Recently, studies on anti-corrosion performance of galvanized steel rebar (hereafter, HDZ bar) have been carried out. HDZ bar is known to perform sacrificial anodic reaction by zinc. While it is subjected to corrosion, zinc coating continuously corrodes itself replacing for steel substrate. Meanwhile, in order to expand the usability of HDZ bar in RC structure, two things are regarded which are durability examination of corroded HDZ bar and a partial application of HDZ bar where active corrosion is predicted.

 Concerning the durability examination, accelerated corrosion test which induces anodic current to the rebar has been used. However, in using this test on HDZ bar, the influence of applied current density has not yet been clarified. Although there was an attempt which applied 0.625mA/cm² of current density, the result showed that HDZ bar performed its anti-corrosion effect to the maximum when it is 43% corroded. About using HDZ bar and ordinary rebar together in concrete, there has been a concern on galvanic corrosion of HDZ bar. Electrical connection between HDZ bar and ordinary rebar causes accelerated local corrosion of HDZ bar at contacted area. Nonetheless, it has not been fully clarified on which area is locally corroded.

 In this study, in order to investigate the suitability of accelerated corrosion test, relatively lower range of applied current density was tested on HDZ bar as series1. A designed corrosion amount (hereafter, DCA) was considered which intends full dissolution of zinc coating. In series2, in order to investigate the galvanic corrosion, segmented rebar was experimentally attempted to evaluate macrocell corrosion behavior. Galvanic corrosion was regarded as macrocell corrosion caused by formation of anode and cathode region.

 As a result, in series1, HDZ bar showed continuous anti-corrosion effect and no corrosion of substrate and concrete crack were found within 100% of DCA. In electrochemical evaluation, HDZ bar showed relatively high polarization resistance (R_p) as the applied current density is lower. Furthermore, R_p was increased with corrosion progress. It is assumed that a formation of Zn(OH)₂ on the surface was decisive because it is known as insulating material. In actual corrosion amount, approximately 1/4 of zinc coating was corroded against DCA.

 In the result of series2, as similar to series 1, continuous anti-corrosion effect of HDZ bar, formation of oxidation film and no corrosion of substrate was identified. Actual corrosion amount of HDZ bar was also smaller than DCA. However, 2/5 of zinc coating was corroded which is assumed by influence of galvanic couple. A potential difference was generated among segments of HDG bar. By this influence, anodic macrocell current was formed where it was the farthest from ordinary rebar. In conclusion, HDZ bar performed adequate anti-corrosion behavior in both corrosion environment in the scope of this study.

 In recent years, although the form materials and construction methods have been diversified into various kinds, conventional formwork method using plywood for concrete on sheathing plates is the most common construction. The conventional form is temporary structural object until the concrete hardened. Additionally, in order to formwork need economy and rationality, plywood reused as much as possible with the conventional formwork. However, influence of form reuse on the quality of plywood and the finish of concrete surface using those reused plywood, has never been studied.

 The purpose of this study is to clarify the quality variation of plywood by form reuse and the finish of concrete surface using reused plywood. In this paper, on the basis of the experimental results, the influence of reuses on the quality of plywood, such as bending strength, bending Young’s modulus, and surface roughness, were examined. Also, the influence of reuses on the finish of concrete surface, such as surface roughness, glossiness, color unevenness, and water absorption, were examined.

 Experiments in this paper, the influence of up to 10 reuses of two types of plywood uncoated plywood and coated plywood was investigated. Furthermore, as an influence of work of the formwork in construction site, factors of this experiment is included the influence of with or without water immersion of plywood by rainwater and the influence of with or without application of release agent.

 The tendency of the influence by reuse of plywood based on experiment are as follows.

 (1) The bending strength and bending Young’s modulus of plywood decreased with reuse, especially uncoated plywood showed noticeable.

 (2) In case of bending strength and bending Young’s modulus of uncoated plywood, the decreasing tendency with reuse is reduced by applying a release agent.

 (3) Although the surface roughness of unpainted plywood is increased significantly with one reusing, thereafter does not variation so much. In case of coated plywood, the surface roughness of plywood is about the same as unused plywood even at time 10 reuses.

 (4) In case of surface roughness of uncoated plywood, the increasing tendency with reuse becomes noticeable when change in moisture content by rainwater, and the increasing tendency with reuse is reduced by applying a release agent.

 (5) The surface roughness of concrete increased with reuse, especially noticeable in concrete using uncoated plywood.

 (6) In case of surface roughness of concrete using uncoated plywood, the increasing tendency with reuse becomes noticeable when change in moisture content by rainwater, and the increasing tendency with reuse is reduced by applying a release agent.

 (7) Although the glossiness of concrete using coated plywood is decreased linearly until at time around 7 reuses, thereafter does not variation so much. In case of concrete using uncoated plywood, the glossiness of concrete surface is less variation with reuse.

 (8) For both types of plywood, the color unevenness and water absorption of concrete surface is increased with reuse.

 Finishing materials installed at the building surface closest to the living space of people often fall down and harm the life, body or property. Ceiling collapse especially occurs by the various reasons, e.g. the rain leaking, dew condensation, wind, traffic vibration, on a daily basis. In addition, the ceiling collapse tends to occur in the region where the seismic intensity is over five lower during earthquakes.

 The authors propose the quantitative evaluation method to assess the degree of risk of the ceiling collapse by comparing the impact force and the proper human tolerance. So far, we dropped the various ceilings having commercial size of 910 mm square from one to twenty meters high using the guide wires to hit to the dummy head remaining horizontal and measured the impact force with the load cell installed underneath of the dummy head. However, the experimental conditions are constrained. Therefore, the examination regarding the efficiency of ceiling size and impact angle is insufficient. To do the parametric study by the numerical analysis, the material properties of the finishing material in a wide range of strain rates are needed. However, few reports are available on the material properties of the finishing materials.

 In this study, a series of compressive tests, i.e. quasi-static tests, dynamic tests and impact tests by split Hopkinson pressure bar method, at various strain rate from 10^-5 to 10³ s^-1 were conducted to examine the strain rate effect on the compressive property of the representative finishing materials: gypsum boards, calcium silicate boards and rock wool sound absorbing boards.

 The results obtained from this study are following:

 1) The X-ray CT images show the internal structure of the finishing materials. The porosity of the gypsum boards is about 25%, and it was also slightly observed in the rock wool sound absorbing boards. On the contrary, it was found that the calcium silicate boards were the solid material with no void.

 2) It was revealed that all materials used in this study have the strain rate dependence of the compressive strength at the strain rate from 10^-4 to 10³ s^-1. In addition, it was found that the strength drastically rose at the strain rate of about 10² s^-1, although it is not confirmed that the strain rate dependence is attributed to the material itself or its foam, i.e. open cell structure.

 3) The constants for Cowper-Symonds constitutive equation (dynamic constants) of the representative finishing materials were identified by least-squares method based on the compressive test results. Due to this, it can be expected to improve simulation accuracy in the future.

 Base-isolated systems have been demonstrated to be an effective structural design to minimize damage to building superstructures during earthquakes. However, when the deformation of a base-isolated story exceeds design considerations under extreme earthquake ground motions, a superstructure may collide with a displacement-limiting device such as the surrounding retaining wall. In this study, collision tests using a shaking table were conducted considering a collision of a base-isolated model specimen with a retaining wall. Considering the relative story displacements, the floor accelerations and impact forces measured on the base-isolated model specimen during the collision, the influence of collision for the superstructure was investigated. Numerical simulation was also carried out to reproduce the measured responses during the collision and analysis results are compared with experimental results to validate the numerical model for collision analysis. In addition, time history analysis using impulse which can reproduce the behavior of superstructure during a collision was proposed. Findings were summarized as follows.

 1) The maximum values of story shear force, floor acceleration and impact force during a collision have an almost linear relationship with collision velocity. The maximum values of floor acceleration and impact force depend on the rigidity of the retaining wall, whereas, the maximum story shear force depends less on the rigidity of the retaining wall.

 2) About the story shear force（Fⁱ）calculated by summing inertial force, the story shear force（Fⁱ）becomes close to the story shear force（F^d）calculated by using relative story displacement in case without a collision. However, in a case with a collision, story shear forces Fⁱ and F^d become different because of acceleration responses including high-frequency components. By considering floor acceleration responses filtered out the range over than the highest natural frequency of testing model, both the story shear forces Fⁱ and F^d during a collision become the similar value.

 3) By calculating the impulse using the first floor impact force measured by a load cell, it was found that the impulse depends less on the wall rigidity. Because, although the impact force increases, the impact duration decreases about the increase of wall rigidity. Therefore, the maximum story shear force of the superstructure depends on the impulse and the maximum floor acceleration depends on the maximum impact force.

 4) The collision analysis including collision spring for MDOF model can accurately reproduce the experimental results. Furthermore, it is confirmed that the response of superstructure during a collision can be well-reproduced by using numerical analysis which applies time history of the impulse to the first floor, and the maximum floor acceleration and story shear force also have good agreement with experimental results. By using this analysis method, the possibility to easily predict the maximum response of superstructure during a collision can be indicated without using collision analysis.

 This paper focuses on the first-mode natural period and damping ratio in the vertical direction for the purpose of grasping the actual characteristics of the vertical vibration of buildings. The identifications of period and damping ratio based on earthquake observation records of multiple buildings are carried out. The targets of the identifications are a total of 18 buildings consisting of reinforced concrete (RC), steel and reinforced concrete (SRC), and steel (S) structures. Auto regression and exogenous (ARX) model is used to identify the first-mode vertical natural period and damping ratio. The identified values are summarized for each structural type, and its characteristics are discussed in view of the input amplitude and the height of buildings and in comparison with its natural period and damping ratio in the horizontal direction as well.

 The main findings obtained here are summarized as follows:

 1) The 1st-mode vertical natural periods of 18 buildings show relatively small dependence on the input acceleration amplitude. On the other hand, the 1st-mode vertical damping ratios are found to be dependent on the amplitude especially in RC buildings. No significant change is found in 1st-mode vertical natural periods and damping ratios of 18 buildings between before and after The 2011 off the Pacific coast of Tohoku Earthquake.

 2) When the effective value (root-mean-square value) of vertical acceleration at the lower point of the building is 2.0 Gal, the 1st-mode vertical damping ratios are about 2~5% in RC and SRC buildings and about 0.5~2% in Steel building.

 3) The gradient of vertical natural period of a building tends to be smaller as the height increases, in contrast to the horizontal period, which is generally proportional to the height of a building.

 4) The above tendency is examined based on the natural period of equivalent single-degree-of-freedom (SDOF) system, and the estimation equation of the 1st-mode vertical natural period which is proportional to the square root of building height is proposed. The coefficient C of the estimation equation is proportional to the square root of the material strength f and the axial force ratio β of the first story, and inversely proportional to the square root of the Young's modulus E of vertical structural members such as columns. The estimation equation shows good correspondence to the identified values in this paper and the data from the past references.

 5) The simple equation of relation between the 1st-mode vertical natural period _vT₁ and the horizontal 1st-mode natural period _hT₁ is proposed.

 6) The 1st-mode vertical damping ratios of 18 buildings are similar to the horizontal 1st-mode damping ratios of the same buildings in the RC and SRC building, and those tend to be small in the Steel buildings.

 In recent years, the adoption of the rectangular ETFE cushion as a roof cladding element has become widespread. It is anticipated that this trend will continue and that the size and shape of the panels made of ETFE will become more diverse. The cushion structure allows non-linear material behavior until the second yield stress, because the supply of air to the cushion can increase the volume of air even if elongation of membrane caused by residual strain has occurred. However, the behavior of the cushion structure exceeding the 1st yield point hasn’t been studied and investigated yet.

 In this paper, the behavior of ETFE sealed air cushion under Excessive Negative Pressure and Cyclic Loading was examined. As a result of comparative studies using experiments and numerical analyses,

 - It was confirmed that there was a trend that both the upper and lower membranes deformed almost equally keeping the initial cushion depth under full monotonous negative pressures.

 - The internal pressure gradually decreased and showed a tendency to become negative after initially maintaining zero Pa under full negative pressures.

 - Although the residual strain remained in the membrane and the initial internal pressure reduced due to the material non-linearity after the excessive load was applied, it was confirmed that the system did not collapse within the scope of this study.

 - Under cyclic loading, the residual strain occurred and the initial internal pressure decreased which is a typical characteristic of the sealed air cushion. However, it was confirmed that the system did not collapse within the scope of this study.

 In addition, because of numerical analyses,

 - By increasing the initial internal pressure, it was able to delay the disappearance of the tension of the upper membrane on the unloaded side and to delay the deformation of the upper membrane into a flat shape. However, regardless of the difference in the initial internal pressure, the load at which the lower membrane reached the first yield point was constant, and the displacement of the lower membrane associated therewith was almost constant.

 - When the aspect ratio was 1.5 or more, the results of internal air pressure of the cushion, membrane displacement and maximum membrane stress showed approximately equivalent values.

 - Even if the panel scale was different, the decrease in internal pressure could be reduced and the cushion stiffness was increased by using thicker film. In the elastic range, it was confirmed that the relationship between the film thickness and the load to the first yield point were approximately proportional.

 1. Introduction

 The cross-sectional area and width of foundation beams under shear walls are required exceed the recommended Architectural Institute of Japan (AIJ) standard for reinforced concrete (RC) structures. However, the occurrence of brittle failure in the foundation beam under shear walls with door openings, similar to that in boundary beams between door openings on upper floors, is also a concern. Therefore, this study conducted static loading tests of beam specimens under shear walls with door openings to investigate their structural performance. This paper reports the experimental results upon the restoring force characteristics and cracking pattern in the beam specimens and proposes an ultimate strength evaluation method for foundation beams under a shear wall with door openings.

 2. Foundation beam specimens

 The assumed foundation beams were based on the AIJ standard for RC structures. The two specimens consist of a foundation beam, two pile-caps, and two walls with boundary columns. Clevis pins were installed at the top and bottom of the specimens. The specimens were designed with the beam axial force ratio at maximum strength set to 0.13 with reference to the results of the Finite Element pre-analysis. The transverse reinforcement ratio pw of the beam was the variable investigated.

 3. Experimental results

 Failure mechanisms were formed in the beam at the cycle of 0.5x10-2 rad due to tensile yielding of the main bars at the top and bottom of the beam in both specimens. However, the observed strength and deformation capacity in a specimen with pw of 0.22% were lower than those with pw of 0.66% due to occurrence of shear failure in the beam. That is, shear deformation of the beam from the opening on the tensile column side to the wall on the compression column side increased significantly in the specimen with pw of 0.22%. On the other hand, a flexural deformation of the beam increased in the specimen with pw of 0.66%. In addition, the compressive axial force of the beam decreased due to shear failure in the specimen with pw of 0.22%, but it increased until the final loading cycle in the specimen with pw of 0.66%.

 4. Ultimate strength evaluation

 The ultimate bending moment of the beam was calculated using the stress block method by American Concrete Institute. The critical faces were assumed to be an end-of-the-pile cap on the compression column side and inside the opening on the tensile column side. A balance of the moment using the ultimate bending moments of the beam, vertical compressive force of the wall on the tensile column side, and tensile force of the vertical reinforcement in the wall on the compressive column side was considered. The shear strength of the beam was calculated using the Truss-Arch equation. The ratio of the observed strength in the specimen with pw of 0.22% and calculated shear strength in the beam was 1.13. On the other hand, the ratio of the observed strength and calculated bending strength in the specimen with pw of 0.66% in the beam was 1.01. This indicated that the ultimate strength evaluation methods for the beam specimens in this study showed good agreement with the observed strength of the foundation beams under the shear wall with a door opening. Furthermore, the above evaluation model of bending strength indicated that the critical face of the foundation beam on the tension column side is located 0.1 to 0.3 times inward of the beam from the opening end.

 1. Introduction

 Under quite strong ground motion, complete collapse of steel moment frames may occur because of strength deterioration by fracture and/or buckling of steel members. To simulate deterioration behavior of square hollow section (SHS) columns, which are most used in Japan, is the most important for revealing complete collapse behavior of steel moment frames. In relevant studies, multi-spring model (MS model) is often adopted for complete collapse analysis because it can achieve high accuracy result at low computational cost. In this paper, procedure to build up MS-model for cold press-formed SHS, which has not been dealt in past researches, is proposed.

 2. Outline of multi-spring model

 MS model consists of an elastic bar element and multi-spring elements, as shown in Fig. 1. The multi-spring element is a component of some uniaxial springs inserted between two rigid plates. Force-deformation relationship of the uniaxial spring is modeled in consideration of following regulations;

 1) Some differences of material properties between flat area and corner area of SHS exist.

 2) On hysteresis curves after local buckling occurring, stiffness during unloading becomes smaller than elastic stiffness and slip behavior occurs.

 3) Plastic deformation of uniaxial spring is transformed from longitudinal plastic strain in plastic region or local buckling region in consideration of equivalence of plastic dissipation energy.

 3. Formularization of stress-strain relationships of stub-columns under axial force

 In order to clarify stress-strain relationships after local buckling of stub columns occurring, finite element method (FEM) analysis of stub columns is conducted under only axial force. Analysis parameters are width-to-thickness ratio, length-to-width ratio and cyclic loading protocols. Based on observation regarding stress-strain relationships under monotonic compression axial force and under cyclic axial force, several parameters to define approximated stress-strain relationships, as shown in Fig. 7 and 11, are formulated.

 4. Transformation for force-deformation relationships of uniaxial springs of MS model

 In order to transform for force-deformation relationships of uniaxial springs from stress-strain relationships as represented in Chapter 3, the following configurations should be considered.

 1) Longitudinal strength at the end of column is larger than the strength under uniaxial stress because of constraint of circumstance strain of the column due to weld.

 2) Strength of uniaxial spring of MS model is amplified from the defined stress at local buckling region, based on bending moment distribution.

 3) Plastic deformation of uniaxial spring of MS model is obtained as to correspond plastic dissipation energy in the entire either plastic region or local buckling region with that of MS model.

 5. Verification of validity of proposed MS model

 In order to confirm accuracy of the proposed MS model, analysis results by MS model are compared with analysis results by FEM or test results in the past research. Most of results by MS model agree well with FEM results and test results, unless width-to-thickness ratio is small.

 6. Conclusions

 As a result, it is clarified that the proposed MS-model is valid for simulating deterioration behavior of cold press- formed SHS column due to local buckling. In the further study, MS-model should be modified to adapt for other kind of SHS columns.

 The brace joint fracture was observed in several severe earthquakes which occurred in the past. That's because fracture prevention of a brace joint wasn't considered by old seismic design code. The important problem to promote seismic retrofit works of a brace joint in the future is fire-less and compact seismic retrofit method which doesn't use a welding and gas-cutting.

 In this paper, the seismic retrofit method with the knee members for preventing the joint fracture of the angle brace was proposed. This method joins a brace, a column and a girder using the knee members and high tension bolt. Therefore it became possible to build this method without using fire at a site of retrofit works. And an obstacle in retrofit works could be avoided by adjusting the location and holding angle of the knee members. This method aims at fracture prevention of a brace joint by adding a load transfer routes of the knee members.

 A series of loading test and a series of analysis have been carried out to investigate strength improvement effect of this method. The strength and stiffness evaluation model of each part in this method were proposed. And a relation between story drift angle of the frame which occurred by the seismic force and the load sharing ratio of the knee members was investigated analytically.

 As a result, the following knowledge was obtained.

 [1] The joint fracture prevention before a brace yielding was confirmed by installing the knee members in both sides of the brace. However, story drift angle of the frame which occurred by the seismic force wasn't considered by this result.

 [2] The load sharing ratio could be expressed by an evaluation model proposed by this paper. However, story drift angle of the frame which occurred by the seismic force wasn't considered by this result.

 [3] Consideration is needed by the load sharing ratio changed by holding angle of the knee members, the installed height of the knee members, stiffness of the knee members and holding angle of a brace to make this method reach the reinforcement target. And it's necessary to consider story drift angle of the frame which occurred by the seismic force.

 Large cross-section beams are used as main beams to design structural members effectively in real large space structures. Long span beams may not possess the plastic strength due to lateral buckling, so that many lateral braces along the beams should be set up to prevent the lateral bucking deformation (AIJ 2010). Most of beams in frames are connected by continuous braces such as folded-roof plates, which are effective to prevent the lateral buckling of beams. However, in the Japanese design code, non-structural members are not considered as the braces. Kimura (2013) clarified the lateral buckling behavior for H-shaped beams with continuous braces when the boundary condition of the beams is simple support.

 On the other hands, in design standard for steel structures (AIJ 2005), the boundary condition of beams against lateral buckling is provided as simple support. Meanwhile, beams are often jointed to box-shaped columns with high torsional rigidity as shown in Fig. 1. Therefore, it is considered that beam-to-column joints have effect of warping restraint on lateral buckling for beams. Suzuki and Kimura (2000) elucidated that the lateral buckling load with the warping restraint of the column is larger than that with simple support and the torsional rigidity of box sectional columns is corresponding to warping fixed support. Furthermore, it is shown that elastic lateral buckling load for H-shaped beams with continuous braces under the uniform flexural moment are expressed as the loads between simple support and warping prevention for Kimura (2018). Actually, H-shaped beams carry the gradient flexural moment, when steel moment frames are subjected to seismic forces.

 This paper evaluates the effect of continuous braces on elastic lateral buckling load for H-shaped beams under the gradient flexural moment, considering warping and Saint-Venant torsional rigidities in energy conservation equations. In this study, two types of loading conditions are considered. The continuous braces rigidities are consisting of the lateral and the rotational rigidities as shown in Fig. 1. In the case of Type A, the lateral rigidity is effective for preventing lateral deformation of the compressive upper flange, whereas in the case of Type B, the rotational rigidity is effective for preventing torsional deformation of the beams.

 This study is conducted by the following procedures:

 1. The equations of the elastic lateral buckling load for H-shaped beams with warping restraint and continuous braces under gradient flexural moment is developed by the energy method and numerical analyses. To simplify the equations, the closed from solution of lateral buckling loads are suggested with reference to the terms of the flexural and torsional rigidities of beams and loading conditions.

 2. The elasto-plastic buckling behaviors of the beams are simulated by elasto-plastic large deformation analyses. The lower bound of the buckling stress for the beams are evaluated by the buckling curve of bending member in Recommendation for Limit State Design of Steel Structure (AIJ) with proposing modified equivalent slenderness ratio with the elastic buckling load of the beams with warping restraint and continuous braces.

 3. The upper bound of rotational stiffening moment and lateral stiffening force which continuous braces is evaluated based on the ratio of the flexural and torsional rigidities of beams, the rigidities of braces, the warping restraint ratio, and the gradient of flexural moment.

 While the embedded column base connection is effective for securing stiffness and strength relative to the other type of column base connections, the bearing resistance of embedded portion of the column due to the prying action, may results in undesirable failure mode such as punching shear failure which lead to significant deterioration in the strength. Previous studies have reported that improved the strength, ductility capacity and decreased the strength deterioration were attributed to the arrangement of U-shaped rebars and hoops surrounding the column. Based on these experimental results, Japanese design provision specified the estimation of plastic flexural strength of exterior column base based on the lateral resistance mechanism including axial yielding of U-shaped rebars. This provision however had not clearly specified the reinforcement details.

 The objective of this study is to provide the recommended details of U-shaped rebars for achieving satisfactory performance and evaluate the ultimate strength of column base connection. Cyclic lateral loading tests were performed with SHS column-grade beam sub-assemblages to investigate the cyclic responses including failure modes and ultimate resistances with emphasizing on contribution of rebars to the lateral resistance. A total of 13 specimens were tested, with embedment lengths of column, the amount of U-shaped and hoop rebars and the spacing between the rebar and column in the loading direction taken as key test parameters (See Table 1, Figs. 1 and 2).

 Based on test results, the influence of the reinforcement details on the fracture behavior of the connection, the contribution of the rebars and resulting ultimate strength were investigated (See Table 3 and Figs. 3 to 5). It was observed that the upper U-shaped rebar yielded or remained elastic at the ultimate load, depending on the details of U-shaped rebars as shown in Fig. 7. The evaluation method of anchorage strength of 90° bending rebars [Ref. 5] was extended for embedded column base connection and used to predict the anchorage strength of U-shaped rebars. The predicted axial force of the U-shaped rebars given by the lesser of axial yielding and the anchorage strength of rebar fairly agreed with the experimental results (Fig. 11). In other words, this method can be applicable to determine proper detailing for ensuring the axial yielding of the U-shaped rebars. Furthermore, the ultimate strengths of the embedded column base connection were evaluated assuming the same resistance mechanism in current provision and using modified axial resistance of rebars given by Eq. (9). The proposed method was able to consistently predict the experimental ultimate strengths, regardless of the details of U-shaped rebars (Fig. 13).

 1. INTRODUCTION

 Although several studies on the structural performance of concrete filled steel tubular (CFT) columns and steel beam-to-concrete filled tube (CFT) connections have been reported, the structural performance of rectangular CFT columns has not been clarified yet.

 Herein, seismic behaviors of the rectangular CFT connections were investigated to evaluate their stiffness, shear strength, and hysteresis characteristics. This paper summarizes the test results and discusses the approach to evaluate the structural characteristics, such as yield and ultimate strength, required in the structural design practice, as well as the hysteresis characteristics of the CFT connections. In addition, the paper presents analytical models for representing the hysteretic behaviors of the CFT connections.

 2. TEST PROGRAM

 Eight specimens were tested to investigate the effects of the sectional shape on the seismic behaviors of the CFT connections.

 3. TEST RESULTS AND DISCUSSION

 All specimens showed very ductile and stable spindle-shaped hysteresis loops. The brittle failure was not significant during testing.

 4. STRUCTURAL PERFORMANCE OF RECTANGULAR CFT PANELS

 The shear strength was calculated according to the design formula of the Architectural Institute of Japan (AIJ). The maximum strengths of all specimens were larger than the calculated shear strengths. It is understood that the experimental shear strength values fairly agreed with the calculated strengths. The skeleton curve of the shear-rotation angle relation of the CFT connections can be expressed using the tri-linear skeleton model.

 5. CONCLUSIONS

 The following findings were obtained from the structural experiments simulating a rectangular CFT column-steel beam connection.

 1) The rectangular CFT connection had an excellent deformation capacity and showed stable behavior up to large deformation.

 2) A simple numerical simulation demonstrated the impact of the difference in the cross-sectional shape on the shear behavior.

 3) The shear resistance of the rectangular CFT connection could be appropriately evaluated using design formulas (1) to (5) and formula (6)’.

 4) The restoring force characteristics of the rectangular CFT connection could be appropriately evaluated by estimating _sk in design equations (7) to (8) using equations (10) and (11).

 In the case of steel-concrete composite beams pinned with steel girders, and the beam ends are rotationally restrained by the girders, reinforcing bars in the concrete slab work effectively so that the hogging moment resistance at the pin joints can be exerted considerably. Considering this effect, there is a possibility that fire protections of the steel beams can be omitted or reduced when the load level is low. However, the hogging moment resistance at pin joints considering rebars and the influence of the rebar specification on it have not been clarified.

 This paper presents the results of load-bearing fire tests of composite beams protected only at the beam end with continuous floor slab conducted to clarify the hogging moment resistance at pin joints considering rebars and the influence of the rebar specification on it. 3 specimens (Pure steel beam, Composite beam using composite slab, Composite beam using RC slab) were tested, and the main test parameter was the specifications of the rebars in the concrete slab (the amount of main rebar, the type of rebar, the fixing method). The load level was 40% of the long-term allowable moment, and the heating was in accordance with ISO 834 standard curve. In this study, the hogging moment resistance of the joints and steel temperature of composite beam in fire were discussed.

 The main conclusions from the test results were as follows:

 1) In the case of the cross section of the steel beam used in this tests, it was confirmed that the temperature of the joints during heating for 1 hour remained at about 600 ℃ by protecting at the joints using the fire resistive materials for 1 hour.

 2) The pin joints have sufficient rotational ability at elevated temperature without fracture of the bolts even in bending collapse.

 3) The pure steel beam with protected pin joints had the same ultimate strength as that of unprotected fixed end.

 4) In the case of composite beam using RC slab, the calculated value by the evaluation model of the hogging moment resistance at pin joints considering the rebars shown in Fig. 17(d) corresponded with the experimental values well.

 5) In the case of composite beam using composite slab, the rebars did not contribute to the hogging moment resistance at pin joints because the rebars were fractured at the initial stage of heating. It is necessary to evaluate the hogging moment resistance at the beam ends as simple steel beam without the effect of rebar when the rebars are fractured prematurely.