Nội dung

Nuclear Engineering and Technology 50 (2018) 1387e1401 Contents lists available at ScienceDirect Nuclear Engineering and Technology journal homepage: www.elsevier.com/locate/net Original Article Dynamic assessment of the seismic isolation influence for various aircraft impact loads on the CPR1000 containment Runyu Mei a, b, Jianbo Li a, b, *, Gao Lin a, b, Xiuyun Zhu c a c State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian 116024, China Plant Site and Civil Engineering Department, Nuclear and Radiation Safety Center, Ministry of Environmental Protection of PRC, Beijing, 100082, China a r t i c l e i n f o a b s t r a c t Article history: Received 25 April 2018 Received in revised form 29 July 2018 Accepted 4 August 2018 Available online 6 August 2018 An aircraft impact (AI) on a nuclear power plant (NPP) is considered to be a beyond-design-basis event that draws considerable attention in the nuclear field. As some NPPs have already adopted the seismic isolation technology, and there are relevant standards to guide the application of this technology in future NPPs, a new challenge is that nuclear power engineers have to determine a reasonable method for performing AI analysis of base-isolated NPPs. Hence, dynamic influences of the seismic isolation on the vibration and structural damage characteristics of the base-isolated CPR1000 containment are studied under various aircraft loads. Unlike the seismic case, the impact energy of AI is directly impacting on the superstructure. Under the coupled influence of the seismic isolation and the various AI load, the flexible isolation layer weakens the constraint function of the foundation on the superstructure, the results show that the seismic isolation bearings will produce a large horizontal deformation if the AI load is large enough, the acceleration response at the base-mat will also be significantly affected by the different horizontal stiffness of the isolation bearing. These concerns require consideration during the design of the seismic isolation system. © 2018 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Keywords: CPR1000 containment Seismic isolation bearing Aircraft impact Dynamic behavior 1. Introduction In the event of an extreme accident, such as an earthquake, a large fire, or an aircraft impact (AI), the safety and integrity of nuclear power plants (NPPs) should be ensured, and the release of radioactive materials should be precluded. Since the 9/11 incident, the United States has promulgated several regulations, such as [1] and has decided that the design of new NPPs must be evaluated against hostile impacts from large-scale commercial aircraft. The research pertaining to AIs on NPPs was initiated by Ref. [2]. The following scholars' research is mainly from the following research perspectives: the global structural damage, local structural damage, and effects of fuel-initiated fires, as well as the functional failure of the structures, systems, and components (SSCs) due to the induced vibrations in the structural members and safety related equipment [3]. b Institute of Earthquake Engineering, Dalian University of Technology, Dalian 116024, Liaoning, China * Corresponding author. No.2 Linggong Road, Ganjingzi District, Dalian City, Liaoning Province, China. E-mail address: jianboli@dlut.edu.cn (J. Li). [4] used the forceetime history method and missileetarget interaction method to analyze the vibration characteristic of the NPP under an impact by a 747-400 aircraft, and the sensitivity of the results on the assumed Riera force-loading area was evaluated. [5] analyzed the impact of the tendon prestressing, impact angle, and impact position on the impact load and dynamic response of the containment. [6] conducted a safety assessment of an A92 reactor building under a Boeing 747 impact by analyzing different parameters through a multistep process. [7] conducted a comprehensive review of the AI analyses of nuclear-safety-related concrete structures. [8] appraised the vibration safety of the internal equipment and components in the primary auxiliary buildings. [9] and [10] analyzed the influence of an induced fire from an AI on the outer containment of an NPP. [11] and [12] have done a study of the mechanical properties of reactor pressure vessels under thermal shock, which can be applied to the thermal stress analysis induced by a fire caused by AI. In an AI evaluation, the NPP is often assumed to be a structure without a seismic isolation system, and the boundary conditions are generally considered to be fixed boundaries. The previous nuclear accidents caused by strong earthquakes, such as the Kashiwazaki Kariwa nuclear accident in July 2007 and the Fukushima nuclear accident in March 2011, have directly led to https://doi.org/10.1016/j.net.2018.08.003 1738-5733/© 2018 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/). 1388 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 a substantial increase in the current seismic fortification standards for NPPs. Thus, the NPP designers doubt the aseismic capabilities of traditional earthquake resistance measures. Seismic isolation is proven as an effective technology for reducing the seismic response of a superstructure from its base to negotiate the destructive ground movement. The guidance for the seismic isolation of NPPs is presently available in ASCE4-16 and other relevant standards. Seismic isolation has been deployed in the NPPs of France and South Africa, thus, generating a new concern regarding the impact analysis of an aircraft on a base-isolated NPP. [13] noted that an NRC-funded research project studied the topic of AI analyses for seismically isolated nuclear structures and stated that the safety of the isolated structures under an AI should be evaluated. Currently, there is little research pertaining to this aspect of AI analyses. For an NPP with a seismic isolation system, the flexible isolation layer weakens the constraint function of the foundation to the superstructure. The seismic energy that is transmitted from the underground region to the superstructure is weakened by isolator, shown in Fig. 1(a). Unlike the seismic load case, the aircraft load case consists of a short-duration impact occurring in a single arbitrary direction in space on superstructure, shown in Fig. 1(b). The impact energy is directly acting on the superstructure. Under the impact of such loads, the influence of isolation bearings on the dynamic behavior of NPPs is needed to be studied, such as the local damage of the containment and the vibration response of the NPPs. In this study, the dynamic characteristics of the base-isolated CPR1000 containment under AI are analyzed, and based on these results, a reasonable analysis model is proposed. This paper is organized into the following sections. Section 2 establishes a representative detailed three-dimensional (3D) model of the CPR1000 containment. A series of six isolation bearings are selected, and the isolation performance of each isolation bearing is analyzed and verified in Section 3. An AI containment analysis is conducted with the six isolation bearings. In Section 4, based on the 3D finite element (FE) model of the containment, three different AI loads are considered. The distributions of the concrete plastic strain, displacement response of the containment, and acceleration response of the containment are analyzed. Finally, the conclusions are presented in Section 5. illustrated in Fig. 2 and is mainly composed of three parts: the concrete containment, steel lining, and prestressed reinforcement. The concrete containment includes the base mat, cylinder, and dome. The inner diameter of the containment is 37 m, the height of the cylinder is 50.11 m, the total height from the bottom of the plate to the dome is 66.68 m, the normal thickness of the dome is 0.8 m, and the normal wall thickness of the cylinder is 0.9 m. Two layers of the circumferential and vertical prestressed reinforcement are shown in Fig. 4 and Fig. 5, respectively. The upper end of the vertical prestressed reinforcement is anchored at the top surface of the ring beam, and the lower end is anchored to the base mat of the containment. In addition, three layers of prestressed reinforcement, which are anchored on the ring beam, are arranged in the dome. The inner side of the concrete containment structure is also connected with a layer of 6 mm thick steel lining, which is connected to the concrete containment with rivets. 2.2. FE modeling of the containment 2.2.1. Concrete constitutive model When conducting a structural dynamic analysis, a proper and reliable model that can reflect the concrete material behavior at high strain rates is needed. [15] used the Winfrith concrete model to simulate the impact resistance of steel-plate concrete and reinforced concrete panels. [16] used the plasticity concrete model to simulate an AI on the concrete containment. [8] evaluated the vibration safety of the primary auxiliary buildings using the CSCM concrete model. In this study, the plasticity concrete model is selected. This model includes the damage and strain rate effects and has been widely used for modeling concrete members subject 2. Geometric and 3D FE modeling of the containment 2.1. Geometric modeling of the containment The structure of the CPR1000 pressurized water reactor is Fig. 2. CPR1000 containment. Fig. 1. The schematic diagram of an NPP with seismic isolation under a seismic load and AI. R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 Dsr ¼ 1389 P ðresidual yield surfaceÞ; a1f þ a2f P (3) In Eq. (4), after reaching the initial yield surface but before the maximum failure surface, the current surface Dsf is obtained as a linear interpolation between Dsm and Dsy . In Eq. (5), after reaching the maximum surface the current failure surface Dsh is similarly interpolated between Dsm and Dsr . The function h(l) varies from 0 to 1 depending on the accumulated effective plastic strain parameter l. This function would normally begin at 0 at l ¼ 0, increase to 1 at some value l ¼ lm, and then decrease to 0 at some larger value of l. lm is defined simply as the value of l corresponding to the first relative maximum of h. Based on this, the elasticity, plasticity and softening deformation of concrete can be qualitatively described. Fig. 3. FE model of the CPR1000 containment. Dsf ¼ hDsm þ ð1  hÞDsy ; (4) and Dsh ¼ hDsm þ ð1  hÞDsr ; (5) The main advantage of this model is that one user input is required, the unconfined compressive strength. Other parameters are automatically generated within the program. The required concrete parameters include the density (r ¼ 2500 kg=m3 Þ, Poisson's ratio ðn ¼ 0:2Þ; uniaxial compressive strength ðf c ¼ 41 MPa), and uniaxial tensile strength ðf t ¼ 2:85 MPaÞ [18]. To consider the strain rate effect of concrete, a dynamic increase factor (DIF) is introduced in this study. The DIF represents the ratio of the dynamic-to-static strength versus the strain rate. The DIF of the compressive strength for CEB is given in Eq. (6) and Eq. (7) [19,20]. Fig. 4. Positive view of the prestressed reinforcement. CDIF ¼ fcd ¼ fcs  ε_ d ε_ cs 1:026a ε_ d  30s1 (6) and Fig. 5. Top view of the prestressed reinforcement. CDIF ¼ to high-velocity impacts, such as projectile perforation, reinforced concrete targets, and AI on a concrete containment. This concrete model decouples the volumetric and deviatoric responses of concrete. The volumetric response is easily captured via a tabulated input which gives the current pressure as a function of current and previous minimum (most compressive) volumetric strain. A three-failure-surface model is used to analyze deviatoric responses, as shown in Fig. 6. During initial loading or reloading, the deviatoric stresses remain elastic until the stress reaches the initial yield surface. The deviatoric stresses can then further increase until the maximum yield surface is reached. Beyond this stage the response can be perfectly plastic or soften to the residual yield surface [17]. a0y , a1y , a2y , a0 , a1 , a2 , a1f , a2f are material constants. Mean stress P ¼  ðs1 þ s2 þ s3 Þ=3. s1 , s2 , s3 are the first, second and third principal stresses respectively. Dsy ¼ a0y þ P ðinitial yield surfaceÞ; a1y þ a2y P (1) Dsm ¼ a0 þ P ðmaximum yield surfaceÞ; a1 þ a2 P (2) fcd ¼ gð_εd Þ1=3 ε_ d > 30s1 ; fcs (7) where f cd is the dynamic compressive strength at the strain rate ε_ d ; fcs is the static compressive strength at strain rate ε_ cs ð_εcs ¼ 30  106 s1 Þ; logg ¼ 6:156a  0:49 and a ¼ ð5 þ 3f cu =4Þ1 ; f cu is the static compressive strength. The DIF of concrete the tensile strength is determined by Eq. (8) and Eq. (9). TDIF ¼ ftd ¼ fts   ε_ d d ε_ d  1s1 ε_ ts (8) and TDIF ¼  1=3 ε_ ftd ε_ d  1s1 ; ¼b d ε_ s fts (9) where ftd is the dynamic tensile strength at the strain rate ε_ d ; fts is the static tensile strength at the strain rate ε_ ts ð_εts ¼ 1Þ; log b ¼ 6d  ‘ ; and f ‘ ¼ 10MPa; f ‘ is the static tensile 2, d ¼ 1=1 þ 8fc‘ =fco co c strength. This constitutive model was used in the latest LS-DYNA software code [14], corresponding to keyword card Mat_Concrete_Damage_Rel3. 1390 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 Fig. 6. Strength model for concrete, (a) failure surface of the concrete material model and uniaxial stress-strain relationship. 2.2.2. Metal structural model Fig. 7 shows the elasticeplastic model with the kinematic hardening material model that is used to model the behavior of the prestressed reinforcement and steel lining. In this paper, kinematic hardening of the metal was considered by setting the parameter b ¼ 0. The prestressed reinforcement is embedded in the concrete using the option *CONSTRAINED_LARGRANGE_IN_SOLID. This method does not require the consideration of complex reinforcement modeling problems, while maintaining good accuracy and stability. The steel lining and concrete structure are closely connected by rivets. The steel liner is simulated by shell elements with six degrees of freedom per node, and the concrete is simulated by solid elements with three degrees of freedom per node. Considering the difference in the nodal degrees of freedom for the two elements, the keyword *CONTACT_TIED_NODES_TO_SURFACE is used to simulate the close connection between the steel lining and concrete surface. In this study, the yield strength of the rebar and structural steel is highly dependent upon the strain rate. The yield strength increases when the strain rate increases. This dynamic yield strength of steel is taken into consideration by the CowpereSymonds formula for uniaxial tension or compression, as shown in Eq. (10).  1 ε_ sd ¼1þ C sy P ; (10) where sd is the dynamic yield strength; sy is the static yield strength; ε_ is the strain rate; C and P are the constants of the CowpereSymonds relation. The material parameters of the prestressed reinforcement and steel lining are presented in Table 1 [21]. 2.2.3. Summary of the FE containment model According to the above analysis, the detailed FE model, shown in Fig. 3, is established by using the ANSYS16.0/LS-DYNA software [14]. The main components of the containment are simulated. The concrete is simulated by solid164 with multilayered solid elements along the radial direction; the steel lining is simulated by shell163; the prestressed reinforcement is simulated by link160, and the prestress load is added by applying an initial stress to the link element [22]. In order to verify the rationality of this concrete constitutive model, metal structural model and the corresponding material parameters, such as the C and P of the steel bar, the impact tests of a 1/7.5 ratio GE J79 engine with 215 m/s impact velocity on 12 cm thick reinforced concrete target plate conducted by Muto in 1989 [23] are simulated. The FE model of the engine and reinforced concrete target plate are shown in Fig. 8. The coupling interaction between steel bar and concrete in reinforced concrete plate adopts the keyword of *CONSTRAINED_LARGRANGE_IN_SOLID. Both the material model of steel bar and engine adopt the Plastic Kinematic model. Lastly, 304,164 elements are used for the entire model. The size of the impact hole obtained by numerical simulation is 198 mm  208 mm, which is very close to the experimental result 185 mm  195 mm, shown in Fig. 9. The residual velocity of the engine after perforate the panel is also very close to the experimental result 54 m/s, shown in Fig. 10. From the comparison results, the reinforcement coupling method used in the CPR1000 finite element model and the selected concrete and reinforcement material constitutive model and the corresponding material parameters are reasonable. 3. Seismic isolators At present, there are several types of new and existing isolation Table 1 Material parameters of the metal structure. Fig. 7. The kinematic hardening material model. Prestressed reinforcement Steel lining rðkg=m3 Þ E/Pa n sy =Pa C P 7850 7850 1.9e11 2.06e11 0.3 0.3 1.77e8 3.25e8 641 40.4 7.3 5 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 1391 Fig. 10. The comparison of engine residual velocity between numerical results and experimental results. Fig. 8. The FE model used in the engine impact tests. systems, such as the low damping rubber bearing (LDRB), lead rubber bearing (LRB), friction sliding isolator (FSI), high damping rubber bearing (HDRB), and friction pendulum isolator. For elastomeric bearings, the behavior of the LDRB, LRB, and HDRB in shear is well established [24,25]. The mathematical model of the shear behavior for these bearings can be simplified into a bilinear restoring force model, as shown in Fig. 11. However, the damping characteristics of these isolating bearings are considerably different. For the LDRB, owing to its low viscous damping, additional damping should be added during the design and construction to improve the damping effect. The HDRB uses a rubber with a high damping property. The LRB mainly relies on the lead core to exert the damping effect of the bearing. When the dynamic analysis is conducted, the damping of the isolation bearing affects the dynamic characteristics of the structure, especially under an AI. This work does not involve the study and design of the isolation bearings for the CPR1000 containment according to the relevant Fig. 11. Bilinear restoring force model. specifications; this study is a qualitative analysis of the influence of the isolation bearings on the dynamic behavior of the CPR1000 containment under different AI loads. The influence of the isolation bearings on the horizontal stiffness is mainly considered, and the additional damping is assumed to be small. In the LS-DYNA code, the isolators can be modeled by a linear spring Fig. 9. Comparison of the damage of front face of panel between numerical results and experimental results (units: mm). 1392 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 analysis result shows that the isolation bearings are entirely experienced compression deformation. [24,25] summed up the important conclusions of other scholars’ experimental work on the tensile properties of elastomeric bearings. One of them shows the loadedeformation behavior in tension is linear up to cavitation with the tensile stiffness approximately equal to the compressive stiffness, followed by nonlinear post-cavitation behavior. So, the vertical tension stiffness coefficient of the isolation bearing is considered to be the same as the compression stiffness coefficient. Lastly, a series of six different isolation bearings are determined by trial calculation. However, prior to the AI analysis, the isolation performance of the selected isolation bearings is analyzed. Thus, combined with the non-isolation condition, seven different boundary conditions are composed. The six different isolation bearings are represented by case 1 through case 6, and the material parameters of these isolation bearings are given in Table 2. Case 7 represents the non-isolation condition. Fig. 12. Layout of the isolators. Table 2 Various cases for the isolation bearings. Parameter Unit case 1 case 2 case 3 case 4 case 5 case 6 Design load Equivalent stiffness Pre-yielding stiffness Post-yielding stiffness Yield force Te kN kN/m kN/m kN/m kN mm 706 882 3091 476 23.6 58 1256 1325 4647 715 41.9 68.6 1963 1480 5187 798 65.4 96 2827 1859 6519 1003 94.2 110 3848 2531 8873 1365 128.2 110 6361 2841 9959 1532 212 162 case 1 case 2 case 3 case 4 case 5 case 6 case 7 For superstructures, the target of the design is to reach “seismic isolation” by designing the stiffness and damping of the isolation layer to reasonably extend the first period of structure, to dissipate the seismic energy, and to moderately control the displacement of the isolation layer. Generally, the structure mostly moves in the horizontal plane with rigid body behavior under seismic loads [26,27,28]. To a certain extent, the motion of the base mat can reflect not only the motion of the whole structure, but also the deformation of the isolation bearings. The decreasing amplitude ratio (DAR) can reflect the decreasing function of the superstructure response under earthquake excitation. 0.71 0.72 0.90 0.89 0.89 1.10 0.94 0.94 1.13 1.05 1.05 1.29 1.21 1.21 1.44 1.30 1.30 1.58 4.09 4.13 6.39 DAR ¼ Table 3 Natural frequency (Hz). Mode no. 1 2 3 3.1. Seismic analysis Case (*MAT_SPRING_ELASTIC) in the vertical direction and nonlinear spring (*MAT_SPRING_ELASTOPLASTIC) and viscous damper (*MAT_DAMPER_VISCOUS) in horizontal directions, as shown in Fig. 12. Tensile deformation in elastomeric bearings has traditionally been considered undesirable. In the preliminary analysis, the vertical deformation analysis of the isolation bearing under AI is carried out (The AI loads will be introduced in section 4). And the Sa  Sb ; Sb (11) where DAR is the decreasing amplitude ratio; Sa is the response of the non-isolated structure; Sb is the response of the isolated structure. Lastly, the maximum resultant horizontal displacements of the base-mat and DARs of the dome vertex for the above seven cases are analyzed and compared. After generating the 3D FE models of the main structures, the modal and seismic analyses are conducted for case 1 through case 7. Table 3 presents the modal results, which indicate that the Fig. 13. Acceleration time histories of RG1.60 in each direction (a) X and (b) Y. R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 1393 Fig. 14. Acceleration time histories of El Centro in the direction (a) X and (b) Y. Fig. 17. Impact location and direction. Fig. 15. The DAR and the maximum resultant horizontal displacement of the containment under different seismic loads. Fig. 16. Reaction time response of the aircrafts. vibration frequency of the structure is associated with the horizontal stiffness of the isolation bearings. For the seismic analysis, two types of seismic waves are considered as the seismic oscillation input. Additionally, only the effects of horizontal earthquakes are considered. One seismic oscillation input is the artificial seismic wave derived from the Regulatory Guide 1.60 reference response spectra (RG1.60), as shown in Fig. 13. The peak ground acceleration for the X and Y directions is equal to 0.3 g. The other seismic oscillation input is the El Centro seismic wave, as shown in Fig. 14. Based on the numerical results, the maximum resultant displacements of the base mat and DAR of the dome vertex for each case under these two seismic loads are shown in Fig. 15. The lateral axis in the diagram represents the first vibration period for each case. Considering these two types of seismic loads, the variation of the DAR and the maximum resultant horizontal displacement are similar. From case 1 to case 6, the DAR and maximum resultant horizontal displacement will increase with an increase in the vibration period. The growth rate of the DAR slowly decreases, while the maximum resultant horizontal displacement demonstrates the opposite trend. One of the isolation system design criteria requires that the 1394 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 under RG1.60 excitation is 66.93%, and 68.27% under El Centro excitation; the maximum resultant horizontal displacement is 191.95 mm under RG1.60 excitation exceeding 3 Te, 163.59 mm under El Centro excitation. The maximum resultant horizontal displacements from case 2 through case 6 do not exceed 3 Te, thus, meeting the relevant requirements. For case 6, the DAR reaches 18.14% under El Centro excitation, which indicates that the isolation bearing corresponding to this case does not have good isolation performance. 4. Numerical simulation and analysis of an AI on the containment Fig. 18. The reference points of the CPR1000. calculated horizontal maximum resultant displacement of the isolated structure must remain less than 3 Te (Te is the total thickness of the rubber, presented in Table 2). For case 1, the DAR There are several different characteristics to consider for an AI on an NPP. The plane has complicated deformation and failure mechanisms, the impact area varies in size and shape with respect to time, and the impact load varies with respect to time. Presently, there are two main analytical methods that can currently be implemented for an AI analysis: the forceetime history analysis method and missileetarget interaction analysis method. The former only needs to specify the impact area and the impact time Fig. 19. The plastic strain distribution of the containment under each aircraft load (a) F4, (b) A320, and (c) 707-320. R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 1395 Fig. 20. The X displacement response of the center node of the impact location and the center node of the base mat under the impact of each aircraft load (a) F4, (b) A320, and (c) 707-320. 1396 R. Mei et al. / Nuclear Engineering and Technology 50 (2018) 1387e1401 Table 4 The maximum X displacement of the center node of the impact location during the aircraft impact process (mm). F4 A320 707-320 case 1 case 2 case 3 case 4 case 5 case 6 case 7 83.746 82.595 99.372 83.559 81.444 100.42 83.811 81.748 98.622 83.891 81.34 97.172 83.679 80.436 101.54 83.922 80.093 96.207 85.122 73.608 84.89 history load and does not need to build a complex threedimensional finite element model, so that the analysis efficiency is greatly improved. The accuracy and reliability of the results generated by this method are sensitive to the loading area and timing of load application. [29] noted that the loading surface of the AI will influence the deflection, damaged area, and reaction force of the slab. There are two main approaches to consider for calculating the impact area. One divides the impact area into three regions: the fuselage, wing, and engine [30]. The other simplifies the impact area into a circular area [31e34]. M.R. Sadique and Mohd. Ashraf Iqbal considered that when the purpose is to evaluate the global response of the containment, simplifications can be made for the determination of contact area. They analyzed the impact area of Phantom F4 (215 m/s), Airbus A-320 (120 m/s) and Boeing 707-320 (103 m/s) is a circular area of 6 in diameter, Boeing 747-400 (120 m/ s) and Boeing 767-400 (120 m/s) is a circular area of 12 in diameter. Taking into account the characteristics of different AI loads, fighter Phantom F4 (215 m/s), commercial aircraft Airbus A-320 (120 m/s), and Boeing 707-320 (103 m/s) are selected, which are referred to as F4, 707-320, and A320, respectively, shown in Fig. 16. For each of the aircraft loads, the impact area is assumed to be circular with a diameter of 6 m. One impact location is considered, which is most critical location located at the junction of the dome and cylinder [33], as shown in Fig. 17. The impact direction is also shown in Fig. 17. Recent research on AIs has mainly focused on the deformation, destruction, and vibration of NPPs during the aircraft collision. However, relatively little research concerning the dynamic response of the structure in the free attenuation vibration process has been conducted. During this process, the vibration characteristics of a base-isolated NPP are likely to be different from those of a non-isolated structure. The isolation bearings play a major role in the attenuation of the vibration response. Based on the results of a trial calculation, the calculation time of the entire process is fixed at 3.0 s. The initial 0.2 s is the stable stage of the prestressed reinforcement with large damping. The time range of 0.2 se3.0 s is for the dynamic calculation process, which is the main consideration of this analysis. This main portion of the analysis contains two processes: the AI process and vibration attenuation process. The influence of the isolation bearings on the dynamic behavior of the containment is analyzed and compared by the distribution of the ultimate plastic strain distribution and vibration response of the containment. Given the stable construction of the containment, which consists of the prestressed reinforcement, cylinder, buttress, ring beam, and steel lining, the AI can damage the structure within the vicinity of the impact area, but the entire structure cannot be damaged. Under the AI loads, the structure will not incur serious damage if the stiffness is sufficiently high. The AI directly acts on the superstructure over a short duration of time, and therefore, the seismic isolation system is unable to influence the plastic strain distribution of the containment. Consequently, the possibility of excessive displacement and acceleration responses may cause internal equipment failure. 4.2. Dynamic deformation analysis and results In the analysis of the local damage characteristics of the containment, the deformation of the impact zone requires consideration. The maximum displacement of the center node of the impact location (in Fig. 18) during the AI process is analyzed, as shown in Fig. 20. Owing to the high rigidity and integrity of the base mat, the horizontal displacement of the base mat is considered consistent. Thus, the displacement response of the base mat can be used to effectively analyze the real motion of the containment and to reflect the deformation of the isolation bearings. The maximum displacements for the seven cases are presented in Table 4. The maximum displacement of the isolated structure is approximately 83.7 mm under the F4 aircraft load, 81.0 mm under the A320 aircraft load, and 99.0 mm under the 707-320 aircraft load. It can be seen from Fig. 20 that during the AI process, the maximum deformation of the rubber bearing does not exceed 50 mm under these three kinds of aircraft loads. The isolation bearings mainly affect the structural response of the free attenuation vibration process and do not produce a notable effect on the response during AI process. As previously mentioned, the additional damping of the 4.1. The plastic strain distribution of the concrete containment Fig. 19 shows the plastic strain distribution of the containment under various aircraft load. The plastic strain distribution for each aircraft is found to be mainly located on the cylinder positioned on both sides of the impact location, at the medial region of the buttress, and on a portion of the buttress. The ring beam and buttress strengthen the overall stability of the containment, and thus, the distribution of the plastic strain is not dispersed to other portions of the cylinder. Regarding the isolated structure, each aircraft load produces a similar plastic strain region distribution. Fig. 21. The X displacement response of the center node of the impact location under the A320 aircraft load for case 2 and case 3 with high damping.