We have developed a technology to inspect rail surface defects and corrosion on the rail base using image data captured by a line sensor camera and a deep learning model. The inspection model for rail surface defects can identify four types of defects shelling, head check, gauge corner cracks, and corrugation from images taken from the top of the rail. The inspection model for rail base corrosion can determine the presence of corrosion on the rail base from images taken at an angle of the rail's side. Both inspection models use a dataset created with soft labels checked by multiple people for the same image, allowing the model's predictions to reflect the variability in human judgment. This approach has enabled us to establish a high correlation between the model's prediction confidence and human judgment confidence.

 Maintenance of railway track facilities is carried out by track engineers who go to the site to measure and to visually inspect the track and facilities, which places a heavy burden on the track engineers. Therefore, there is a need to streamline this work using measurement and image processing technologies, which have been rapidly evolving in recent years. In this study, images of the rail head and inside of the gauge were acquired by a profile camera (3D camera) installed on a cart. We proposed an algorithm to measure the joint gap and the gap of the filling layer around the circular stopper of the slab track from the images. The proposed algorithm was applied to the images acquired while running a cart on a commercial line and a maintenance base line, and the results were compared with manual measurements. The results showed that the proposed algorithm can measure the joint gap with an accuracy of around 0.5 mm and the gap of the filling layer around the circular stopper with an accuracy of around 1 mm.

 In a society with a declining working-age population and an aging workforce, there is a need for research and development of labor-saving measures for railway track maintenance work using digital tools. Against these circumstances, we proposed a smartphone application software that supports track conditions inspection during train patrols and its application to a low-cost track condition management. In this study, we applied SfM-MVS technology to train forward view images captured by a smartphone application software, and reconstruct the 3D-spatial model around the railway tracks. As a result, we found that the 3D-spatial models reconstructed from train forward view images captured by a smartphone has the potential to be used to an alternative or complementary means for labor-intensive work, such as maintenance of railway track alignment, construction surveying of railway facilities around the tracks.

 We have developed the image analysis engine for detecting obstructions within the structure gauge and environmental change along the railway track. On the other hand, there is a strong demand for labor-saving and efficient checking of the structure gauge when constructing a new Shinkansen. In order to confirm the structure gauge of the Shinkansen in dark areas, we acquired multiple images around the track at stations and tunnel sections of the Shinkansen under construction to comfirm the accuracy of the engine's detection of obstructions to the structure gauge. As a result, we confirmed that the engine can detect obstacles within an average error of about 10 mm when the image is taken at a speed of 30 km/h or less by performing a sharpening process on the images acquired using multiple near-infrared light projectors and then analyzing them.

 The maintenance of railway tunnels requires mandatory periodic inspections. Inspectors determine the soundness through visual inspection and sound inspections, to ensure proper maintenance. With a predicted decline in the labor force, there is a growing need for the digitization and automation of tunnel inspections. In this study, we developed two technologies to address this need. The first is the "Automatic Tunnel Condition Assessment System," and the second is the "Critical Area Projection System." This paper describes the overview and verification results of each system.

 Maintenance of earthworks that support the infrastructure of railroad spaces ranging from flat land to sloping terrain, hills and mountains is essential for a safe and sustainable railroad system. In this study, 3D point cloud data acquired by railroad MMS and aerial LP are deployed in GIS with the aim of contributing to labor saving in maintenance and systemization of disaster risk management. Then, we verify (1) the generation of 3D vector data showing the characteristics of earthworks and their arrangement relationships, (2) the classification of 3D point clouds obtained by irradiating laser beams onto objects in contact with the ground surface, (3) the synthesis of point cloud orthoimages showing reflection intensity and microtopographic representation images and (4) the integration of railway MMS point clouds that cut out the cross section of the structure gauge and aerial LP point clouds that cover the vast space along the railroad line. Through these trials, we visualize the three-dimensional structural characteristics of earthworks. This paper proposes two methods by using these spatial data: one method links existing records in real space to a digital twin, and the other method assumes that the edges of the structure are immovable and captures relative changes over time.

 Tokyo Metro conducted a study to improve the efficiency of hammering tests to prevent concrete spalling in subway tunnels. The study focused on evaluating the accuracy of AI-based defect detection using different types of tunnel images. Two types of images—high-saturation color images and grayscale images—were used to create defect distribution maps, which served as ground truth data for training AI models. Accuracy tests were performed using various combinations, including applying a model trained on high-saturation color images to analyze grayscale images. The results showed that detection accuracy tended to be higher when the image types used for training and evaluation were the same. Furthermore, the study confirmed demonstrated that even grayscale images, which lack color information and generally have lower resolution, can still be used effectively for defect detection with sufficient training. These findings indicate that further research on grayscale image-based AI detection is worthwhile.

 Faced with aging and population decline, streamlining labor-intensive visual inspections is essential. This study explored the use of AI for extracting deformations in civil structures within Shinkansen track areas and developed equipment for continuous image capturing. Results showed that AI models trained on images taken under specific conditions could extract deformations with reasonable accuracy. Additionally, the newly developed equipment successfully captured images continuously and func-tioned effectively during tests on operational Shinkansen lines, enabling efficient data acquisition.

 In the current continuous welded rail management, axial force of rail is managed by using the temperature equivalent of the axial force of rail added by rail creeping and track maintenance work, and by appropriately updating the change in neutral temperature after resetting. However, even under such management, track buckling has occured, so it is possible that axial force of rail exceeding expectations is generated due to the temporal change in neutral temperature after resetting. This study analyzed the changes in rail creeping amount focusing on the track structures and measured the axial force of rail and the rail temperature over a long term from the time of installation of welded rail to understand the changes in neutral temperature characteristics. As a result, the authors confirmed the quantitatively impact of items considered to be factors of rail creeping and the possibility of large axial force of rail occurring due to the temporal change in neutral temperature after resetting is small.

 Preventing track buckling is important for safety and reliability of transportation. For that reason, inspections are carried out to prevent track buckling. If the inspection result is poor, we will implement restressing of continuous welded rail. However, in recent years, track buckling have not been eradicated within JR East's jurisdiction. Therefore, in this study, we focused on alignment as a supplementary in-spection. We analyzed how the precursors of track buckling affect the alignment. As a result of the study, in the process of track buckling, alignment is characterized by increasing with the increase in rail tempera-ture. Additionally, based on the data from the rail axial force sensors installed at the locations analyzed for alignment displacement, we report on the relationship between alignment displacement and rail axial force.

 Railway turnouts involve multiple departments, including stations, rolling stock, tracks, and signals, while also having mechanically movable parts, necessitating periodic maintenance by each department. Therefore, we are aiming to optimize inspection cycles and prevent turnout failures in advance, by transitioning from conventional periodic maintenance to condition-based maintenance (CBM). Implementing CBM requires sensors for condition monitoring; however, NS-type electric point machines lack such mechanisms. In this study, we aimed to develop long-term installable sensors for turnouts equipped with NS-type electric point machines in operating railway lines, ensure stable data acquisition, and establish a system change for turnout inspections. Additionally, to obtain sensor values during turnout failures, we conducted verification tests under controlled conditions, simulating various adjustment defects and failure scenarios.

 To construct a sustainable railroad maintenance system, one of the important issues is the sharing and passing on of equipment maintenance knowledge (tacit knowledge) associated with individual engineers. In this study, we attempted to formalize the tacit knowledge of engineers involved in periodic inspections of electric point machines based on questionnaire surveys of engineers. As a result, when inspecting lock, it was found that while many engineers adjust with a 3:7 ratio, some engineers adjust with a 4:6 ratio, and that engineers who adjust with a 3:7 ratio tend to leave a misalignment amount, considering temperature changes, irregularities, and rattles. The percentage of engineers who leave a misalignment amount showed no correlation with age, suggesting that tacit knowledge may not be sufficiently handed down and shared through field experience and OJT alone.

 Labor-saving in equipment maintenance and management is an urgent issue for regional railways. Rail bonds, which ensure electrical connections between rails, are installed in large numbers and require frequent inspections. Therefore, rationalizing their management based on actual conditions is essential for improving efficiency.However, due to the ledger management unit and missing data, it has been difficult to accurately assess anomaly frequencies using the general exponential hazard model. In this study, we propose a method for constructing an integrated dataset that combines inspection and ledger data from both the electrical system, which manages rail bonds, and the track maintenance system. By applying this approach to 2, 092 rail bonds on actual railway lines, we empirically demonstrate that cross-system data integration significantly improves the accuracy of anomaly frequency analysis, even when using the same exponential hazard model.

 In Japan, the general practice for repainting steel railway bridges is to remove deteriorated coatings and repaint while leaving the active coating. For this reason, in steel railway bridges that have been in service for a long time, the coating thickness increases, and internal stress increases. In addition, the coating hardens and becomes brittle over time, which may cause cracks and peeling. However, there are few cases where the adhesion and hardness of coatings have been compared over multiple periods. Therefore, we conducted appearance surveys, coating history surveys, and adhesion evaluation tests on actual bridges over two periods on steel railway bridges that were in service in the 1960s and have been repainted frequently, and measured the indentation hardness of each coating layer in laboratory tests. As a result, although slight coating cracks were confirmed, most of the coatings had sufficient adhesion, and no peeling of the coating was confirmed over a large area. On the other hand, we found that the adhesion of the coatings was declining, and the red lead rust prevention paint in the bottom layer was hardening.

 Several movable bearings in the Tokaido Shinkansen steel bridges do not follow the statical expansion of main girders due to temperature changes. It may cause stress concentrations at the bearing and results in cracking at end stiffeners to lower flange connections. However, there are difficult cases to find deterioration of movable bearings only by close visual inspections. Therefore, it was investigated to identify the bearing parts with deteriorations that static measurements during three days on lower flange stress in front of sole plates, movable bearing movement, and member temperature. The relationships between stress and temperature changes could evaluate its soundness. In this paper, the investigation method was verified by the measurement of actual two bridges, which have deteriorated movable bearings, before and after the bearing re-installation. As a result, the method could be used to identify deterioration of movable bearings.

 To understand the influence of nonlinear structural behavior on seismic running safety, we performed numerous nonlinear dynamic analyses on structures with typical vibration characteristics and analyzed the RSI values, which serve as an index for evaluating seismic running safety. As a result, we found that structural nonlinearization reduced the influence of acceleration, had no effect on the influence of angular rotation, and led to a decrease in the final RSI value, which integrates both effects. Additionally, we examined the influence of differences in yield seismic intensity between adjacent structures and found that differences in yield seismic intensity increased angular rotation. Differences in yield seismic intensity between adjacent structures should be minimize

 The evaluation of track buckling stability typically focuses on static buckling under thermal loads without considering train loads. However, many railway operators still experience significant track misalignment incidents. Considering these circumstances, there is a growing need for more precise evaluation methods for the buckling stability of continuous welded rail. To conduct such an evaluation, it is essential to understand the actual static rail behavior under thermal loads and the characteristics of lateral resistance of the ballast under dynamic influences during train passage, both of which have not been fully elucidated. Therefore, this study conducted long-term measurements of rail displacement in a continuous welded rail in sharp curve section to grasp the actual static behavior of the rails. Additionally, to understand the actual dynamic influences, field measurements were conducted focusing on track uplift and vibration. Specifically, the uplift of unsupported sleepers and the vibration acceleration at rail irregularities in the long rail sections were measured on-site.

 Corrosion is one factor in rail fracture. Due to corrosion, there is a noticeable loss in volume at the side of the rail bottom where the rails contact the components of rail fastening systems, particularly at the spot where leaked water drops onto the rails in a mountain tunnel. A high-accuracy prediction of the fatigue strength of corroded rails is required because the fatigue strength decreases according to the decrease in rail volume. This study attempts to evaluate the fatigue strength of rails, focusing on actual stress and a stress gradient derived from a calculation using the finite elements analysis model, which reproduces the accurate surface shape of the whole rail bottom using a three-dimensional measurement device. The result reveals that the rail strength prediction accuracy of the proposed method against the fatigue test is 0.7 to 2 times at the cycle number when a rail breaks due to a fatigue crack originating from the surface of the rail bottom.

 In current rail defect detection, ultrasonic testing from the rail head surface has the problem of the side of the rail bottom becoming a shadow zone. Therefore, as a method for detecting defects on the side of the rail bottom, we proposed a new defect detection system focusing on changes in magnetic flux leakage caused by the presence or absence of defects using a non-contact sensor placed above the side of the rail bottom and conducted a basic study through indoor tests. A coil to generate a magnetic field was installed on the rail head surface, and a magnetic sensor to measure the magnetic field was placed above the side of the rail bottom. The output of the magnetic sensor around the artificial defects created on the side of the rail bottom was assessed. As a result, it was suggested that the presence or absence of defects and the depth of the defects could be determined by changes in the magnetic sensor output in the absence of rail fastening systems. Furthermore, it was observed that even when the rail was secured with rail fastening systems, the presence of defects could be determined by changes in the magnetic sensor output.

 In order to prevent the progression to the squat, it is effective to remove the initial micro cracks. We have been considering the introduction of eddy current testing to detect micro cracks on the rail surface. In this study, we verified the accuracy of detecting micro cracks at high speeds (approximately 40 km/h) both indoors and on-site in order to perform on-board inspections with an eddy current testing device installed on a ultrasonic rail inspection car. As a result, we confirmed that the detection accuracy was sufficient. It became possible to exclude unnecessary responses from the data obtained by eddy current testing on a ultrasonic rail inspection car and extract only the responses of micro cracks. In addition, we analyzed the occurrence of micro cracks in commercial lines and clarified their occurrence tendency. We also developed an eddy current testing mechanism that expanded the detection area on the rail surface from the conventional width of 5 mm to 20 mm, making it possible to measure an area equivalent to that of ultrasonic testing.

 To accurately evaluate the structural performance and train mobility of railway bridges, it is essential to properly understand the dynamic response of the entire system and its components, including stiffness and weight. In railway bridges, non-structural members, such as track components, significantly affect the stiffness of the pri-mary structure, such as the main girder. In this study, two Shinkansen composite girder bridges and one Shinkansen SRC girder bridge were subjected to image measurement using a high-resolution camera to evaluate the overall stiffness of the composite girder. The deflection and natural frequencies were measured as trains passed by. The stiffness was calculated based on beam theory, incorporating non-structural members and cross-sectional variations along the railroad track direction, and these calculations were compared with the image measurement results. The analysis results indicate that the effect of non-structural members dominates the actual stiffness of the bridge compared to the influence of cross-sectional changes in the main structure along the track direction. Fur-thermore, when the contribution of non-structural members is considered, the stiffness values calculated by beam theory agree with the measured values.

 To accurately evaluate the dynamic response of railway bridges during train passage, it is necessary to consider the behavior of trains crossing the bridge. This study aims to establish an evaluation method for the dynamic response of bridges that accounts for dynamic interaction through numerical analysis and theoretical investigation. First, dynamic interaction analysis revealed that rail camber has a minimal effect on the dynamic response of typical concrete and composite railway bridges while reducing the acceleration of train bodies. Furthermore, for a 50-meter composite girder, it was demonstrated that increasing the damping coefficient of the train’s air springs reduces the impact factor during resonance. Additionally, a simplified evaluation method was developed using an equivalent damping coefficient corresponding to the energy dissipated by the train on the bridge. It was confirmed that this method can reproduce both the results of dynamic interaction analysis and measured dynamic responses.

 Many of Japanese bridges are facing deterioration issues due to aging, and there is an urgent need to develop efficient health monitoring methods. However, monitoring methods that utilizes attached sensors to bridges require a lot of effort and cost for maintenance, making it difficult to apply to a large number of bridges. In this study, we focus on track elevation measurement values as a method of utilizing data obtained from the vehicle, and examine its applicability based on actual measurement data obtained from representative bridges. It was shown that it is possible to estimate bridge deflection by performing differential processing on track elevation measurements taken at the front and rear axles when passing over the bridge. Additionally, we conducted a comparative study of the trends in on-board measurement data from bridges that had previously been damaged and had undergone repairs, and it was shown that it may be possible to estimate from the on-board measurement data whether flapping had actually occurred in the bridge supports.

 Some drive-by structural measurement methods based on track geometry, which have been developed in recent years, clarifies that a difference of two loaded track geometries is an important physical quantity. Because its calculation error is directly related to the accuracy of the drive-by measurement method, we theoretically derived an error contained in it when there is a misalignment in its calculation process. In addition. based on the derived theory, we proposed a new calculation method using modification of a sampling interval in its previous calculation process. It was found that the proposed method reduced the error of the difference of two loaded track geometries through a comparison of the difference of two loaded track geometries obtained by an actual track inspection vehicle using the proposed method with FEM analysis. Finally, the error derived by the theory is also valid for the actual track displacement and becomes larger as gradient of track geometry and misalignment increase or a wavelength of track geometry decreases.

 The improved glued-insulated rail joint is designed to prevent adhesion failure between the rail and the fishplate by pre-inserting a Teflon sheet within the adhesive layer, thereby creating an artificial interface within the adhesive. However, in recent years, there have been repeated occurrences of fishplate fractures originating near the end of the Teflon sheet. Therefore, the stress distribution occurring in the fishplate during train operation was measured, and existing improved glued-insulated rail joints were retrieved to assess the extent of fishplate corrosion and the residual adhesive strength between the rail and the fishplate. Based on the results obtained, a new glued-insulated rail joint was manufactured using an improved adhesive and an enlarged Teflon sheet. To verify its performance, tensile tests and fatigue tests using a rail bending fatigue testing machine were conducted, and it was confirmed that there were no practical issues.

 At present, the design loads for wheel loads and lateral forces of rail fastening systems of general track are used for verification in accordance with design standards for the components of turnouts. However, no design loads have been established to account for such turnout structures, despite the fact that impact wheel loads and lateral forces are generated at the point and crossing transfers. In order to establish a method for calculating the design loads that takes into account the turnout structure, an estimation equation for the wheel load and lateral force generated in the turnout was investigated. Specifically, the wheel load and lateral force generated in the turnout were calculated using simulation of a vehicle travelling on the turnout with multi-body dynamics. Multiple regression analysis was carried out using the maximum wheel load generated at the point as the objective variable and the number of turnouts, vehicle weight, and running speed as explanatory variables. Furthermore, the wheel load generated at the point was calculated using the proposed estimation equation, and the results were in agreement with the real measured values.

 The authors have studied methods of track management using small information terminals to estimate track irregularity in longitudinal level from vertical vibration acceleration. A local railway operator uses an information terminal to frequently measure vehicle vibration and manages track based on vertical vibration acceleration. However, there have been no reports of methods of estimating alignment irregularity using information terminals. The number of locations where alignment irregularity exceed the maintenance target value is less than those of irregularity in longitudinal level, but vehicle vibration is prominent in curved and transition curved sections. Therefore, the authors decided to grasp the track condition, focusing on curved sections, and to consider methods of managing alignment irregularity. The results of their analysis, a track management method based on yaw angular velocity, and the effectiveness of the track management method are reported.

 This study aims to clarify the strength of rail fastening systems and the fracture behavior of track components in Shinkansen slab tracks under significant lateral forces from vehicles during earthquakes. For this purpose, static horizontal loading tests were conducted on the heads and bottoms of fastened rails using a full-scale Shinkansen slab track as a test specimen. We performed a loading test using a single fastening assembly with a short-length rail to evaluate the fundamental mechanical properties of the rail fastening system tested. We conduct the loading test with eight fastening assemblies on a single track slab to evaluate the properties under an actual usage environment.

 Analyzed the correlation between the girder water level gauges installed by river administrators and their own girder water level gauges. The regression equation developed from the analysis allowed us to estimate the river level at the railroad bridge. Estimated river water level at the railroad bridge point based on the calculated flow rate by the RRI model and actual measurements from the water level gauge installed by the river administrator. Both methods were able to estimate river levels at the railroad bridge with good accuracy.

 Most of the railway damage caused by natural disasters was caused by sediment inflow from mountain streams. In this study, a risk analysis was conducted based on geomorphological analysis using a digital elevation model (DEM). The focus was on sediment inflow caused by mountain streams. Field observation was conducted on the railway lines affected by the torrential rain disaster in Yamagata Prefecture in July 2024. Based on the observations and analysis results, a method is proposed to identify the indicators that contribute most significantly to the occurrence of sediment inflow by means of a Type II quantification method.

 In recent years, many railway facilities have suffered damage due to the increasing severity and frequency of natural disasters. Local railway operators, in particular, face a significant shortage of engineers, making it difficult to respond quickly. As a result, restoration takes time, and the functioning of the transport infrastructure that supports the lives of local people can be disrupted for extended periods. The early restoration of damaged railway facilities requires the involvement of engineers with specialized knowledge. In cooperation with the MLIT, JRTT established RAIL-FORCE in 2023 to assist in surveying damaged railway facilities and planning their restoration. Additionally, as part of local railway support, we provide technical assistance to railway operators even outside of disaster situations.

 Third-sector railways of the parallel conventional line serve as freight train routes, but their financial situation is poor and disaster prevention measures are insufficient. Therefore, they will suffer great damage in the event of a large-scale disaster, and the suspension of freight train services could have a major impact on the whole of Japan. In this study, we used an input-output table to analyze the economic damage caused by the suspension of freight train services when third-sector railways of the parallel conventional line sector railways on parallel conventional lines are affected by a disaster. The analysis showed that the damage would be particularly severe on three lines connecting Hokkaido and the Tokyo metropolitan area: the Dounan Isaribi Railway, the Aoimori Railway, and the IGR Iwate Galaxy Railway. The economic damage calculated in this study is useful for cost-benefit analysis of disaster prevention projects, but there is still room for further research on cost-benefit analysis of disaster prevention projects for railways, so at this point in time, it is expected to serve as a reference for the government's prioritization of disaster prevention measures.

 High-manganese steel is used in railway crossings due to its excellent wear resistance. While previous studies have focused on the wear resistance of manganese steel, recent observations have identified rolling contact fatigue (RCF) damage in manganese crossings. To assess wear and RCF resistance, a comparative analysis was conducted with pearlitic steel. The results indicated that the initial wear rates of both materials were similar, but over time, high-manganese steel exhibited superior wear resistance. Similarly, initial hardness values showed no significant difference; however, high-manganese steel underwent more pronounced work hardening during the early stages of rolling contact. EBSD measurements revealed that both materials exhibited reduced IQ values in the surface layer, indicating the formation of a strain accumulation layer, with surface cracks being observed. These findings suggest that although high-manganese steel has lower initial hardness, its work hardening effect under rolling contact helps suppress wear progression. However, as the number of rolling cycles increases, strain accumulation occurs, which is considered a contributing factor to the initiation of rolling contact fatigue damage.

 We investigate the influence of material parameters and shape dimension of grid-type sleeper (GSLP) on settlement and vibration of ballasted railway track, with a FE-based simulation. The simulation method is a coupling method of track vibration analysis and ballast settlement anaysis. The present simulation for a suspended joint with a GSLP needs to consider rail surface profile around a railjoint. The use of soft under sleeper pads (soft USP) to GSLP leads to level the settlement and to suppress progression of permanent deformation of ballast layer.