Clay wall on bamboo lathing is a traditional Japanese method of constructing walls using a timber framework, bamboo lathing, and plaster produced by mixing clay and straw. Although the use of clay walls is environmentally friendly, it is necessary to evaluate the performance of the walls, especially their structural resistance, so that they can be used widely and safely.
 This paper aims to clarify the influence of the behaviors of plastering layers of clay walls upon their mechanical characteristics and fracture behaviors. Two experiments were conducted, that included diagonal loading onto small element specimens, referred to element experiments, and the loading of horizontal force onto structural specimens, referred to structural experiments.
 First, the element experiments were conducted to clarify the influence on the mechanical properties of three factors; (1) degree of looseness around the end of “bamboo furring”, that is inserted into holes in the sides of neighboring framework members, (2) composition of wall clay mixture for each plastering layers, base-coat layer and middle-coat layer, and (3) ratios of thickness between the base-coat and middle-coat layer. As a result, we clarified the followings;
 (1) In case of without looseness around bamboo furring, the load tends to small because larger delamination between the face and back of the base-coat layer is created.
 (2) The initial stiffness was high with the specimens using mixture had large elastic modulus and compressive strength for the middle-coat layer, so that this layer started to resist at an early stage of loading. Then around the maximum load point, after the corners of the base-coat layer started to resist, the load was large with the specimens using mixture had large absorbed energy for the base-coat layer.
 (3) The initial stiffness was higher when the middle-coat layer was thicker, so that this layer started to resist at an early stage of loading.
 Secondly, the structural experiments were conducted to confirm the results of the element experiments. The specimens employed had the same cross-sectional dimensions of framework members as those used in real buildings. However, the specimens were proportionally smaller in height. The specimens had different factors based on the result of element experiments. Horizontal force was loaded to the positive and negative sides, alternating every three cycles until the prescribed deformation angle was reached. As a result, we clarified the followings;
 (1) From the observation of the destructions, we confirmed four elements contributed to resistance to the horizontal forces. These resistance elements exist different layers, two base-coat and two middle-coat layers. Therefore, resistance of those elements cause the different behaviors of each layers and cause the delamination at the interface of layers that are difficult to repair.
 (2) Clay walls of high initial stiffness can be achieved by avoiding gaps around the resistance elements. However, such walls tend to develop the delamination at the interface of layers.

View full abstract

 From the viewpoint of the durability of coating materials for buildings, the estimation of their weatherability is critical. JIS A 6909 has regulated the weatherability types for coating materials based on the accelerated weatherability tests. However, the relationship between the deterioration progress in coating materials under the accelerated weatherability test and that under the outdoor exposure test is not clear yet.
 An outdoor exposure test on coating materials for 13 years in Sapporo, a literature survey on the results of the outdoor exposure tests and accelerated weatherability tests (using the sunshine and Xenon lamp weather meters) were carried out to elucidate the relationship. In a previous study, the authors proposed an equation for estimating the gloss retention decrease, which is an indicator of deterioration in coating materials. In this study, the applicability of the proposed equation to the experimental results of the outdoor exposure test and literature survey was investigated. The results showed a high correlation between the estimated results and the measured results for gloss retention though there are some exceptions.
 To investigate the estimation method of the gloss retention after the long term exposure at the initial material age, we also compared the ten years gloss retention measured values to estimated values using two years values. The estimation of the long-term outdoor exposure results from the short-term results was difficult. The outdoor exposure test results from multiple regions showed that the weatherability was mainly affected by the coating material type.
 A method for the estimation of the gloss retention decrease during the outdoor exposure test using the accelerated weatherability test results was proposed. The average gloss retention was determined using the gloss retention obtained from the accelerated weatherability test. The dispersion for the average gloss retention showed the regional differences. The gloss retention was estimated by a probabilistic model using the average gloss retention and standard deviation obtained from the experimental results. Consequently, the obtained experimental results were almost within the significance level of 5%.
 Finally, the standard deterioration of the gloss retention for each resin type of the coating materials was also analyzed. The standard degradation progress of gloss retention and this definition enabled to calculate the service life of each resin type. The results showed that the weatherability types 1, 2, and three regulated in JIS A 6909 were equivalent to the service lives of fluorine, urethane, and acrylic resins respectively. The estimation results correlated with the empirical service life of the coating materials.

View full abstract

 A parametric study is applied to determine the dynamic cross interaction (DCI) of two closely spaced adjacent buildings focusing on the mean power of superstructures. Analyses are done for mat and pile foundations taking place on layered elastic soil and elastic half-space considering different embedment depths. Adjacent buildings are analyzed by the substructure method consisted of the thin layer method and finite element method in 3 dimensions. Consequently, 3 main factors are found to determine the DCI effect. First one is mass ratio of adjacent buildings that has detrimental consequences on the lighter adjacent building. Second one is rocking restriction that is beneficial effect especially on mat foundations and tall buildings. Third one is horizontal vibration power transfer between foundations having different embedment that induce due to the different horizontal movement of adjacent foundations at the same frequency. This detrimental effect is observed on identical tall adjacent buildings placed on mat foundations. However, this effect cannot be observed on pile foundations. Moreover, more detailed parametric analyses of height and mass of building and soil properties are applied for mat foundation placed on half-space soil.

View full abstract

 Seismic floor system, which can decrease seismic responses (e.g., story drift, absolute acceleration on the floor, and stress of members) by means of inserting visco-elastic materials (VEMs) between concrete floor slab and beams as shown in Fig. 1, has been developed to enhance not only seismic performance but also workability of dismantlement of buildings. The effects of the floor system on the seismic responses have been studied by both time history analysis and shaking table tests of single-story simplistic elastic structure model. And also the validity of evaluation method of elastic response was confirmed in Ref. 2. On the other hand, it was verified that the effect of the seismic floor system deteriorates if the VEMs support the floor slab directly (Ref. 2). Then, in this paper we propose laminated VEM, which is made by pasting alternately visco-elastic material sheets and plated thin steel plates, in order to increase the compressive stiffness of VEM. And this paper describes compression and shear characteristics of the laminated VEMs based on compression test and shaking table test respectively.
 Firstly, two kinds of compression tests are conducted to confirm both short term compression characteristics and long term compression characteristics. Test parameters are ratio of paste area to surface area of VEM, aspect ratio of paste area of VEM, and compressive stress. From the short term test, it is revealed that static compressive strain becomes smaller as the ratio of paste area to surface area of VEM becomes larger. Furthermore, the compressive strain can be estimated based on the material properties by using the compressive stiffness proposed in Ref. 6 and 7. On the other hand, from the long term test, it is revealed that creep strain becomes smaller as well as the static strain. However creep coefficient varies widely, and maximum total compressive strain can be obtained as three times of the static compressive strain.
 Next, shaking table test of single-story steel frame is conducted to confirm shear characteristics of laminated VEMs. Comparison between the test with linear sliders not to support the floor weight and the test without linear sliders to support the floor weight by laminated VEMs, both results almost agree because compression deformation of the laminated VEM is small even when the VEMs supports the floor weight. And it is confirmed that the maximum responses, which are the story drift and the shear deformation of VEM, can be estimated by means of modifying only to use the thickness after deforming from the evaluation method of shear characteristics of VEM proposed by Ref. 2. Furthermore, the laminated VEM behaves stably under bi-directional input as well as uni-directional input, and the responses under bi-directional input can be obtained by combining the responses under uni-directional input respectively when the frame remains elastic and does not have eccentricity.
 As a result, it is revealed that great compressive stiffness can be achieved by using the laminated VEMs, and high reduction effect of seismic responses can be obtained even if the other mechanisms of supporting the floor weight are not used.

View full abstract

 In nonlinear structural analyses, it is very important to use an accurate hysteretic model. However, there are some materials and structural members for which the human development and application of a hysteretic model are too complex. Such situations can apparently be addressed using a recurrent neural network (RNN) system. An RNN is a type of neural network (NN) that can be changed by a system based on previous hysteresis inputs and outputs.
 The objective of the present study was to confirm the applicability of RNN modeling to the normal bilinear modeling of very complex nonlinear hysteresis. First, we identified the appropriate long short-term memory (LSTM) RNN model using training data comprising bilinear historical displacement as input and force as output data. It was confirmed that the identified RNN model could be applied to the system, and that it was equivalent to a theoretical bilinear model.

 Following is a summary of the study findings:
 (1) A three-layered RNN model with LSTM units is required for a bilinear system. This was determined using training data comprising theoretical bilinear input-output data. Comparison of the output of the identified RNN model with that of a theoretical model revealed good agreement. The changeability of an RNN by a system based on previous hysteresis inputs and outputs was confirmed in the case of basic hysteretic bilinear model, which is often used for structural seismic design.

 (2) The sparsity of the weight matrix of the identified RNN model was investigated. The matrix was not found to exhibit sparsity when its layer contained “drop out”. This indicated that the RNN model was very large and had many weight coefficients, which would make its analysis very difficult.

 The findings of this study confirm the applicability of RNN modeling to structural hysteresis. There is, however, the need for further improvement of the procedure.

View full abstract

Horizontal hybrid structure which involves stiff cores to enhance the seismic and fire resistance is sometimes employed as a solution of large wood buildings. However, it makes it difficult to evaluate the three-dimensional vibration mode and the connection force between the two different structures. The authors have proposed dynamics model of horizontal hybrid structure considering continuous body based on an assumption that shear walls are equally distributed in the plan. In this paper, the former model is developed to consider various arrangements of shear walls, and new method to predict maximum displacement mode and member forces is proposed. The accuracy is demonstrated by comparison with numerical analyses, and they give close agreement.

View full abstract

 A model is presented for pushover analysis of reinforced concrete (RC) beam-column joints in soft-first-story frames. In such frames, the sections of first-story columns are usually much larger than those of second-story columns to prevent story collapse (Fig. 1). Based on the experimental results of specimens O-1 and O-1t (Fig. 2a), which are reported in Reference 3, this study proposes new equations for calculating the strength and the yield deformation of such frames with columns extended toward outside of frames and subjected to opening load.
 Figure 1a shows the column failure mode, which is considered in the conventional model. Figure 1b shows the beam failure mode, which is proposed in this paper. To represent these two modes, the proposed model includes a rotational spring of first-story column and that of beam (Fig. 2b). Taking the result of strut-and-tie analysis (Fig. 9) into account, we determine the non-linear characteristics of these two springs which may be applied to actual design.
 The strength of column spring, _yM_c1, mainly consists of the moment capacity of the first-story column of which depth is l_b, where l_b is the development length of the beam bottom bars. Moreover, adding the contribution of the hoops in the joint, _yM_c1 is calculated from Eq. 3. The column spring is located at the center of the beam bottom bars (Fig. 2b). The strength of beam spring, _yM_j, is evaluated as the sum of the flexural resistance of the beam (_yM_b) and that of the second-story column (_yM_c2) (Fig. 11). _yM_b is calculated as the moment capacity of the section shown by hatched lines in Fig. 11 (Eq. 4). The beam spring is located at the center of the joint (Fig. 2b).
 The yield deformation of each spring is determined based on the observed strain distribution. Figures 14d and 14f show the deformations related to the column and beam spring, respectively. It is assumed that the strains of the bars in the dark regions in Figs. 14d and 14f reach the yield strain when the column or beam spring reaches the yield deformation. Considering the elongations of the longitudinal bars caused by the rotational deformations (Figs. 14d and 14f), the yield rotations of the column spring is given by Eq. 13, and that of the beam spring is given by Eq. 14.
 The model defined in Fig. 2b is integrated into the model as shown in Fig. 2d with a single spring at the end of the column.
 The load-drift relationships obtained from the proposed model (the chain lines in Figs. 4 and 5) agreed with the test results much better than the conventional model (the broken lines). The ultimate strengths and the yield deformations given by the new model accurately correspond to the test results with the ratio of 1.02 to 1.21 (Figs. 16 and 17). It is indicated that the conventional model may provide the higher strength than the observed value and less than a half of the observed deformation due to neglecting the beam failure mode.

View full abstract

 In Japan, the structural performance of RC shear walls with openings has been directly evaluated by several strength reduction factors implemented in AIJ standards, while the strength reduction due to door openings (AIJ 2010) has not been verified thoroughly, owing to the lack of test date from reliable experimental simulations on fundamental behavior of shear walls with door openings. This paper reports a series of laboratory tests conducted on multi-story shear walls with variable layout of door openings, investigating the performances and behaviors of the specimens.
 On the other hand, a new loading method was proposed, which was verified through two-dimensional FEM analyses of three-story specimens with different loading conditions. The analyses considered W1-D and W1-C subjected to triangularly distributed loads on the beam of each floor and a couple of concentrated loads at the equivalent height (7/9H) simulating equivalent shear force and bending moment on the bottom of the W1-D, respectively. Consequently, the shear force vs. drift angle relationships and the minimum principle stress contours showed good agreements between two models.
 Three 1/4 scale shear wall specimens were designed with variable layout of door openings: WN0, WV1 and WD1, respectively without openings, with eccentrically-aligned openings and with diagonally-aligned openings. The specimens were subjected to static cyclic lateral loads induced by two horizontal jacks connected to two separated stubs at the height of 7/9H on both sides of the specimens. PC bars were provided between the stubs to prevent tensile failure at the stub-wall boundaries.
 WN0 showed compression concrete failure at the base of a column (Fig. 12), and failed in shear. While WV1 and WD1 specimens with openings failed mainly with concrete crushing of the beams around openings, and showed compression failures at their bases as well, occurring differently under the positive and negative loads due to the unsymmetrical layout of the openings (Figs. 20 and 25). As a result, the seismic performance, especially the maximum strength of the specimens decreased due to the existence of door openings: 500kN for WN0, 183kN and -241kN for WV1, 365kN and -299kN for WD1, respectively. Therefore, the lateral strength estimated for WV1 with eccentrically-aligned openings based on the reduction factor r3 overestimated the maximum strength obtained from the test, while WD1 with diagonally-aligned openings based on the reduction factor r2 overestimated the maximum strength under the negative loads and underestimated the maximum strength under the positive loads.
 As observed during the tests, the failure patterns that high compression was applied to the beams around openings (WD1 and WV1), and to the bottom of the wall with the tensile boundary column (WV1), should be taken into consideration in practical design for multi-story shear wall with door openings. Meanwhile, the experimental results obtained in this study, especially the different seismic performance of specimens under the positive/negative loads due to the unsymmetrical layout of the openings, can be set as a premise for future researches on the same subject and/or a base to improve the design code of AIJ standards.

View full abstract

 It has been shown that a slab has an effect to restrict lateral buckling of the steel beam, and many studies have been done. However, the study on evaluation of the elastic lateral buckling strength is the main constituent, and there are few studies on plastic deformation capacity. In late years the experimental studies on composite beam have advanced, and a lot of data are accumulated. Depending on this, the studies paid its attention to plastic deformation capacity are accomplished. But complicated numerical analysis and evaluation model are used, and the expression is not practical.
 This paper shows a method to evaluate plastic deformation capacity in consideration of a slab buckling restriction effect in a simple and easy expression and suggests a design method to examine lateral bracing.
 The end boundary condition and the action moment of the beam to intend for assumed it two cases of the next.
 Case1: Both ends fixed support and double curvature moment
 Case2: One end fixed, another end simple support and single curvature moment
 A previous evaluation is used for the plastic deformation capacity, but therefore a good elastic lateral buckling strength of the precision in consideration of a slab buckling restriction effect is necessary. According to the existing study, it is thought that the transverse of the upper flange is completely restricted as for the buckling restriction effect.
 The elastic buckling strength in consideration of a slab restriction effect M_cre becomes the middle value not to consider rotational restriction M_cr1 and rotation fixed M_cr2. Therefore elastic lateral buckling strengths not to consider rotational restriction and rotation fixed are calculated by eigenvalue analysis on FEM. Using the analysis results, the elastic lateral buckling strength of both values M_cr1, M_cr2 are evaluated in a simple and easy expression using an elastic lateral buckling strength applied equal bending moment under each boundary condition M_cr0.
 So, the elastic lateral buckling strengths are calculated by eigenvalue analysis under rotational rigidity changed to infinite from zero in succession sequentially. Using the analysis results, the elastic buckling strength Mcre is expressed by a logistic function using M_cr1, M_cr2 and torsional rigidity ratio of slab and beam.
 In this paper, torsional rigidity of the slab is calculated with an effective width being represented by a column width. According to this, the torsional rigidity ratio becomes around 15 from 5 and shows practical value. This expression secures good precision in this range. Then, applying this to a previous expression, a plastic deformation capacity is evaluated. The validity of the evaluation is shown by a comparison with the experimental value.
 Finally the design method to compare the plastic deformation capacity by shown evaluation with the demand capacity decided by member angle of the beam is shown. The required deformation capacity is set low so that a beam becomes slim. Therefore, it is shown that the lateral bracing becomes needless with most beams without a slab restriction effect.

View full abstract

Seismic motion amplification ratio corresponding to fundamental period, Gs₁, is a very important factor used for evaluating site effects. In this paper, a simple procedure for evaluating Gs₁ of layered soil profiles is proposed. Firstly, a theoretical formula for Gs₁ of a two-layer soil profile on elastic bedrock, is developed. For a multi-layer soil profile, the multiple soil layers are approximately replaced by equivalent two layers, then Gs₁ of the multi-layer soil profile can be estimated by that of the equivalent two-layer soil profile on bedrock. To demonstrate the validity of this development, Gs₁ of 68 representative soil profiles are evaluated, it is shown that the results by the proposed procedure agree well with those by the wave propagation method. Moreover, applying the proposed method for Gs₁, free field response spectrum of 4 soil profiles are evaluated, and compared with the results by program SHAKE, good agreement between results by the two methods is observed.

View full abstract

 The success of simulating the collapsing behavior of steel columns under fire conditions depends on how accurately time-dependent mechanical properties of structural steels can be predicted at elevated temperatures and high stress levels. In the field of fire resistant design of steel buildings, high-temperature creep behavior is not often included in constitutive relationships, but has a significant influence on the fire response of the structures. In this paper, the multiplicative viscoplasticity model was selected for time-dependent mechanical model of structural steel at elevated temperatures, and the material constants of the model were identified by using the stress-strain curves of tensile tests at two different tensile speeds at elevated temperatures.
 All experimental results have been obtained using specimens of SM490 steel that were made in Ref. 7. Uniaxial tensile test of SM490 under temperature (350, 400, 500, 550, 600, 700°C) and tensile speeds (1.0 and 0.05 mm/min), corresponding to the strain rates of about 1.5 %/minute (FAST tests) and 0.08 %/minute (SLOW tests) respectively, were carried out. The experimental results showed that the flow stresses of SM490 were both temperature and strain rate dependent. Due to the rate dependency at high temperature, the ultimate stresses of FAST tests were higher than those of SLOW tests at 500~700°C
 A multiplicative viscoplasticicity model was selected to represent the uniaxial monotonic behavior of SM490 steel, and the set of material constants of the model was identified by the results of uniaxial tensile tests. The fitting parameters were determined with a nonlinear least squares optimization method through a Gauss-Newton iterative procedure. Comparing the material constants of strain-rate sensitivity in reference 9 with those derived from the mechanical model in this study, the strain-rate sensitivity between them was almost the same.
 By integrating the viscoplasticity model under constant stress condition, the Norton-Bailey creep law can be explicitly obtained. Therefore, we can simulate the actual creep curves by the creep model, which was derived from the viscoplasticity model, and we were able to evaluate the validation of the mechanical model for estimation of actual creep response behavior. Correlations between measured and predicted creep strains for SM490 steel at elevated temperatures were obtained and a relatively good correspondence was found. Especially, the simulated creep curves at the highest applied stress level corresponded with the actual creep responses, and the predicted creep curves at low strain levels gave a little variation. This relation could depend on the fact that the stress response levels of the tensile tests were higher than those of actual creep experiments in reference 7.
 From the comparison of creep responses, it appears that one can predict the creep response of structural steels reasonably well from strain controlled tensile data. This is clearly useful considering that the strain controlled tensile tests can be accomplished rapidly compared with a set of creep tests.
 In summary, the following results were obtained:
 1) The effect of strain rates on the stress-strain curves of tensile tests was expressed reasonably well with the use of the multiplicative viscoplasticity model. Comparing the strain-rate sensitivity in reference 9 with those derived from the mechanical model, the material constants of strain-rate sensitivity between them were evaluated and found to be almost the same.
 2) In the case of creep tests at over 500°C and highest stress level, the creep model derived from the multiplicative viscoplasticity model approximated to the experimental result of creep tests reasonably well. Finally, the accuracy of creep model at relatively low stress levels did not obtain satisfactory results.

View full abstract