First, we will introduce the background of this Forum, and then describe the architecture of a global manufacturing enterprise and the role of humans. Furthermore, uncertainty and complexity in the modeling of Human/Society will be discussed and the classification and modeling of human knowledge (scientific, empirical, tacit) will be presented. Finally, we will discuss the replacement of the human knowledge by the surrogate models using AI for building Digital Twins.

The brain always predicts the world. We see brain’s predictions as the world. Sensory observations are the only source of correction for predictions. Our perception is an estimate based on predictions and sensory observations. This estimation follows Bayes’ theorem. Prediction error is calculated as information surprise. The brain recognizes the world by minimizing prediction errors. This essential function of the brain and its mathematical model explains and predicts human perception, emotions and self-awareness. We discuss how to apply the brain model toward human-centered society.

Role of digital twin in manufacturing is discussed. The benefits of introducing digital twin appear when three aspects of digital twin are realized. They are, existence of physical and cyber spaces, Data flows connecting them and so-called digital thread that runs through all the initial design to production processes of manufacturing. Tokyo University of Science, under the financial support and technical collaboration with Hexagon M. I., opened Hexagon/TUS Digital Twin Laboratory in April 2020 after half a year preparation. The objectives and activities of the laboratory are presented in this manuscript. The essence of Digital Twin exists not in the software system but in the strong motivation and the real activities by the people having awareness of the same issues toward establishment of the efficient manufacturing process.

The future image of machining process control utilizing a digital twin is presented, and the roles of AI and humans are described. By entrusting a certain degree of adaptive control to the digital twin + AI, people can take charge of more processes than ever before, leading to higher productivity. In addition, production in multiple remote locations can be managed and supported by a single skilled operator. Discussions on how and what information should be communicated to humans and digital twin + AI according to their roles, including the development of remote monitoring methods, will become important in the future.

Design Engineering Systems Division has played a central role in planning and implementing a Japanese-German symposium series on IoT design, systems and applications, which were held in Germany, Japan and online, and offered a forum for lively discussion between Japanese and German researchers. Various topics were discussed in those symposiums, especially, interesting opinions were exchanged on the topic of what lies beyond Industry 4.0 and Society 5.0. This paper overviews the history of this symposium series and proposes an idea of real-world modeling beyond Industry 4.0 and Society 5.0. Using a case study which were jointly held by Okayama Prefectural University and Hannover University of Applied Sciences and Arts, this paper describes an application example of this real-world modeling based on the proposed idea.

This paper describes the solution approaches for horizontal value creation using IoT and digital twin data. First, the various lifecycle stages that a product goes through are described. A classification of digital twins is then carried out. On this basis, the horizontal and vertical data integration is described, which forms the basis for future value creation.

Machine learning offers powerful algorithms for various application domains and objectives. Therefore, this paper presents a systematic overview of applications in injection molding science and industry. Scopus and Web of Science were used to search for articles in which authors indicated they had used machine learning. 192 papers spanning nearly three decades were reviewed and analyzed, with 68 papers selected for this review. All papers were analyzed in terms of their objective, machine learning algorithms used, learning type, and data source. The most common objective was quality or defect prediction, with neural networks being used in most cases. Supervised learning was applied to the most common data source, machine and process data. Although there are already a large number of publications on this topic and this area has grown rapidly in recent years, there are only a few industrial applications in injection molding so far.

This article presents the cooperation between the Institute of Mechanism Theory, Machine Dynamics and Robotics (RWTH Aachen University) and the Mechanical Systems Design Laboratory (Tokyo Institute of Technology), whose fundamental idea is the exchange between students and scientific staff in the course of international research relations. Within this cooperation, three consecutive research projects, all funded by the DAAD PAJAKO program, haven been established. The first two projects focused on methods for the kinematics, dynamics, and stiffness optimization of Delta robots, whereas the third and ongoing project is dedicated to the dynamic balancing of Delta robots, to be able to use their full potential in practical applications.

The Low temperature differential Stirling engine for a craft workshop was improved. Fine adjustment during assembly is not necessary. Here, the explanation for attendees of a 5-hour-course craft workshop is shown by some pictures.

The New ☆ Energy Contest is a contest conducted by the Technology and Society Division as one of the activities of environmental and energy education, with the aim of raising awareness of environmental and energy problems. The 1st New Energy Contest was held at Tamagawa University on December 20 (Sat) and 21 (Sun), 2008 with the cooperation of the Technology and Society Division. In this manuscript, we will report on the birth of this contest.

The Alternative ☆ Energy Contest is a new contest on "energy utilization" mainly for university and technical college students, which started in 2008 as an event in which the technology and society departments of the JSME cooperate. The mark "☆" in the title of the contest indicates something that goes beyond the preconceived notions of alternative energy, and the essence of this contest that appeals for novel ideas appears there. From 2012, the year after the Fukushima nuclear accident caused by the Great East Japan Earthquake, a stationary event was held at the Nihon University College of Engineering located in Koriyama City, Fukushima Prefecture, as an event to watch over the reconstruction of Fukushima. In recognition of the efforts up to the 11th (September 7, 2018), we received the Japan Society of Mechanical Engineers Education Award under the title of "Promotion of Engineering Education on Sustainable Energy Utilization through the Alternative ☆ Energy Contest". The 12th meeting (October 19, 2019) had to be abandoned face-to-face due to the enormous flood damage caused by Typhoon No. 19. Since the application was accepted and the summary was printed, the summary was published as scheduled, and the poster was confirmed to have been announced. Despite the sudden change to remote screening based on poster data, the screening by the sponsoring companies and organizations responded promptly, and the award was given as usual. After the 13th (October 17, 2020), due to the influence of the spread of COVID-19 infection, we had to change to the WEB holding. We were able to reproduce a poster session using Zoom by making use of our experience of web lessons at the university. It was a memorable event to receive an application by a member of society for the first time in the contest during this period. The 15th meeting (October 15, 2022) is scheduled to be held on the website with the College of Engineering, Nihon University as the operating base. We are grateful for the opportunity to give an overview of this contest at this civic forum in this paper.

Friction stir incremental forming (FSIF) process was applied to join a metallic foam with a resin sheet for fabrication of porous metal–nonporous resin composite. In this process, a rotating rod-shaped tool was vertically pushed and horizontally fed against the sheet on the foam. The sheet was frictionally heated and incrementally deformed by the rotating tool, while the cellular matrix of the foam was not plastically deformed. The bottom of the sheet was mechanically interlocked to the pores of the foam by plastic flow of the frictionally heated sheet. In this study, (1) the joining characteristics of an open-cell type nickel foam and a polymethyl methacrylate (PMMA) sheet and (2) the compressive characteristics of the nickel foam/PMMA sheet specimen were introduced.

The advantage of laser peening is the possibility of fine execution management and the capability to introduce deep compressive residual stresses on the material surfaces. It is well known to be highly effective in inhibiting stress corrosion cracking and fatigue cracking on material surfaces. In addition, laser peening has an excellent effect on improving the fatigue strength of welds, which compensates for the disadvantage of high strength steels when welded. Laser peening has a high potential for enhancing material surface, but the high-power laser used requires clean room facilities, large equipment and severe operating conditions. Therefore, the application of the laser peening has been limited to the mitigation of high cycle fatigue of jet engine fan blades and stress corrosion cracking of nuclear reactor structures. If microchip lasers, which are small and easy to handle, could be used as a light source for laser peening, it would be possible to apply them not only to production processes in factories but also to existing structures such as bridges, airplanes, etc. to which conventional lasers have been difficult to apply for the above reasons. In this presentation, we report the results of the laser peening treatment of A7075 (aluminum alloy) and HT780 (high strength steel) with the laser pulse energy of less than 10 mJ. Even at such a low pulse energy, compressive residual stress was imparted and fatigue properties were improved.

This article summarized the recently published papers or books on the design for additive manufacturing (DfAM) from the standpoint of the numerical simulation or computer aided engineering (CAE), focusing on the laser powder bed fusion (LPBF) technique for metals. Chapter 2 described the state-of-the-art of the process simulation and topology optimization. Micro-CT image-based simulation is useful for the evaluation of real products, but for predictive simulation before manufacturing to be used in the design phase, probabilistic simulations are presented in Chapter 3. Finally, some recent reports on the Digital Twin are shown in Chapter 4.

There are various technical issues in aiming at the productization of metal additive manufacturing parts. One of them is the issue of quality assurance. In particularly, it is necessary to thoroughly consider the inspection method and assurance method for internal structures and complex shapes.

The X-ray CT equipment is one of the effective means to solve this problem. The X-ray CT equipment can confirm the internal structure and internal defects of the product in a non-destructive manner. In recent years, various developments of Xray CT have progressed, and their utilization methods are increasing. Various utilization methods are also being examined in metal additive manufacturing. In this talk, we will summarize and report on the utilization method of X-ray CT equipment and the current issues in metal additive manufacturing.

The "Workshop of Advancement of Steam Flow Measurement" started its activities in April 2017 within the Power and Energy Systems Division of the JSME and finished as planned in March 2022. The industrial sector has a steam demand of 1.12 million TJ/year, which is the second largest energy source after electricity, and steam accounts for about 80% of the heat energy use in the industrial sector and is widely used in large quantities in the main production processes: sterilization, drying, cleaning and distillation. Steam is thus a major heat source in the industrial sector and there is a strong need for flow measurement from the perspective of energy conservation. However, flow measurement of steam is very difficult due to its high temperature and susceptibility to wetness. In addition, there are very few measurement standards. Therefore, the study group has made efforts to clarify the condition of application and accuracy of each measurement method with regard to the flow measurement method of steam. The number of committee members is 30 (chief: Shuichi Umezawa (Tokyo Electric Power Company Holdings, Inc.), vice-chair: Ryo Morita (Central Research Institute of Electric Power Industry), secretary: Tatsuya Funaki (National Institute of Advanced Industrial Science and Technology)), who come from universities, companies, research institutes and others. The frequency of its activities is as follows: it holds study groups once or twice a year, summarizes its results and publishes a report. The predecessor of this study group, the "Workshop of Wet Steam Flow Measurement" (April 2012 - March 2017), also published a report summarizing its overall activities. On the basis of these results, an academic standards development committee has been set up this year and is preparing guidelines for the correction of wet steam flow rate indications from the perspective of thermal energy utilization.

In large gas turbines, the cooling air flow rate is an important parameter for analyzing the temperature and efficiency of the stators and the blades, so measurement is desired. However, ultrasonic flowmeters are difficult to use because the temperature of the measuring section usually exceeds 300°C, while insertion-type flowmeters are difficult to apply because they require modifications to the piping for installation and cause a pressure drop. In this study, an attempt was made to apply the heater method to the measurement of cooling flow rates. In this method, an annular heater is attached to the outside of the pipe and the temperature distribution along the pipe axis is measured using thermocouples. The temperature distribution is affected by the flow velocity of the fluid in the pipe. The flow velocity can therefore be analyzed from the temperature distribution in the pipe. Measurements were conducted for two kinds of cooling air pipe at an advanced combined cycle power plant. In one, cooling air is extracted from the 13th stage of a compressor and supplied to the 2nd stators of the gas turbine; in the other, cooling air is extracted from the 9th stage of the compressor and supplied to the 3rd stators of the gas turbine. As a result, it was clarified that the cooling air flowrate had a positive dependence on atmospheric temperature at the plant. The dependence was also compared to that of a tit 1,600 degrees centigrade-class combined cycle power plant, with a power output of 700 MW.

In industrial fields, as the steam is commonly utilized for heat supply, drying process and so on, it is important to be aware of the steam flow rate in the view point of energy management. However, the steam utilized in the factory is usually wet steam condition. Though it is well known that the wetness affects the flowmeter reading, it is difficult to clarify the effects of the wetness in the steam flow quantitatively in actual plants and factories, and thus far, there has been no established method for estimating the error caused by the wetness of steam flow. Therefore, we have started to investigate the effect of the wetness to the steam flow meters.

The characteristics of several steam flow meters in the wet steam were described in this report, and the relationship between the steam wetness and supply heat of the wet steam has been investigated. As a result, in the wet steam, the indicated flow rate underestimated the actual flow rate and the flow rate difference between them became bigger when the wetness became larger. The supply heat ratio evaluated from indicated flow rate was also smaller than that from actual flow rate, however, the degree of the difference was smaller (less than 5%).

In general, it is practically difficult to separate the gas and liquid phases of a steam flowmeter for simultaneous measurement, given its measurement principle. Therefore, based on our past research activities and results, we have reorganized the steam flow rate value for the gas phase, which contributes to understanding the amount of useful heat energy, by making useful corrections to the flow rate indication value when measuring wet steam in which the gas and liquid phases are mixed. In addition, the relationship with related international standards is discussed, and future prospects are discussed.

Clamp-on ultrasonic flowmeter is a powerful tool for measuring flow rate in an existing pipe. However, measuring flow rate of wet steam flow has not been established yet. It is important to investigate the various error factors caused by the wetness fraction. Based on our previous investigations, it was clarified that the error of the flow rates tends to increase with the wetness fraction because of the changes in the velocity distribution in the wet steam and the effect of the liquid holdup. Therefore, it is important to estimate the flow regime of the wet-steam flow in the pipe. This study proposed a flow pattern recognition method using a machine learning based on the guided wave that is the propagated ultrasound through the pipe wall. The guided wave changes depending on the flow regime due to the fluctuations at the liquid-gas interface. Using the received ultrasonic signals of the guided wave in stratified flow, annular mist flow, and transition between them were used as the teaching data, and the other conditions were tested for the flow pattern recognition. Flow patterns were accurately predicted in adiabatic air-liquid two-phase and wet-steam flows in horizontal pipes. Furthermore, it was shown that the selection of the guided wave region is important to predict the flow pattern in various pressure conditions using the same teaching data.

This paper introduces two concepts, KEIROKA assistance and human augmentation, which aim to maximize and utilize human abilities, and describes the Smart Suit as a technology to realize these concepts. The expected functions of actuators used in the technologies for KEIROKA assistance and human augmentation are also described.

The number of people requiring nursing care is increasing in Japan. This is because individuals in this older segment of the population tend to suffer more falls and other accidents than others. This background highlights the need for the development of a fall prevention method. Various research for fall prevention has been studied for this problem. We also have proposed a new method for fall prevention based on the phenomenon called Light Touch Contact (LTC), in which lightly touch to an object with the fingertips reduces the body sway. This paper introduces the development of a wearable device for fall prevention based on the LTC phenomenon and a fall risk assessment index based on large-scale experimental data of more than 1,000 subjects as examples of our past efforts.

As the size of offshore wind turbines increases, the maximum height of the blade tips will soon reach 300m.

The external conditions premised in designing a wind turbine are modeled in the design requirements, and are not necessarily premised on future enlargement. Therefore, we will introduce examples of external conditions that should be considered for wind turbines, which are expected to grow in size in the future, based on recent research. -

Each country in the world has been proactively discussing to reduce emission of carbon dioxide, and renewable energy is expected to be introduced as a main driver. Wind energy is the most potential driver because wind blows everywhere. Wind energy will also extend from onshore to offshore, which results in skyrocketing as well as higher generation capacity turbines will be introduced. This paper presents both forecast of wind market and technical activity on high reliability of main shaft bearing mounted in wind turbine.

The Japanese government stipulates that they will promote the maximum introduction of renewable energies and plan to place them as major power source in the 6th Strategic Energy Plan approved in a Cabinet meeting in October, 2021. Wind power is one of the most expected renewable energies, however, wind turbine repair tends to take a long time in Japan due to the location far from Europe where wind turbine manufacturers are mainly present. FZG micropitting test and modified FZG scuffing test were conducted in accordance with a hypothesis considering a turbulence impact on gear damage of gearbox, and it was demonstrated that a suitable gear oil for Japan’s climatic factors suppressed gear tooth surface wear better than a gear oil approved by European wind turbine gearbox manufacture. In addition, Idemitsu participates in an international joint working group for revision of IEC 61400-4 as a representative of Japanese committee and acts so that the standard will take wind conditions of Japan into account. Currently we are challenging a NEDO project to develop next generation lubricants for wind turbine which contribute stable operation and maintenance cost reduction of Japan’s wind farm by making full use of some state of the art technologies.

This paper focuses on a parallel wire-driven system and a musculoskeletal system which have a multi-degree-of-freedom mechanism driven by unidirectional forces. The characteristics and applications of the parallel wire-driven system are described. After pointing out the analogy between the two systems, the musculoskeletal potential method is explained.

By applying wire suspended structures to a robotic system, it is possible to realize a large working space and a payload in comparison with its weight. However, it is necessary to keep the tension positive during its motion so that the suspended part is controllable. Here three types of wire suspended robotic systems are introduced which can position their objects rigidly against external forces applied to the object.

There are exoskeletons using hard frames and assistive suits using soft materials. Each has completely different characteristics. However, there was little technology available to combine them. Wire-assist technology could be a technological bridge between the two suits.

With the aging of social and industrial infrastructure becoming a social problem, various new technologies such as the utilization of robots and drones and the automatic extraction of damage by deep learning are being studied. Among them, inspection robots that utilize wires are attracting attention as a method that has less risk of crashing and can be approached without directly contacting the structure to be inspected. However, when using a wire, it is necessary to consider a method of fixing the end to the structure, and it is also necessary to consider incorporating the extension of the rope itself into the control system and putting it into practical use. This paper introduces a device for practical use and an approach for achieving the purpose of the Rope Stroller, which inspects without touching the structure hanging the wire.

This paper introduces the automation of transport operations in a logistics warehouse and excavation and loading operations in civil engineering by utilizing networked robot systems, which are robot systems based on networked control systems. The paper shows that safe and highly efficient automation can be achieved by relaxing the constraints on the robot’s perception and intelligence through communication networks.

To improve the reliability and safety of autonomous drone navigation, we are conducting research and development with the aim of realizing "no-fall/safe-to-fall" drones equipped with deep learning AI technology. It consists of autonomous operation AI technology that detects people and vehicles and avoids flying over them, failure diagnosis AI technology that detects malfunctions and unstable behavior based on onboard drone sensors and telemetry information, and emergency landing AI technology that moves to a safe area and lands when a failure is diagnosed. Techniques for finding a safe area for a drone to land during an emergency landing and approaches to ensure object detection performance are described.

The Model Based Development (MBD) is the adoption of digital technology to transform real machinery to virtual machinery. MBD of Autonomous Transportation and Logistics machinery is investigated.

It is expected to develop and advance treatments for diseases such as hematologic and solid cancers that have been difficult to treat with existing pharmaceuticals. The use of pharmaceuticals based on genetically engineered cells and macromolecules, rather than conventional low-molecular-weight compounds, is rapidly advancing. To spread these therapies, several issues must be resolved. Specifically, the elucidation of disease biology, the establishment of drug safety, and cost reduction are necessary. The challenges include the complexity of cells, their fragility, and a large amount of processing involved, which are difficult to solve based-on conventional engineering alone. We try to solve these problems and work on the following two technologies: (1) autonomous drug discovery using disease models and (2) processing and screening of more than 100 single cells. In this presentation, I will present the status of our laboratory's works.

Recently, flexible/stretchable electronic devices, such as flexible sensor, display, and solar cell, has attracted much attention. To achieve such a flexible/stretchable electronic device, many researchers have used organic materials, such as organic light-emitting diodes (OLEDs), organic transistors, and organic electrodes (e.g. PEDOT:PSS). As compared with metal or semiconductor materials, organic materials might have good properties for mechanical deformation, however, the electrical properties are usually poor. Therefore, we focused on achieving “flexible/stretchable” electronic devices using “rigid” materials/components with good electrical properties. In this paper, as research based on this idea, we introduce research on “self-healing metal interconnect” as an interconnect for stretchable electronic devices and research on “origami/kirigami structures” for stretchable electronic devices.

In recent years, the demand for digital transformation (DX) in the industrial sector has been extremely high. In response to this high demand, a variety of smart devices have been invented and offered. Especially in the research field, flexible or bendable smart devices have been devised by using thin-film polymers such as polyimide and parylene as substrates. Micro/nanofabrication is a fundamental technology in the research and development of these smart devices. The presenter's laboratory has conducted research on flexible smart devices for social implementation based on micro/nanofabrication. Based on the knowledge obtained from flexible smart devices, we have also studied stretchable devices, which are expected to be the next generation of smart devices. In this presentation, we report an overview of our research on flexible smart devices and stretchable systems. In particular, we present our research on flexible smart devices and the development of smart devices for use in medical and animal experiments. In addition, research on stretchable sensors using liquid metal and stretchable systems are reported. Stretchable devices based on micro- and nano-fabrication might be further accelerated toward social implementation in the future.

Cell cryopreservation is the only way to stop biological activity and is widely used to maintain backups of cells. Conventionally, at least a cryoprotectant agent (CPA) must be added to avoid forming damaging ice crystals and to vitrify water on the inside and outside of the cell. We developed a CPA-free cryopreservation method by superflash freezing based on inkjet cell printing, in that cells were ejected as micro-droplets and deposited on the liquid nitrogen-cooled substrate. The method cryopreserved mouse fibroblast cells at a viability almost as high as the conventional method using dimethyl sulfoxide. To improve the reproducibility and viability further, the thawing process is important as well as the cooling process. Next, we also developed an automatic thawing apparatus that transports the vitrified cells rapidly into a prewarmed medium using a spring hinge. The experimental results indicate that the combination of superflash freezing with the rapid thawing process using the apparatus would become a possible approach to circumvent the problems associated with the addition of CPAs.

Fine bubbles with diameters of less than 100 μm are currently attracting considerable attentions because of their properties such as a large surface area per unit volume, low rising velocity, and self-pressurization due to surface tension. We found that fine bubbles with diameters of less than 100 μm can be easily generated by using a hollow cylindrical ultrasonic horn. Compared with to the conventional fine bubble generation method, which are wherein gas is supplying supplied gas to porous media and atomizing the gas is atomized with shear flow, this fine bubble generation method can generate small-sized air bubbles and mass circulation of the liquid phase is unnecessary. Therefore, this device is capable of generating MBs in any fluid, even in high-temperature, high-viscosity and high reactive materials. In this paper, we show the summary of fine bubble generation method using hollow ultrasonic horn, and also the example application of this method to fabrication process for hollow microcapsule and porous metals.

Our researches aim to develop biohybrid system with functional cellular tissues and artificial devices by the integration of microfabrication and microfluidic techniques with biological techniques. The developed biohybrid system is applicable in a wide range of fields, from biotechnology and medicine such as in vitro cellular models for biological studies and drug development without laboratory animals, to robots and sensors using the sensory and motor functions unique to living organisms. To form functional tissues, we have shown that cellular beads and fibers can be used as building modules for in vitro tissue construction. Furthermore, as a demonstration of the biohybrid system using functional cultured tissues, we developed a biohybrid robot with a pair of antagonistic skeletal muscle tissues and a biohybrid sensor with sensor cells expressing olfactory receptors. In this presentation, we will introduce cultured tissues fabricated by in vitro 3D cell culture and biohybrid systems usable cellular functions by integrating cultured tissues and devices.

In this study, temporal fluctuation of the boiling heat transfer was simply modeled based on measurements, and unsteady heat conduction simulations were conducted using this model as a boundary condition. As a result, it was found that the wall surface temperature fluctuation varies significantly depending on the thermophysical properties of the heat transfer surface. It was also shown that the amplitude of the surface-averaged wall temperature fluctuation over a typical area varies inversely proportional to the thermal effusivity √cρλ. However, the simplified model of heat transfer fluctuation is used to predict local temperature fluctuations, the prediction accuracy is degraded due to thermal diffusion in the plane direction. Therefore, in order to improve the accuracy to predict local wall temperature fluctuations, it is necessary to properly model the surface distribution of heat transfer coefficient in addition to the temporal fluctuation.

A creep constitutive equation incorporating the effect of microstructure coarsening was introduced into the Finite Element Method analysis using a user subroutine to predict thermal fatigue life of solder joint. Dispersed particles in the microstructure grew due to thermal diffusion with increasing fatigue cycles, which resulted in the inelastic strain range and the inelastic strain energy density range increased. This increase in the inelastic strain range reduced the fatigue life by approximately 77% compared to the case where no microstructure coarsening effect was incorporated. Thus, it is necessary to incorporate the microstructure coarsening effect in the fatigue analysis of solders. Thermal fatigue life was found to be significantly affected by the microstructure coarsening effect. Thus, it is necessary to incorporate the microstructure coarsening effect in the fatigue analysis of solders.

The degradation process of the crystallinity of grain boundaries in Al thin-film interconnections was visualized by applying EBSD analysis under high current density. The grain boundary diffusion was accelerated by electromigration only along random grain boundaries with low crystallinity. The crystallinity (IQ value obtained from EBSD analysis) of grain boundaries varied drastically as strong functions of manufacturing process such as deposition method, temperature of thin-film deposition, gas pressure during the deposition, substrate material, and so on. The effective activation energy of self-diffusion of component element along random grain boundaries was found to be dominated by the crystallinity of random grain boundaries. The measured activation energy obtained from the low-crystallinity interconnections was much lower than that obtained from the thermodynamically stable bulk material, and therefore, the diffusion constant due to electromigration was drastically increased, resulting in the drastic decrease of the lifetime of the interconnections. The lifetime of the interconnections was, therefore, improved by increasing the crystallinity of grain boundaries.