Several Nuclear Power Plants established 5 decades ago have been faced on decommissioning process in Japan. The amount of concrete wastes generated from demolished Nuclear Power Plants are about 5 hundred thousand tons for a typical 1,100Mw class unit. Basically Nuclear Power Plants were used advanced and controlled concrete which met Architectural Standard Specification JASS 5N ”Reinforced Concrete Work at Nuclear Power plants”. Even if demolished concrete include quality controlled aggregates, they will be reclaimed without using for load material.

 Research and developing about recycle aggregate concrete started at 1974, first almost all of studies were connected to mix proportion and strength. But studies about quality controlled of recycle aggregates have increased in perspective of durability of concrete with recycled aggregate. Architectural Institute of Japan standardized the use of high quality recycle aggregate in Architectural Standard Specification, JASS 5 “Reinforced Concrete Work” in 2003, and Japanese Standards Association specified Japan Industrial Standard, JIS A 5021” Recycled Aggregate for Concrete Class H” in 2005.

 This paper presents that 5 high-quality recycled coarse aggregates were produced from two types of demolition concrete, and the characteristics of different methods and the effects of raw material were investigated through experimental studies.

 ・Five types of recycled coarse aggregate met to JASS 5N or JIS A 5021.

 ・The grain size distribution of recycled aggregate differs depending on the method, and the volume around 15 mm increased more than other sizes.

 ・Despite the ratio of adhered mortar was less than 20%, there was a strong correlation between the oven-dry density and the ratio of adhered mortar, between the ratio of water absorption and the ratio of adhered mortar.

 ・Only the mortar particles were contained in the recycled aggregate because the strong mortar particles were not destroyed and eventually generated mortar particles with friction surface.

 ・Measuring the amount of mortar particle was efficient and convenient to check the quality of recycled coarse aggregate in manufacturing process.

 ・Demolished concrete from Nuclear Power Plants was appropriate for raw material to produce high quality recycle aggregate due to confirm profile of original concrete and characteristic of original aggregate.

 The results were found to be acceptable for Nuclear Power Plants and the use of high quality recycled aggregates have been specified as a Special Concrete in Section 15.4 “Recycled Aggregate Concrete” of JASS 5N in 2013.

 Autoclaved lightweight aerated concrete (ALC) is a building material with tobermorite as the major constituent mineral, and made by Portland cement, silica and lime. On account of the most striking characteristics of thermal insulation properties and fire resistance, ALC is widely used as walls, floors and roofs. Tobermorite and xonotlite are well known as typical calcium silicate hydrates, in particular, xonotlite has long been used as a heat insulating material for high temperatures. However, it was difficult to synthesize xonotlite using cement because aluminum ions were the factor of inhibiting xonotlite formation and its crystal growth.

 The authors are aiming to manufacture xonotlite-based ALC using cement as a raw material in order to further improve the thermal stability of ALC. So far, we have studied the autoclaving conditions for the synthesis of xonotlite. In addition, we proposed a method to understand the state of xonotlite formation by measuring the heat shrinkage by thermomechanical analysis (TMA). In this study, the mixing, casting and autoclaving conditions for the production of xonotlite-based ALC were examined.

 The following conclusions were obtained.

 1) If the water content ratio is constant, the slurry temperature and the slurry viscosity during casting are almost linear, and the higher the slurry temperature, the higher the slurry viscosity. Additionally, with the slurry temperature is fixed, the water ratio and the slurry viscosity are almost linear, and the slurry viscosity decreases as the water content ratio increases.

 2) After casting of the slurry, the bearing strength of the green cake of the xonotlite-based ALC becomes higher and higher, especially when the water content ratio is smaller. This behavior has a close relationship with the maturity, and the relationship between the maturity and the bearing strength can be predicted for each water content ratio.

 3) The autoclaving condition at 240℃ for 24 hours is necessary to sufficiently synthesize a xonotlite and reduce the heat shrinkage rate.

 4) The compressive strength of xonotlite-based ALC is proportional to the density. However, in comparison with the previous studies on the relationship between the density, compressive strength and Young's modulus of ALC mainly composed of tobermorite, the effect of density on compressive strength of xonotlite-based ALC was smaller than that of tobermorite-based ALC.

 Headquarters for Earthquake Research Promotion (2017) compiled an official procedure for evaluating fault parameters for strong motion prediction. It adopted an asperity model as a physical basis, and defines the asperity as a strong motion generation area (SMGA) and explaines that the SMGA is a large-slip area and a high-stress-drop area. These three areas coincide on the same region on the fault of the first-stage crustal earthquakes.

 However, the 2016 Kumamoto earthquake and the 2011 off the Pacific coast of Tohoku earthquake, categorized as second-stage earthquakes, showed that the shallow part of the ruptured fault had large slip but did not generate strong motions.

 The technical terminology of asperity has been used in the field of the strong motion prediction, and has different definition such as a locked area, a high-stress-drop area, a strong-motion-generation area, and a large-slip area. I summarized the historical evolution of the technical terminology of asperity. Then, I compiled some technical terminologies relevant to asperity in recent research papers. Finally, I proposed that the high-stress-drop area should be used instead of the asperity, and it corresponds to a strongly locked area and an SMGA. Then, I named each part of the fault by categorizing the slip distribution into six types of faulting as follows:

 1) First-stage crustal earthquakes without surface breakings such as the 1997 Kagoshima-Ken-Hokuseibu earthquake (M_W 6.1). The fault consists of the high-stress-drop area and background within the seismogenic layer. The location and the size of the high-stress-drop area is almost the same as those of the large-slip area.

 2) Transition-stage crustal earthquakes with partial surface breakings such as the 2000 Tottori-Ken-Seibu earthquake (M_W 6.6). The fault consists of the high-stress-drop area and background in the seismogenic layer and the small-slip area in the surface layer. The location and the size of the high-stress-drop area is almost the same as those of the large-slip area.

 3) Second-stage and third-stage crustal earthquakes with long and clear surface breakings such as the 2016 Kumamoto earthquake (M_W 7.1) and 1999 Kocaeli, Turkey, earthquake (M_W 7.6). The fault consists of the high-stress-drop area and background in the seismogenic layer and the large-slip area (=LMGA: Long-period Motion Generation Area without strong-motion generation) and small-slip areas in the surface layer. The location and the size of the high-stress-drop area is almost the same as those of the large-slip area in the seismogenic layer.

 4) First-stage subduction plate-boundary earthquakes without surface breakings such as the 2003 Tokachi-Oki earthquake (M_W 8.1). The fault consists of the high-stress-drop area and background within the seismogenic layer. The location of the high-stress-drop area is almost the same as that of the large-slip area, and the size of the high-stress-drop area is almost half of that of the large-slip area.

 5) Transition-stage subduction plate-boundary earthquakes without very-large-slip areas such as the 2015 Illapel, Chile, earthquake (M_W 8.3). The fault consists of the high-stress-drop area and background in the seismogenic layer and the large- and small-slip areas in the shallow ruptured zone.

 6) Second-stage subduction plate-boundary earthquakes with very-large-slip areas such as the 2011 off the Pacific coast of Tohoku earthquake (M_W 9.0). The fault consists of the high-stress-drop area and background in the seismogenic layer and the very-large-, large-, and small-slip areas in the shallow ruptured zone.

 In June 2016, MLIT (Ministry of Land, Infrastructure, Transport and Tourism) has announced a countermeasure against the Nankai trough long period ground motions (LPGMs). MLIT also announced the region divided by input level of LPGMs in the Kanto (KA1), Chukyo (CH1–3), Shizuoka (SZ1–3), and Osaka (OS1–3) areas. However, the response spectrum method, indicated by the Ministry of Construction notification Vol. 2009, is out of the countermeasure, because this method was considered to be difficult to reflect the influence by LPGMs. Also, we proposed a vertical distribution of design shear force coefficient for seismically isolated buildings (SIBs) using the equivalent isolation ratio which has applicability to various dampers.

 In this paper, we analyze the influence of LPGMs on shear force coefficient distribution of SIBs based on knowledge that response amplification is excited by higher modes. And based on the analytical results, we propose a new equivalent amplification factor and variation correction coefficient. The acquired conclusions are as follows;

 1) The equivalent amplification factor of the shear force coefficient changed depending on the region divided by input level of LPGMs. In addition, it was confirmed that the number of cases where the equivalent amplification factor falls within the range of evaluation of Eq. (10) varies depending on the region divided by input level.

 2) Based on the analysis results of nonlinear modal analysis, as region divided by input level of LPGMs decreases, the 1st mode response decreases, and the influence of the higher mode increases relatively.

 3) By quantifying the relationship between the equivalent isolation ratio and the equivalent amplification factor, and the variation of the equivalent amplification factor for each region, it was possible to appropriately evaluate the tendency that the equivalent amplification factor differs depending on the region divided by input level of LPGMs.

 In time-history analysis of building structures, damping is commonly modeled by classical viscous damping models among which Rayleigh damping is favored due to its computational efficiency and ability to assign near-constant damping ratios over a wide range of frequencies. However, researchers recognize that Rayleigh damping can produce unrealistically large damping forces in elastic-plastic analysis, and, in turn, lead to underestimation of computed response. Now that ever sophisticated nonlinear analysis with high-fidelity hysteresis models is at our disposal, more attention should be directed towards damping models.

 This paper describes a fundamental study on how the time-history response of elastic-plastic systems can be affected by the selection of damping models. Comparison is made between 13 damping models listed in Table 1, including massor stiffness-proportional models, Rayleigh model, Caughey series, modal-damping proportional model, and modifications of these models. Some models require step-by-step update of the damping matrix. Models 4, 7 and 13 require step-by-step eigenvalue analysis, which by nature is computationally expensive. In the post-elastic state, the damping matrix may become nonclassical, remain classical, or always retain the original damping ratios. Fig. 2 compares the models for a partially-yielded state of a 5-DOF system, plotting the effective modal damping ratio defined in Eq. (4) against the updated eigenfrequencies. Summation was taken over all modes for the Caughey series (Model 9) and modal-damping proportional models (Models 12 and 13). The Rayleigh model (Model 5) led to large damping ratios, while the modified Rayleigh models, Models 7 and 8, suggested by Charney and Hall, respectively, retained the damping ratios near the original target value. Although computationally expensive, Model 13 retains the damping ratio under any state, and may therefore be viewed as an exact model.

 Time-history response was compared through a 5-DOF system with bilinear kinematic hysteresis (see Fig. 3) subjected to a strong earthquake ground motion. Integration over time was conducted using the central difference method, expressed by equations (10) to (12), and a time increment of 0.002 s. The resulting story drift distribution varied substantially in both elastic analysis (see Fig. 5b) and elastic-plastic analysis (see Fig. 6). Mass- or stiffness-proportional models or their variations (Models 1 to 4) resulted in large deviation from Model 13. The Rayleigh model (Model 5) resulted in somewhat smaller response than Model 13. Fig. 7 illustrates, for the third-story response, how the damping force from Model 5 was substantially greater than Model 13 when the system yielded. Fig. 9 illustrates that the damping force in Model 5 was consistently dictated by the first mode response although the normalized effective modal mass (defined in Eq. (6)) for the first mode fluctuated from the original 0.85 to a minimum 0.2 and maximum 1.0. Fig. 11 correlates damping energy (computed using Eq. (13)) and damping ratio averaged over all modes and time. Most models overestimated the damping energy while Models 7 and 8 matched Model 13 very closely.

 Time-history analysis was repeated for a 20-DOF system (see Fig. 12). Fig. 13 shows that, as in the 5-DOF system, the modified Rayleigh models (Models 7 and 8) yielded very similar results to Model 13. Fig. 14 shows that at least 5 of 20 terms must be taken for Model 13 to yield a result as accurate as Models 7 or 8.

 The non-linear finite element analysis is useful to evaluate the structural safety during a big earthquake. Constitutive models with high reproducibility against the hardening behaviors have been required to obtain the high-precision result of the finite element elasto-plastic analysis. Some effective constitutive models have been proposed, however they had high calculation cost because they employed many plastic internal variables and assumptions. The neural network is useful to overcome the disadvantages of the conventional constitutive models because of its high performance for the regression analysis and the low calculation cost. Many neural network constitutive models have been proposed by many researchers, where their inputs are the current total strain and past variables, and their output is the current stress. This type of neural network constitutive model can perform the high-precision regression analysis within the range of the training data while its performance of the regression analysis becomes low outside the range of the training data. Thus, this disadvantage makes it difficult to apply the neural network model to the finite element analysis because it is difficult to assume whether the strain and stress at integration points for finite elements is ranged within the range of training data before the finite element analysis.

 In this study, a stress calculation method for constitutive models is proposed by embedding the neural network into the return mapping algorithm, where the neural network is in charge of the calculation of the plastic corrector. The feedforward neural network is employed where the input layer with a unit, the output layer with a unit and the hidden layer with ten units are utilized. To hinder the fluctuation of the input and output values of the neural network, the variables used for the input and output layer are the normalized trial value of the yield function and absolute value of the increment of the plastic strain respectively. The mean squared error is employed as the cost function for the learning. The neural network is optimized with the optimization algorithm ADAM on mini-batchs of size 240 which is the one-fifth of the training data sets. The training datasets are generated from the stress calculation of the constitutive model using the return mapping algorithm, where the loading path is a cyclic loading with the constant strain amplitude and the total strain increment is generated from the uniform distribution. The comparison between the stresses calculated by the return mapping and the pre-trained neural network under the random loading are carried out. As the result, the maximum difference between the stress calculated by the return mapping and the pre-trained neural network is 1.4%, thus the validation of the present method is shown.

 During the 2003 Tokachioki earthquake, seven oil storage tanks of floating-roof type located at Tomakomai, Japan were seriously damaged due to liquid sloshing. Immediately after the earthquake, the Fire and Disaster Management Agency of Japan^{2), 3)} has issued the amended notification of the Fire Defense Law (FDL), in which the stress evaluation formulas for a floating roof of oil storage tank under long-period earthquake motion have been newly regulated. However, these are the empirical formulas prioritizing the practical usage, and it is of great practical significance to confirm their validity through comparison with the exact nonlinear solution for the coupled fluid-structure system.

 In the present paper, the hybrid analytical and finite element methood (FEM) proposed by the authors^{11), 12)} is employed for carrying out nonlinear sloshing analysis of coupled liquid-floating roof system. The tank is composed of a rigid cylindrical wall and a flat bottom, while the floating roof is treated as an elastic plate undergoing large deflection. The contained liquid is assumed to be inviscid and incompressible, and the flow is assumed to be irrotational. The method of analysis is based on representation of the liquid motion by superposing the analytical solutions that satisfy the Laplace equation and the rigid wall and bottom conditions. This requires only the discretization of the liquid surface and the floating roof into finite elements (see Fig. 3), thus leading to a computationally very efficient method compared with full numerical analysis. In order to model the pontoon of thin box-shaped cross-section precisely use is made of the eccentric plate and beam elements in this study (see Fig. 1). Numerical results are presented for the case of three oil storage tanks with single-deck type floating roof damaged during the 2003 Tokachioki earthquake (see Table1 and Fig. 2). Comparison is made between the results predicted by the present nonlinear analysis and the evaluation formulas notified in the FDL.

 Conclusions arising from the present study can be summarized as follows:

・As for liquid surface elevation, the contribution of linear oscillation modes with circumferential wave number 1, especially the contribution of fundamental sloshing mode with radial wave number 1, is dominant (see Figs. 8-10). However, the existence of nonlinear oscillation modes with circumferential wave numbers 0 and 2, caused by finite liquid surface elevation as well as large deformation of the floating roof can never be ignored (see Figs. 11-13).

・There exist two other sources of nonlinear oscillation modes which are excited by the internal resonant oscillation with the linear second order mode. One is the bi-harmonic resonant oscillation mode at half the fundamental sloshing period, as observed in Model 3A tank of 30,000m³ capacity and Model 4A tank of 40,000m³ capacity (see Figs. 11, 12). The other is the nonlinear oscillation mode with circumferential wave number 3, as recognized in Model 10B tank of 100,000m³ capacity (see Fig. 13).

・The nonlinear oscillation modes observed in the present study produce excessively large stresses in the pontoon, which are far beyond the evaluation based on the empirical formulas notified in the FDL (see Figs. 14-16).

・It should be emphasized that the stress evaluation formulas in the amended Notification of the FDL, taking a part of these nonlinear oscillation modes out of consideration, tend to underestimate the pontoon stresses significantly and should be improved (see Fig. 17).

 In the case of repair and reinforcement of an R/C structural frame, and installation of new members for extension and reconstruction, it is necessary to anchor such as main bars and wall reinforcement bars of new members to the existing frame. In this case, the anchor is required to reliably transfer the tension of the rebar to the existing frame. The purpose of this research is to propose a post-installed headed rebar anchor method with high reliability both in structural performance and workability, and to evaluate the anchorage performance of the method.

 In the second part of this paper, experiments in a relatively long embedment length were carried out to clarify the effects of factors such as the concrete edge distance and the number of anchors for the bond behavior of the rebar and the bearing behavior of the small headed plate. The long embedment length assumes a situation where bond deterioration near the joint surface (near the load end) and bond resistance in the length direction shift for actual use.

 The following results were obtained.

 (1) The bond strength on rebar surface of proposed method is equal to or higher than that of the cast-in-place deformed bar under the seismic load and the long-term load. The small headed plate is useful for ensuring uniform bond resistance over the entire length of the embedment length and contributes to an increase in anchor strength due to bearing resistance.

 (2) Bond strength decreases when anchoring near the concrete edge or multiple rebars are close to each other, but by using a long anchoring length, it is possible to ensure the tensile strength of the anchoring bars. In addition, there is no difference in performance depending on the construction direction.

 (3) It can be said that the bond strength on rebar surface of this method is estimated to the allowable bond stress of AIJ reinforced concrete calculation standard and bond strength of AIJ design recommendation for composite construction. The small headed plate contributes 10 to 30% increase in anchor performance.

 (4) Anchorage behavior due to long embedment length can be simulated by reducing local bond strength in consideration of a concrete cone failure strength and edge and group effects.

 Introduction

 In past earthquakes, non-structural members such as RC non-structural walls were damaged, and the damage makes it difficult to continue using the building after earthquake. For this reason, new structural systems that can be used continuously after an earthquake will be investigated. As one of them, the stress and deformation of wall are reduced by rocking behavior and energy dissipation with dampers, respectively. This paper presents the analysis model focusing on R/C wall rocking behavior, and the validity of the model is investigated.

 Outline of experiment

 The test specimen is composed of top and bottom stubs and RC box-shape multi-story structural wall, which allows rocking behavior of the wall. The dampers are attached to the wall ends and the number is changed according to the experiment stages (Experiment I：0, II / II'：4, III：8, IV：16 ) . As the number of damper increases, rocking behavior of the wall is restricted and yielding behavior of wall dominated. As for the experimental results, the rocking behavior was confirmed in Experiment I, the damping effect was confirmed in Experiment II' and Experiment III, and the collapse type transition to the bending yield of the wall was confirmed in Experiment IV.

 Outline of analysis

 As parameters of analytical models, two models were investigated: The pseudo 3d model using line element by one column, and 3d model using line element by three columns. For the pseudo 3d model, the initial stiffness considering orthogonal wall effective length is investigated, and maximum strength considering orthogonal wall effective length or full length is investigated. In addition to the conventional model, 3d model was also applied which the structural wall was vertically divided into two elements and to express antisymmetric moment distribution as confirmed in the experiment. A normal model showed lower bending stiffness than that of the experimental specimen. This is probably because when the structural wall is divided into the central column and the pin columns on both sides, the central column has a smaller cross-sectional area. Therefore, a model in which bending stiffness was calculated from the cross section of the full length of the in-plane wall was also created.

 Analysis results and Conclusions

 Regarding of the pseudo 3d model in both directions, uplifting displacement can be evaluated by the uplift and damper springs. This model can evaluate initial stiffness considering orthogonal wall effective length properly. Also, this model can evaluate uplift strength, and maximum strength of the wall considering orthogonal wall effective length properly. Regarding of the 3d model in X directions, this model can evaluate initial stiffness calculated from the cross section given by the conventional method. The evaluation of uplift strength can be evaluated same as the pseudo 3d model. And, this model can evaluate maximum strength of the wall considering the axial spring located at center of the orthogonal wall to the effective length. Regarding of the 3d model in Y directions, which the in-plane wall is a coupling wall, structural wall is vertically divided into two elements to express antisymmetric moment distribution. This model can evaluate initial stiffness considering gross cross section of in-plane wall. Uplift strength and maximum strength of the wall can be evaluated same as the result of X direction for the 3d model.

 1. Introduction

 The latest earthquakes such as the Tohoku/Kumamoto earthquake in 2011⁴⁾/2016⁵⁾ revealed that RC columns with wing walls having high stiffness suffered from serious damage, which resulted in restoration and demolition. To discuss this issue, the authors have proposed and verified a new rein. arrangement for columns with wing walls which removes the vertical wall rein. anchorage to let the wing walls resist the only compression, and then clarified the effectiveness for damage control⁶⁾. This new structural detail is expected that the short-term allowable bending moment may be improved because the tensile exterior rein. in the wing walls is not anchored, resulting in no yielding. Therefore, static loading tests using two columns having wing walls with/without vertical wall rein. anchorage were conducted. Furthermore, the current paper discusses specific conditions the short-term allowable bending moment increases/decreases through parametric analyses.

 2. Test plans

 Two 1:2 scale specimens were designed as classified into type FB based on AIJ Standrad⁷⁾ by modifying confining rein. ratio and vertical wall rein. ratio based on the specimen in the authors’ past study⁶⁾. Specimen CWJ2 did not have vertical wall rein. anchorage, while Specimen CWJ2A anchored vertical wall rein. to the lower stub. (Figs. 2-3 and Table 1). Static cyclic loads were applied to the specimens (Figs. 4-5 and Table 4).

 3. Test results

 In specimen CWJ2/CWJ2A without/with vertical wall rein. anchorage, the concrete/vertical wall rein. reached the short-term allowable stress first. The curvature of specimen CWJ2 was larger than that of specimen CWJ2A because the vertical wall rein. of specimen CWJ2 was not under tension by removing the vertical wall rein. anchorage (Fig. 7). The shear force of specimen CWJ2 at the short-term allowable bending moment was equivalent to that of specimen CWJ2A because the longitudinal rein. strain of specimen CWJ2 was larger than that of specimen CWJ2A (Fig. 8-9).

 4. Parametric analyses

 A method for bending analyses was presented to simulate the experimental behavior of the specimens (Figs. 10-11). Based on the analytical method, a series of parametric analysis was performed with five parameters of wall thickness, concrete strength, column longitudinal rein. ratio, wall length, and vertical wall rein. ratio.

 In the case of the column with wing walls on both sides, the short-term allowable bending moment ratio 𝑀_r (Fig. 15) was effectively improved by the wall thickness and concrete strength because the increase of wall thickness/concrete strength increased the curvature of the specimen without anchorage, which resulted in the increase of column rein. stress (Figs. 17-18). On the other hand, the reduction of the short-term allowable bending moment was limited under the compressive axial force ratio beyond 0.1, except for specific conditions. In the case of the column with a wing wall under tension, 𝑀_r was effectively improved by concrete strength and vertical wall rein. ratio (Fig. 22). In the case of the column with a wing wall under compression, there was little difference in 𝑀_r due to the parameters because the tensile exterior rein. in both specimens was the column longitudinal rein., which resulted in almost no difference in the curvature at the short-term allowable stress.

 5. Conclusions

 The present paper investigated the structural characteristics of the column having wing walls without vertical wall rein. anchorage proposed in the authors’ previous study⁶⁾ mainly focusing on the short-term allowable bending moment. The findings from the experimental and analytical studies are summarized.

 To strengthen reinforced concrete structures against seismic events, design methods are being established to fit the interiors of existing structures with steel braces or shear walls as well as to retrofit existing structures with steel braces on the exteriors of existing frames for seismic protection. For example, indirect connections using post-installed anchors are frequently employed when steel-braced frames are used for seismic retrofits, with many post-installed anchors required for each existing structure. The connection with existing structures and reinforcement elements in seismic retrofitting of concrete construction must be effective for distributing stress. If the strength of the connection is not adequate and if the details are not properly designed, the structure as a whole is not likely to display effective seismic performance.

 We proposed using shear connecters with a steel shear key of frustoconical shape and post-installed anchors as a new method of connecting existing concrete to extended concrete. A prior study on the strengthening method using steel angles and steel disks as shear keys reported that connections using steel shear keys improve shear strength. However, bearing failure often occurs in the concrete of a connection with steel shear keys, with drastic strength deterioration after the maximum carrying capacity of the reinforced structure and minimal ductility. We hypothesize that steel shear keys of frustoconical shape reduces the bearing failure of concrete. This shear connecter used together with a post-installed anchor is a “’funnel shape post-installed anchor’’.

 In the present study, we first confirmed the mechanical behavior of the proposed shear connecter through structural tests under a cyclic shear force. The presence of the shear key, shape of the shear key (double or single funnel shape), and central angle of the frustoconical section (30° or 45°) were used as experimental parameters. It was found from the structural tests that a double-funnel-shaped post-installed anchor with a 45° shear key improves shear strength in comparison with other shear keys. Moreover, when the frustoconical-shaped shear key is used, an anchor bar is always required for spacing on the concrete joint surface. Second, three-dimensional nonlinear finite element model (FEM) analyses were conducted using ‘‘FINAL’’. The analytical values closely predict the experimental results of the ultimate strength of the test specimens, with the stress distribution and the deformation of the anchor bar, shear-key, and concrete further examined. From the FEM analysis, it is clear that the rotation of the shear key influences shear strength.

 Finally, we proposed an evaluation method for shear strength of funnel shape post-installed anchors based on the stress transfer mechanism around the shear key, and the calculation accuracy of the proposed shear strength to the experiment value was shown.

 An economic and convenient seismic retrofitting technique based on the thick hybrid wall (THW) technique reported by Yamakawa¹⁾ is proposed. In the proposed technique, a cast-in-site partial hybrid wing-wall is built using additional concrete sandwiched by steel plates and high-strength steel bars (PC bars) prestressing. The aim of this technique is to enhance the lateral strength, stiffness, and ductility of soft-first story reinforced concrete (RC) buildings that are vulnerable to large seismic excitations. In the THW technique, the retrofitted section consisting of an additional wing-wall with short depth and the existing RC column are unified together as one unit using channel-shaped steel plates and tightened with PC bars. Since the additional wing-wall is not reinforced by longitudinal or transverse bars, the technique is convenient and cost effective. The important structural aspect of the THW technique is increasing the flexural strength as well as ductility by ensuring that all the longitudinal bars in the existing RC column yield in tension due to the increment of the internal moment lever arm, which results from the increase in the neutral axis depth into the additional wing-wall. To verify the efficiency of the proposed THW technique from the perspective of flexural strength, the equations to evaluate ultimate moment resistance in the retrofitted THW column section was proposed³⁾ based on the ACI stress block parameters, which consider the condition that all longitudinal bars yielded under tension in the existing RC column, and the additional wing-wall was in the compression side. Furthermore, the equation to calculate the minimum additional wing-wall length ratio was also proposed to estimate the affordability of the THW technique in Ref. 3).

 This study aimed to experimentally investigate the shear resistance and shear strength of the arch mechanism of the RC column retrofitted by the THW technique. From the test results of the retrofitted RC column showing a flexural failure mode, the proposed equations of the ultimate moment resistance³⁾ of the THW technique were verified.

 Experimental investigations were conducted on six specimens. In this study, two types of specimens were considered. One was a retrofitted RC column with no bonding force between the concrete and embedded longitudinal bars, thereby generating the arch mechanism. The other was a retrofitted RC column with bonded longitudinal bars to evaluate the flexural strength. In brief, the conclusions are as follows: (1) Bonded specimens for which the THW technique is applied showed flexural behavior with high ductility involving the tension yielding of all longitudinal bars in the existing RC column, and the calculated results of proposed equations are in good agreement with the test results. (2) The application of the THW technique not only creates a connection between the RC column and additional wing-wall, but also increases the shear resistance greatly. (3) In the unbonded specimens, the compression zone of the RC column for the arch mechanism was greater than 0.5D, and the zone was distributed from 0.8D to 1.0D. (4) Based on the test results and observations, an equation was proposed to evaluate the shear strength in the case of the THW technique following the proposed concept of the shear resistance (arch) mechanism with a nonuniform section of compression strut. The calculated results of the proposed equation are in good agreement with the test results showing shear failure mode.

 The behavior of RC frame consisted with shear walls and slab is influenced by slab, when earthquake happens. So, the behavior characteristic of slab was studied. On this study, the evaluation of effective region of slab is very important. At first, this evaluation method was developed. Secondly, with this method, the calculating method of restoring force characteristic of slab was developed.

 By experiment and FEM analysis results, the following knowledge was obtained.

 1. The region of slab main bars yielding was very wide. It was almost slab width.

 2. The right angles bars of slab with main bars were yielding.

 3. The elastic stiffness of slab was not changed, if the slab width changed.

 Assumptions of the proposed method are the following.

 1. The elastic effective region of slab at critical section is determined by the flexural stiffness of slab length direction and the torsional stiffness of slab width direction at shear wall side. The upper limit of this region is half of inside measurement length of slab.

 2. The effective region of slab at flexural yielding is equal to the elastic region without the upper limit.

 3. The yielding of slab right angles bars with main bars is considered.

 4. In calculation method of restoring force characteristic of slab, the flexural cracking and the flexural yielding are considered.

 It was found that the calculation results by the proposed method for the elastic stiffness, the flexural yielding moment and the restoring force characteristic of frame consisted with slab and shear wall corresponded well with experiment and FEM analysis results.

 Vertical haunch and horizontal haunch H-shaped beams have been widely employed in steel structures because the vertical haunch H-shaped beams have the advantages of improving economic efficiency by reducing the amount of steel, reducing building weight, and preventing brittle fracture of beam-column joints. Also, these beam ends treatment have the advantage that they can effectively resist external force by changing the cross section according to the acting stress, and that brittle fracture can be prevented by reducing the stress at the beam end weld. However, there have only been few studies on the mechanism and strength of the local buckling of the vertical haunch H-shaped steel beams. Therefore, the purpose of this study is to demonstrate the influence of various factors on the local buckling, such as haunch shape, the thickness of the component plate element, loading direction.

 In this study, the influence of the haunch shape on the local buckling strength of vertical haunch H-shaped beams is investigated by conducting eigenvalue analysis and experiment. Then, the evaluation method of elastic local buckling strength considering the effect of stress distribution of the component plate element is proposed. Also, the influence of haunch shape on the elastic-plastic behavior of vertical haunch H-shaped beams is investigated by performing the half- scale experiment and numerical analysis based on the finite element method. Finally, formulae for calculating the maximum strength and the plastic deformation capacity are proposed.

 From this research, the followings are found.

 1) The stress distribution formula for the component plate, considering geometric properties due to taper gradient, was derived. The validity of the formula was confirmed by performing the static loading experiment.

 2) The elastic local buckling behavior of the vertical haunch beam was investigated by using eigenvalue analysis. When the flange with the taper gradient is compressed, web crippling is likely to occur due to the increase of the compressive stress acting on the haunch starting point. To evaluate these effects, the evaluation formula of buckling strength, which is the function of the haunch beam shape factor ρ, was proposed. The haunch beam shape factor ρ can reflect the magnitude of the compressive stress.

 3) The collapse type, considering the expansion of the plasticized region in the haunch part, is classified into three types (haunch starting point type, beam end type, and haunch part type). Based on these three types, an equation for evaluating the full plastic moment of the vertical haunch beam was derived.

 4) The full-scale experimental monotonic loading tests were performed to investigate the effect of haunch shape, the web thickness, and loading direction on the elastic-plastic behavior. When the tapered flange is subjected to a compressive force, there is a possibility that the web crippling occurs because of the compressive stress act from the haunch starting point.

 5) The evaluation formulae for the maximum strength and plastic deformation capacity were proposed, and these formulae can be applied when the haunch side flange is subjected to a compressive force, which is often dangerous.

 The design of steel buildings in Japan and the US differ in a number of aspects, including the material utilized, structural design methods, and approaches. I-shaped columns with limited moment connections between the columns and beams (called the “perimeter frame system (PFS)”) are used in the US and other countries, while most beam-to-column connections are moment connections with square hollow structural section (HSS) columns (called the “space frame system (SFS)”) in Japan. In this paper, a popular elastic building design procedure in the US, the equivalent lateral force (ELF) method, is assessed from a Japanese structural design perspective. A design example of a 12-story steel moment-frame office building shown in FEMA P-1051 is studied. Furthermore, the required lateral strength by ELF and the Japanese allowable design procedure is compared for moment-frame and braced-frame buildings with respect to the natural period. The findings are as follows:

 (1) The lateral stiffness of the FEMA P-1051 design example required by ELF is significantly lower than that required in Japanese buildings. The first natural periods are 3.48 and 3.13 sec., while the approximate periods estimated for the same building height are 1.56 and 1.40 sec. in the US and Japanese practice, respectively. Furthermore, the first natural periods of the building designed in Japanese system, SFS, in accordance with Japanese design standards are 1.70 and 1.75 sec. It is confirmed in this example that moment resisting steel buildings with significantly lower lateral stiffness are designed and constructed even in California (CA) State, a seismic zone in the US.

 (2) The steel volume in the example in FEMA P-1051, designed with PFS, can be reduced to 85%, if it is designed in Japanese style (SFS) using square HSS columns and moment connections in almost all beam-to-column connections. This is because the cross-sectional area of the columns and beams needed for the relatively low lateral stiffness is not significantly larger than that needed for the strength against the gravity load. All frames are lateral frames in SFS and they contribute to the lateral stiffness. When the building is designed in accordance with Japanese design standards with a higher lateral stiffness requirement, the steel volume in PFS is close to that in PFS. It may be true that concentrating additional steel in limited perimeter frames in PFS is more effective for the higher requirement of lateral stiffness.

 (3) The required lateral strength by ELF and the Japanese allowable stress design procedure is compared. Two building sites in California are selected for the comparison. One is Stockton, which is an FEMA P-1051 example location, and the other is Stanford, where the seismic load is almost the maximum in the US. The ratios of required lateral strength in the US, Q_US, with respect to that in Japan, QJP, is defined as α_Q (=Q_US/Q_JP), and are calculated for moment-frame and braced-frame buildings with various natural periods. Assuming that the natural periods of braced-frame buildings are shorter than 1 second, then α_Q can be as large as 1.4 in highly seismically active regions in the US. Contrarily, assuming that the natural periods of moment-frame buildings are higher than 1 second, then α_Q is mostly less than 1, i.e., the required lateral strength of moment-frame buildings in the US is lower than that in Japan.

 Many of the newly constructed buildings are becoming intelligent, thus required equipment for building service such as air conditioning, electrical and information lines is diversified. In addition, for the sake of reducing construction costs, planning a floor height as low as possible is a common practice. In order to cope with this, circular holes are widely provided on the web of a beam composed of H-shaped steel to maximize the efficiency of ceiling spaces. In such a case, it is more likely that more holes will be installed to accommodate design changes during construction or future facility renewal than the minimum requirements of the initial design. Perforated beams may have lower strength and stiffness than ordinary non-perforated beams because parts of the cross section of the web are cut off. In the general structural design at normal temperature, some measures have been taken to prevent an excessive reduction in strength and stiffness.

 On the other hand, a perforated beam that is heated during a fire demonstrates a different deformation behavior from that at normal temperature due to the deterioration in yield strength and elastic modulus, thermal expansion and thermal stress. Regarding the fire resistance of beams, although the fire resistance test method based on the ISO 834 series is specified, test method specialized for perforated steel beams is not established. For rational fire resistance design of steel buildings, sufficient knowledge on the fire resistance of perforated beams especially having circular opening in the web is essential.

 In this study, two full-scale protected perforated steel beams with different opening spacing were used for fire resistance test to grasp the behavior of H-shaped steel beam with circular holes at high temperature. The hole spacings of two specimens were 1.5D (1.5 times the diameter of hole) and 2.5D (1.5 times the diameter of hole) on center respectively. Collapse behaviors of the steel beams including collapse time, distribution of steel temperatures, vertical deflection and out-of-plane deformation were reported and comparatively discussed.

 The following findings were obtained.

　(1) At high temperature, the shear yield of web posts affected the beam collapse in both the 1.5D and 2.5D specimens. Also, the deformation of holes occurred due to local buckling by the shear yield. It was found that the 1.5D specimen had more clear local buckling and larger deformation of the hole than the 2.5D specimen, thus the narrower the hole spacing, the greater the collapse of the perforated beam were affected by the shear yield.

 (2) At collapse, the average temperatures of the web were 570 Deg.C for the 1.5D specimen and 660 Deg.C for the 2.5D specimen. The difference was about 90 K. The 1.5D specimen with the narrow hole spacing and the larger area loss of the web due to holes collapsed at a lower temperature.

 (3) The perforated H-shaped steel beams showed collapse behavior due largely to shear buckling of the web, which was different from the general collapse behavior of non-perforated beams.

　著 者 小野里憲一

　表 題 せん断変形を求めるための形状係数に関する研究

　掲載誌 日本建築学会構造系論文集、第84巻、第762号、1065-1071 頁、2019年8月

 本論文は取り下げたい旨の申請が著者からあり、これを承認した。

 日本建築学会 論文集委員会