Ryoji Takaki

Large-eddy simulations of an ideally expanded supersonic jet at Mach number 1.8 are performed by a high-order unstructured grid solver for aiming at the validation for aero-acoustic applications. The developed solver is based on the flux reconstruction scheme to discretize unstructured hexahedral grids. The localized artificial diffusivity scheme is employed to capture shock waves robustly and to retain high-order accuracy away from the shocks. The turbulent jet flow is obtained by the FR-LAD solver using the degree of polynomial up to p=4. The computed results are in reasonable agreement with the experimental and numerical data in the literatures for time-averaged and fluctuating quantities.

© 2014, American Institute of Aeronautics and Astronautics Inc. All rights reserved. A cell-centered finite volume flow solver LS-FLOW that has been developed by JEDI/JAXA is applied to benchmark RANS simulations, and the solution accuracy and the computational efficiency are investigated. LS-FLOW can handle unstructured grids consisting of arbitrary polyhedral cells. A high-order flux reconstruction (FR) scheme on unstructured hexahedral grids has been developed and devised in LS-FLOW recently. The developed high-order unstructured grid solver is also assessed for the benchmark cases. Comparisons of the results obtained by the second order FVM and the higher-order FR method indicate that the FR solver can be competitive with the FVM in terms of the computational time by employing similar basic solution algorithms such as LU-SGS scheme for the same number of degrees of freedom, and also that the higher-order method can provide accurate result more efficiently for certain case of steady RANS simulation.

The JIEDI (JAXA's Ion Engine Development Initiative) tool has been developed to assess the ion acceleration grid erosion of an ion thruster. The sensitivity analysis of the input parameters of the JIEDI tool is conducted in this paper. We compare the simulation results of the JIEDI tool and show that it successfully reproduces the full lifetime test of the μ10 prototype model. Through sensitivity analysis, we found that the sticking factor is the most sensitive input parameter when estimating the accelerator grid erosion and electron backstreaming time. For the worst case scenario of grid erosion (without re-deposition), the uncertainty in the neutral mass flow rate through grid holes is important when estimating the accelerator grid erosion. A 30% change in the neutral mass flow rate caused by a 6% change in the propellant utilization efficiency, corresponds to about a 20% change in the accelerator grid mass loss as well as the increasing rate of minimum potential on the axis of the ion optics. In contrast to the accelerator grid erosion, the uncertainties in the discharge voltage and grid gap (that affect the trajectory of the mainstream ions), are important when estimating decelerator grid erosion. A 40% change in the discharge voltage corresponds to about a 90% change in the decelerator grid mass loss. In addition, a 30% change in the grid gap between the screen grid and the accelerator grid corresponded to about a 40% change in the decelerator grid mass loss.

A high-order flux reconstruction (FR) scheme on unstructured hexahedral grids is developed for aerospace flow simulation. In order to use computational grids generated by the body-fitted Cartesian (BFC) method, the grid which contains polyhedral cells with hanging nodes are subdivide into hexahedral cells first. Then, to perform computation by the FR scheme on the non-conformal hexahedral grid, the mortal element method (MEM) is employed to determine the numerical flux at cell interface. As a shock capturing scheme, the localized artificial diffusivity (LAD) method is developed for the FR scheme. The developed scheme is tested for typical benchmark problems including smooth and shocked flows. © 2013 by the authors.

Space Transportation FY2012 (January 17-18, 2013. Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency (JAXA)(ISAS)), Sagamihara, Kanagawa JapanSolar sail is a next-generation spacecraft that has a large-scale membrane to utilize the solar radiation pressure as its thrust. Hence, stable deployment of the membrane during a space flight is a critical issue to maintain the thrust performance of the spacecraft. In this paper, we have numerically estimated the electrostatic force on the membrane due to spacecraft charging as one of the possible factor to cause the deformation of the deployed membrane. We obtained spatial distribution of the electrostatic force and the maximum magnitude of the force on the membrane in solar wind plasma at 1.0 AU. We also made structural analysis for the deployed membrane with the electrostatic force. As a result, we can hardly recognize the deformation of the membrane due to the electrostatic force in this case.Physical characteristics: Original contains color illustrationsPhysical characteristics: PDF

PC cluster is installed to evaluate the parallel and remote visualization of large-scale computational results produced by the supercomputer in the remote site. It is revealed that more than 60 % of the time for visualization is required for reading the binary data of the computational result. Partitioned file format such as VTK multi-block is desirable for reading large-scale data in parallel. High scalability more than 80 % is obtained by using the parallel rendering and image compositing. Methodology to extract the region of interest (ROI) based on the VTK multi-block format file is developed, and reduction of reading data size is realized for large-scale computational result without losing the parallel I/O performance. The knowledge obtained here will be utilized for the post-processing system of the large-scale computations done by remotely connected supercomputers.

A non-deterministic analysis method of thin membrane dynamics is presented. The thin membrane is modeled based on the tension field theory, which represents wrinkling phenomenon as a compressive stress relaxation. The generalized α method is applied to formulate a numerical dynamic analysis of wrinkled membrane. A Monte-Carlo approach, where the probability distributions of uncertainties parameters contained in the membrane model are approximated by an ensemble set, is presented to evaluate statistical properties of membrane dynamics behaviors. A simple numerical example is presented to verify the effectiveness of the presented approach.

In this report, we first describe the acquisition and installation of the Numerical Simulator III, which has started operation at October 2002 in the former National Aerospace Laboratory, and has been operated until October 2008 as part of the JAXA Supercomputer System even after the consolidation of three space organizations into JAXA. Next, by clarifying what we were able to achieve or fail by the acquisition, what we learned from the operation experience with using the performance evaluation data and the operational statistics, we draw and visualize the important technical aspects in aerospace supercomputing, and the know-how and implicit knowledge for the large equipment operation. In particular, we address the performance characteristics of the JAXA applications in terms of hybrid parallelism which come from the architectural features of the central computing engine, i.e. a SMP cluster. Finally, with those materials, we discuss what the JAXA's supercomputing system should be, and the critical issues to the future supercomputing in order to earn useful views for the next-generation practitioners.

Aerodynamic characteristics of reusable observation vehicle are computationally investigated under subsonic and supersonic flows using the RANS (Reynolds-averaged Navier-Stokes) simulations. The initial investigation for the concept design is done with the light optimization using the light CFD. The results show that the simulations using coarse grid estimate the axial force coefficient and the lift to drag ratio accurately except some cases. The results indicate the correlation between the supersonic lift to drag ratio and the axial force coefficient. The results show the correlation between the y-coordinate of the design variable and the volume. The required knowledge for the concept design in the near future is obtained.

Evaluation of ablation of nozzle wall in solid rocket motor is studied numerically on three solid rocket motors. A coupled analysis of fluid dynamics and surface recession simulates a total ablation amount. The analysis consists of the two-dimensional axisymmetric fluid analysis and the estimation of ablation amount using a correlation equation of surface recession rate. The simulation results of total surface recession amount agree qualitatively with experimental results in Nozzle-A case. The numerical simulations estimate the erosion rate on the safe-side. The effect of shield for reactant gas from nozzle surface reaction is estimated by a simple model. This model works very well for the agreement to experimental results. However, the parameter of the model has to be decided from experimental data. This model is not sufficient for prediction of unknown rocket motor. The coupled analysis was performed on Nozzle-B and C cases in order to understand flow field and erosion mechanism.

Evaluation of ablation of nozzle wall and natures of three-dimensional flow in solid rocket motor are studied numerically. Three kinds of simulations are carried out to reveal the features of ablation. At first, a coupled analysis of fluid dynamics and surface recession simulates a total ablation amount. The analysis consists of the two-dimensional axisymmetric fluid analysis and the estimation of ablation amount using a correlation equation of surface recession rate. The features of nozzle shape can explain heat flux distributions and surface recession amount for each nozzle. The simulation results of total surface recession amount agree well with experimental results. Most of solid rocket motors use the different materials in the throat and nozzle parts because the throat diameter should be unchanged as possible as it could throughout the whole burning period. Therefore, a backward-facing step is formed at the boundary between those materials because the ablation rates of walls are different. Next, three-dimensional fluid steady analysis in solid rocket motors is carried out. Three-dimensional grid is made based on the results of the axisymmetric coupled analysis. Behind the step, longitudinal vortices and streak of heat flux appear. An effect of shape irregularity at nozzle inlet nose Is considered. The shape irregularity makes the fluctuation with low frequency and high amplitude on the heat flux. Lastly, three-dimensional fluid unsteady analyses around step on the nozzle wall are carried out to reveal the relation between longitudinal vortices and streaks of heat flux. The artificial steps with various heights are set on the nozzle wall. The number of longitudinal vortices depends on the step height. The non-uniformity of shear layer in circumferential direction affects the generation of the longitudinal vortices. The collisions of longitudinal vortices to the wall make the high heat flux.

Mixture fraction is a very important quantity for nonpremixed combustion modeling and is a conserved scalar whose conservation equation does not include any source terms even in reacting flowfields. In this paper, the DNS data of a hydrogen / air jet diffusion flame is analyzed using mixture fraction and its dissipation rate (scalar dissipation rate). PDF of mixture fraction in the mixing core region shows a β-distribution like feature. The distribution of mixture fraction is complicated in unburnt mixing region and smooth in the flame. The simulated distribution of sealar dissipation rate validates the theoretical quenching criterion defined by sealar dissipation rate.

This paper presents a computational parametric study showing how the amount of preceding heat release affects aerodynamic characteristics of an axisymmetric blunt body in hypersonic flow. A chemically non-equilibrium viscous flow is assumed. For this purpose, seven species and finite-rate chemical reactions are considered to treat chemically non-equilibrium phenomena. Results show that the increase of the amount of heat release can reduce both the drag and the amount of heating rate decrease monotonically. However, as the heat release becomes large, the heating rate increases with the heat release.

This paper presents a computational parametric study showing how preceding heat release position affects aerodynamic characteristics of an hemisphere in hypersonic flow. A thermally and chemically non-equilibrium viscous flow is assumed. For this purpose, seven species and finite-rate chemical reactions are considered to treat chemically non-equilibrium phenomena. Park's two temperature model is also used to take account of thermally non-equilibrium phenomena. Results show that the increase of the distance between heat release and hemisphere causes both the drag and the amount of heating rate to decrease monotonically. These reductions saturate when the distance becomes longer than 14 times the length of the hemisphere radius.