The relation between the predominant period of H/V spectrum ratio of microtremor and the layer thickness of the ground with different landform evolution was arranged and summarized in a relational expression. Effect of each landform evolution as to geomorphic provinces is reflected in the rational expression. The validity of the relational expression was then confirmed using actual ground motion records. When developing a hazard map of the ground, it is proposed to divide it into geomorphic provinces that show the same landform evolution as basic zoning data, and to subdivide the geomorphic provinces by the H/V spectrum ratio and sedimentary layer thickness.

In the stability evaluation of ground for important structures such as nuclear power plants, the stress of the ground during an earthquake is calculated by two-dimensional FEM analysis, and the safety factor for slip is estimated from the driving stress and the resisting stress on the assumed slip lines. However, the setting of the slip line requires a lot of preliminary study and engineering judgements. Therefore, it is difficult to set the critical slip line, which has the minimum safety factor in complex ground models. In this paper, we developed a method to find the critical slip line using particle swarm optimization based on the stress of the ground evaluated by FEM analysis. This method is applicable to search the critical slip line through a discontinuity plane such as a fault plane. To confirm the validity of the method, searches are conducted for several geologic models, and the results are compared with previous studies and analyzed in detail. It is shown that the method enables searching the critical slip line appropriately and efficiently based on the result of FEM analysis in complicated geologic model including a discontinuity plane.

To support an efficient and rapid investigation of seismically induced damage to buildings, this study applies a deep learning algorithm to on-site photos taken after earthquakes. The convolutional neural network (CNN) is employed to detect the collapsed wooden houses after the 2016 Kumamoto earthquake. The accuracy of discrimination was improved when the on-site photos of collapsed buildings after the other Japanese earthquakes were included in the image dataset. The dropout layer was also very effective to prevent overfitting. The overall accuracy was 80.8%, and this study is helpful to support damage investigation performed by municipalities soon after an earthquake.

The accurate evaluation of story-collapsed buildings is important for human casualty estimation in seismic risk assessment and disaster response. This study investigates the story-collapsed buildings in Mashiki Town, that are identified based on damage survey data of the local government, due to the 2016 Kumamoto, Japan, earthquake. As the result of analysis from the viewpoints of structural material and construction period, the story-collapse ratio of wooden buildings was found to be much higher than that of other non-wooden buildings, and it becomes larger as the construction period gets older. By comparing this result with the estimated distributions of the peak ground velocity (PGV) and the Instrumental JMA seismic intensity, the fragility curves for story-collapsed buildings were developed. The developed fragility curves for Mashiki Town showed lower story-collapse ratios than those developed for Nishinomiya City in the 1995 Kobe Earthquake.

To improve the repairability of reinforced concrete (RC) buildings after an earthquake, it is desirable that the damage area of the members be small. One method to control the damage area is de-bonding of longitudinal rebars. The de-bonding of longitudinal rebars can suppress the damage area of an RC beam member, but the strength and energy dissipation performance are lower than those of a normal beam member. In this study, to suppress the damage area and secure the restoring force characteristics, the structural behavior under antisymmetric bending of RC beam with a guaranteed hinge position having de-bonding of longitudinal rebars and its anchoring at the center of the span was examined. First, finite-element analyses of the RC beam member with the parameters of the de-bonding position of the longitudinal rebars, the anchoring method of the longitudinal rebars at the center of the span, and the presence or absence of the slab were conducted. The analysis demonstrated that restoring force characteristics equivalent to those of a normal beam can be secured and the damage area can be suppressed by de-bonding of the longitudinal rebars only at the four corners, along with anchoring with a steel plate at the center of the span. Next, a cyclic loading test of an RC beam with a slab manufactured using that de-bonding and anchoring method was conducted. The test confirmed the stable restoring force characteristics. In addition, cracks were concentrated at the assumed hinge position, and the damage area could be suppressed. The maximum residual crack width of the beam was small enough that the building could continue to be used with minor repairs.

We identified the story stiffness and damping from a shaking table test of 5-story steel building with various dampers, considering the rocking motion of the table. The identification method is based on the first modal response of substructures successively truncated at lower stories from the top. This method was verified by comparing shear-force vs story-drift relation obtained by measurements and the identification. Neglecting the rocking motion of the table in the identification process, smaller story stiffness and larger story damping can be erroneously evaluated. In the 5-story steel building model case, the complex damping was better to explain the second peak of the transfer function than viscous damping.

In this study, we integrated the disaster-victim certificate data by the Mashiki Town government and the Uki City government, Kumamoto Prefecture, due to the 2016 Kumamoto earthquake, and constructed building fragility curves with respect to the structural material and construction period in combination with the estimated seismic ground motion distribution. The correlation coefficient showed a strong positive correlation of around 0.9 in all classifications of the building fragility curves covering a wide range of ground motions. The validity of the proposed fragility curves was demonstrated, compared with the existing fragility curves that were developed for different data sets and different damage category.

The purpose of this study is to generate a nationwide map of average shear-wave velocity of ground upper 30 m (Vs30). Before estimating the nationwide Vs30, we predicted local Vs30 for expanding the data from a shallow velocity profile. Therefore, two types of predictions, which are nationwide Vs30 using terrain features and local Vs30 using a shallow velocity profile, were executed in this study. These predictions utilized machine learning methods: Support Vector Machine, Random Forest, and Gradient Boosting Decision Tree. As a result, relatively high accuracy was achieved. To explain the interpretability of the models for nationwide Vs30, three analyses were implemented. Analyses include important features for prediction; correlations between prediction and terrain features; and model sensitivity to slope and elevation. The important features are elevation, slope, distance from the tertiary mountains and geomorphologic classification similar to the explanatory variables utilized in Matsuoka et al. (2005). Correlation between results and features is reflected in the physical hypothesis with the exception being the result for tertiary mountain. In the sensitivity analysis results for elevation, there was a contradiction to the physical hypothesis in some geomorphologic classification such as tertiary mountain. Finally, the nationwide Vs30 map was generated after substituting the Vs30 estimated by Matsuoka et al. (2005) for the case of unintended results in the geomorphologic classification.

We examined the scaling relations proposed in the previous studies (the 3-stage model; Leonard, 2014; Hanks and Bakun, 2008; Thingbaijam et al., 2017) by using the database composed of the source models of the inland crustal earthquakes (M_w4.5-8.0) which occurred in the foreign countries. At first, we trimmed the periphery of the source inversion models with the small slip, which would not affect the seismic moment. Then, we validated the scaling relations between length, width, fault area, average slip, and the seismic moment. The 3-stage (HERP, 2020) and 2-stage (Hanks and Bakun, 2008) scaling relations assuming the saturation of fault width and the 1-stage (Leonard, 2014; Thingbaijan et al., 2017) scaling relations assuming with self-similarity show more or less similar levels of agreements to the source parameters. We extracted asperities from the trimmed fault models, and we found the combined area of asperities was 23% of the trimmed fault area on average. We found that 22% (HERP, 2020) of the 3-stage scaling relation of the fault area shows best agreement to the combined area of asperities in the wide range of the seismic moment. The natural logarithmic residuals between the source parameters and the scaling relations of fault area and asperity area have a tendency to be smaller in the larger seismic moment (M₀ > 10¹⁹ Nm), and the average and the standard deviation of the residuals of the 3-stage scaling relation and the one of Thingbaijam et al. (2017) are smaller than others. Furthermore, we selected the source models inverted from strong motions, and we validated the scaling relations by using the source parameters extracted from the selected source models. The natural logarithmic residuals of the selected models were not significantly different from those of the whole source models including models inverted from not only strong motions but also teleseismic data, InSAR data, and so on. Taking into account the correspondence of the dataset and the different schemes of the methods to estimate the scaling relations, we validated the scaling relations without Thingbaijam et al. (2017). It is found that the 3-stage scaling relation provided the smaller average and standard deviation of residuals at all stages in terms of both fault and asperity areas, and that the simple scaling relations of Leonard (2014) also provide similarly small residuals within the wide M_w range.

When creating input ground motions based on the response spectrum in the seismic design for nuclear facilities in Japan, for horizontal components, simulated ground motions, which are composed of two mutually orthogonal components, are to be created using one target response spectrum. In this case, the characteristics of two ground motions are distinguished by the randomness of the phase angle given by the uniformly distributed random numbers and/or the difference in the phase characteristics of the two different components of the observation records. On the other hand, according to the US Nuclear Regulatory Commission standards, when performing seismic response analysis for nuclear facilities using the method of simultaneous input of three earthquake ground motion components, it is stated that the three components should be shown to be statistically independent of each other, and an absolute value for correlation coefficient which was proposed by Chen (1975) is introduced as a criterion. In this paper, focusing on the correlation coefficient by Chen (1975), we analyzed the correlation coefficients between two orthogonal components of strong-motion records observed in Japan after 2000, and performed statistical analyses of these correlation coefficients and analyzed the impact of various earthquake-related parameters upon them. In addition, we actually created simulated ground motions via a common practice based on the response spectrum, and analyzed their correlation coefficients.