Kurume Institute of Technology

• A geometric multigrid preconditioner is proposed for a Bi-CGSTAB solver applied in the ISPH method for multi-GPU environments. • The first restriction operation is conducted from the free-moving particles to a background cell followed by standard coarsening restrictions. • We proposed several techniques on multi-GPU programming including communication-computation overlap and communication-hiding. • The computational efficiency was evaluated in 3 supercomputer clusters yielding speedups of about 2.3 times and weak scalability around 0.85. • The method’s potential was verified by simulating the Fukushima Daiichi nuclear power plant tsunami with 97 million particles in 16 GPUs. This paper presents the implementation of a multigrid-preconditioned asymmetric solver within the context of Incompressible Smoothed Particle Hydrodynamics (ISPH) method. The restriction operator maps particles to the background grid used for neighbor searching, followed by standard grid coarsening. Key contributions include the adjustment of the multigrid preconditioner to multi-GPU cluster environments, incorporating communication-computation overlap and communication-hiding techniques, and adapting a dynamic load balancer to respect the coarse-grid hierarchy for efficient domain decomposition using the slice-grid technique. The present work also shows the evaluation of the computational efficiency of the proposed method through dam-break simulations on three supercomputers–JAMSTEC’s Earth Simulator, Tongji University’s supercomputer, and the University of Tokyo’s Miyabi)–demonstrating weak-scaling efficiencies between 0.82 and 0.89. The multigrid preconditioner substantially improved solver performance, reducing the average number of iteration for solving the pressure from over 300 to nearly 60 across problem sizes ranging from 5 to 320 million particles, resulting in an overall speedup of approximately 2.3 times. Furthermore, we demonstrate the method’s capability for real-world engineering applications through the simulation of the 2011 tsunami inundation at the Fukushima Daiichi Nuclear Power Plant, involving 97 million particles on 32 GPUs. While the preconditioner proves highly effective in accelerating solver convergence, the study concludes by identifying limitations associated with coarse-grid constraints on load balancing and communication overhead, providing clear directions for future research.

ABSTRACTAircraft cabins experience translational accelerations along three axes and rotational accelerations around three axes during flight, leading to uncomfortable motion and vibrations. To mitigate these effects, this study proposes a Motion Reduction Cabin concept. To establish its feasibility, fundamental data on commercial jet transport dynamics were measured and analyzed. The measured pitch angle variation (−5° to 22°) and maximum Z‐axis acceleration (4.95 m/s2) were used to define the operating range of the Motion Reduction Cabin system. These values served as key parameters in determining the required actuator response time and displacement limits. To implement the Motion Reduction Cabin, aircraft motion is categorized into angular and translational components. Angular motion can be stabilized using a gimbal mechanism, while translational motion—particularly horizontal and lateral movements—may be mitigated using a system analogous to an air caster mechanism. However, vertical motion presents a major challenge. This study proposes a novel vertical motion compensation system, which actively counteracts vertical displacements in real time. Furthermore, a preliminary MATLAB simulation was conducted to observe the basic behavior of the system. The results suggest that the proposed system has the potential to mitigate vertical disturbances, providing promising insights for the feasibility of the Motion Reduction Cabin concept.

In Japan, guideline values for indoor air quality for a total of 13 VVOCs (highly volatile organic compounds) and VOCs (volatile organic compounds) were published by 2002. However, approximately 20 years have passed since the publication of the guideline values, and the use of new chemical substances has increased as substitutes for the regulated substances, and sick building problems caused by unregulated substances have been reported. The aim of this study is to investigate sick house issues by measuring existing regulated and unregulated substances in newly built wooden houses and evaluating indoor air quality. A total of 10 rooms in three houses were targeted for measurement and air quality was measured using active methods. The substances measured were 13 regulated substances and 16 uncontrolled substances. The results of the measurements showed that, among α-pinene and limonene, which are commonly detected in wood, the concentration of α-pinene exceeded the LCI reference value, while for limonene it was less than half of the LCI reference value. The concentrations of unregulated substances other than α-pinene were below the LCI standard values.

Ground reaction force (GRF) analysis is critical to evaluate gait and running performances. However, installing force plates on the ground surface is costly, and their transportation is challenging owing to weight constraints and restricted measurement areas. This study proposes the method by Extended Kalman filter (EKF) to estimate the three-axis force components in the GRF, utilizing vertical force data from insole sensors, angular velocity, and acceleration data from inertial measurement units (IMUs). Shear forces were calculated using the acceleration in global coordinates, with the equivalent mass calculated by the vertical force and acceleration. The force components were then compensated for sensitivity and interference using a matrix. Moreover, the method was extended to include the center of pressure and unit position vector. The experimental protocol comprised two phases: calibration and measurement, conducted on force plates installed on the ground surface. The error compensation for force components was based on actual gait motion data and the results were compared with the force plates' outputs. The findings confirmed the effectiveness of this method. This method will be expanded to include motion measurements during long-distance walking and outdoor environments.

Methods for evaluating the fluctuation of texture patterns that are essentially regular have been proposed in the past, but the best method has not been determined. Here, as an attempt at this, we propose a method that applies AI technology (learning EfficientNet, which is widely used as a classification problem solving method) to determine when the fluctuation exceeds the tolerable limit and what the acceptable range is. We also apply this to clarify the tolerable limit of fluctuation in the “Kurume Kasuri” pattern, which is unique to the Chikugo region of Japan, and devise a method to evaluate the fluctuation in real time when weaving the Kasuri and keep it within the acceptable range. This study proposes a method for maintaining a unique faded pattern of woven textiles by utilizing EfficientNet for classification, fine-tuned with Optuna, and LightGBM for predicting subtle misalignments. Our experiments show that EfficientNet achieves high performance in classifying the quality of unique faded patterns in woven textiles. Additionally, LightGBM demonstrates near-perfect accuracy in predicting subtle misalignments within the acceptable range for high-quality faded patterns by controlling the weaving thread tension. Consequently, this method effectively maintains the quality of Kurume Kasuri patterns within the desired criteria.

To evaluate the integrity of a building, many experts and engineers have evaluated the damage classification of a building based on superficial visual information through field surveys. On-site surveys are hazardous and require several years of experience and expertise. In this study, a system for detecting the presence or absence of cracks and calculating their lengths was developed using image processing technology. The accuracy of the system was examined using crack image data obtained from shear force experiments. For crack detection, a crack detection method was developed using canny edge, threshold, and HSV color detection. The detection of the presence of cracks was proposed to be coupled with image segmentation to improve detection accuracy. A method for calculating the crack length using image processing was also developed. In this study, we proposed a method to calculate cracks as straight lengths, and obtained results with 98.1% accuracy. However, for curved cracks, it was necessary to rotate or segment the image.