We here propose a new nonlinear numerical method of local and global corrections for accurately evaluating thermal efficiency as result of spatially-integrated thermodynamic quantities, which is deeply related to total amount of CO2 exhausted, and also pollutant emissions including NOx in large eddy simulation of subsonic and supersonic flows of power systems. Detailed chemical reaction models alone are not enough for evaluating emissions. We here show computation results when the new nonlinear correction method is applied for single component flows, especially those in a new type of high thermal-efficiency engine proposed by Naitoh et al.

Recent developments have enabled us to employ machine learning techniques for a wide range of research. We propose here to classify mathematical problems by its difficulty using a simple 3-layer Neural Net learning algorithm. We applied the algorithm to learn simple binary addition and the Mackey-Glass equation, which gave us results with good precisions. On the other hand, learning of the prime number distribution posed a fair difficulty. Further, learning for the next ((n + 2)-th) Collatz-Kakutani minimal cycle length from odd number n and the associated cycle length showed us no sensible predictions. We view that this result is caused by the reflection of the difficulty of the problems in learning performances of the learning algorithm. This indicates that the levels of difficulties associated with mathematical problems may be measured by learning performances of the machine learning models.

Symmetric and asymmetric size (mass) ratios observed in biological particles in nature and child atoms of nuclear fission are explained by using a stochastic momentum conservation law and weakest stability principle, proposed by Naitoh (J. of Physics, 2012). The three-dimensional momentum conservation law considers vector quantity (i.e. velocity), while conventional theories started from Bohr have been based on energy conservation law.

 Furthermore, we have shown that the size and mass ratios of those particles are predicted more accurately by using multi-dimensional Taylor expansion in the previous report (Kobayashi and Naitoh, JASSE, 2019).

 In this report, we clarify the relationship between the mass distribution of child atoms generated in nuclear fission and the energy level of neutron provided for the nuclear fission. We classified the terms in the stochastic momentum conservation law obtained by using the multi-dimensional Taylor expansion. As a result, it can be explained that mass distribution of atoms generated in the nuclear fission of uranium 235 depends on the energy of neutron collided to the uranium nucleus because neutron having higher input (impact) energy produce stronger flow in the deep region inside the particle, which leads to a symmetric size ratio of the child particle pair.

Visualization is an extremely important technology for exploring and presenting various physical phenomena included in the simulation results and many scientific visualization applications have been developed since the latter half of the 1980s. In recent years, virtual reality (VR) visualization, which represents physical phenomena with complicated three-dimensional (3-D) structures in stereoscopic images to further facilitate their understanding, has attracted attention with a prevalence of VR technologies. However, most of them display the visualization results, which were exported as 3-D model data from the existing visualization application, on the VR devices. For this reason, it is necessary to re-export the 3-D model data every time the visualization parameter is changed. Therefore, we have developed a framework for completing a visualization workflow such as reading simulation data and applying visualization algorithms on the game engine which has been used for the development of applications for VR devices in recent years. The framework proposed in this paper is possible to build a VR visualization application without programming by connecting the visualization module appropriately on the visual interface of the game engine.

This paper investigates structural coloration by a multilayered dielectric grating with non-ideal multilayers where the optical thickness of two media differ greatly. We formulate the matrix eigenvalues method to analyze the three-dimensional scattering problem of the structure for an unpolarized light incidence. Getting the specularly reflected diffraction efficiencies and transforming the efficiencies into the chromatic coordinates in standard RGB (sRGB) color space and the chromaticity coordinates, we numerically study the structural coloration. From the numerical results, we obtain the following characteristics of the structural coloration by the structure. The coloration requires a large number of layers to enforce interference reflection, vivid high-saturation structural coloration with a narrow reflection peak can be obtained, and periodic structure has a small effect on the structural coloration.

Recently, accompanied by the dramatic performance improvement of computers, acoustic analysis has been used for acoustic design inside and outside the room. Up to now, the authors have proposed a large-scale steady-state and non-stationary acoustic analysis method based on the parallel finite element method. The finite element method has the merit that can handle complex shapes such as real problems using a small number of elements, compared to the difference method. However, large-scale analysis with hundreds of millions to 10 billions of degrees of freedom is also required due to the increase in analysis space and increase in analysis frequency. The previous version of all parallel acoustic analysis modules have been developed in a 32-bit environment, which is a constraint for large-scale analyses, and tens of millions of degrees of freedom is the upper limit of the number of degrees of freedom that can be calculated. Therefore, in this research, all parallel acoustic analysis modules consisting of I/O library, domain decomposer, and acoustic analysis solver are newly converted to 64-bit. and the scale is 100 million degrees of freedom, which is unprecedented in the world. As a result, we have realized large-scale analyses with a scale of 100 million degrees of freedom such as the calculation of the head shape transfer function that requires analysis of this scale, and in real world problems such as high-frequency live music club analysis at over 1[kHz], so we report here.