Nagai Lab, Hokkaido University

Estimation on corrosion distribution from surface crack using RBSM integrated with machine learning

In reinforced concrete (RC) structures, the rebar corrosion estimation is crucial to evaluate the residual performance and perform maintenance. A simulation system for estimating the distribution of corrosion along the rebar based on surface crack widths is proposed.

The system integrates the rigid body spring model (RBSM) with machine learning methods. The inputs are surface crack widths from surrounding positions, and the desired output is the distribution of corrosion-induced expansion. A large dataset of training samples for machine learning is generated by running RBSM simulations and after training with the dataset, the network model correlates the input and output, allowing it to estimate an expansion distribution from given cracking data. With estimated corrosion result, surface cracking is reproduced with RBSM, and an optimization process is introduced by making use of the cracking error.

The applicability of the proposed system is verified by model with single rebar and with multiple rebars. During the estimation process, the internal cracking propagation and stress condition can also be simulated, which serve as important information for evaluation on structural performance.

Coupled Mechanical–Transport Modeling for Investigating Moisture Transport in Reinforced Concrete

Reinforced concrete is the most commonly used composite material in civil engineering, and its durability plays a critical role in determining the service life and maintenance costs of structures. Moisture ingress and the penetration of aggressive ions such as Cl^-, SO₄^2-, and CO₃^2- are key factors driving steel corrosion, freeze–thaw damage, and chemical deterioration. Although the importance of moisture transport is well recognized, its multiscale mechanisms remain insufficiently understood, especially under the influence of cracking, reinforcement, and pore structure variations in advanced concretes such as UHPC. Existing models often rely on simplified assumptions and struggle to reproduce realistic transport behavior under complex boundary and material conditions. Numerical simulation provides a promising alternative to direct measurement, offering a means to predict internal moisture transport and support durability assessments.

This study aims to develop a coupled simulation framework based on a self-developed 3D RBSM Conduit model, enabling high-resolution prediction of moisture distribution within reinforced concrete. By integrating mechanical damage, crack geometry, and interfacial transition zones (ITZs) into a multiphysics transport model, and validating it through experiments, the research provides a novel and practical tool for durability design. The results are expected to improve the understanding of moisture-induced degradation and contribute to more reliable service life predictions and structural optimization.

Mesoscale Simulation of Precast Joints and 3D–Printed Concrete Structures with Weak Interfaces Using a Discrete Analysis Model

Industrialized concrete systems–precast concrete structure and 3D-printed concretes (3DPC)–promise rapid construction and quality control, yet their structural performance is often governed by weak interfaces: mortar joints, steel–concrete bonds, layer–to–layer interfaces, and localized adhesive transition zones. These interfaces drive crack initiation, slip, and failure localization, especially in loop/headed-bar joints, grouped studs, and reinforced 3DPC with anisotropic pore distribution. Conventional continuum approaches, such as FEM, struggle to predict crack path propagation, bond degradation. A mesoscale, mechanism-explicit framework is therefore needed to reproduce observed failure modes, capture reinforcement anchorage and interfacial fracture, and translate insights into rational detailing for industrialized construction.

This study develops a unified mesoscale simulation framework using 3D Rigid Body Spring Model (3D RBSM)–based discrete analysis model to resolve concrete, reinforcements, and multiple interface types with traction–separation laws in tension, shear, and compression. For precast slab joints, loop/headed–bar anchorage is modeled with explicit bar–concrete interfaces, enabling prediction of load–slip responses, crack paths, and stress transfer path; parametric studies quantify the effects of bar spacing a, overlap length H, transverse confinement design, and joint geometry. For reinforced 3DPC, anisotropic behavior and adhesive transition zones properties are embedded to simulate compressive strength anisotropy and bar pull–out behavior, linking crack path evolution to bond-stress distributions. The framework is calibrated and validated against targeted experiments and then used to deliver design–oriented maps of anchorage efficiency, crack-width control, and interface performance. The outcomes provide a physics-grounded basis for detailing precast joints and reinforced 3DPC, supporting performance–based design, reliability assessment, and optimization in industrialized concrete structures.

Understanding localized corrosion effects in RC half–joints through 3D RBSM modeling

RC half–joints were popular during infrastructure expansion period because they offer advantages in construction and structural analysis. As time flies, deterioration problems have been reported and come into consideration as they were taking part in the bridge collapse due to half–joint failure in the past two decades. Since an RC half–joint contains complex stress flow, and corrosion effects locally arise, the utilization of 3D discrete mesoscopic simulation is considered proper.

The frameworks of 3D–RBSM were employed to analyze the behaviors of corroded RC half-joints. By using RBSM simulation, localized phenomena such as strain increment during corrosion process was found to increase due to bond deterioration while the structure was in service condition. This event is quite difficult to physically measure on corroding rebars. Moreover, cracks and the corresponding strain distribution could be precisely captured.

The complex stress distribution allows several localized failures to happen and progress. The study also focused on the corrosion happening in several areas of an RC half–joint to determine criticalities, damage and preventing mechanisms. The results show that corrosion in different locations varies the structural behaviors. Corrosion arising in the areas containing the fundamental cracks are critical. Contrastingly, other locations are not very dangerous regarding some preventive mechanisms that allow the fundamental failure mode to happen. For practical use, the summarized simulation results can potentially support bridge visual inspections.

Bond investigation of corrosion reinforcement

Reinforced concrete structures especially those exposed to chloride environment are susceptible to reinforcement corrosion which significantly deteriorates bond between reinforcement and concrete thus reducing the serviceability of the structure. Understanding the impact of corrosion on bond is challenging, as corrosion causes concrete damage and steel damage which complicates failure mechanism.

An innovative experimental scheme has been developed in which the reinforcement is corroded first and then placed inside concrete thus eliminating concrete damage. The re-casted specimens have a trapezoidal window which allows continuous observation and recording of displacements using high resolution camera while standard axial tensile loading is applied. Digital Image Correlation technique, an innovative, non-contact method of tracking the pixels and building up strain maps is used for accurate and reliable examination of crack initiation and propagation using sequential images obtained from the recorded video.

From the DIC analysis we can analyze strain patterns, cracking phenomenon and identify zones of mechanical and frictional failure. One major advantage of using DIC is that local interface opening and sliding information can be measured and how this relationship changes for corroded reinforcement can be studied with DIC. Using the displacement information extracted from DIC, effect of rib height and corrosion layer on stress transfer between reinforcement and concrete has been studied at a very local level.

A mesoscale discrete model for mechanical performance of concrete damaged by ASR/DEF

Durability issues arising from phenomena such as the alkali-silica reaction (ASR) and delayed ettringite formation (DEF) may threaten the serviceability and safety of concrete structures due to premature cracking. In many cases, these types of deterioration can even occur simultaneously and promote the development of others. Although they both show surface map cracking, the nature of the damage caused by each is different and the damage process varies also. To evaluate or predict the mechanical property degradation precisely, it is important to understand internal cracking behavior under the combined effects and numerical analysis is required for this since the information is difficult to obtain through experimentation.

In our research, a three-dimensional rigid body spring model (3D RBSM) is used to simulate the cracking and mechanical property degradation of concrete under coupled ASR and DEF. Concrete expansions induced by ASR and DEF are introduced in the model by applying initial strain on the springs of interfacial transition zone (ITZ) and mortar elements as damage history respectively. The effect of the percentage of reactive aggregate is studied in the cases of ASR damage, while different intensified expansion areas are considered in the case of DEF. Surface cracking and internal crack numbers were analyzed to correlate the mechanical reduction.

Meso-scale modeling of non-uniformly corroded reinforced concrete using 3D discrete analysis

Corrosion is one of the most serious issue in maintaining reinforced concrete structure. Once corrosion occurs, it can lead to several structural issues such as concrete covering crack, loss of reinforcement and bond deterioration. The main problem is that the residual capacity of the structure remains unknown. Therefore, the objective of this research is to develop the numerical model for predicting the residual capacity of the structure.

In simulating corrosion problems, a discrete analysis model is an appropriate model because local corrosion cracking can be directly represented by the displacement of elements and the rebar geometry can be modeled in the actual manner.

The function for replicating the corrosion behavior were developed and implement into RBSM. In this study, an innovative pull-out test is used to study changes in bonding in reinforced concrete after rebar corrosion, then an equivalent equation for modification of spring properties is developed to simulate the observed bond behavior. The expansive strain method is used to simulate the expansion behavior of the corrode reinforcement. The properties of steel elements is modified every 5 mm along entire the bar length to represent the non-uniformity of corrosion pattern of rebar. Finally, all developed functions were combined and implement into RBSM model, and use to simulated the behavior of corrode beam under monotonic loading for verification the applicability of the numerical system.

Evaluation of Crack-bridging Performance of Steel Fibers under Flexural Fatigue for SFRC Structural Beams

Among the several modes of failure that occur in reinforced concrete (RC) structural members such as bridge slabs and girders, the fatigue failure is associated with progressive, permanent, and localized internal changes in the material caused by repeated stresses. Consequently, a major concern in the fatigue design and safety of an RC structure is to ensure that it can sustain millions of repeated cyclic loadings over a specified design lifetime. Undesirable degradation of material strength leading to structural failure, and particularly the rupture failure of ordinary steel rebars, must be avoided. The incorporation of steel fibers in concrete enhances its flexural fatigue performance, providing higher strength and longer fatigue life. The additive toughness and fracture energy brought by the inclusion of fibers, which bridge cracks within the tensile stress zone, result in reduced rebar stress compared to normal concrete (NC).

Under cyclic loading, the crack-bridging ability of steel fibers degrades over the fatigue life of a structure as bond strength is lost. This leads to significant increases in rebar strain levels and reduces fatigue life. The ability to evaluate crack-bridging strength over the fatigue life is crucial to understanding the fatigue performance and estimating the fatigue life of SFRC structural beams. The characteristics of crack-bridging strength associated with their degradation and evolution mechanisms were investigated and driven based on back analysis.

A large-scale computational model for deteriorated reinforced concrete structure
(3D RBSM and Truss Network)

Aging of infrastructure has become a social problem in Japan because the year for maintenance will be reached soon since the post-war reconstruction in the past half century. In addition, it was predicted that an enormous earthquake will occur in the future and the response prediction of the existing deteriorated structures due to this enormous earthquake is necessary.

Numerical simulation can be a beneficial tool for this purpose. In our research group, a simulation system of 3D Rigid Body Spring Model (RBSM) has been developed. As a continuation of the current active research on 3D RBSM, the future research will be focused on the application to the real reinforced concrete structures which are the deteriorted RC structures at site. In reality, RC structures are subjected no only to mechanical loading but also to environmental conditions. To broaden the applicability of 3D RBSM, "Truss-network" is being developed in our research group for modeling substance movement due to the environmental exposure. A combination of both simulations could be a powerful tool for assessing the deteriorated RC structures at site and predicting the life span of the deteriorated structures.

Conversion of Concrete to Sludge around Deformed Bars due to Liquid Water and Repeated Loads and Its Effect on Pull-out Failure

In the actual environments, civil engineering structures are frequently subjected to various repeated loading and at the same time are subjected to liquid water from the external surface (hereafter referred to as external liquid water), such as rainwater or snow water. In Japan, problems of fatigue in the RC deck slabs of road bridges have become serious since the 1960s, especially in the Tohoku Region, and in recent years a granulation phenomenon has been found in which the cementitious component of concrete is washed out and separated from the aggregates. Against this background research has been carried out on the fatigue properties of reinforced concrete (RC) deck slabs focusing on the effect of external liquid water.

Research into the effect of external liquid water on the fatigue properties at the material level in the past was carried out, in which the fatigue properties in water have been investigated, and it has been reported that the fatigue life of concrete in water is greatly reduced compared with the properties in air. On the other hand, no research could be found investigating the effect of external liquid water on the bond fatigue properties of reinforcement. In our research group, In this research, fatigue pull-out tests of deformed bars have been carried out using the presence or absence of external liquid water, matrix type and strength, loading method, and the presence or absence of a mechanical anchorage as experimental parameters, in order to determine the effect of external liquid water on the bond fatigue properties of deformed bars in concrete. The characteristics of the pull-out failure mode of deformed bars associated with formation of sludge were investigated by measuring the load, the pull-out displacements, and the strain distribution of deformed bars.

Cross-ministerial Strategic Innovation Promotion Program (SIP)

The Cross-ministerial Strategic Innovation Promotion Program (SIP) is a Japanese project led by the Cabinet Office's Council for Science, Technology and Innovation. The project was founded to promote scientific and technical innovation through management that extends beyond the boundaries of existing fields and government departments, ministries and agencies. The project will tackle ten major global issues that are also social issues of truly critical importance to the Japanese people and issues that can help to revitalize the Japanese economy. Each issue will have a Program Director (PD) who will exercise firm leadership and play a central role in promoting collaboration among industry, academia, and government, as well as pursue comprehensive research and development that takes into consideration the entire process from basic research through practical application and commercial development. SIP will provide a powerful impetus to the scientific and technical innovation that is the driving force behind economic growth and dramatic change in society.

Scientific and technical innovation are essential for revitalizing the Japanese economy and achieving sustainable economic growth. Under the leadership of the Prime Minister and the ministers in charge of science and technology policy, the Council for Science, Technology and Innovation plans and coordinates policy for basic and comprehensive science, technology, and innovation based on a broad perspective of science and technology in Japan. The Cross-ministerial Strategic Innovation Promotion Program (SIP) was established as part of this effort with the aim of strengthening the management functions (headquarter's functions) of the Council for Science, Technology and Innovation. It is designed to be one of the three main pillars of policy along with the Strategic Formulation of Overall Governmental Science and Technology Budget program and the Innovative Research and Development Promotion Program (ImPACT) program.

Our research group participates of international activities of "Infrastructure maintenance, renewal, and management". The following activites have been conducted.
(1) Development of educational programs of engineers for infrastructure maintenance in Vietnam
(2) SIP Special Session held in EASEC-14
(3) The 3rd SIP International Seminar in Phnom Penh, Cambodia

Science and Technology Research Partnership for Sustainable Development (SATREPS)

With Myanmar and its cities under large-scale development, the risk of disasters increases due to expansion of the urban population and climate change. This project monitors changes in the ground, terrain, and urban environment associated with the development process, and develops a system for assessing vulnerabilities to potential disasters in Myanmar. The project aims to identify disaster risks in advance to contribute to the formulation of regional development planning and disaster prevention countermeasures as well as to support the strengthening of the Myanmar government's disaster response capabilities.

Our research group participates of international activities as "Infrastructure Group" for SATREPS project. The primary activities can be clasified into:
(1) Proposal of simple monitoring systems
(2) Deterioration mechanism and countermeasures for the bolt-ruptured bridge
(3) Evaluation of road roughness in Yangon
(4) Development of @the existing bridge database with GPS
(5) Verification of infrastructure management system

A Study on Retrofitted Damaged Reinforced Concrete Corbel Using Various Methods

Reinforced Concrete (RC) corbel is low span cantilever structure used basically to support direct vertical load, and horizontal load (lower than vertical load) due to thermal shrinkage. It is recommended that outer edge of bearing pad should not be projected beyond the straight portion of main tension bar of corbel and depth of corbel at corresponding position should never be less than half of its depth at the base (BS, 1997 & ACI, 2008). Mainly these provisions are provided to ensure desired failure modes in corbel avoiding premature failures. But to such tiny details, it might be common not to obey such provisions, which may lead to unsatisfactory structural performance. Design and construction with poor structural detail is more likely to be seen in underdeveloped countries with exception in developed one.

In a field study, several corbels were found to be failed at lower load than their corresponding design values due to faulty design of bearing pads. Bearing pad in such typical case was found to be extended to the edge of corbel resulting premature outward splitting parallel to the bent bar. Such case resulting premature local failure in corbel is targeted in this study and scenario is termed as 'Local Failure Criterion'. Practically, we can't simply demolish such corbels considering its structural importance and cost. Hence, it is necessary to propose appropriate retrofitting solution for such corbels, and equally important to identify parameters that governs the behavior of retrofitted corbels. In our research group, further retrofitting method which is eficient, simple, and easy is proposed throught simulations and experiments.

The Mechanical Performance of High-Performance Fiber-Reinforced Cementitious Composite Concrete under Multiaxial Stress Conditions

In Japan, concrete is a dominant structural material in the engineered construction. It is well known that concrete has more advantages than other materials. Concrete is strong in compression but weak in tension. Thus, cracks develop whenever the tensile stresses, caused by loads, excess the tensile strength of the concrete and the brittle failure mechanism tends to develop while concrete is subjected to the tensile stress. Generally, steel reinforcements are used to improve the tensile performance of the concrete. In recent years, fiber-reinforced cementitious composites (FRCC) has been developed as an alternative for the replacement of conventional steel reinforcements.

High-Peformance fiber-reinforced cementititous composite (HPFRCC) is a class of fiber-reinforced cementitious composites (FRCC) that exhibits multiple cracks and pseudo-strain-hardening behaviour under tension stress. One example of the HPFRCC is Engineered Cementitious Composite (ECC) that contains fine-graded materials and a moderate amount (2-3%) of short random polymeric fibers. Many experiments have been conducted to investigate the tensile behaviour of ECC and prove that the ductile failure mechanism can be achieved, due to the occurance of multiple cracks. In the other hand, the performance of the crack-shear transfer of ECC is low. As the result, the use of ECC is still limited in retrofitting. Therefore, some researches are still needed to improve the shear performance of ECC. Further research through experiment and numerical analysis is developed in our laboratory.

A Rational Design Method for Reinforced Concrete Beam-Column Joint

As many buildings experienced failures due to the previous earthquakes, the performance requirements of the building in the seismic code are more stringent than the requirements in the previous code. To satisfy the requirements, a large number of steel reinforcements must be placed in the concrete member. As the result, a reinforcement congestion, especially in beam column joints where reinforcements meet from many directions, is occurred. A reinforcement congestion can cause difficulties during the compaction of the concrete and furthermore it can result a poor quality of construction. In the other hand, the behaviour of reinforced concrete where the reinforcements are arranged multidimensional that cause the complex stress condition and crack propagation, has not been clarified well yet. So that, there is still a possibility to reduce the reinfocement congestion, especially in the beam column joints. Our laboratory conducts the research by numerical simulation at meso scale to propose a more rational design method in reducing reinforcement congestion based on the mechanical behavior..

Mesoscopic Simulation of Failure of Mortar and Concrete by RBSM

Two-dimensional and three-dimensional numerical simulations of failure of mortar and concrete are conducted using the Rigid Body Spring Model (RBSM). This analysis method is useful to simulate discrete behavior like concrete fracture. In the analysis, concrete consists of mortar and coarse aggregate. Since cracks initiate and propagate along boundaries between elements, mesh arrangement may affect fracture direction. To avoid formulation of cracks with unarbitrary direction, fine and random shape element is introduced using a Voronoi diagram.