tag:forum.featool.com,2006:forum-343Nabble - Suggestions and Feature Requests2024-03-28T10:04:17ZPost suggestions and feature requests here, and if possible include a reason why you think a specific feature would be a good idea to add to the toolboxes.tag:forum.featool.com,2006:post-1555Polymer flow simulation2023-10-23T02:13:23Z2023-10-23T02:13:23Zmhmhussein
Hello please suggest a solver tool for Fluid-Structure Interaction (FSI) research problems for polymer production and elastic Turbulence of viscoelastic fluid polymer models (i.e., Giesekus; Rolie-Poly and FENE-P models).
tag:forum.featool.com,2006:post-1523Re: Being able to remove equations from a model2023-08-20T15:47:05Z2023-08-20T15:47:05ZPrecise Simulation
Thank you for the suggestion, functionality to remove physics modes should be available in the next update.
tag:forum.featool.com,2006:post-1521Being able to remove equations from a model2023-08-17T12:24:44Z2023-08-17T12:24:44Zmsandink
Hello,
<br/><br/>If I'm not mistaken, there is currently no way to remove equations from the equation settings. This has caused considerable confusion, as deactivating equations causes eigenvalues to repeat themselves for a mode analysis I'm trying to do. In that sense this is both a suggestion and bug report. I hope this can be resolved.
<br/>
tag:forum.featool.com,2006:post-1429Re: Transient Euler Bernoulli Problems2023-04-04T03:52:34Z2023-04-04T03:52:34ZPrecise Simulation
Dear Prof Akay,
<br/><br/>Thank you for trying the software, and taking the time to leave constructive feedback. That is very much appreciated.
<br/><br/>It is certainly planned to support hyperbolic equations as well as many other physics equations and systems. Unfortunately with (very) limited resources it takes significant amount of time and effort to implement, test, and validate new features. Hopefully, you're willing to try the software sometime again in the future when we can disappoint a little bit less.
<br/><br/>Thank you again for trying.
tag:forum.featool.com,2006:post-1427Transient Euler Bernoulli Problems2023-03-31T08:20:38Z2023-03-31T08:20:38ZH.U.Akay
I am surprised that we can not run transient Euler-Bernoulli problems with FEATool. There is no time integration solver for the second-order time derivatives (hyperbolic) in the code.
<br/><br/>I tried to use the Custom Equation mode by introducing a new variable forming two parabolic time-dependent equations, as it is done in the tutorials for the wave equation, but the Custom Equation mode does not support Hermite interpolation functions in this case. I am disappointed that the developers of FEATool do not support the solution of hyperbolic time-dependent problems by using the well-known Newmark's scheme.
<br/><br/>The situation is the same with the transient elasticity equations too. You can calculate eigenvalues, but no correct structural dynamics response can be calculated under dynamic loads.
<br/><br/>Hope this deficiency will be somehow removed.
tag:forum.featool.com,2006:post-1335Re: How to get intermediate variables in function solvetime2022-10-20T09:36:22Z2022-10-20T09:36:22ZPrecise Simulation
The built-in solvers currently can't return FEM matrices, however you can access and manipulate the system matrix and load vector using solver hooks
<br/><br/><a href="https://www.featool.com/doc/solver#solver_hooks" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/solver#solver_hooks</a><br/><br/>Alternatively, you can use the assembly functions to define your own custom solver (and thereby access the matrices)
<br/><br/><a href="https://www.featool.com/doc/Classic_PDE_01a_poisson1#tut_pde01_cli3" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/Classic_PDE_01a_poisson1#tut_pde01_cli3</a><br/><br/><a href="https://www.featool.com/doc/assembleprob_8m" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/assembleprob_8m</a><br/><a href="https://www.featool.com/doc/assemblea_8m" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/assemblea_8m</a><br/><a href="https://www.featool.com/doc/assemblef_8m" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/assemblef_8m</a>
tag:forum.featool.com,2006:post-1332How to get intermediate variables in function solvetime2022-10-19T16:47:15Z2022-10-19T16:47:15Zlusongno1
I call ex_navierstokes6 function to solve time-dependent flow around a cylinder and it works fine.
<br/>However, I want to take out the assembled matrix (such as stiffness matrix, mass matrix, etc.) during the time step, but it is embedded in the FEA kernel. I can only see this line of code which call the embedding functions:
<br/>````
<br/>if ~(nargin||nargout),help mfilename,return,end,varargout=cell(1,nargout);[varargout{:}]=featool('feval','solvetime',varargin{:});if ~ nargout,clear varargout,end
<br/>````
<br/>Is there any way I can get these intermediate data? It is important for my research.
tag:forum.featool.com,2006:post-1305Re: Export plots as .pdf file2022-08-11T23:07:32Z2022-08-29T02:23:22ZPrecise Simulation
Thank you for the feedback.
<br/><br/>One way to plot PDFs is to recreate the plot on the Matlab command line, and then use the <a href="https://www.mathworks.com/help/matlab/ref/print.html" target="_top" rel="nofollow" link="external">print command</a> as:
<br/><br/><pre>
print test.pdf -dpdf
</pre><br/>PDF export is now also available in FEATool version 1.16.
tag:forum.featool.com,2006:post-1304Export plots as .pdf file2022-08-09T09:09:10Z2022-08-09T09:09:10ZVengatesan
Dear all,
<br/><br/>The tool is excel in solving and much simpler approach.
<br/><br/>I would like to export the plot as .pdf format in the post processing. Please let me know if anybody having suggestions to do it..
<br/><br/>Thanks in advance
tag:forum.featool.com,2006:post-911Re: Import FEA struct from MATLAB2020-11-10T05:10:09Z2020-11-10T05:10:09Zrandress
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">I think the two step method will not be attempted by users who don't know it, or kind of know what they are doing. You can potentially always write your own script to hijack the buttons and automate things.
</div>
</div></blockquote>
Understood. As as co-worker of mine used to say "if you are dumb enough to do it, you'd better be smart enough to know how" :-)
<br/><br/>Kind regards,
<br/>Randal
tag:forum.featool.com,2006:post-910Re: Import FEA struct from MATLAB2020-11-10T03:25:53Z2020-11-10T03:25:53ZPrecise Simulation
I have of course considered this, but I think it is currently safer not to allow this as it is unclear what could happen if invalid data is found in many stages. I initially had a command terminal where one could execute Matlab code, scripts, and modified the fea data, but I disabled it just in case (as well as the Mathworks not allowing such things for compiled Apps at least).
<br/><br/>I think the two step method will not be attempted by users who don't know it, or kind of know what they are doing. You can potentially always write your own script to hijack the buttons and automate things.
tag:forum.featool.com,2006:post-909Import FEA struct from MATLAB2020-11-09T11:27:05Z2020-11-09T11:27:05Zrandress
Perhaps there is good reason not to.... but could there be a "File" menu option to import the same struct that is exported via the menu "Export FEA struct to MATLAB" option.
<br/><br/>This would seem to accomplish the same thing as the two step process:
<br/><pre>
>> save path_to_my_model.fea fea -v7
</pre>...followed by FEATool Gui "File" menu "Load Model/GUI Script" path_to_my_model.fea
<br/><br/>Perhaps a GUI option would make it too easy to mistakenly "pollute" FEATool with an illegal problem structure. And since there is already a two step method (above) to do this, which by is very nature carries with it a certain amount of caution, it may not be worth additional risk.
<br/><br/>Kind regards,
<br/>Randal
<br/>
tag:forum.featool.com,2006:post-888Re: Multi-boundary selection2020-10-28T04:21:17Z2020-10-28T04:21:17Zrandress
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">Multiple-select has been specifically disabled when having internal subdomains/boundaries active (enabled with only external boundaries), as a user could then possibly select both internal and external boundaries which typically have different number and types of boundary conditions causing a mismatch (the Matlab GUI toolkit does not allow for multiple select by groups).
</div>
</div></blockquote>
...just the kind of mistake I would make :)
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">However, the "Continuity" condition should be the default boundary condition for internal boundaries and should not need to be selected manually.
</div>
</div></blockquote>
I do not think I have always found this to be the case. For example, here is one I just created:
<br/><br/><a href="http://forum.featool.com/file/n888/multiselectboundary.fea" target="_top" rel="nofollow" link="external">multiselectboundary.fea</a><br/><br/>#1-#104 boundaries are the internals:
<br/><br/><img src="http://forum.featool.com/file/n888/multiple_boundary_selection2.png" border="0"/><br/><br/>Note that the default for boundary #1 is "Flux discontinuity".
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">You can of course also manually hack the listbox to allow multiselect with something like ( <a href="https://www.mathworks.com/help/matlab/ref/matlab.ui.control.uicontrol-properties.html#propname_Max" target="_top" rel="nofollow" link="external">https://www.mathworks.com/help/matlab/ref/matlab.ui.control.uicontrol-properties.html#propname_Max</a> )
<br/><br/><pre>
set(findall(0,'Tag','list_seldom'),'Min',0,'Max',2)
</pre><br/>but there is not guarantee that it won't lead to errors in Gui and model, in particular if you happen to select incompatible conditions.
</div>
</div></blockquote>
Thanks for the tip, but I do not think I will be doing this as a general rule (maybe for some special case)... I like the protection provided by inhibiting the multi-selection.
<br/><br/>Kind regards,
<br/>Randal
<br/>
tag:forum.featool.com,2006:post-887Re: Multi-boundary selection2020-10-27T23:50:01Z2020-10-28T01:22:26ZPrecise Simulation
Multiple-select has been specifically disabled when having internal subdomains/boundaries active (enabled with only external boundaries), as a user could then possibly select both internal and external boundaries which typically have different number and types of boundary conditions causing a mismatch (the Matlab GUI toolkit does not allow for multiple select by groups).
<br/><br/>However, the "Continuity" condition should be the default boundary condition for internal boundaries and should not need to be selected manually.
<br/><br/>You can of course also manually hack the listbox to allow multiselect with something like ( <a href="https://www.mathworks.com/help/matlab/ref/matlab.ui.control.uicontrol-properties.html#propname_Max" target="_top" rel="nofollow" link="external">https://www.mathworks.com/help/matlab/ref/matlab.ui.control.uicontrol-properties.html#propname_Max</a> )
<br/><br/><pre>
set(findall(0,'Tag','list_seldom'),'Min',0,'Max',2)
</pre><br/>but there is not guarantee that it won't lead to errors in Gui and model, in particular if you happen to select incompatible conditions.
tag:forum.featool.com,2006:post-886Multi-boundary selection2020-10-27T19:51:26Z2020-10-27T19:51:26Zrandress
It would be convenient to be able to select multiple (especially internal) boundaries which require the same condition and coefficient.
<br/><br/>For example, a model with 104 boundaries is enclosed in a Block.
<br/><br/><img src="http://forum.featool.com/file/n886/multiple_boundary_selection.png" border="0"/><br/><br/>In this case, boundaries 7 through 110 need to be set the same: "Continuity". Each one must be selected and checked, then changed if necessary. But, for example, being able to left-click on number 7, select "Continuity" from the drop-down, scroll down and SHIFT+left-click on 110, then "Apply" would be very nice.
<br/><br/>Kind regards,
<br/>Randal
tag:forum.featool.com,2006:post-885Re: Selection of Objects and/or faces2020-10-27T14:43:37Z2020-10-27T14:43:37Zrandress
This works really great! Thanks,
<br/>Randal
tag:forum.featool.com,2006:post-882Re: Selection of Objects and/or faces2020-10-27T00:03:19Z2020-10-27T00:03:19ZPrecise Simulation
FEATool Multiphysics version 1.13 allows for selecting/de-selecting entire geometry objects with a mouse click, or in 3D selecting/de-selecting individual faces with holding down the <i>Control button</i> and performing a mouse click on a face. Note that mouse selection is not possible is any of the "Zoom-in", "Zoom-out" ,"Pan" or "Rotate 3D" toggle buttons in the upper toolbar are active.
tag:forum.featool.com,2006:post-881Re: Geometry "Delete" tool reversal: UnDelete2020-10-26T23:58:09Z2020-10-26T23:58:09ZPrecise Simulation
FEATool Multiphysics version 1.13 now supports recovering/un-deleting of the "last" deleted object or objects (as in for the geometry object(s) the last delete operation was performed on).
<br/><br/>To such recover deleted objects first ensure no geometry objects are selected or highlighted (in the <i>Selection</i> listbox) and then press the <b>Revert/Undo</b>. Press "Yes" at the question prompt asking whether the last deleted objects should be restored.
<br/><br/><a href="https://www.featool.com/doc/geom.html#op_undo" target="_top" rel="nofollow" link="external">https://www.featool.com/doc/geom.html#op_undo</a>
tag:forum.featool.com,2006:post-829Selection of Objects and/or faces2020-10-04T23:30:39Z2020-10-04T23:30:39Zrandress
The "Geometry Objects" list is essential in being able to locate and select the desired objects (and now faces). However, I also liked being able to select using mouse-over-and-left-click that appeared with v1.13 beta2.1 (and disappeared with beta3 :-) A selected object remains selected as others are selected. Clicking on an already-selected item de-selects that item. Being able to select Objects and faces in this manner is quite convenient when possible (when the items are not hidden by objects that display "over" or "on top of" them).
<br/><br/>The only thing I was going to comment on (in beta2.1) was that when these selections were made using the mouse, the corresponding item in the "Geometry Objects" list was not selected. This was confusing.
<br/><br/>One problem I anticipate is how to distinguish between selection of object vs. face. Perhaps let "click" designate face and Crtl+Click designate object. Then it might work as follows:
<br/><pre>
- A click on an un-selected face selects the face (adds to set of selected items)
- A click on a selected face de-selects the face (but leaves undisturbed any existing object selection)
- A ctrl+click on an un-selected object selects the object (adds to set of selected items)
- A ctrl+click on a selected object de-selects the object (but leaves undisturbed any selected faces)
</pre><br/>But (unless that same method were extended to the method of selection within the "Geometry Objects" list, then the use of "Ctrl" to distinguish between face and object might be confused with the use of ctrl when making multiple selections (adding to items already selected)...but I think it might be workable because in the "Geometry Objects" list it is clear that you are selecting either the Object or one of its faces, so using the Ctrl+click to add (or subtract if clicking on an already selected item) in the usual way would make sense.
<br/><br/>An additional benefit would be that the [Ctrl+]click could be used to idenfity the object or face tag of an area of the model where it is hard to distinguish objects and/or faces.
<br/><br/>Just an idea.
<br/><br/>Kind regards,
<br/>Randal
<br/><br/><br/><br/><br/>
tag:forum.featool.com,2006:post-820Geometry "Delete" tool reversal: UnDelete2020-10-01T20:31:33Z2020-10-01T20:31:33Zrandress
The Geometry mode's "Revert/Undo" tool reverses <b>any operation</b> on an object that has been previously modified or transformed (The creation of an object can't be "undone" - but it may simply be deleted :-).... <b>any operation</b>, that is, except "Delete".
<br/><br/>An Undo-delete tool would be convenient. If the creation and modification/transformation operations and parameters of a deleted object have not been recorded, re-creating it could be frustrating, if doable at all. It is true that if one of the "File-->Save As ..." operations happened to have been used before the deletion, the deleted object's pedigree and specifications may have been preserved. Still, it would be convenient and stressless to just undo an accidental or hasty mis-step.
<br/><br/>If that is a not easily done in the near term, then possibly a "do you really mean to delete this object" dialog might be a good alternative. Since some users (or any user from time to time) might consider this reminder to be an aggravation, it should include a "don't show me this again" check box or sub-option.
<br/><br/>If the Undo-delete tool <b>were</b> added, several functional alternatives come to mind:
<br/><br/>1) an "Undo-last-operation" (only good as an un-delete <b>if</b> "Delete" were the last operation done). To clarify, I am not suggesting that this function would be able to reverse-build a model from finish to start, but that it would always and only be capable of a single operation: to undo the very last operation on a geometry object. The undoing of operations previous to the immediately preceding one would not be possible.
<br/><br/>2) an "Un-delete" that restored the last deleted object (regardless of what object it was or when it had been deleted),
<br/><br/>3) an "Un-delete-designated-object" that restored a deleted object designated by its tag (e.g., "E3"). Since a user my not remember the tag of the deleted object, this function should probably prompt the user with a list of tags of deleted objects (from last deleted to first deleted).
<br/><br/>Note: The parent objects that are removed when transformed, such as "B1" and "C1" in the transformation"B1+C1==>TF1", are not intended to be considered "deleted" objects in the above function descriptions.
<br/><br/>Kind regards,
<br/>Randal
<br/><br/>PS A third alternative to "UnDelete"or a "Delete Warning Dialog" might be a frequent save to a temp geometry or .fea file which could be viewed at any time and reverted to if desired. The frequency and location (and name perhaps) could be configurable.
<br/><br/><br/>
tag:forum.featool.com,2006:post-808Default ellipsoid axis vector2020-09-26T10:39:39Z2020-09-26T10:39:39Zrandress
The dialog for the creation of the geometry object ellipsoid includes an "axis" field that requires an [x y z] vector which is defaulted to [1 0 0].
<br/><br/><img src="http://forum.featool.com/file/n808/default_ellipsoid_dialog.png" border="0"/><br/><br/>The default settings produce the following:
<br/><br/><img src="http://forum.featool.com/file/n808/default_ellipsoid_axis_1_0_0.png" border="0"/><br/><br/>The "axis" vector determines the orientation of the ellipsoid. For these default settings, this vector [1 0 0] is the axis of the largest radii (z, in this case). However, it seems to me that the more intuitive default setting for this "axis" would be [0 0 1] so that the x, y, and z radii settings would be seen/measured along the x-, y-, and z-axis:
<br/><br/><img src="http://forum.featool.com/file/n808/modified_ellipsoid_axis_0_0_1.png" border="0"/><br/><br/>The ellipsoid can then be oriented as needed using the "axis" vector:
<br/><br/><img src="http://forum.featool.com/file/n808/reoriented_ellipsoid_axis_1_2_1.png" border="0"/><br/><br/>Maybe I am missing the significance of the "axis" vector... but until I figured it out, I was struggling to design an ellipsoid with the x/y/z/ dimensions that I intended.
<br/><br/>Kind regards,
<br/>Randal
<br/><br/><br/><br/><br/><br/><br/><br/><br/>
tag:forum.featool.com,2006:post-548Option to disable automatic postprocessing following solution2020-06-29T20:46:59Z2020-06-29T20:46:59Zrandress
I should say at the outset that this is a small thing. But it is a small thing that many might appreciate :-)
<br/><br/>It seems (so correct me, please) that the last postprocessing which was done on the current problem or on the previous problem solved is automatically done for the current problem immediately following the execution of the Solve step. For problems and plots which plot quickly this is not really an issue - after it is done, postprocessing can be set as desired and re-run.
<br/><br/>However, for some problems/plots it can take long enough to be annoying, especially when thinking (as I do - correct me if I am wrong) that it is performing unnecessary processing.
<br/><br/>I think I have discovered a crude workaround in creating an "fea" file "template" that has already set the Postprocessing Settings to not specify any plot at all. If I begin a problem by opening this file then the initial Solve execution does not do a plot but reports an error that there is nothing to plot :-)
<br/><br/>On subsequent re-solves if I again specify no plot before solving, then no plot is made and I am notified of the fact.
<br/><br/>I think what I would suggest is an option (not sure where) that, if selected, would display the Postprocessing Settings but would not generate the plot or simply place the Gui in he post processing mode.
<br/><br/>As I said this is not a biggie. Just a thought.
<br/><br/>Kind regards,
<br/>Randal
tag:forum.featool.com,2006:post-518Re: Slice-plane tool in Geometry Mode2020-06-20T08:34:11Z2020-06-20T08:34:11Zrandress
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">Thank you for the suggestion Randall. I am working on updating the geometry engine with support for more CAD tool functionality and will eventually get around to incorporating a proper slicing tool, but it might take some time for this to be implemented and made available.
</div>
</div></blockquote>
Good to hear.
<br/><br/>I must confess that what prompted me to suggest this was the trouble I was having with the dup&sub technique as reported here ( <a href="http://forum.featool.com/Adding-an-interior-boundary-by-duplication-and-subtraction-tp506.html" target="_top" rel="nofollow" link="external">http://forum.featool.com/Adding-an-interior-boundary-by-duplication-and-subtraction-tp506.html</a> )
<br/><br/>Now that you have explained that my issue there was not my inability to properly create the boundary AND have provided a work around, this feature does not seem so urgent any more :-)
<br/><br/>Kind regards,
<br/>Randal
tag:forum.featool.com,2006:post-514Re: Slice-plane tool in Geometry Mode2020-06-19T02:33:55Z2020-06-19T02:33:55ZPrecise Simulation
Thank you for the suggestion Randall. I am working on updating the geometry engine with support for more CAD tool functionality and will eventually get around to incorporating a proper slicing tool, but it might take some time for this to be implemented and made available.
tag:forum.featool.com,2006:post-508Slice-plane tool in Geometry Mode2020-06-18T09:48:43Z2020-06-18T09:48:43Zrandress
As explained in another thread ( <a href="http://forum.featool.com/Integration-of-Current-Density-over-Slice-Intersection-tp180p182.html" target="_top" rel="nofollow" link="external">http://forum.featool.com/Integration-of-Current-Density-over-Slice-Intersection-tp180p182.html</a> ) currently, a planar slice cannot be added during post processing such as might be used for integration:
<br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message">Boundary integration is only available on defined (external and internal) boundaries (not general cut planes). So in your case you would have to split the cylinders in half (or at the plane you want to perform the boundary integration). This should after meshing create internal boundaries that you now can select in the "Boundary Integration" dialog box (Switch to "Boundary" mode to see the boundary numbering, and toggle the "Interior Boundaries" check box in the lower left hand side of the "Boundary Settings" dialog box to see the numbering of internal boundaries).
<br/><br/>The reason for this (limitation) is that boundary integration is for both practical and accuracy reasons performed on selections of grid cell faces (or edges in 2D). And the way to ensure the grid generation algorithm does indeed mesh a reasonably smooth plane/boundary is to first define it in the geometry.
</div>
</div></blockquote>
In another thread ( <a href="http://forum.featool.com/Export-of-STEP-geometry-to-Gmsh-cut-with-plane-and-re-import-STEP-tp423p426.html" target="_top" rel="nofollow" link="external">http://forum.featool.com/Export-of-STEP-geometry-to-Gmsh-cut-with-plane-and-re-import-STEP-tp423p426.html</a> ) a method is outlined whereby the equivalent functionality may be accomplished in the Geometry mode:
<br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Precise Simulation wrote</div>
<div class="quote-message shrinkable-quote">If your desired cutplane is aligned with the x, y, or z axis you should already be able to do this with the existing capabilities, as for example:
<br/><br/>1) Make an in-place copy of the object you want to split using the "Copy/Transform Object..." button or menu item.
<br/><br/>2) Create two blocks that touch along the plane you want to split, but extend a bit beyond the object in the other 5 directions/planes.
<br/><br/>3) Subtract one block from the orginal object, and the other from the copy.
<br/><br/>4) You should now be left with two halves of the original object split along the plane where the two blocks touched.
</div>
</div></blockquote>
Could this technique be implemented by a tool added to the Geometry Mode menu which would, for example, provide for the definition of a plane's perpendicular vector (x, y, z) and a contained point (x, y, z) which would then be used to bisect any/all selected objects? Perhaps a similar tool would be appropriate for the 2D and 1D modes.
<br/><br/>Kind regards,
<br/>Randal
tag:forum.featool.com,2006:post-394Re: Suggestions for future release2020-04-27T19:41:40Z2020-04-27T19:41:40ZPrecise Simulation
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">tkdiff wrote</div>
<div class="quote-message">Where would I enter the commands you list to make this happen?
</div>
</div></blockquote>
The commands should be entered in to the Matlab "Command Window" if you are running FEATool as a <i>Matlab toolbox</i>.
<br/><br/>Note that the <i>stand alone</i> version of FEATool does not support command line interface (CLI) use.
tag:forum.featool.com,2006:post-393Re: Suggestions for future release2020-04-27T13:29:11Z2020-04-27T13:29:11Ztkdiff
That is interesting. Where would I enter the commands you list to make this happen?
tag:forum.featool.com,2006:post-384Re: Suggestions for future release2020-04-26T19:17:02Z2020-04-26T19:17:02ZPrecise Simulation
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">tkdiff wrote</div>
<div class="quote-message">2. I'm aware I can export the model to Matlab and control axis display from there, and have done that, but it would be nice to be able to do something in your app as well, in 1D as you say. The options only constrain the data displayed, and don't change the axes, which is what I'm seeking. Not a high priority, but would be nice.
</div>
</div></blockquote>
Thank you for elaboraing on your suggestions, regading the axis although not as convenient as built-in buttons, if you really want to change the GUI axis there is another workaround by extracting and directly manipulation the corresponding Matlab graphics handles, for example (assuming you have the FEATool GUI up with a plot)
<br/><br/><pre>
h_ax = findall(0,'type','axes');
set( h_ax, 'XScale', 'log', 'YScale', 'log' ) % Set log-log scale.
xmin = 0; xmax = 1; ymin = 0; ymax = 1;
axis( h_ax, [xmin xmax ymin ymax] ) % Set axis limits.
</pre>
tag:forum.featool.com,2006:post-380Re: Suggestions for future release2020-04-26T13:24:18Z2020-04-26T13:24:18Ztkdiff
Just to clarify my thoughts per your response:
<br/><br/>1. When I start the app (v1.12.1), a splash screen appears while its loading with nothing to check, after the app is loaded, the splash screen changes to the one requiring that you accept or decline the license. In previous versions, only the license accept splash screen appeared.
<br/><br/>2. I'm aware I can export the model to Matlab and control axis display from there, and have done that, but it would be nice to be able to do something in your app as well, in 1D as you say. The options only constrain the data displayed, and don't change the axes, which is what I'm seeking. Not a high priority, but would be nice.
<br/>Regarding the verbosity of the command log, it no longer slows down the calculation since you have fixed that, but additional control over what is output to the log would be nice. Again, not high priority.
<br/>Thanks,
<br/>Tom
tag:forum.featool.com,2006:post-374Re: Suggestions for future release2020-04-24T06:42:59Z2020-04-24T06:42:59ZPrecise Simulation
Thank you for the suggestions Tom,
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">tkdiff wrote</div>
<div class="quote-message">1. Is it necessary for the splash screen asking to decline or accept the license for licensed customers? I'm wondering if it could be eliminated if you have paid.
</div>
</div></blockquote>
Yes certainly, it actually worked like that before, but it might have been lost amongst other changes.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">tkdiff wrote</div>
<div class="quote-message">2. Add the ability in postprocessing to permit modifications to the axes, such as semilog, log, scale, label, etc. I have found that if you change the pp value at all, it throws the plot off screen, and you cannot change the axes or display to retrieve it. For example, if you run a problem using large numbers and want to plot semilog instead of linear, you cannot do this
<br/><br/>3. If possible, an option to control/limit the output to the command log, to decrease verbosity of the output if desired. If repeating time step calculations, the log gets quite large.
</div>
</div></blockquote>
I'm aware of the issues you have had with 1D simulations and will certainly try to improve it when I get a chance. In the meantime you can always use the command line interface to customize your simulations. Regarding the axis you can also use the "Options" menu to manually set the axis limits, that might also work better at the moment than the "automatic" axis settings for 1D problems.
tag:forum.featool.com,2006:post-373Re: 3 suggestions (display while computing / periodic b.c. / superimpose grid)2020-04-24T06:37:38Z2020-04-24T06:37:38ZPrecise Simulation
Thanks again for the suggestions.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message">1. Boundary settings dialog box : Add « periodic boundary conditions » (for a couple of faces or edges of equal size)
</div>
</div></blockquote>
Unfortunately, periodic boundary conditions are not quite easy to implement and automate in the general cases as one must have some kind of intermediate grids to map between the boundaries (as they rarely match perfectly up in higher dimensions). But it is certainly something that is planned for the future.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message">2. During a time-dependent computation, it would be nice if the toolbox could display the current solution at each time step: if something wrong or unwanted happens during the computation we can decide to abort it before the end.
</div>
</div></blockquote>
Yes, some more solution monitoring in various forms are on the roadmap.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message">3. Postprocessing: option to superimpose the grid to the displayed field.
</div>
</div></blockquote>
At the moment you can export the "fea" data struct to Matlab, and call the function:
<br/><br/><pre>
plotgrid(fea)
</pre><br/>this should essentially overlay the grid on the current plot. As is often the case, using the Matlab programming and scripting functionality you can achieve most things that are not possible in the GUI (the GUI will naturally always be more limited than the CLI interface).
tag:forum.featool.com,2006:post-372Suggestions for future release2020-04-23T17:32:15Z2020-04-23T17:32:15Ztkdiff
Here are a few suggestions for future updates:
<br/><br/>1. Is it necessary for the splash screen asking to decline or accept the license for licensed customers? I'm wondering if it could be eliminated if you have paid.
<br/><br/>2. Add the ability in postprocessing to permit modifications to the axes, such as semilog, log, scale, label, etc. I have found that if you change the pp value at all, it throws the plot off screen, and you cannot change the axes or display to retrieve it. For example, if you run a problem using large numbers and want to plot semilog instead of linear, you cannot do this
<br/><br/>3. If possible, an option to control/limit the output to the command log, to decrease verbosity of the output if desired. If repeating time step calculations, the log gets quite large.
<br/><br/>Thanks,
<br/>Tom
tag:forum.featool.com,2006:post-3703 suggestions (display while computing / periodic b.c. / superimpose grid)2020-04-23T05:38:47Z2020-04-23T05:38:47ZFrederic Moisy
Hi,
<br/><br/>Here are 3 new suggestions for future releases:
<br/><br/>1. Boundary settings dialog box : Add « periodic boundary conditions » (for a couple of faces or edges of equal size)
<br/><br/>2. During a time-dependent computation, it would be nice if the toolbox could display the current solution at each time step: if something wrong or unwanted happens during the computation we can decide to abort it before the end.
<br/><br/>3. Postprocessing: option to superimpose the grid to the displayed field
<br/><br/>Thank you!
<br/>Frederic
tag:forum.featool.com,2006:post-347Re: Suggestions for grid and postprocessing2020-04-21T01:08:18Z2020-04-21T01:08:18ZFrederic Moisy
Thank for your answers and for the new forum subject, which I am happy to inaugurate!
Let me clarify my suggestion regarding the "Refine" option in the grid mode.
Suppose that you have a 1x1 square domain. The default grid size is 0.1, so that I start with 10 meshes along each side of the square. If I click on "Refine", it computes and display a new grid of size 0.05 (20 meshes), but the "Grid size" field still indicates "0.1", which is strange. It would be more natural (at least to me) if, when clicking on "Refine", it automatically updates the "Grid size" field to its new value.
Regarding the "Next/Prev" buttons suggestion: Having them directly in the main window instead of the "Postprocessing" dialog box would be even more convenient! And... adding "play/pause" buttons, to enjoy the movie, like in standard movie players?
Frederic
tag:forum.featool.com,2006:post-344Re: Suggestions for grid and postprocessing2020-04-20T21:04:14Z2020-04-20T21:04:14ZPrecise Simulation
<blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message">Hi, I am a new user and enjoy very much the toolbox. After a few days of use (with the Navier-Stokes solver only), I would like to make the following suggestions for future releases
</div>
</div></blockquote>
Thank you for your kind words and feedback. I have created a new area for suggestions.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message"> - in "Grid" mode, when clicking on "Refine", the displayed value in "Grid Size" (and the location of the cursor) is not updated.
</div>
</div></blockquote>
I'm not quite sure what the issue with the cursor location is here? The "grid size" is only used when calling a mesh generator (using the "Generate" button), while "refine" simply splits all edges in half (irrespective of "grid size"). So they are not really connected. If you want to know the mesh statistics, min/max/average sizes etc. you can select <a href="https://www.featool.com/doc/gridstat_8m.html" target="_blank" rel="nofollow" link="external">"Grid Statistics..."</a> from the "Grid" menu.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message"> - in "Postprocessing" Settings: it would be nice to add a "Scale Factor" for the arrow plot
</div>
</div></blockquote>
Yes, this is a good idea.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message">- Display streamlines
</div>
</div></blockquote>
Sure, this is somewhere on the list but requires some more functionality to integrate the paths other than basic expression evaluation.
<br/><br/><blockquote class="quote dark-border-color"><div class="quote light-border-color">
<div class="quote-author" style="font-weight: bold;">Frederic Moisy wrote</div>
<div class="quote-message"> - "Next/Prev" buttons or cursor to easily nagivate through time-dependent solutions (the "Solution" selection in the Postprocessing dialog box is not very convenient for this...)
</div>
</div></blockquote>
This is a good idea, I like it although I would probably implement it as buttons around the "Apply" button in the "Postprocessing Settings" dialog box (as you can always make it smaller and just show the last row of buttons), something like:
<br/><br/><pre>
[OK][<][Apply][>][Share][Cancel]
</pre><br/>Where </> would let you step between solutions if available (as the functionality is essentially "apply" but with another selected solution).
<br/>