ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
LIST OF MODIFICATIONS FROM V2 TO V3
2
3
4
MODULEPRIORITY/STATUSISSUE/OBJECTIVESOLUTIONDETAILS
5
GENERALDONEPython version: MIGRATE TO PYTHON 3 (let a version in 2.7 and 3.9)
6
DONEOverall speed-up: Check if the VTK algorithms used in Epsilon are multithreaded. See vtkSMP (Shared Memory Parallelism)vtk algorithms are already automatically parallelized, thus, not implementedhttps://www.youtube.com/watch?v=RJohCYPIids&list=PL1tk5lGm7zvRXfQeDlzWrob3rLaTlc4Yf&ab_channel=cscsch https://www.youtube.com/watch?v=Pr9xb1cqFf4&ab_channel=cscsch https://vtk.org/Wiki/VTK/Parallel https://hal.inria.fr/hal-00789814v2/document https://vtk.org/Wiki/images/3/3b/VTK_SMP_Guide.pdf
7
DONEGUI: Delete "Script" input in GUI of all filters???? See if this is not annoying for re-Applying a filterIt seems to be still practical for re-applying the filter
8
DONEEsthetic: Add decorative logo image in Epsilon_Aero_DataNot implementedADDING IMAGES IN PROG FILTER IS NOT SUPPORTED BY PARAVIEW (in Phyton-based programmable filters but yes in C++/Qt) https://discourse.paraview.org/t/programmable-filter-include-image-in-the-properties-panel/3906/3
9
DONEEsthetic: Add icon to the filters (in the sources/filters menu)<Hints> <ShowInMenu category="EPSILON" icon="C:\EPSILON\Icons\Epsilon_Symbol.png"/></Hints>https://discourse.paraview.org/t/icon-for-programmable-source-not-working/9327
10
DONEEsthetic: Change icon to the filters in the Pipeline Browser<Hints><PipelineIcon name="C:\EPSILON\Icons\Epsilon_Symbol.png" /></Hints>
11
EPSILON AERO DATADONEPre-procesing: Avoid using MergeBlocks before AeroData (integrate this vtk into AeroData). Test "vtkGeometryFilter" (see EpsilonWakeSurvey, line 1083)The VTK operation was included. No user interaction is required: it detects if the input is multiblock or not. If not, no conversion is madehttps://discourse.paraview.org/t/vtkmultiblockmergefilter-usage/7017/3 use the standard GeometryFilter or DataSetSurfaceFilter to convert UGs to PD. Those filters do each block separately and don’t bother appending them together – which saves time. https://discourse.vtk.org/t/vtkmultiblockdataset-and-parallel-pipelines/3137/3 https://discourse.paraview.org/t/programmable-filter-algs-is-not-working-properly-on-paraview-5-5-0/143 https://public.kitware.com/pipermail/paraview/2018-April/042564.html
12
DONEPre-procesing: Extract blocks --> Check if PV5.10.0 deals better with adapted grids. Otherwise, include vtkExtractBlock into AeroData. This is also required for input multiblock data where the blocks are of different type (volume mesh, surface mesh, etc) where merge blocks gives a VTK object that cannot be manipulated by Epsilon. In this case, only the fluid data must be extracted and all the boundaries ignoredA new input was included in the GUI. The user has to indicate the desired Block ID of the flow field. It then ignores the block of the refined region.https://vtkusers.public.kitware.narkive.com/tbuOhjUc/contour-of-3d-exodus-data https://vtk.org/pipermail/vtkusers/2012-October/076990.html
13
DONEPre-procesing: Cell data to point data --> apply only to cell arrays (Automatic detection of cell-based arrays)The original code was retained but an IF check was addedSee C2P example https://public.kitware.com/pipermail/paraview/2018-April/042564.html
14
DONEReaders: Include IC3 format
15
DONEReaders: Include turbulent Pr input as advanced option in GUI
16
DONEReaders: R and GAMMA varies from solver to solver. Define values in readers instead of AeroDataImplemented in the XML as an advanced option to be modified by the user. It was impossible to include it inside each data format readerActive advanced options in the GUI by clicking on the gear icon (upper-right of the properties panel). The XML advanced visibility is based on: https://www.paraview.org/Wiki/ServerManager_XML_Hints
17
DONEReaders: load additional arrays. Add option to let pass all the desired input variables besides those required by Epsilon (so the user can use IntegrationTool with that variable too)
18
DONEReaders: load vorticity data from imput (to avoid cleanup of masked regions from PIV and to avoid wall vorticity issues from CFD)
19
DONEReaders: Simplify the reader files and the procedure required to add new formatsNo thermodynamic calculations are now required. Each reader must now read the input arrays only. If an array do not exist at the input, it is defined as 'None'. This 'None' case and the aero-thermo calculations are now treated in the AeroData code.
20
DONEReference values: Get automatic reference values with VTK to define upstream flow conditions. Also, let the user to choose the current method or the automatic method.The most upstream point along a slice plane is used to get the usptream reference values. The slice plane is defined as the plane passing through the origin and inclined depending on the angle of attack
21
DONETransformation matrix: review transformation equationReviewed. Seems to be OK
22
DONEAero parameters: allow computing pressure, temperature and density when the related arrays are missing from the input. Also include possibility of assuming isoenergetic flow (see paper 3AF)
23
DONEAero parameters: Check Ts calculation for 5HP from Pt. Not working…
24
DONEAero parameters: Check Lambda2 parameter --> not working fineImplementedParaview's native "eigenvalue" calculator do not support non-symmetric tensors. It is solved by using "np.linalg.eigvals"
25
DONECustom variables: Facilitate insertion of new user defined aero variables (variable declaration and output)A new code section was included where the user can define UDV (user-defined varaibles)
26
DONEOutput variables: Allow selecting especific aero-analysis methods (Meheut, Maskell, etc.) from a given category. Add multiple boolean options with categories/sub-categories at the end of the GUI ("FFM: Maskell, FFM: Meheut, …"). Test all the output possibilities in order that the required variables are computed ! (i.e., if AERO parameters are disabled but FFM: Meheut is activated)
27
DONEFieldData: Spreadsheet view is not well displayed in some ParaView versionsIt works fine in ParaView 5.10.0
28
DONEGUI: Add option --> save output in VTK VTP AerodataOnly VTU format was impemented
29
DONEAnalysis methods: Meheut --> Include u* and ubar in the output of Aero_Data
30
DONEAnalysis methods: Arntz --> implement divergence of wave anergy masked by Q-criterion
31
DONEReaders: Include SU2 reader https://su2code.github.io/Implemented ISAE-DCAS project lead by Juan ALONSO VERGES in 2021
32
DONEGUI: add panel_visibility="advanced" to show extra optionshttps://www.paraview.org/Wiki/ServerManager_XML_Hints https://www.paraview.org/pipermail/paraview-developers/2017-June/005429.html
33
EPSILON AXESDONEGUI: Do not hide AeroData while executing EpsilonAxes.A hint was included in the XML file: <Visibility replace_input="0" />
34
EPSILON NEAR FIELDDONEArchitecture: integrate the geometrical operations of 2D/3D pipeline into this filter by using VTK tools
35
DONEOutput: compute integral of forces & momentsintegrated values are available as Field Data. The extrusion lenght is taken into account for 2D, extruded 2D and extruded 3D cases,
36
DONEproblem: wall velocity is not zero after CellDataToPointDataSet V=0 at the extracted surface points. Include option to force non-slip condition (even if input data is point based). Rotational bodies must be taken into account
37
DONEproblem: wall vorticity is half of the real value when computed in Paraview. Check pertinence of the correction factor of 2 implemented in Epsilon V2Vorticity can now be computed from the corrected velocity field with the non-slip condition imposed. However, the factor of 2 still is required and also correct. NFM=CD_Arntz verified with Fluent CASE vs CGNS filesWallVorticityFactor=2 when vorticty is computed by ParaView to take into account the lack of cells on the other side of the wall, which reduces its computed value to the half.
38
DONEOutput: make consistent the dimensionalization/non-dimensionalization of the output integrated values based on Epsilon_Aero_Data setup
39
DONEOutput: Include surface integral of new user defined variables UDV variables from EpsilonAeroData can be loaded. Also, new UDV variables can be computed and integrated.
40
EPSILON WAVE DRAGDONEinterface: Improve interface. It is not that intitive the first time you use it. The condition required to find the wave region must be clearer. Use a new input box for Exergy method. Find a better method for the initial guess of the thresholds. Each method can now be selected independently avoiding any confusion. Also, the clipping thresholds are now easier to interpret since a slidebar was added. The user must reduce the first threshold value and then the second threshold value. So, the bars says explicitly that thresholds must be reduced (it is intuitive) and also shows the step order.
41
DONEupdate: Allow decomposing any parameter. This parameter must be split in two parts (one lying on the viscous zone and the other lying on the wave zone), but the parameter used to breakdown it must be "ARNTZ_Anergy" (for Aguirre method) and "KUSUNOSE_profile_drag" (for Kusunose method)A generic mask is based on (ARNTZ_A+ARNTZ_Aphi) for Aguirre method and (AERO_Vorticity_Mag+KUSUNOSE_profile_drag) for Kusunose method. This mask is then applied to the desired variable.
42
DONEupdate: Deprecate "Absolute/relative" option. Let just absolute values because the relative vorticity or Aphi values are too small
43
DONEupdate: Deprecate "jet" decomposition since a negative drag is not necesarily the jet region due to the mixing region between the jet and the viscous wake
44
DONEProblem: Take into account the variable output from EpsilonAeroDataAn alarm is printed if the user didn't activate Arntz and/or Kusunose in Epsilon_Aero_Data (which is required for Epsilon_Drag_Breakdown)
45
DONEProblem: The viscous and wave regions are not 100% complementary. There are some cells missing somehow…Solved. It was the masking method that was not correct for the viscous part. Now it detects the wave region and then the viscous part is masked with Masking array that is the perfect complement of the domain.
46
DONEProblem: Test breakdown for 2D and 3D. During PhD i found a different behaviour for the two casesTested. No problems found with the new code (maybe because the masking procedure has been modified)
47
DONESlice: Include a slice tool for analyzing 2D/3D data. The breakdown must be implemented either to the slice or to the entire volume (3D)/surface (2D).
48
DONENew module: Include Arntz method to add cells around shockwave for Aw integrationNot implemented. I was considered that the current decomposition of ARNTZ_Aw_dic_MASK_Qcrit is enough and equivalent to the original Arntz methodIf it is implemented in future versions, check if vtkConnectivityFilter is useful. Maybe it is better to use distance filter (vtkImplicitPolyDataDistance) with the distance defined by slide bar from 0m to 0.5m
49
EPSILON INTEGRATION TOOLDONEProblem: Extremely large execution time when integrating volume dataChange VTK filterThe original volume clip used "vtkClipDataSet" because it works in all versions. It was replaced by "vtkTableBasedClipDataSet" (400 times faster)
50
DONEProblem: sweep mode along mesh X-axis. Wrong normal definition. It should use the normal to the upstream flow
51
DONEProblem: differenciate the sweep direction (of the survey plane) from the plane normal direction. This was implemented for sweep plane mode only. For 3D to 2D mapping it wasn't implemented because it difficults the interpretation of the mapped data
52
DONENew capability: Include standard "integrate variables" filter + single slice plane tool, in order to simplify the working pipeline (integrate in a single survey plane)
53
DONEInput data: Read additional arrays --> Add option to let pass all the variables available on the input data (so the user can also use IntegrationTool with that user defined variable created in AeroData).Impose UDV_Sup or UDV_Vol nomenclature in Epsilon_Aero_Data????
54
DONEClip data: for spanwise distributions of wake-reduced formulations, add slice plane and clip by scalar !!! Include option to visualize clipped region before integrating
55
DONEClip data: Define clip direction by using upstream flow direction (Axial sweeps only). Let "Flip Clip" option anyway for spanwise or vertical sweep cases.The user can select the normal = aerodynamic x-axis
56
DONEClip data: Implement Juan's method to compute exergy when survey plane slices a body --> Modified control volume: CV = plano up + plano down + usptream body surface (UBS). In the UBS all the exergy terms are zero except Ep. The UBS must also be used in Momentum Conservation Method. See if it is required by other methodsUse python code structure from Epsilon_Near_FieldOnly Ep must be obtained from this method!!!! The other terms are affected by the non-zero wall velocities. Let this for Epsilon V4
57
DONEAnalysis methods: implement inclined/custom survey plane possibility (using "V.n" instead of "u"). This plane might intersect the fuselageMultiply each variable by (LocalNomal dot x-aero normal). This corrects the local flux depending on the local plane inclination.
58
DONEClip data: Arbitrary integration surface (Attention: all AeroData equations must be in (V.n) format beforehand !!!). Use VTK/STL reader for reference geometry line to be extruded normal to upstream direction.Test XML hint SelectionInput Use Compute Wall Distance + Contour to create a custom slicehttps://kitware.github.io/paraview-docs/latest/cxx/PropertyHints.html
59
DONEOutput: only integrate the parameters that have physical sense
60
DONEGUI option: Add save output in VTP (attention: NOT VTU)
61
DONEVisibility: Do not hide the filter at which Integration Tool is applied
62
EPSILON CIRCULATION TOOLDONECirculation tool: implement computation of circulation by both approaches: velocity curvilinear integral and vorticity flow. Compute it for multiple slices (X or Y axes) for Integrated Circulation Method and axial lift variationVorticity flow is already included when the AERO parameters are integrated spanwise in Epsilon Integration Tool
63
DONELifting line method: include lifting line parameters like lift, induced angle of attack and induced drag distributions. Provide spanwise integrated values as well as the spanwise distributions
64
DONEOutput: make consistent the dimensionalization/non-dimensionalization of the output integrated values based on Epsilon_Aero_Data setup
65
DONEOutput: make consistent with 2D data (for both circulation and Cl)
66
DONEProblem: Slice do not work when the control station is placed outside the CFD volume! Add warning message or fin the way to ignore the STA that lies outside the CFD domain
67
EPSILON POISSON SOLVERDONENew module: Add Green functions for wind tunnel data (wall bounded flows), i.e., image method by layers
68
DONENew module: Add solution by panels method (wall panels method --> see Phd report Meheut)
69
DONENew module: Add ground wall by Green functions
70
DONEGUI option: integrate and print vortex drag
71
DONENew module: add slice tool for CFD data
72
DONENew module: add clip the slice by scalar to get the wake only
73
DONEProblem: solve error with Phi gradient. Prevents induced velocity calculation"algs.gradient(Psi)" failed to compute gradients properly. It was replaced by "vtkGradientFilter"
74
DONEGUI option: add save output in VTK VTP
75
EPSILON WAKE SURVEYDONE
76
DONEGUI: Simplify GUI by allowing deploying branches of GUI depending on user selection (XML code improvement)
77
DONEMesh generation: Local mesh refinement --> apply refinement also to survey line generator and mapped survey plane generator
78
EPSILON 5HPDONEAutomatize: read and process data from input file
79
DONEReal time: add option to process data during WTT run (convert to live source?)
80
EPSILON PIVDONE
81
DONEGUI: Simplify GUI by allowing deploying branches of GUI depending on USER selection (XML code improvement)
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100