Corrosion process of reinforcements of RC building group based on super long-term combined deterioration is integrated into the seismic assessment and changes of human and building damage in virtual earthquakes causing one fatality at the new condition are analyzed.
The number of fatalities in the group with maintenance and deterioration by the scenario oriented for Economy does not change for 200 years, but maintenance could not eradicate the increase of severely injured people and completely collapsed buildings.
Risk of buildings in super long-term should be reviewed for the group unit stochastically because of the uncertainty of seismic performance.

抄録全体を表示

 Finishing material that is bonded to the wet slab peels off at high probability. We are researching to propose the moisture management index of slab surface. In our previous thesis, targeting at resilient floor coverings, the relation among bond strength and water content of slab surface was quantitatively examined. At that time, it was proved that both of two methods to grasp physical properties of bond strength regulated JIS A 5536 (Adhesives for resilient, textile and laminate floor coverings) are insufficient to use them alone, due to the fact that force application modes are too simplified and unrealistic.
 Actual fails were investigated in hearing for engineers of covering maker. As a result, it was found that poor adhesion part became to be "Bulding", "Peeling", "Joint tenting" or "Shrinkage". Then, we observed those fails in an actual building, and grasped the outline of force application modes of them.
 When the fail which require the development of a device to simulate the actual force application mode was chosen, it takes a lot of labor and time to prepare the experiment. It was seen that joint tenting (reference to Photo 2) is suitable for our research object, because it might be reproduce immediately in our experimental environment. It can be realized just by adjusting the temperature.
 By a preliminary experiment, joint tenting was able to be generated intentionally by adhering cooled coverings on specimen slabs (reference to Fig. 1) at the middle of the night in winter which was 5 °C outdoors and moving specimens to a warm laboratory. Size of specimen slab and specifications of testing covering were regulated appropriately (reference to Photo 4), joint tenting test method was fixed.
 At last, while drying specimens, joint tenting tests were carried out. The target which we chose was compressive strength of concrete which is 27 N/mm² and slump of concrete which is 18 cm (reference to Table 2). We used the working technique for 2 concrete surface finishing that is set according to the typical condition of actual construction sites (reference to Table 3). The typical tile covering was chosen for test, because hard tile covering was thought to be tented easily than soft seat covering (reference to Table 4). Two kinds of adhesives were chosen (reference to Table 5). One is an acrylic adhesive, representative of the water soluble. Another is an epoxy adhesive, representative of the one which is not soluble in water but is unexpectedly influenced by water in the chemical reaction. Water content was grasped by reading of moisture meter (reference to Photo 3). Then, the probability of occurrence of joint tenting at interval of 1 moisture meter reading was calculated. Due to the result, relationship between probability and moisture meter reading was considered (reference to Fig. 4). Then, the probability changed drastically at 5 of the moisture meter reading. This result proved covering engineers' rule of thumb to be correct, and the possibility that the moisture meter reading was able to be applied to the moisture management index.

抄録全体を表示

 So many human damages have been reported during past earthquake. Some of them got injured due to falling and crashing to furniture. Critical structures such as chemical plants and nuclear power plants, it is required to take countermeasures after accidents to prevent harmful substance leakage in the event of a severe accident caused by an earthquake. However, if the workers are injured during earthquake, the retrieval activities might become difficult for the workers. In addition, in order to predict the indoor human damage due to huge earthquake such as the Nankai Trough earthquake, it is important to grasp the behavior of human during an earthquake. To predict injuries resulting from human falling and crashing to furniture due to earthquake shaking, it is necessary to evaluate the seismic response of the human body.
 From the view point of the above background, in order to construct a seismic response analytical model of a human body, we conducted shaking table tests with human subject. The seismic response characteristics of the human body during shaking were investigated. The behavior of human subject during excitation was observed by 3-D motion capture system. The center of pressure (CoP) on the floor of the human subject was measured by a force plate during the excitation. Then, the mechanisms of postural control of the human subject were investigated. The human subject controlled his posture by moving the CoP to place the center of gravity (CoG) within the base of support (BoS). When CoG deviates from BoS, the human subject controlled his posture by widening BoS through the stepping strategy.
 Next, the seismic response analytical model in the forward-backward direction (sagittal plane direction) of a human was constructed based on the cart-type single-link inverted pendulum model with state feedback controller. The displacement of CoP, CoG, velocity of CoG and head could be evaluated by the model. The frequency response function of human subject during excitation in the experiment and that of the analytical model were compared in order to set the appropriate feedback gain of the controller. The displacement of CoP, CoG, velocity of CoG and head evaluated by the analytical model showed good agreement with those of the experimental results.
 Finally, the validity of the model was verified by the simulation analysis with other input motion. The maximum displacement of CoP and CoG could be well reproduced. On the other hand, the velocity of CoG was slightly overestimated compared to the experimental result, and the velocity of the head was slightly underestimated compared with the experimental result. Improvement of accuracy of the seismic response model (e.g. modeling with double inverted pendulum or using other control methods, etc.) is a future task.

抄録全体を表示

 In this paper, we conduct a finite element analysis of a square concrete-filled-tube (CFT) column to replicate cyclic shear-bending behavior under constant axial load. CFTs are used as columns of architectural and civil engineering structures. CFTs usually have large deformation capacity due to lateral constraints from steel tubes providing ductility. There are several existing studies on numerical analysis of CFTs. To the best of authors' knowledge, however, no numerical analysis method for simulating cyclic bending behavior of CFTs, in which interaction between steel tube and filled concrete, softening of concrete and local buckling of steel tube are simultaneously considered, has been proposed.
 The authors have been developing a detailed finite element analysis system called E-Simulator. In this research project, we simulate structural behavior only by combining detailed solid finite element model and material constitutive laws, which can represent damages and fractures of materials. By using E-Simulator, we can consider interaction between filled concrete and steel tube without any additional constraint effects.
 The concrete constitutive model, which has been proposed by the authors, is updated and applied to the analysis. The concrete model is a combination of extended Drucker-Prager model and a simple damage criterion. Here, we improve the concrete constitutive model so that damage progresses only in compressive stress state.
 A cyclic shear-bending loading experiment of a square CFT column is replicated. The height and width of the column are 1440 mm and 240 mm, respectively. A pseudo-static cyclic lateral displacement and a constant vertical load are given at the top of the column. The ratio of lateral displacement amplitude to the column height is increased from 1/400 to 1/33, gradually. The CFT consists of high strength concrete and steel, whose nominal strengths are 150 N/mm² and 780 N/mm², respectively. The material parameters are identified from material tests. The load-deformation relation obtained from the numerical result is compared to the experimental one. The interaction between the filled concrete and the steel tube is also discussed.
 The results of the analysis are summarized as follows:

  1. The load-displacement relationships of the analysis and experiment have been almost linear and elastic until the distortion angle has reached 1/100. The stiffness of the analysis has agreed well to that of the experiment. The load measured in the experiment is not proportional to the displacement. In positive loading direction, as a result, the load of the analysis is larger than the experimental one. The ratio of load from the analysis to the experimental one is 0.979 – 1.142 in positive loading direction, 0.715 - 1.004 in negative loading direction.
 2. When the distortion angle has exceeded 1/100 first, strength deterioration due to local buckling of the tube and damage of the concrete has been seen in the numerical result. In the experiment, local buckling of the steel tube has been observed in the same loading cycle. Although the strength deterioration in the analysis has occurred later than that in the experiment, the load-deformation relation of the experiment after plasticizing has been replicated, qualitatively. The ratio of load from the analysis to the experimental one is 1.076 – 1.154.
 3. The contact between the filled concrete and the steel tube due to the local buckling of steel tube has been first observed. After that, the large bending deformation has caused another contact. The contact pressure gave the concrete confinement, which made the equivalent stress higher than the uniaxial compressive strength of the damaged concrete. From this result, it can be concluded that interaction between filled concrete and steel tube can be represented by using detailed solid FE mesh model and E-Simulator.

抄録全体を表示

 Progress of parallel computing enables us to conduct large-scale finite element (FE) structural analysis using high fidelity finite element mesh of a structure with complex geometry. In the present study, a high fidelity solid element mesh of a 31-story super-highrise steel building frame is generated, and a seismic response analysis is conducted. A series of researches have been conducted by the present authors since 2007. A parallel finite element structural analysis code, E-Simulator, which has been developed at the National Research Institute for Earth Science and Disaster Resilience (NIED), Japan, is used in the analysis.
 The 31-story frame is a center-core-type office building whose total height is 129.7 m, and the size of the plan is 50.4 m × 36.0 m. The FE mesh has 15,598,662 elements, 24,220,688 nodes, and 72,662,064 DOFs. The number of multi-point constraints (MPCs) that are used to connect component meshes is 2,832,402. Plates such as the flanges and webs of beams are divided into at least two layers of solid elements in the thickness direction. The mesh is generated manually using the mesh generation module in a 3D solid modeler. A concept of mesh generator for steel frames using hexahedral solid elements is proposed in which a database of component solid element meshes for beams, columns, and connections is constructed and used. A mesh for a frame is made by assembling the component meshes.
 A parallel FE structural analysis software package, ADVENTURECluster is used as a platform of E-Simulator. The algorithm of the analysis code is based on the domain decomposition method. The Coarse Grid Conjugate Gradient (CGCG) method has been developed originally for the ADVENTURECluster as a powerful linear solver. In the present study, E-Simulator is implemented on K computer, which was one of the fastest supercomputer in the world when it started operation in 2011.
 The seismic response analysis of the super-highrise frame subjected to a simulated ground motion of great Nankai Trough earthquakes is conducted. The computation is performed on K computer using 256 computation nodes (2048 cores). The large strain elastic-plastic analysis is conducted. Time increment is taken to be 0.1 s for observing the response due to lower order eigenmodes excited by the long-period ground motion. Duration of the analysis is 125 s. The vibration due to coupled two lowest modes in the longitudinal and transverse directions and a torsional mode is continued many times. Distribution of equivalent plastic strain is visualized. Region in plastic state spreads gradually as the number of cycles of the oscillation increases. The equivalent plastic strain more than 30 % is observed. A downward vertical drift is observed in the time history of up-down displacement, which is due to plastic deformation of the whole structure. Discussions on computation performance on K computer are also described.
 Concluding remarks are as follows. 1) Mesh generation of super-highrise frame using hexahedral solid elements takes very long time. However, a concept of mesh generator for steel building frames is proposed. Component meshes for beams, columns, and connections are re-used and assembled in the system. 2) Both local yielding in the members and global plastic deformation of the entire structure in the vertical direction are analyzed under the long-period ground motion that continues more than two minutes. 3) Efficient visualization can be performed on K computer using an offline rendering code. 4) Although computation performance on K computer is better than those on previous supercomputers, further improvement of computation performance is necessary for the analysis code. Use of too many MPCs has an unavoidable effect on the performance.

抄録全体を表示

 In our previous paper, the rotational stiffness and the flexural strength of the timber-steel hybrid connections for the lattice shell roofs were clarified by static experiments and the effects on the buckling behavior of the entire roof were revealed. In this paper, we examine the influence of the rotational stiffness at connections on natural oscillation characteristics and earthquake response characteristics of the roofs with supporting substructures. Also, the accuracy of the response spectrum evaluation using CQC method and the evaluation with the equivalent static earthquake load are verified, followed by the simplified evaluation method proposal for the effect of rotational stiffness at connections.
 As in the past researches, we study with the steps using roof models without supporting substructures then models with substructures. The roofs are categorized as single-layered lattice roofs and double-layered lattice roofs with depth-span ratio being 1/60, where the rotational springs obtained in the previous experiment were incorporated at the both ends. First, we investigate the influence of connection stiffness on the seismic response characteristics of the roof model, comparing the response spectrum method with the time history response analysis to verify the accuracy of the response spectrum method. Here, the applicability of the equivalent static load using response amplification factors proposed by the past researches are also verified. Next, a model with a supporting substructures were set and the earthquake response characteristics were verified. The influence of the rigidity of the supporting frame on the earthquake response was analyzed, and accuracy of the equivalent static load was verified. Here, we also analyze the influence of the rotational stiffness of the connections as well as the roof model. From these investigations, it was confirmed that there was no significant influence of the connection stiffness on earthquake response characteristics were observed, if the natural periods of principal modes are correctly evaluated. Therefore, it was considered that the influence of the rotational stiffness at the connections on the earthquake response can be represented by the natural period ratio R_T. It was also found that the natural period of the antisymmetric mode including connection stiffness can be expresses as the function of normalized rotational stiffness κ, and easily evaluated from the eigenvalue analysis with rigid connections.
 From above studies the following knowledge was obtained.
 1) In the roof models with the rotational stiffness at connections, the influence of the rotational stiffness on the vibration characteristic is not significant where normalized rotational stiffness κ >7..
 2) Seismic response of a lattice shell roofs with the connection of normalized rotational stiffness κ = 3 ~ ∞ (rigid) is can be well evaluated if the correct roof natural period with the rotational stiffness at connections are obtained and used. It was confirmed that the accuracies of the equivalent static loads with amplification factors are almost the same as the those with rigid connections.
 3) The effects of connection rotational stiffness on the principal natural period of the roofs can be expressed as a function of κ. Therefore, the natural period including the connection stiffness for equivalent load estimation can be easily obtained by multiplying the proposed formula to the natural period of the roof model with rigid connections. The accuracy of the proposed method is verified, and confirmed to provide equivalent accuracies.

抄録全体を表示

 Several studies based on structural properties of plywood walls and wooden semi-rigid frames have been conducted. However, the complex behavior of a wooden semi-rigid frame composited with structural plywood has made it difficult to establish its structural design method. The wooden semi-rigid frame joint has the problem of displaying semi-rigid performance. Therefore, it is difficult to satisfy the necessary stiffness of 1/120 radian, which is the limit value of the deformation angle in the elastic design of buildings. In this case, it is possible to compensate for the initial stiffness by incorporating a structural wall into a wooden semi-rigid frame. In general, an independent wall framework is set up within a wooden semi-rigid frame. A plywood board is nailed onto this framework. If it can be nailed directly onto the columns of the semi-rigid frame, there is no need to place a post in the framework, and the designing process can be more cost-effective.
 Due to the complex behavior of the wooden semi-rigid frame composited with structural plywood, the structural design method cannot be established. This study aims to clarify the properties of the wooden semi-rigid frame with structural plywood with tests and build its structural design method.
 A wooden semi-rigid frame with a lag-screw-bolt (LSB) for the joints was used in this study. The structural properties of the semi-rigid frame with structural plywood were estimated by conducting static racking tests for four types of specimens. The four types of specimens were as follows: “Rigid” wooden semi-rigid frame; “Outside” wooden semi-rigid frame with a structural plywood wall nailed on the outside edge of the column; “Inside” wooden semi-rigid frame with a structural plywood wall nailed on the inside of the column; and “Wall-O” and “Wall-I” structural plywood wall only.
 Nuts were used to fill the empty 20 mm. The structural models of these specimens were simplified. At the bottom of each LSB, load cells were installed to measure the tension side axial force.
 We examined the following: the shear-rotational angle curves of the wooden semi-rigid frame of “Outside”, where “Inside” was determined from the supporting point reaction; the shear-rotational angle curves of the plywood wall of “Outside”, where “Inside” was found by taking the load-rotational angle curves of “Wall-O” or “Wall-I”. The tests showed that there is a certain difference in the shear-rotational angle curves subjected to the nail position.

抄録全体を表示

 H-shaped steel beams in frame structures are rotationally restrained at both ends by connecting columns or beams. It is well known that the elastic lateral buckling strength is affected by both the boundary conditions at the beam ends and the moment distribution. Although there have been many reports on the elastic lateral buckling strength of such beams, the combined effects on lateral buckling of restraints on the amount of lateral bending and beam-end warping are still unknown.
 The purpose of this study is to develop a formula for evaluating the lateral buckling strength of H-shaped steel beams with arbitrary restraints on the amount of lateral bending and beam-end warping. Firstly, we propose a useful displacement function which is applicable to approximation analysis based on the energy criterion. An elastic lateral buckling strength approximation formula is then derived for the case of a uniform moment distribution using a Galerkin method. Finally, a lateral buckling strength formula that is valid for any boundary conditions, sectional shape or moment distribution, is proposed.
 The results obtained in this study are as follows.
 1) Displacement functions were derived for structural members undergoing flexural buckling under arbitrary boundary conditions, including pinned-pinned, fixed-fixed, and pinned-fixed members. These functions were found to accurately describe the buckling mode shape for members with different rotational rigidities that are supported at both ends.
 2) A parameter R, determined by the sectional shape and the member length, was introduced to evaluate the lateral buckling strength of H-shaped steel beams. It was shown that as R increases, the effect of the warping restraint on the buckling strength decreases. By including this parameter, the lateral buckling strength could be determined with Eq. (34) for arbitrary boundary conditions by Eq. (35).
 3) An additional moment gradient modification factor C₁ was also introduced for the cases that the rotational stiffness of one end having larger moment is also larger than the other end with Eq. (48) through Eq. (51).

抄録全体を表示

 In Japan, steel moment frames, which consist of H-shaped beams and square hollow section columns with through diaphragms, are often used for building structures. It is commonly known that panel zones of these steel moment frames often yield under strong ground motion and have high deformation capacity. In order to take into account high deformation capacity of panel zone for seismic design, it is needed to calculate the full plastic strength of panel zone. Several methods for calculating the full plastic strength of panel zone have been proposed by previous researches [Ref. 5]-7]]. For example, in the most frequently used method, which is adopted in the recommendation [Ref. 3]], only axial force and shear force is considered, and the effect of bending moments which act upon the top and bottom of panel zone is ignored. Thus, this method can only be applied to the panel with aspect ratio up to 1.6 [Ref. 3]]. On the other hand, most methods proposed by previous researches can not correspond to arbitrary directional shear force. Therefore, this paper proposes a new method to calculate the full plastic strength of square hollow section panel zones with through diaphragms under axial force, bi-directional shear forces and bi-axial bending moments.
 Firstly, cyclic loading test was conducted in order to investigate effects of aspect ratio, axial force and input direction of shear force on elasto-plastic behavior of square hollow section panel zones, as described in Chapter 2. From the test results, it is clarified that ultimate states are classified as ductile crack, shear buckling, and local buckling. And the deformation capacity with shear buckling is larger than the cases with other ultimate states. Furthermore, it is verified that maximum strengths of panel with small aspect ratio and small axial force ratio under 45 degrees input are larger than those in other cases.
 Secondly, finite element analysis (FEA) was conducted in order to verify full plastic strengths with various parameters more than the cyclic loading test. The validity of FEA model is confirmed by comparing between FEA results and loading test results, as discussed in Chapter 3. Strength interaction curves of bi-directional shear force, which are obtained by FEA, are shown in Fig. 12-14. From these figures, while the aspect ratio and axial force ratio are small, the strength interaction curve looks like square with a bit round at corner because shear yielding mainly occurs at all over the panel. On the other hand, while either aspect ratio or axial force ratio is large, the strength interaction curve is circlelike due to yielding under shear stress and normal stress at ends of panel.
 Finally, a method to estimate the full plastic strength of square hollow section panel with through diaphragms, considered with correlation among axial force, arbitrary bi-directional shear force and bi-axial bending moment, is proposed. The full plastic strength is obtained as the smallest one between strength at center section and strength at end section. In Fig. 19 and 20, calculation results by the proposed method are compared to analysis results from Chapter 4. From these figures, in case aspect ratio is over 2.5 and input direction of 0 degrees without axial force, calculation results by proposed method underestimates full plastic strength of panel as much as 20%. However, aspect ratio of panel zones, which is used in steel moment frames, are usually under 2.0, in this case, calculation results by proposed method corresponds well to finite elements analysis results.

抄録全体を表示

 Cold-formed steel members were composed of thin plate elements thus the members were easily occurred the local buckling under compression. The local buckling strength significantly affects the maximum strength of the members, therefore providing an accurate evaluation method for the local buckling strength is one of the key issues in their design. The conventional design method was assumed that the plate elements behave as a simply supported plate element. However, the longitudinal edges of plate elements were restrained their rotation by adjacent plate elements, and this restraining might lead an increase of local buckling strength of cold formed members both in the case of elastic and inelastic.
 In this paper, we investigated the restraining effect on rectangular section under combined compression and bending condition. First, we conducted the eigenvalue analysis by finite strip method. Through the comparison between numerical analysis results and theoretical results which assumed simply supported condition, we found that the restraining effect was appeared not only under axially compressed loading condition but also under combined loading condition.
 Then, we considered evaluating the elastic local buckling strength under combined compression and bending condition by the equations which we had developed for axially compressed members. Through the comparison between eigenvalue analysis results and calculated results, we found that the elastic local buckling strength of rectangular sections under bending and compression could be evaluate the developed equations accurately with a tiny modification.
 Finally, we conducted a finite element analysis to investigate the restraining effect on inelastic strength (i.e. maximum strength) of rectangular section members. The maximum strength of rectangular sections was mostly higher than the nominal strength which assumed the boundary condition as a simply supported condition. Furthermore, we re-evaluated their maximum strength with a consideration of restraining. The re-evaluated results were agreed well with the numerical analyses, thus we found that the maximum strength of rectangular sections was evaluated accurately to consider the restraining effect between adjacent plate elements compared with the conventional method which regard the plate elements as a simply supported plate.

抄録全体を表示

 The advantages of concrete-filled steel tube (CFT) columns include high strength and remarkable ductility, since the steel tube provides confinement to the concrete while the concrete inhibits local buckling of the steel tube. CFT column composite frame systems with steel beams have been widely used in moment-resisting frame systems for mainly office buildings. The CFT columns of the building involve short columns with small ratios of buckling length to depth of cross section at typical floors and slender columns with a large ratio of buckling length-depth at the entrances of lower floors. CFT columns from short to slender have been used in buildings. The stress states of CFT columns applied to buildings are in a state of flexural-shear stress, and the structure designs are made for the predominant flexural yielding type. It is thought that accurate evaluation of flexural behavior in the evaluation of flexural-shear behavior leads to design of columns that have the features of CFT columns. In addition, there is a tendency to increase the strength of CFT columns due to higher rises and larger spans of buildings. For the moment curvature relationship, which is the basis of the restoring force characteristic model of CFT columns from short to slender, it is important to devise a simple model with a range up to high strength material.
 This paper proposes a new simplified model of the moment-curvature relationship for square CFT beam-columns under monotonic bending moment loading to estimate the elastic-plastic flexural behavior of beam-columns from short to slender. The model of square CFT beam-columns involves a range from ordinary-strength to high-strength materials. The proposed simplified model is a multi-linear model having a crack strength point, a yield strength point, an ultimate strength point, an ultimate strength holding point and strength reduction points. For the main points of the proposed model, each flexural strength is calculated by the general superposed strength method, and new multiple regression formulas for each curvature are proposed based on results of monotonic bending moment loading tests from previous researches. Predictions from the proposed model of the moment-curvature relationship are found to agree approximately with experimental results up to large flexural deformations.

抄録全体を表示