Utilization of collected and accumulated data of the bridge inspection results for maintenance is required. Although the qualitative evaluation items of the bridge have previously been often used, in this research the opinion of bridge inspection by technical expert is focused. Topic model is probabilistic modeling of term frequency occurrences in documents, which can be used to classify documents of the opinion of bridge inspection. And apply a topic model for the text data to detect the characteristics of bridge damage not recorded as database evaluation items. The topics corresponding to the characteristic damage condition of the bridge are organized. A topic has been organized by high probability words about deterioration due to ASR (Alkali-Silica Reaction), which is a characteristic deterioration in the Hokuriku region. Focusing on the opinion of bridge associated with high probability topic, it is possible to do early detection and preventive maintenance. While these effective results were obtained, some topics were organized by characteristic writing by the inspection company. This means that in order to accumulate the inspection data while maintaining quality, it is necessary to decide the criteria of opinion method of listing and contents.

The pulling out of existing pile is one of the critical issues in recent years with the aging of infrastructures, and it is difficult to clarify that the construction has been completely carried out because it cannot be directly observed. In recent years, two types of a wire rope type and a chucking type have been generally applied for pulling out existing piles. The chucking type is developed as a new method of the pulling out method to overcome the issues of the wire rope type for which impossible to pull out the damaged pile and the filling material cannot be injected from the deepest part at the same time. However, the pulling out methods is a method based on the rule of thumb, and it is not shown that the filling material surely shows its enough strength. Therefore, this research finds grounds for establishing the pile tip chucking type existing pile pulling out method, and statistically analyze the test results of the filler material samples collected after the construction of the chucking type existing pile pulling out method. The reliability of the pulling out method was evaluated.

The S-N curve regression parameters of structural steels were obtained by analyzing “Database on Fatigue Strength of Metallic Materials” published by the Society of Material Science, Japan. According to authors’ recent researches, it was found that S-N curves for structural steels can be estimated from the tensile strength of the respective steels. In such estimations, it was also found that the estimation accuracy could be improved by putting some classifications of the kinds of steels depending on the specimen fabrication process and fatigue testing conditions. Thus, the useful classification index (C_i) based on the S-N curve regression parameters and tensile strength _σB is proposed to estimate the S-N curve for structural steels with the high accuracy. It was finally confirmed that C_i had the consistency with the “semi-logarithmic bilinear model”, and that the classification by using C_i has a sufficient validity on structural steels compiled in the above database. Another finding is that the specific number of stress cycles at the knee point of the S-N curve (N_w) has a clear dependence on both C_i and _σB. That is, N_w tends to increases with decrease of C_i under the condition of same strength, while N_w tends to increases with increase of _σB within the steel group giving the same level of C_i.

Recently, assessing the integrity of the structures accurately and reliably has become extremely important in various fields in order to increase operational lifetime and improve safety. Detecting surface damage such as cracks in evaluating the soundness of the structure is particularly important as it is one of the major factors causing deterioration and destruction of the structure. In this study, we develop a two-step screening system for structural surface crack by using object detection and recognition techniques based on deep learning. The accuracy of object detection depends largely on the quality and quantity of images given as training data in advance. In this study, we also construct the system to make up for the lack of training data by converting the detection results into teacher data.

The combined finite-discrete element method (FDEM) has attracted significant attention for numerical simulations of complex fracture process of rock-like materials as one of the promising hybrid methods. The mainstream of FDEM simulators developed to date is based on the intrinsic cohesive zone model (ICZM) in which cohesive elements are inserted into all the boundaries of continuum solid elements at the onset of simulations and an artificial elastic behavior must be incorporated to model the intact deformation of rock-like materials. However, the effect of introduction of the artificial elastic behavior on the precision of intact stress wave propagation has not been discussed in previous literatures and this paper discusses this issue. As an alternative for the ICZM-based FDEM, we apply the FDEM based on the extrinsic cohesive zone model (ECZM). An advantage of the ECZM-based FDEM is presented through the 3-dimentional (3-D) numerical modelling of dynamic tension test. In addition, the effect of considering the anisotropy of wave propagation in granite, which has been neglected in all the previous works using FDEM, is investigated through the ECZM-based 3-D FDEM simulation of dynamic Brazilian test with a split Hopkinson pressure bar apparatus. Through the presented numerical simulations, it can be concluded that the ECZM-based FDEM may be an alternative for numerical simulations of complex dynamic fracture process of rock-like materials instead of the ICZM-based FDEM.

A numerical analysis model considering mineral distribution of targeted rock to predict the behavior of fracture evolution and tensile strength that is obtained through the Brazilian test is proposed. The model considered the mineral distribution obtained from image analysis. In order to confirm the validity of the model, Brazilian tests employing granite specimens were conducted. Subsequently, by utilizing the model, the numerical analysis coupled with damage theory was performed to reproduce the results of Brazilian test. Simulated results are in good agreement with fracturing process and tensile strength obtained from experiments, and validity of the model was achieved.

In general, the method used to obtain the tensile strength from a Brazilian test depends on the elastic theory for a pair of concentrated forces. However, the contact areas between the specimen and loading platens are observed in experiments. Several attempts have been tried for the boundary conditions in modeling of an isotropic or orthotropic circular disk in the Brazilian test. The forces in only the normal direction to the loading platen or specimen have been dealt with in most of these models. Therefore, a dent in the contact area between specimen and loading platens is observed in numerical results from most models. However, no dents have ever been observed in experiments using loading platens. In this study, not only normal loads but also shear loads on the loading platens are adopted as boundary conditions. The depth of the resulting dent from the present model is smaller than those seen in previous studies. In addition, huge tensile stresses are observed near both edges of the contact area in the present model. Furthermore, cracks are also observed near both edges in experiments. In brittle materials, cracks are caused by tensile stress. Therefore, the accuracy of calculations is expected to be improved using the present model for Brazilian tests.

Strain softening is the mechanical behavior of the soil and rock materials and is important in understanding the soft rock foundation. To investigate the mechanical behavior of siltstone which is one of the sedimentary soft rocks, the consolidation tests using constant-strain rate loading were conducted using the consolidation ring to constrain the lateral deformation. Using the Quaternary siltstones distributed in the Boso Peninsula, central Japan as specimens, the strain-softening was confirmed in the consolidation process for some specimens by two test machines in Kyoto University and Nagoya Institute of Technology, respectively. Just before the yielding, stress decreased suddenly at increasing strain. The stress at the time of the softening were different even for specimens taken from the same formation. Furthermore, micro-focus X-ray CT images taken before and after the tests indicated that the specimen had no macro crack inside. It suggests that the strain-softening is not due to brittle failure in local areas, but due to the softening of the framework structure of the siltstone itself.

In-situ orientations of drilled core samples are valuable for a broad area of geological and geophysical researches (e.g. the determination of in-situ principal stress directions or the analyses of geological structures). In many cases, however, the information on in-situ orientation of core samples is not available. In this paper, a method for restoring the in-situ orientation of drilled whole-round core samples and its application to hemipelagic ocean sedimentary rocks are presented. The method is based on natural remanent magnetization (NRM) of rocks. As a case study, we applied this method to the 15 oceanic sedimentary soft rock samples collected at the toe of the Nankai trough, SW Japan, during International Ocean Discovery Program (IODP) Expedition 370. We developed a new sample preparation procedure which enabled us to get more data and then to evaluate the NRM measurement results well. Some samples were not able to be reoriented due to magnetic overprints associated with the drilling operations. In order to evaluate the magnetic overprints caused by drilling operation and to assess the data quality of the core reorientation by this method, we propose an evaluating system including three ranks defined for the different levels in which NRM results were affected by drilling operations. This evaluating system could be useful to assess the data quality of core reorientation by NRM method in the other similar applications.

High strength and ultra low permeability concrete (HSULPC) is being considered as a material used to package transuranic (TRU) waste for disposal in geological repositories. Therefore, information on the permeability of HSULPC is essential. Permeability tests need to be highly accurate to determine the hydraulic conductivity of HSULPC because of its ultralow permeability. In our study, we measured the permeability of HSULPC samples using the transient pulse method. The temperature of the concrete was finely controlled and held constant. The hydraulic conductivities were determined from the measurements to be around 10^-13 to 10^-12 m/s for confining pressures between 2 and 10 MPa. The pore pressure was a constant 1 MPa. The results further showed that the permeability of HSULPC had a hysteretic dependence on the effective confining pressure. We found that the hydraulic conductivity of HSULPC is comparable to or less than that of intact Toki granite obtained from Gifu Prefecture in central Japan. It was also considered that the hydraulic conductivity of HSULPC stabilized at around 10^-13 m/s after being buried and stressed. The high density and impermeability of HSULPC would enable it to effectively confine ¹⁴C radionuclides found in TRU waste.

It is known that icing on the blade in jet engine causes serious accidents in operation of airplane. Thus, anti-icing technology has been developed extensively. It is necessary to clarify the icing process and its deposition mechanism as providing the basic knowledge to develop the advanced technology. This study focuses on the deposition process of a single water droplet and the adhesion strength on aluminum alloy (A2017) substrate at low temperatures (-20°C and -10°C). In addition, influences of surface roughness and thermal conductivity of substrate on the deposition process are investigated. As a result, it was shown that the deposit area and shape of the frozen droplet is strongly affected by impact energy. The deposit area of the frozen droplet was, moreover, decreased with increasing the surface roughness and thermal conductivity. Observation of the deposition process by a high-speed camera revealed that the shape of the frozen droplet depends on wave motion which is generated on a surface of the deposited water droplet. The deposition and frozen process of a water droplet was also discussed based on these results and a simple theoretical model. All the results indicated that the deposition and frozen process are significantly affected by heat conduction between a water droplet and a substrate. From the results of an adhesion strength test, the adhesion strength of the frozen droplet was estimated to be approximately 0.1 MPa regardless of substrate temperature.