/MTU c. Low burnup rate (H) 25 GWd/MTU (7) Interim cooling period 10 y
refers to penetration and perforation of the building structure caused by a stiff element impact, is considered in this study. The reference case study is a regional load from an indepen-dent jet engine impact that perforates the reinforced concrete (RC) wall and directly hits a freestanding, single, fully loaded metal cask being the worst-case scenario.Fig. 1illustrates the impact scenario and sequence of events. Readers are referred to the authors' previous work[12]for further details on the justification and development of the aircraft crash scenario.
Table 1 provides the model specifications that are employed in the analysis. Specifications of concrete strength class, wall thickness, elastic recovery of seal, and U-235 burnup rate are separated into three sets and analyzed accordingly. These parameters are regarded as safety relevant
components, categorized in respect to increasing or
decreasing potential risk. The RC class range is from C16 as a low-resistance level (L) to C25 as a high-resistance level (H), which is commonly used in ISF design[1,2]. Thicknesses of existing ISFs' walls vary from 70 cm to 120 cm[13]; therefore, 70 cm is considered as a low-resistance structure (L), 95 cm as
a median-resistance structure (M), and 120 cm as a high-resistance structure (H). Regarding the cask's metal seal, it may lose seating pressure and sealing forces due to relaxation effects and creeping; these degradation mechanisms could cause a decrease in the elastic recovery of the seal during long-term operation, as investigated by Wolff et al.[14]. The minimum reduction of the elastic recovery distance as a function of holding time and temperature is around 0.03 mm, while the elastic recovery distance of a new seal is around 0.25 mm [14]. Thus, this study defines the elastic recovery distances for the low- (L), median- (M), and high-performance (H) seals as 0.03 mm, 0.15 mm, and 0.25 mm, respectively. Three burnup rates are also considered in the consequences analysis, as follows: a low burnup rate at 25 GWd/MTU, an intermediate burnup rate at 45 GWd/MTU, and a high burnup rate at 60 GWd/MTU. The rupture strain of rod failure is found to decrease when the burnup rate increases, due to the effect of the fuel burnup on the cladding embrittlement (see Section 3for further explanations); therefore, the low burnup rate is considered to indicate a high-resistance structure (H), and the
Fig. 3 e Aircraft impact speed histogram.
high burnup rate a low-resistance structure (L). The general approach of this study is outlined schematically in Fig. 2, which addresses the structural response and radiological consequence analyses with inputs and outputs. The results of this study are intended to act as a basis of discussion and to provide comparisons between the three sets of specification levels.
3.
Structural response analyses
The structural response analysis of this study is divided into three parts. The first part is a fragility analysis of the facility wall based on an empirical model, followed by a structural
response analysis of the cask lid closure, with the last part investigating the fuel rod failure under a mechanical impact load. Detailed methodologies and information related to the fragility analysis of the facility wall and the lid gap analysis are not covered in this paper, as details of these analyses can be found in the authors' previous work[12,15].
3.1. Facility RC wall damage analysis
As a best deterministic estimation, with a bias toward conservatism, as suggested in the NEI (Nuclear Energy Insti-tute, U.S.) report[16], the empirical formulas employed in this study can predict the minimum RC wall thickness required to prevent local damage caused by the normal impact of an
Fig. 5 e Weibull probability distribution functions of residual engine velocity for the three wall structures. CDF, cumulative distribution function.
aircraft engine and its residual velocity. Impact speeds and compressive strengths of concrete are treated probabilistically as major sources of uncertainty.Fig. 3displays a frequency histogram of AI speeds and the limited set of data that has been collected from the literature[17, 18]. The mathematical model of the probability density function, as proposed by €Oztemel and S‚ensoy [19] using the lognormal distribution model, is applied to represent the uncertainty in the compressive strength of concrete for the three classes (C16, C20, and C25) shown inFig. 4.
The obtained distribution curves for residual velocity var-iables are presented inFig. 5for the three structural levels of the RC wall. Utilizing the Monte-Carlo method, the cumulative distribution function for perforation thickness F(tw) is gener-ated based on 10,000 trials, as shown inFig. 6. The failure
probability of a given wall thickness (Pw) is calculated from the complementary cumulative distribution function. Therefore, the estimated failure probabilities for the three grades of wall are as follows: Pw(LG)¼ 0.98, Pw(MG)¼ 0.8, and Pw(HG)¼ 0.27. It is noticeable that the failure probabilities of the facility wall for low- to median-resistance structures are expected to be very high compared with those for high-resistance structures. Therefore, selection of high-level design properties for a robust wall, namely, thickness and concrete strength, is rec-ommended to enhance ISF capacity to withstand any stiff element impact from an aircraft crash.
3.2. Metal cask damage analysis
For the engineecask impact analysis, the impact loadetime history function and effective loading contact area for the relevant aircraft engine are needed. This paper does not discuss the derivation of the loadetime function since the
Fig. 7 e Finite element model for lid gap analysis.
Fig. 8 e Summary of the calculated leakage path areas for all impact conditions.
Table 2 e Cask status results for each impact condition. Case no. O-ring recovery
distance
Impact velocity (m/sec) 40 60 80 100 120 140 160 Case 1 H (0.25 mm) RD RD RD RD RD RD SD M (0.15 mm) RD RD RD SD SD SD SD L (0.03 mm) RD SD SD SD SD SD SD Cases 2, 3 H (0.25 mm) RD RD RD RD RD SD SD M (0.15 mm) RD RD RD SD SD SD SD L (0.03 mm) RD SD SD SD SD SD SD Case 4 H (0.25 mm) RD RD RD SD CD CD CD M (0.15 mm) RD RD SD SD CD CD CD L (0.03 mm) RD SD SD SD CD CD CD Case 5 H (0.25 mm) RD SD SD CD CD CD CD M (0.15 mm) RD SD SD CD CD CD CD L (0.03 mm) RD SD SD CD CD CD CD CD, containment damage; H, high-resistance level; L, low-resistance level; M, median-low-resistance level; RD, recoverable damage; SD, seal damage.
analytical approach and the assumptions in the derivation of the jet engine impact load were discussed in detail previously [12,15]. The applied impact velocities of the engine are 40 m/s, 60 m/s, 80 m/s, 100 m/s, 120 m/s, 140 m/s, and 160 m/s; these values have been selected to be within the range of the ob-tained residual velocities after concrete shield perforation.
The previously developed finite element (FE) model explained in Almomani et al.'s work[15]is shown inFig. 7; this model consists of a carbon steel cask body with one bolted upper lid closure without impact limiters, and a solid inner dummy weight representing the interior structure and spent fuel assemblies. The purpose of this model was to analyze the dynamic response of the lid closure system in order to calculate the leakage path area between the lid and the flange, since the lid seal plays an essential role in preventing the escape of fine particulates and radioactive gases from the cask. Therefore, unnecessary details of the cask system, such
as the interior structures, were not modeled. The explicit time
integration solver LS-DYNA [20] was used to address the
impact problem effectively. In the lid gap analysis, it is assumed that a closure opening greater than the pre-compression state of the metallic seal will cause leakage. Three elastic recovery distances are examined in the analysis representing the performance of the metallic seal, as mentioned in Table 1. The cask failure criteria, defined by Almomani et al.[12,15], are as follows: recoverable damage is defined as the damage that occurs when the opening gaps are less than the elastic recovery distance (<er); seal damage is defined as the damage that occurs when the opening gaps exceed the elastic recovery distance (>er); and containment damage is defined as the damage that occurs when an irre-versible deformation occurs on the cask containment.
Fig. 8summarizes the leak areas estimated in this study, andTable 2shows the response status of the cask for each
Table 3 e Conditional probability calculations for the cask response statuses.
Seq. # Case # Boundary criteria (Va, Vb) (m/sec) State Pc
H M L H M L 2 Case 1 V V140 V V80 V V40 RD 0.9999 0.83876 0.06911 3 V140 V V160 V80 V V160 V40 V V160 SD 1.11Ee04 0.16124 0.93089 4 Case 2 V V120 V V80 V V40 RD 0.99989 0.83876 0.06911 5 V120 V V160 V80 V V160 V40 V V160 SD 0.00011 0.16124 0.93089 6 Case 3 V V120 V V80 V V40 RD 0.99989 0.83876 0.06911 7 V120 V V160 V80 V V160 V40 V V160 SD 1.11Ee04 1.61Ee01 0.93089 8 Case 4 V V80 V V60 V V40 RD 9.77Ee01 0.83876 0.06911 9 V80 V V100 V60 V V100 V40 V V100 SD 2.05Ee02 0.36125 0.72271 10 V100 V V160 V100 V V160 V100 V V160 CD 2.16Ee03 0.04315 0.20818 11 Case 5 V V40 V V40 V V40 RD 5.71Ee01 0.28617 0.06911 12 V40 V V80 V40 V V80 V40 V V80 SD 4.06Ee01 0.55259 0.45376 13 V80 V V160 V80 V V160 V80 V V160 CD 2.26Ee02 0.16124 0.47713
H, high-resistance level; L, low-resistance level; M, median-resistance level.
impact condition. Further explanation of the estimation method for leakage path areas and the cask structure response results have been presented in the authors' previous work[15]. From the results of numerical simulations, probabilities of cask response (Pc) in each impact case were estimated by integrating the area under the residual velocity distribution curves WðVRÞ, shown in Fig. 5, with the defined boundary
criteria (Va, and Vb) shown inTable 3.
3.3. Fuel cladding damage analysis
The dynamic impact characteristics of a fuel assembly inside the cask are analyzed using the explicit nonlinear FE code
ABAQUS[21]. Any method to evaluate fuel assembly response under mechanical impact load requires a detailed FE model. A generic cask model, consisting of a carbon steel cask body and a detailed interior structure without the bolted upper lid closure system, is constructed (Fig. 9). The interior structure includes a fuel basket, fuel basket supports, and one fuel as-sembly placed at the location of engine impact. The other basket cells, which include 20 fuel assemblies, are filled with solid dummy bodies to correct the package weight. The pur-pose of this model is to analyze the dynamic response of the fuel assembly and estimate the fraction of rods that fail during the impact as an essential parameter of the radiological con-sequences. All the materials are modeled as isotropic
Table 4 e Mechanical properties of cask components.
Part Material type Mass density (kg/m3) Elastic modulus
(MPa)
Poisson’s ratio
Yield stress (MPa)
Cask bodya SA516 Gr. 70 7,850 210,000 0.3 334.92114
Basket, support beams, space discsb
Stainless steel (SS304) 7,580 200,000 0.3 334.921
Top nozzleb Stainless steel (SS304) 3,867.67 200,000 0.3 215
Bottom nozzleb Stainless steel (SS304) 1,887.77 200,000 0.3 215
Guide tubesb Stainless steel (SS304) 6,550 114,000 0.296 215
Grid spacersa Zircaloy-4 strap 6,550 114,000 0.296 379
Fuel rodsa Zircaloy-4 3,897.7 72,000 0.296 348
Dummy fuel assembliesc
e 5,350 200,000 0.3 322.486
a Tensile strength tests.
bMATWEB (http://www.matweb.com/). c Artificial material.
Table 5 e Element characteristics of the cask model.
Part Element type Shape function
Cask body, dummies, nozzles Eight-node brick solid element Linear hexagonal Basket, space discs, grids Four-node shell element Linear quadrilateral Guide tubes, support beams, fuel rods Two-node beam element Linear line
elasticeplastic with hardening; material properties are listed inTable 4. In addition, element characteristics are listed in Table 5.
Simplified spacer grids were modeled as a box structure with springs and dimples that are connected to the fuel rods. The guide tube is connected to the spacer grids by a tie constraint. The linear spring stiffness value is 99.1 kN/m and the dimple value is 494.45 kN/m. In addition, mesh refinement studies for the areas of interest in both models have been carried out to obtain a steady discretization error. To achieve stable time integration, mesh distortion, severely distorted element dele-tion, reduced integradele-tion, and hourglass control are all acti-vated in the analysis. As the cask is freestanding on a concrete pad, automatic surface-to-surface contact was implemented between the interfaces of all components. The friction coeffi-cient is reasonably assumed to be 0.2 at the contact surfaces.
The impact analysis studied five impact orientations: lateral impacts on the lower (Case 1), center (Case 2), and upper parts of the cask (Case 3); an impact on an upper corner (Case 4); and a vertical impact on the lid closure (Case 5). The impact loading as a function of time is tabulated, evenly distributed, and perpendicularly directed to the nodes at the impact area. Fig. 10 illustrates the impact load locations applied to the cask body model.
Since it is difficult to justify the impact orientations, it is assumed that the impact angle is equally distributed between 0and 90on the cask body; this is done in order to provide a comprehensive and systematic evaluation of all the impact cases without the need for an exact clue of the possibility of occurrence. The jet engine of a B747 was chosen for this study to provide an upper bound impact load on the cask body. Whereas the scenario itself is a general impact scenario that can represent both commercial airplanes and military or
hobby aircrafts. The maximum possible impact angle of a large commercial aircraft is known to be less than 15 [18]. Thus, the range from 0to 15is considered as the probable angle of lateral impact, while the range of 15e75represents possible impact angles that depend on uncertain sequences of the scenario, with the range of 75e90representing possible impact angles from military or hobby aircrafts. Furthermore,
Fig. 11 e Illustration of the estimation method of fuel damage ratio for one fuel bundle.
Table 6 e Fuel damage ratio results of one fuel assembly for each impact condition.
Impact orientation Impact velocity (m/sec) Fuel damage ratio 25 GWd/ MTU Fuel damage ratio 45 GWd/ MTU Fuel damage ratio 60 GWd/ MTU Case 1 40, 60, 80 0.01 0.01 0.01 100 0.01 0.01 0.2 120 0.01 0.01 0.63 140 0.01 0.14 0.8 160 0.01 0.3 0.8 Case 2 40, 60, 80, 100, 120, 140 0.01 0.01 0.01 160 0.01 0.01 0.5 Case 3 40, 60, 80 0.01 0.01 0.01 100 0.01 0.01 0.5 120 0.01 0.07 0.8 140 0.01 0.2 0.8 160 0.01 0.44 0.86 Case 4 40, 60, 80 0.01 0.01 0.01 100 0.01 0.01 0.5 120 0.01 0.01 0.75 140 0.01 0.2 0.75 160 0.01 0.5 0.86 Case 5 40, 60, 80, 100, 120, 140, 160 0.01 0.01 0.01
vulnerable impact areas on the cask body are divided into three areas: lateral (60%), corner (20%), and vertical (20%). Therefore, the impact orientation probabilities (Po) of lateral, corner, and vertical impacts for the five defined impact areas on the cask body are as follows:
▪ Lateral impact orientation (side): Po_lat¼ (15/90 0.6) ¼ 0.1 ▪ Inclined impact orientation (corner): Po_cor¼ (60/90 0.13)
¼ 0.11
▪ Vertical impact orientation (on the lid closure): Po_ver¼ (15/90 0.2) ¼ 0.03
Fig. 12 e Illustration of fuel assembly deformation at impact velocity of 160 m/s for different impact locations. (A) Case 1. (B) Case 2. (C) Case 3. (D) Case 4. (E) Case 5.
For the fuel assembly response analysis, the cladding strain ratios have been calculated numerically. The strain failure criteria that would induce a cladding rupture with respect to burnup rate level are taken from SNL (Sandia Na-tional Laboratory) reports[22,23], as follows:
High burnup (55e60 GWd/ MTU): strain failure criterion (1%) Intermediate burnup (40e45 GWd/ MTU): strain failure
criterion (4%)
Low burnup (20e25 GWd/ MTU): strain failure criterion (8%) From the above criteria, it is assumed that when the scaled strain level is equal to or greater than the strain failure crite-rion, the fuel rod will fail. Furthermore, this is a conservative approach; it is assumed that 1% of the fuel rods will fail and leak at any impact load that causes strains less than the strain failure limits. Thus, the failure probability of the fuel assembly is deemed to be 1. However, fuel damage ratios (FDRs) for the impact conditions that induce strains exceeding the criteria were calculated using the practical FE modeling technique and used in the source term calculations. The steps of FDR estimation for one fuel bundle are illustrated inFig. 11, and the assumptions in the determination of the FDRs are listed as follows, with some conservatism:
▪ To investigate the maximum possible deformations, the reference fuel assembly is located close to the impacted load area.
▪ Calculated FDR and the deformed shape of the reference single fuel assembly are assumed to be the same for all other fuel assemblies.
▪ Pelletecladding interaction is not modeled, and all the masses of the pellets are lumped in the cladding without increasing rod stiffness.
The calculated FDRs for the 35 impact conditions are summarized inTable 6;Fig. 12illustrates the maximum
in-plane principal strain results, as commonly used in fuel
response to loads [1], for the five cases at the maximum
impact velocity (160 m/s).
4.
Radiological consequence analysis
The release fraction coefficient is used to estimate the amount of radioactive materials suspended in the air due to physical stress from an accident; this parameter is an essential input for radiological consequence analyses[11]. A source term is used to describe the accidental airborne release of radioactive materials in atmospheric dispersion modeling. Descriptions of the estimation process for asso-ciated radioactive material releases from the release fraction coefficients and the source term of radioactive inventory are
described in the authors' previous work [12]. Three
in-ventories for the radionuclides involved in the scenario that dominate the inhalation dose for all the chemical elements classes (volatiles, fission gases, and solid particles) are calculated using the ORIGEN-ARP of SCALE v.6.1.3 code[24] and presented inFig. 13.
Calculations for the element Co-60, called Chalk River Unidentified Deposits (CRUD), are conducted separately because this element would form by spallation of the chemi-cal deposits on the surface of fuel rod cladding during reactor operation. The activity density of CRUD is around 140mCi/cm2 on the rods from pressurized water reactors, with the total surface area of the rods being around 1,200 cm2[22]. Thus, the Co-60 inventory per rod is 2 105mCi/rod, which is added to the radionuclide inventory for the fuel assembly CE 16 16 as 51.2 Ci. The release fraction of CRUD from the cask to the environment is assumed to be 0.001, taken from the data of US Nuclear Regulatory Commission[25]and the work of Einziger and Beyer[26].Table 7displays the obtained release fractions multiplied by FDR for each event tree sequence condition and burnup rate.
The computer code HOTSPOT v.3.0.2 [27], which was developed based on the Gaussian dispersion plume model, is used in this study to estimate the radionuclide spread from the accident. The reference person is presumed to have a breathing height of 1.5 m and a breathing rate of 3.47Ee4 m3/s. The wind speed reference height is 80 m and the release height is ground level. Hourly information from 1-year mete-orological data, including wind speeds and directions, was utilized for a hypothetical site to determine the wind speed frequency groups in the scenario. In this study, three bound-aries of interest surrounding the accident location are considered: an exclusion area boundary at 560 m, a low-population zone at 5.7 km, and a low-population center distance at 7.6 km. The total effective dose equivalent at the three boundaries of interest is calculated by HOTSPOT for each event tree sequence, as shown inTable 8.
5.
Risk analysis framework
The doseerisk elements are the product of the probabilities of the impact condition and the corresponding responses (Pw, Po, and Pc), and the fractions that lead to the release of radioactive materials that cause radiological consequences of magnitude (C), as illustrated inFig. 14. The measure of risk is given by Eq. (1), which is applied in sequence line i:
Ri¼ PwPoPc Ci (1)
InFig. 14, the initiating frequency of the aircraft crash ac-cident (Column 1) is assumed to be 1, representing an inten-tional aircraft crash. As the storage wall building is the first barrier to protect the internal spent fuel storage casks from the AI, the results of the probability of perforation (Pw) for facility walls (low, median, and high) are inserted accordingly into Column 2. Then, the probability of each impact orienta-tion is added in Column 3, as discussed in Secorienta-tion3. Finally, the probability results of the cask response status, as defined in Table 3, for the three structural performance levels of metallic seal, are inserted into Column 4. All probability data are multiplied, as shown in the Probability column. The maximum possible radiological consequences for each sequence line, as shown in Table 8, are inserted into the Consequence column for the three rates of fuel burnup. Finally, through Eq.(1), the expected hazard to the public can be estimated in units of mSv/accident. Risk calculations for the three structural-resistance levels of facility/cask are summarized in the event tree for the three site boundaries of interest, as can be seen inFig. 15. Here, it should be noticed that the most severe damage consequence for the high-resistance structures is from Sequence Number 8, while for the median- and low-resistance structures it is from Sequence Number 10. Investigations reveal that an impact of a heavy foreign component on the upper part or upper corner of the cask would be the most serious one among the impact con-ditions. Further studies to enhance the design of the upper part of the cask by adding additional mitigation systems are recommended.
Overall, the risks from the potential individual peak doses on the population due to these accident conditions expected for a single cask are far below the regulatory limits at any
Tabl e 7 e Ca lculat ed release fracti ons for ea ch sequ ence and bu rnup rat e. Seq. # Case # Statu s Frel,k ¼ [RF C-E RF] RF R-C FDR Respira ble par ticulate s Volat iles Gases CR UD Lo w burn up fue l Interm ediat e bu rnup fue l Hig h burn up fuel Low bur nup fue l Interm ediat e burnup fue l Hig h bur nup fuel Low bu rnup fuel Interm ediat e burn up fue l High bu rnup fue l Same for all 2 Case 1 R D 1.03E e 10 1.03E e 10 1.3E e 09 1.911E e 11 1.91E e 11 4.45E e 10 9.6E e 04 9.6E e 04 9.6E e 04 0.001 3 S D 1.93E e 09 1.945E e 07 8.54E e 07 8.91E e 10 2.33E e 07 2.32E e 06 9.6E e 04 0.042 0.083 0.001 4 Case 2 R D 1.03E e 10 1.03E e 10 4.02E e 10 1.91E e 11 1.91E e 11 9.01E e 11 9.6E e 04 9.6E e 04 9.6E e 04 0.001 5 S D 8.95E e 10 2.61E e 09 3.71E e 07 2.54E e 10 1.64E e 09 8.08E e 07 9.6E e 04 9.6E e 04 0.048 0.001 6 Case 3 R D 1.03E e 10 1.03E e 10 7.54E e 10 1.911E e 11 1.91E e 11 2.01E e 10 9.6E e 04 9.6E e 04 9.6E e 04 0.001 7 S D 4.14E e 10 4.89E e 08 4.71E e 07 9.33E e 11 1.93E e 08 8.06E e 07 9.6E e 04 0.029 0.076 0.001 8 Case 4 R D 1.03E e 10 1.03E e 10 8.02E e 10 1.91E e 11 1.91E e 11 2.18E e 10 9.6E e 04 9.6E e 04 9.6E e 04 0.001 9 S D 1.49E e 10 6.98E e 10 2.16E e 07 2.89E e 11 1.81E e 10 2.53E e 07 9.6E e 04 9.6E e 04 0.048 0.001 10 CD 3.36E e 08 1.68E e 06 2.91E e 06 1.5E e 07 7.5E e 07 1.29E e 05 9.6E e 04 0.048 0.083 0.001 11 Case 5 R D 2.11E e 10 1.07E e 09 4.57E e 09 4.22E e 11 3.31E e 10 5.76E e 09 9.6E e 04 9.6E e 04 9.6E e 04 0.001 12 SD 3.64E e 09 6.03E e 09 9.63E e 09 3.41E e 09 1.05E e 08 3.4E e 09 9.6E e 04 9.6E e 04 9.6E e 04 0.001 13 CD 3.36E e 08 3.36E e 08 3.36E e 08 1.5E e 07 1.5E e 07 1.5E e 07 9.6E e 04 9.6E e 04 9.6E e 04 0.001 CD, containment damage; CRUD, Chalk River Unidentified Deposits; FDR, fuel damage ratio; RD, recoverable damage; SD, seal damage.
distance above the low-population zone (as the dose limit for the general public is 1 mSv/y); further, the current case sce-nario represents a very low risk of incurring latent cancer
fa-tality. This study indicates the applicability of the
methodology to ultimately reveal the risk from an aircraft crash through a realistic scenario, with some conservatism. Notably, the real risk of nuclear facility accidents includes the potential risk of land contamination and associated societal impacts, rather than the human health risk of radiation exposure only. Therefore, the model developed in this study can be used to assess these types of consequences once a specific site is determined. Moreover, new safety-related barriers or mitigation systems could be accommodated easily by the proposed framework. Thus, this study is ex-pected to provide useful information for regulators and li-censees concerning the most severe impact conditions. The results of this study demonstrate that a proper design of ISF physical protection systems, namely, concrete strength and thickness of the facility wall, as well as the seal and fuel cladding material in the storage cask, would enhance the reliability of storage facility/cask for a safe enclosure of radioactive waste materials and, consequently, may greatly reduce the risk of human-induced external events.
6.
Concluding remarks
The aim of the present article has been to confirm the flexi-bility of the proposed framework, which was developed in the authors' previous research via an examination of three com-binations of design structures of facility and cask for a hypo-thetical reference facility that presents integrated damage sequences from an aircraft crash scenario. In addition, a best-estimate structural analysis, including fuel assembly damage, has been incorporated using practical FE modeling techniques to simulate accurately and efficiently the structural response under energetic loading. Subsequently, the radiological con-sequences for a hypothetical site were calculated, and indi-vidual risks for each possible sequence were obtained through a quantitative event tree analysis. From the probability anal-ysis, which takes into account parameter uncertainties with assumptions for three different facility capacities, the conclusion is that an AI on a single fully loaded storage cask is not expected to cause a radiological impact that exceeds reg-ulatory limits under the worst-case scenario for the three sets of design structures studied here. To achieve the final goal of complete risk assessment, as detailed in the Introduction, a fire scenario and a secondary mechanical impact scenario on Table 8 e Calculated total effective dose equivalents (mSv).
Seq. # Total effective dose equivalents (mSv)
Low burnup fuel Intermediate burnup fuel High burnup fuel
EAB LPZ PCD EAB LPZ PCD EAB LPZ PCD
2 6.00Ee02 8.38Ee04 5.52Ee04 6.16Ee02 8.60Ee04 5.69Ee04 9.45Ee02 1.32Ee03 8.73Ee04 3 8.13Ee02 1.14Ee03 7.48Ee04 4.48Eþ00 6.25Ee02 4.12Ee02 2.33Eþ01 3.25Ee01 2.14Ee01 4 6.00Ee02 8.38Ee04 5.52Ee04 6.16Ee02 8.60Ee04 5.69Ee04 7.04Ee02 9.83Ee04 6.50Ee04 5 6.92Ee02 9.67Ee04 6.37Ee04 1.18Ee01 1.65Ee03 1.08Ee03 1.02Eþ01 1.42Ee01 9.34Ee02 6 6.00Ee02 8.38Ee04 5.52Ee04 6.16Ee02 8.60Ee04 5.69Ee04 7.99Ee02 1.12Ee03 7.37Ee04 7 6.36Ee02 8.88Ee04 5.85Ee04 1.21Eþ00 1.67Ee02 1.10Ee02 1.29Eþ01 1.80Ee01 1.19Ee01 8 6.00Ee02 8.38Ee04 5.52Ee04 6.16Ee02 8.60Ee04 5.69Ee04 8.11Ee02 1.13Ee03 7.49Ee04 9 6.05Ee02 8.45Ee04 5.57Ee04 7.49Ee02 1.05Ee03 6.92Ee04 5.96Eþ00 8.32Ee02 5.48Ee02 10 4.56Ee01 6.35Ee03 4.18Ee03 3.82Eþ01 5.33Ee01 3.51Ee01 7.91Eþ01 1.10Eþ00 7.26Ee01 11 6.12Ee02 8.55Ee04 5.63Ee04 8.33Ee02 1.16Ee03 7.69Ee04 1.83Ee01 2.55Ee03 1.68Ee03 12 1.01Ee01 1.41Ee03 9.32Ee04 1.95Ee01 2.72Ee03 1.79Ee03 3.19Ee01 4.44Ee03 2.93Ee03 13 4.56Ee01 6.35Ee03 4.18Ee03 8.22Ee01 1.15Ee02 7.54Ee03 9.74Ee01 1.36Ee02 8.94Ee03 EAB, exclusion area boundary; LPZ, low-population zone; PCD, population center distance.
Fig. 15 e Aircraft event tree progression of probability, consequence, and risk estimation (mSv/accident) results for one cask. EAB, exclusion area boundary; LPZ, low-population zone; PCD, population center distance.
multiple casks can be added to the proposed framework in order to evaluate accurately the overall associated radiological consequences to the public and the environment following an aircraft crash.
The recommendations of this study give insight
regarding the epistemic uncertainty in the available data for impact conditions and the sensitivity of the design param-eters, which should be handled in a more practical manner in future research. Additional consideration of the material degradation properties of the spent fuel during the storage period should be given to control and regulate the trans-portation process and storage condition criteria, as well as to identify potential operating risks for enhancing safety operating procedures within different hypothetical accident conditions.
Conflicts of interest
The authors have no conflicts of interest to declare.
Acknowledgments
This work was supported by the Nuclear Safety Research Program through the Korea Foundation of Nuclear Safety (KOFONS), granted financial resources by the Nuclear Safety and Security Commission (NSSC), Republic of Korea (No. 1305032).
r e f e r e n c e s
[1] U.S. Nuclear Regulatory Commission, A pilot probabilistic risk assessment of a dry cask storage system at a nuclear power plant, NUREG-1864, 2007.
[2] NUS Corporation, Review of proposed dry-storage concepts using probabilistic risk assessment, EPRI NP-3365, 1984. [3] S. Lee, W.-S. Choi, K.-S. Seo, Safety assessment of a metal
cask under aircraft engine crash, Nucl. Eng. Technol. 48 (2016) 505e517.
[4] J.-H. Yoon, W.-S. Choi, S. Lee, K.-S. Seo, Arising technical issues in the development of a transportation and storage system of spent nuclear fuel in Korea, J. Nucl. Eng. Technol. 43 (2011) 413e420.
[5] K. Shirai, K. Namba, T. Saegusa, Safety analysis of dual purpose metal cask subjected to impulsive load due to aircraft engine crash, J. Power Energy Syst. 3 (2009) 72e82. [6] B. Droste, H. Volzke, G. Wieser, L. Qiao, Safety margins of
spent fuel transport and storage casks considering aircraft crash impacts, Radioact. Mater. Transp. 13 (2002) 313e316. [7] G. Marchaud, L. Vilela, S. Nallet, Designing a radioactive
material storage cask against airplane crashes with LS-DYNA, 3rd International LS-DYNA Users Conference, Dearborn, Michigan, USA, 2014, pp. 1e11.
[8] G. Wieser, L. Qiao, A. Eberle, H. Volzke, Thermo-mechanical finite element analyses of bolted cask lid, Packag. Transp. Storage Secur. Radioact. Mater. 15 (2004) 223e230.
[9] G. Wieser, L. Qiao, H. Volzke, D. Wolff, B. Droste, Safety analysis of casks under extreme impact conditions, Packag. Transp. Storage Secur. Radioact. Mater. 15 (2004) 141e147. [10] J. Stepan, J. Maly, I. Holub, Consequences of the large
commercial aircraft crash into the interim spent fuel storage facility, Proceedings of SMiRT 18, Beijing, China, 2005. [11] U.S. Department of Energy, Accident analysis for aircraft
crash into hazardous facilities, DOE-STD-3014, 2006. [12] B. Almomani, S. Lee, D. Jang, H.G. Kang, Probabilistic risk
assessment of aircraft impact on a spent nuclear fuel dry storage, Nucl. Eng. Des. 311 (2017) 104e119.
[13] B. Thomauske, Realization of German concept for interim storage of spent nuclear fuel¼current situation and prospect, Proceedings of WM03, Tucson, AZ, USA, February 23e27, 2003.
[14] D. Wolff, M. Jaunich, U. Probst, S. Nagelschmidt,
Understanding the long-term behavior of sealing systems and neutron shielding material for extended dry cask storage, Probabilistic Safety Assessment and Management PSAM 12, PSAM12, Honolulu, Hawaii, June 2014.
[15] B. Almomani, S. Lee, H.G. Kang, Structural analysis of a metal spent-fuel storage cask in an aircraft crash for risk assessment, Nucl. Eng. Des. 308 (2016) 60e72. [16] Nuclear Energy Institute, Methodology for performing
aircraft impact assessments for new plant designs, NEI 07e13, Revision 8P, 2011.
[17] E. Kausel, Speed of Aircraft, Ch. 3 [Internet]. 2002. Available from:http://web.mit.edu/civenv/wtc/. [Accessed April 2016]. [18] G. Kessler, A. Veser, F.-H. Schlu¨ter, W. Raskob, C. Landman,
J. P€asler-Sauer, The Risk of Nuclear Energy Technology, Springer, Berlin Heidelberg, 2014. http://doi:10.1007/978-3-642-55116-1. ISBN: 978-3-642-55116-1 (eBook).
[19] A. €Oztemel, S. S‚ensoy, Mathematical model for the probability distribution of in-situ concrete compressive strength in north Cyprus, in: 29th Conference on Our World in Concrete& Structures, Singapore, 2004, pp. 387e395. [20] Livermore Software Technology Corporation, LS-DYNA
Keyword User's Manual, Version 971, 2007. California. [21] Dassault Systemes, ABAQUS Unified FEA: Complete
Solutions for Realistic Simulation [Internet]. 2014. Available from:http://www.3ds.com/products-services/simulia/ products/abaqus. [Accessed November 2016].
[22] Sandia National Laboratory, Reexamination of spent fuel shipment risk estimates, NUREG/CR-6672, 2004.
[23] Sandia National Laboratory, Data and methods for assessment of the risks associated with the maritime transport of radioactive materials results of the SeaRAM program studies, SAND98e1171/2, 1998.
[24] I.C. Gauld, S.M. Bowman, J.E. Horwedel, ORIGEN-ARP: automatic rapid processing for spent fuel depletion, decay, and source term analysis, Oak Ridge National Laboratory, Oak Ridge (TN), 2006. ORNL/TM-2005/39.
[25] US Nuclear Regulatory Commission, Spent fuel transportation risk assessment, NUREG-2125, 2012.
[26] R.E. Einziger, C. Beyer, Characteristics and behavior of high-burnup fuel that may affect the source terms for cask accidents, Nucl. Technol. 159 (2007) 134e146.
[27] S.G. Homann, HotSpotdHealth Physics CodesdVersion 2.071dUser's Guide, National Atmospheric Release Advisory Center, Lawrence Livermore National Laboratory, 2010.
