To address groundwater issues during mountain tunnel construction, it is necessary to estimate groundwater flow rapidly and accurately. Although an evaluation method was developed to predict the amount of water inflow and to reflect it in the construction phase, the predicted value is larger than the observed one, and the over-design of the turbidity treatment facility has been pointed out. One of the reasons is that the method does not consider the effect of vertical seepage of water inflow and the effect of lowering the groundwater level in front of the tunnel face, which occurs in real tunnel construction. In this study, we proposed a new prediction method of water inflow based on the conventional method extended by the physical model to take these effects into account and verified the method using the data obtained from an already constructed tunnel. The simulated results showed that the newly proposed method could predict the amount of water inflow more accurately and could be reflected efficiently in the construction phase.

This study presents an identification method for rockfall sources, using remote sensing and machine learning technology, to reduce overlooking during the inspection cycle and improve slope management efficiency. In road disaster prevention inspections, rockfall damage from overlooked points has been reported, and improvements are essential to ensure that road disaster prevention inspections are carried out. In such cases, the use of remote sensing technology is an effective measure. In particular, effective inspection can be expected by utilizing the terrain analysis map created from the measurement data. In this research, we propose a system that applies machine learning to object detection to a terrain analysis map and detects the source of rockfall on the desk. Based on the wavelet analysis map, which is a type of terrain map, object detection by a support vector machine was applied. As a result, the proposed method succeeded in detecting all existing inspection points. We showed its effectiveness as a tool for improving the efficiency of road disaster prevention inspections.

This study presents three-dimensional slope stability analyses that reproduce actual slope failures induced by a heavy rain and an earthquake. Hovland method, which is one of the three-dimensional limit equilibrium methods, is employed to evaluate the stability of slopes over wide area. While forces acting on soil column are ignored in the method, its low calculation cost enable us to utilize the method in wide-area slope stability analysis. In order to confirm the accuracy of the method, slope failures actually caused by the heavy rain and earthquake are reproduced. The obtained results are compared with distribution maps of the real slope failures with the help of the concept of a ROC curve. According to the results of the slope stability analyses, it was confirmed that the Hovland method outputs relatively lower values of safety factor than actual slope failures. In addition, the effect of multi-hazard condition of earthquake and heavy rain is numerically examined.

To predict the strength development of steel slag-dredged soil mixture, the test methods to evaluate the elution of silica from dredged soil as well as one of calcium from steel slag are proposed in this research. To determine the calcium eluted from steel slag, the regulatory leaching test method (Notification No.46 from Japan Environment Agency) was modified by using the slag samples without grain size adjustments and subjected to the artificial seawater as solvent under the 30-minute shaking time. Silica elution from dredged soil was determined by the method, which was proposed by Toda et al. (2018) but slightly modified in terms of the stirring period - where approximately 7-minute stirring of the sample in water was applied to effectively obtain the results for a shorter period of testing time. Analysis of these elution test results versus the unconfined compressive strengths on 6 different steel slags and 15 dredged soils concluded that the strength developments can be predicted by the elution values obtained by the proposed methods, which can be utilized as primary assessment of strength development in practice.

This study focuses on how diffusion leaching characteristics of heavy metals in clay are influenced by the changes in ground temperature caused by geothermal heat utilization. Diffusion tests were conducted on clays containing geogenic heavy metals and using temperatures of 10, 20, 35, and 50°C. The results showed that diffusion leaching increased with elevated temperatures and lower dry densities. Diffusion coefficients calculated using total content and initial pore water concentration, respectively, were compared, and the latter was found to reflect the effect of temperature better. Furthermore, it was suggested that the viscosity of the pore water might be a factor affecting the temperature dependence of the leaching flux. The results indicate the importance of safety assessment for the introduction of geothermal technology, taking into account the effects of temperature and clarifying the acceptable temperature change.

The adsorbents used in the attenuation layer method contact with water multiple times or over a long period of time after construction, but at present, the performance is evaluated by a single batch adsorption test. In this study, to evaluate the adsorption performance of the adsorbents when they are in contact with water for multiple times or fora long period of time, we conducted a water immersion pretreatment, in which the adsorbent is immersed in water for a certain period, and a serial batch fluoride adsorption test, which is a repeated adsorption test, for four types of adsorbents: LDH, Fe-hydroxide, Mg-oxide and ZVI. As a result, changes in the crystalline phase were observed in the Mg-oxide and ZVI based materials, and the adsorption performance also changed. Therefore, it was suggested that the change in solid phase due to water contact could affect the adsorption performance.

Expressway embankment has an individual characteristic based on embankment material, supporting foundation ground and construction procedure, etc. The difference with respect to the deterioration with time is also induced by the initial condition of the embankment. Therefore, in order to diagnose road embankment soundness, we have to visualize the initial condition of the inside the embankment. In this study, the automatic technology for surveys and evaluations of embankments by using surface wave exploration and electric resistivity exploration has been performed. As a result, we could understand the initial distribution of S wave velocity which is closely related to the stiffness of the embankment. It was also confirmed that location with different embankment materials and different ease of compaction had large variations in S wave velocities. In addition, by using the developed cart and impact power induced by the FWD test, the exploration speed was significantly improved, and the measurement was performed in a very short period of time before expressway opening.

This paper revealed the relationship between the precision for displacement measurement and the slope length based on the measured displacement until failure in model slopes and natural slopes. Displacement interval and time interval were examined as the precision of the displacement measurement. Examination of the relationships between the displacement intervals normalized by the displacement until failure and the deviations from the normalized displacement interval to measured displacement clarified that the normalized displacement interval should be less than 0.05 for precise measurement of the displacement. The time interval normalized by the time until failure was also revealed to be less than 0.05 in the same way. Combining the relationship between the slope length and the displacement until failure to the normalized displacement interval indicated that the displacement interval should be less than 22 mm for the displacement measurement in the slope of the length less than 50 m on which the shallow landslide is expected to occur. Time interval was revealed to be less than 400 seconds for precise measurement of the displacement in the slope of less than 50 m.

The stability analysis for sliding collapse of hillside residential land, which is regulated by laws and regulations, should be performed by the two-dimensional slice method according to the national technical guidelines. Sliding collapse damage can occur by landslide deformation, which is a residual deformation phenomenon, as well as by slip failure. Therefore, creating an appropriate model according to the causal factors of predicted damage forms is a challenge for stability analysis. In order to predict the occurrence of landslide deformation accurately, the present study examined a method for constructing a practical stability analysis model that takes into consideration the causal factors of landslide deformation. In the hillside residential lands where landslide deformation occurred in Sendai City, soil and groundwater characteristics were evaluated and causal factors were analyzed to help establish a model of banking section that properly reflects those factors having a large impact on fluctuations. In addition, the concept of appropriate excess pore water pressure ratio and strength parameters in the stability analysis is presented in consideration of the response acceleration spectrum and safety factors obtained from the simulation analysis based on the dynamic effective stress FEM and the circular slip method.