Sonic Boom Prediction With FaSTAR And HexaGrid - NASA
1SciTech2014, National Harbor, MarylandSonic Boom Prediction with FaSTAR andHexaGridMasashi Kanamori, Atsushi Hashimoto and Takashi AoyamaNumerical Simulation Research Group, Institute of Aerospace Technology, JAXAMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
2Introduction / Outlineü The objectives of 1st SBPWü to assess the techniques for predicting signatures suitablefor sonic boom predictionü to compare the solutions of each participantsü Models provided from workshop are:ü SEEB-ALR Body of Revolutionü 69-Degree Delta Wing Bodyü Lockheed Martin 1021 Full Configuration : not consideredü Two models are analyzed with FaSTAR andHexaGrid:ü for GIVEN gridü for OWN gridMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
3About Flow Solverü FaSTAR (FAST Aerodynamic Routines)ü fast flow solver for unstructured grid (1.4 hour/case using 100cores, 10M grid)ü sufficiently accurate results obtainedü many options availableGoverning equationDiscretizationInviscid fluxGradient evaluationLimiter functionHigh resolutionTurbulence modelTransition modelParallelizationCompressible Euler/Navier-Stokes equationFinite volume method (Cell center/Cell vertex)HLLEW, Roe, HLLE, AUSM -UP, SLAULeast-Square, Green-Gauss, GLSQHishida, Venkatakrishnan, Barth-Jespersen, minmodMUSCL, U－MUSCLSA，SST，EARSM，DES，LESForced, Natural(γ-Reθt)MPILU-SGS (steady/unsteady, local/global time step),Time integrationPreconditioning for low-speed flowConvergence acceleration Multigrid(FAS), Krylov method (GMRES)Tetrahedra, Pyramid, Prism, HexahedraElement typeGridgen，Pointwise, HexaGrid, MEGG3DGrid inputFieldview(FV-UNS), Tecplot, Paraview(VTK)Result outputfrom 4th Drag Prediction Workshopü Calculation conditions:ü 64core@JSS(JAXA Supercomputer System)ü Inviscid calculationü HLLEW scheme with 2nd order MUSCL interpolation with Hishida’s van Leer type limiterü LU-SGS time integration with local time steppingü CFL number is set to 10.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
4GIVEN GRID ANALYSESMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
5Summary of Grids and Convergence Historiesü Mixed element type were chosen as given grids.ü Convergence histories indicate that thousands of iteration needs to be converged.Summary of the given gridSEEB-ALR69deg. DeltaCoarse200s200sMedium156s150sFine100s125sVery fine080s100sü Calculation stops when theresidual reaches theplateau.ü Time to converge is about 1to 1.5 hours.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
6Results of SEEB-ALRü Signatures from the result of different resolution show good convergence.ü Only the signature of 156s at 42 in. is different from others.H 42 in.H 21.2 in.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
7Difference on Signature at H 42 in.ü Something is wrong with the signature of 156s at 42 in., not with that at 21.2in.ü Compared with the signatures at different azimuthal angles, namely 90 and 180 deg., sucha discrepancy is not observed.ü This indicates that problematic cells are locatedbetween H 21.2in. and 42in.H 42 in.H 21.2in.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
8Results of 69-Degree Delta Wing Bodyü H-variationsü shows good grid convergence.ü includes nothing suspicious unlike SEEB-ALR.H 31.8 in.H 31.8 in.H 24.8 in.H 21.2 in.H 24.8 in.H 21.2 in.H 0.5 in.H 0.5 in.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
9Results of 69-Degree Delta Wing Body Cont’d.ü φ-variations for H 24.8 asashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
10Results of 69-Degree Delta Wing Body Cont’d.ü φ-variations for H 31.8 in.30deg.0deg.0deg.30deg.60deg.90deg.60deg.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
11OWN GRID ANALYSESMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
12HexaGrid – Full Automatic Grid Generator –Full Automatic grid generation using a few control parametersü Domain (max and min of x, y, z)ü Max and min cell sizes on the surfaceü Layer parameter (thickness of first layer, expansion factor) CAD data(STL format)Algorithm of HexaGridRefinement Boxü Locally refined region can bedesignated.ü Start with one big element( computational domain)ü Cartesian grid is generated by means of successivelocal refinementü Each refinement divides a cell isotropically into eightchild cellsü Refine the element using 3 criteriaü Cartesian grid is generatedin the domain.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXABody-fitted layer grid is generatednear the solid surface(Layer grid is hexahedral)01/12/2014
13Outline for Own Grid AnalysesOur StrategyCartesian grid away from the modelü The model is rotated by Mach angle µ sothat the grid lines become parallel to theshock waves.ü Refinement box, or locally refined regionis designated along shock propagationpath, which guarantees a sufficientresolution for predicting near-fieldsignatures.ü Only the signaturesµ rot.Refinement Boxü for 69-degree delta wing bodyü on the plane of symmetryare considered in this presentation.Grid lines parallel to the shock wavesMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
14Comput. DomainParametric Study SettingsM ü The sensitivity of the nearfield signature ismodelinvestigated by changing the size of RefinementBox parametrically:zü longitudinal length Lü spanwise length Bü L 150mm is about the size of the model in xdirection.ü B 70mm is twice of the span of the model.ü The number of grid point: min. 20M, max. 90MLongitudinalLB150mm200mm250mm70mmcase 1case 2case 3Spanwise100mm150mmcase 4case 7case 5case 8case 6case 9Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXAxyRefinement BoxLBSize of the modelSpanwise 35mmyz01/12/2014zxLongitudinal 141mm
15Example of Own GridsRefinement Box1min. size 0.25Overviewü Body-fitted grid around the body andRefinement Box2min. size 1Cartesian, cubic cell elsewhereü Resolution ( Min. size of the cell) is fixed asshown in this slide.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
16Example of the Resultü Highly accurate signatures can be obtained by thecombination of model rotation and local refinement of thegrid.Refinement BoxH 21.2 in.H 31.8 in.H 21.2 in.Convergence historyH 24.8 in.between shock waves and grid lines.ü connection of the cells of different levelalso causes the waviness.ü computation stops when the movingaverage becomes flat.yxH 24.8 in.H 31.8 in.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXAü looks wavy because of the relation01/12/2014
17Effect of Longitudinal ResolutionLongitudinalSpanwiseLB15020070Graph 1100Graph 2150Graph 3Graph 1250ü Signatures obtained at H 31.8in.ü L variation with fixed Bü Compared with three signatures for fixed B, theresolution along longitudinal direction is not so sensitive. Setting L (size of the model in the x direction) issufficient for predicting nearfield signatures.ü Such a trend is observed at the other locations.Graph 2Graph 3※ ref. Obtained from given grid (100s)Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
18Effect of Spanwise ResolutionLongitudinalSpanwiseLBGraph 1250Graph 3150Graph 2100200ü Signatures obtained at H 31.8in.ü B variation with fixed Lü Signatures corresponding to large B show goodGraph 170150agreements with the reference result. Large span is necessary for accurate prediction of thenearfield signatures.ü Such a trend is observed at the other locations.Graph 2Graph 3※ ref. Obtained from given grid (100s)Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
19Effect of Spanwise Resolution Cont’dGraph 2Graph 3Graph 1ü Pressure rise at the front shock is not sensitive to the spanwiseresolution, but finer grid is necessary to capture the peak (Graph1).ü Pressure rise at middle and tail shock are highly affected by thevariation of B. (Graph 2 and 3) Circumferential effectü Wide Refinement Box will result in the correctnearfield signatures.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
20Circumferential Distribution of Pressure SignalCP Distribution on sliced planeCase 3Case 6Case 9Refinement BoxShock from StingB 70B 100wavyB 150ReferenceShock from Modelü Signatures are somewhat wavy in the circumferential direction.ü Waviness still remains even in case 9, but is not sensitive to the signatureitself. B 150 (case9) is sufficient for this model.ü The width of the Refinement Box B becomes more than 4 times larger than thatof the model !Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
21Concluding Remarksü Predictions of nearfield signatures are conducted with FaSTAR.ü SEEB-ALR and 69-degree Delta Wing Body are calculated.ü For given grids,ü extracted signatures show good convergence against grid resolution.ü only the result of SEEB-ALR has some different tendency compared with otherresults, which is mainly due to the grid itself rather than the flow solver.ü For own grids,ü With HexaGrid, or automatic grid generation tool, nearfield signatures are predictedcorrectly by the combination of model rotation and local grid refinement.ü The effect of refined region on the nearfield signature is investigated.ü longitudinal length is not important.ü sufficiently large spanwise resolution is important for the prediction of nearfieldsignatures.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
22Thank you for your attention!Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
23APPENDIXMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
24Signatures at Location 1Spanwise Blarge largeLongitudinal Lsmall smallMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
25Signatures at Location 2Spanwise Blarge largeLongitudinal Lsmall smallMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
26Signatures at Location 3Spanwise Blarge largeLongitudinal Lsmall smallMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
27Signatures at Location 7Spanwise Blarge largeLongitudinal Lsmall smallMasashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
28Pressure Contours for 69deg. Delta-Wingü Waves emanating from wing causes a generationof disturbances in the circumferential direction.ü Capturing such disturbances isimportant to predict nearfieldwaveform correctly, which will bediscussed later.Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA01/12/2014
fast flow solver for unstructured grid (1.4 hour/case using 100cores, 10M grid) ! sufficiently accurate results obtained ! many options available Masashi Kanamori, Numerical Simulation Research Group, IAT, JAXA 01/12/2014 3 from 4th Drag Prediction Workshop ! Calculation conditions: ! 64core@JSS(JAXA Supercomputer System) !
The utilization of the sonic boom electronic data file in quantifying effects of atmospheric turbulence and molecular absorption on sonic boom waveforms is a major thrust. The purpose of this report is to provide a ha_d copy summary of the measured sonic boom signature data derived from the 39 supersonic flights of the XB-70 (43 passes) over .
Liftcrane Boom Capacities 888 SERIES 2 Boom No. 22EL 179,100 Lb. Crane Counterweight 44,000 Lb. Carbody Counterweight 360 Degree Rating Meets ANSI B30.5 Requirements Oper. Rad. Feet Boom Ang. Deg. Boom Point Elev. Feet Boom Capacity Crawlers Retracted Pounds Boom Capacity Crawlers Extended Pounds 100 Ft. Boom *b *b *b * * * * Oper. Rad. Feet .
REFRESHER Pima County FASTAR Administrative Order No. 2017-116 creates Fast Trial and Alternative Resolution Program (“FASTAR”) Pilot Program 11/1/17 to 10/31/2020 “The new court rules and lower jurisdictional limit establish the framewo
SONIC BOOM DESIGN TECHNOLOGY Masahisa Honda*, Kenji Yoshida* *Japan Aerospace Exploration Agency honda.masahisa@jaxa.jp; yoshida.kenji@jaxa.jp Keywords: D-SEND, sonic boom, balloon, drop test, Esrange . MASAHISA HONDA, KENJI YOSHIDA 2 two axisymmetrical bodies of the D-SEND#1 simply free-fall from a high altitude balloon and .
An application running in SONiC NOS, solving specific use case Integrate with SONiC infrastructure i.e. SAI, Redis DB No impact on SONiC - Using SAI flex interface REDIS DB All enhancements over SONiC are automated & auto generated Combination of application needs and dataplane capabilities
Level Crossing controlled by Flashing Lights and Half -Boom Gates. Level Crossing controlled by Flashing Lights and Four Quadrant Half -Boom Gates. In this Standard the term Half-Boom Gate shall be synonymous with the terms Boom Barrier or Boom Gate. Four quadrant gates shall refer to application of half boom barriers arranged to
Link‐Belt Cranes 218 HSL Main Boom Working Range Diagram 40' TO 230' TUBE BOOM Notes: 1. Boom geometry shown is for unloaded condition and crane standing level on firm supporting surface. Boom deflection, subsequent radius, and boom angle change must be accounted for when applying load to hook. 2.
Abrasive water jet machining experiments conducted on carbon fibre composites. This work reported that standoff distance was the significant parameter which - reduced the surface roughness and the minimum of 1.53 µm surface roughness was obtained [31]. Garnet abrasive particles was used for machining prepreg laminates reinforced with carbon fiber using the epoxy polymer resin matrix (120 .
