Tohoku Institute of Technology

The effective detection of earthquake precursors is pivotal for mitigating the hazards of earthquakes. However, its feasibility has been controversial for decades and remains unanswered. In general, observation systems must be installed to predict physical phenomena. An ideal observation system has not been established globally for earthquake phenomena. In this context, the Japanese government has developed various earthquake-related observation systems over the past few decades. Among them, the GEONET is an extensive and high-density system deployed throughout Japan using continuous GNSS technology. Using this system, we developed methods for analyzing the spatiotemporal evolution of crustal strains and examining their ability to detect precursors for the recent devastating earthquakes. We focused on the tensor analysis of strains rather than the vector analysis of the observed displacements from GEONET, by using the Finite Element Method to estimate the crustal strain from the displacement. The spatiotemporal evolution of crustal strains due to the 2024 Noto Peninsula Earthquake of Mw7.5 was examined in detail in this study. Through the present analysis of this representative earthquake, we demonstrated that the spatiotemporal evolution of crustal strains showed clear signs of earthquake occurrence.

To evaluate the applicability of a cylindrical rotating detonation (RD) combustor as a torch, combustion experiments were conducted in a scramjet model combustor. The results were compared with those obtained under low back pressure conditions to assess the performance of the RD torch. The torch had a combustion chamber with an inner diameter of 10 mm and a length of 173 mm, operating at a mass flow rate of 4.0 g/s and an equivalence ratio of 1.15. The total temperature, total pressure and Mach number of the main air were 641–719 K, 0.6 MPa and 2.8, respectively. Comparing the results under the two conditions, we concluded that the torch operated in detonation mode for up to 9.0 seconds under supersonic flow, even during main H2 combustion. The detonation wave speed was ranged from 2.07 to 2.14 km/s, corresponding to 84–86% of Chapman-Jouguet velocity and 156–160% of the sound speed in the burned gas. At a main H2 mass flow rate of 1.1 g/s, the scramjet combustor operated in scramjet mode, achieving both ignition and flameholding using the RD torch during scramjet-mode operation. Furthermore, we identified the effects of the main H2 combustion on the torch. Continuous propagation of detonation waves, consistent trends in the axial time-averaged pressure distribution even after the exhaust transitioned to a subsonic jet, pressure oscillations induced by the combustion in the cavity, and an increase in detonation wave speed all suggest new potential applications for the RD torch and directions for future work.

With the change in global climate and environment, different types of disaster weather occur more and more frequently, leading to a large number of power security accidents. Considering the fault risk of power systems under disaster weather, the self-test information of protection devices and weather data are introduced to propose a novel fault diagnosis method for power systems based on a random self-regulating algorithm. First, a diagnosis analytic model is established based on the alarm, weather and self-test information of power outage areas. Second, three types of self-regulating trust factors are introduced into the analytic model to strengthen its fault diagnosis ability in the context of disaster weather. Finally, a random self-regulating algorithm is proposed to solve the analytic model to obtain a fault hypothesis. Case studies verify the feasibility and effectiveness of the proposed method, which also has good fault tolerance to the abnormal action of protection devices and the distortion of alarm information.

Aneuploid cells are known to increase with age. Previously, we demonstrated an increased number of aneuploid fibroblasts isolated from aged mice due to chromosomal instability (CIN), which is caused by oxidative stress. It is unclear whether this phenomenon also occurs in human cells, which are more resistant to oxidative stress than mouse cells. Here, we found that fibroblasts from aged individuals exhibited an increase in aneuploid cells. The frequency of chromosome missegregation and micronuclei increased in these cells, indicating CIN. A DNA fiber assay revealed the presence of replication stress, accompanied by an increase in 53BP1 nuclear bodies and ultrafine bridges. Increased levels of reactive oxygen species derived from mitochondria, along with reduced mitochondrial membrane potential, imply that these cells experienced oxidative stress due to mitochondrial functional decline. Antioxidant treatment reduced the frequency of chromosome missegregation and micronuclei, suggesting that oxidative stress causes CIN. Oxidative stress also causes replication stress, which precedes CIN. Spindle microtubules were stabilized in fibroblasts from aged individuals, which was alleviated by antioxidant treatment. Taken together, these findings suggest that aging-related CIN in human fibroblasts is caused by oxidative stress associated with mitochondrial dysfunction, which induces replication stress that in turn causes CIN through microtubule stabilization. Although human fibroblasts are more resistant to the ambient oxygen environment than mouse fibroblasts, our findings showed that they undergo oxidative stress that causes CIN with age in a manner similar to mouse fibroblasts, revealing a conserved phenomenon in mammalian cells.

As artificial intelligence (AI) integrates into extended reality (XR), or known as XR×AI, a central design challenge emerges: how to balance AI assistance with user autonomy. This challenge presents a core dilemma: unassisted interfaces risk creating "hollow autonomy," where overwhelming choice undermines user confidence, while fully automated systems can impose a "tyranny of efficiency" that erodes agency. To investigate this tension, we conducted a study with 32 participants, comparing four distinct XR×AI strategies along a spectrum of intervention, from passive information display to direct recommendation, across two levels of task complexity. Our findings reveal a clear user preference for a collaborative middle ground that avoids the pitfalls of the extremes. While high intervention was decisively rejected for undermining agency, minimal support proved ineffective under complexity, leading to high cognitive load and low confidence. The most successful interfaces were those that acted as a "cognitive scaffold," collaborative partners that curated the decision space to reduce cognitive load and prevent errors, while leaving the final, empowering act of choice to the user. These results demonstrate that the goal of XR×AI design is not to create a system that thinks for us, but a partner that helps us think better.