ilion.
Then i sketched the outline curves in rhino with a few control points. The building is symetric so i only draw one side. But i'm not sure what is better for a voroni. a sharp or a soft surface? Or dose i need points?
So i have some questions:
1. how can i loft the curves correctly? My problem is that if i divide my curves for more control points, grasshopper automatically change my curve. thats ok but than i've the problem with a short curve, which fit bevor with the large one, but after the devision it can't connect.
So i tryed to duplicate the long curve and split it but with the shatter battery it dosen't work. It always cut the curve somewhere.
2. my next problem is, the curves in rhino should be my main construction, which is always visible. so i decided to offset the curves that i got a colum. but i don't know how to orient the offset curves in the xyz axis.
3. hopefully if i have the surfaces, how can i build a voroni which is offsetet, and has maybe some different thicknesses? :D
Would be really great if s.o. can help me. I tried a lot but not every thing is simple.
Sorry for my bad english.
Thx max
Here are my files:
FCP_MAX_GH_konstruktion_1.3dm
FCP_MAX_GH_konstruktion_1.gh
…
iangular element properly discretizes the area continuum forces, so they are independent of meshing density, unlike simply using a network of 1d springs.
The warp and weft stresses can also be set separately allowing greater control of the shape (making them equal will give minimal surfaces).
Because the soap film elements alone do not have any in-plane stiffness, it can often be useful to have some spacer elements to keep the nodes well distributed.
Also, if mesh edges follow geodesics on the surface, it helps keep the strips straight when unrolled, allowing more efficient use of material.
The G-string component can be used for both - keeping the nodes well spaced, and aligning edges with geodesics. It pulls each node toward a combination of its neighbours, but taking only the part of the force tangential to the surface, so it does not interfere with the shape of the surface, only affecting how the nodes are distributed on it.
The "GeoIndex" input lets you choose which neighbours will be used here. In this example, a quad mesh is used, and index 0 and 2 give the neighbours in the warp direction, while 1 and 3 are in the weft direction. Note that it is the triangulation of this quad mesh that is used for the actual soap-film elements.
There is also a "spacing" option. If this is true, the nodes will try and space out evenly along the geodesic, while if it is false, only the direction is affected. In this example it is set to false for the warp and true for the weft.
The example also includes use of the stripper and unroller components to get the flat strips. I have shown the result of splitting in either direction, and as you can see, only one of these is straight.
Finally, if all of this sounds overly complex, don't worry - for quick studies you can still use the simpler approach of just turning all edges of a mesh into springs, and provided you have a decent starting mesh, the result will be very similar to using the 2d element method given here. This is just provided for those that want to take things to greater degree of accuracy and further towards fabrication.…
s that I just can't quite figure out.
Zoom in of the lower left hand corner...
1. First off, I am not sure why two versions of the rectangles are showing, the original and the scaled versions from the image. This doesn't appear to be affecting my final results, so not a big deal, but would like to understand why there are two and get rid of the original rectangles if possible.
2. I would like to change the scaling factor of the the rectangles in the Y direction. They are currently scaling too small based on the image I provided. Is there a way to set a lower value minimum so that the rectangles are not quite so small. Please not that I am only wanting to scale the rectangles in the Y direction. I want the X to currently remain constant.
3. Lastly, I would like to offset the grid points in the y direction randomly so that the image does not seem so gridded. These should shift no more than 3/8 of the distance between the grid from the center point.
Desired Result:
Any help is greatly appreciated!…
Added by Josh Sawyer at 10:48am on September 17, 2017
- nickname is rather the best approach - and not on active group, but that's irrelevant anyway).
Step back (assuming that you are talking about the "Tens_from_random_blah_blah" definition):
1. Engineering is the art of demystifying (or we are promising that anyway, he he). This means that you start defining (better: outlining) some topology for things based on some "generic" rules (like the ones applied for the masts,cables,cones etc etc). These things are kept in some kind of structure (Lists, DataTrees etc). Things are few in 99.99999% of cases (i.e. : even the biggest membrane "module" has, say, 20-50 masts per "module").
2. Then ... handling things "individually" (mostly modifying) becomes the most critical part. See this (an x "possible" solution by combining a myriad of "options" : a no cones membrane solution, in plain English):
3. But the above is impossible (for more than obvious reasons). You should deploy masts in some high/low sequence in order to achieve some meaningful convex/concave formation that could work.
4. This "works" : 5. This doesn't:
6. This works partially (the formation at the back is "flat" == undo able):
7. This is utterly kitsch (and faulty as the case6 - the back portion):
So it's quite obvious that without a (quite complex) capability to individually control things (in this occasion : mast heights) the whole definition is a waste of computer time. Additionally the more the solution is "demystified" (some curve is defined, some random points are created, some masts are in place, some cables appear etc etc) the more additional constrains are required in order to "narrow" the possibilities (In plain English : sliders should control other sliders as regards their min/max values, true/false, you/me etc etc).
Remember that we are talking about ONE (mast height) out of a myriad things that you should control "manually" (it's utterly pointless to mastermind some kind of "generic" rules - or use naive attractors etc etc) .You'll see the difference when I'll completely reform the definition by adding individual control upon anything.
PS: what about the blocks? (the real life stuff that actually make any solution possible). Can you imagine a 2nd set of "restrictions" imposed by "a child to his parent"? (Assembly/Component modeling , that is).
more soon
…
uick answers. Below you will find some suggestions, but don't think of them as rules and especially don't think of them as guarantees.
1. Choose a descriptive title for your post
Don't call your question "Help!" or "I have a problem" or "Deadline tonight!", but actually describe the problem you are having.
2. Be succinct but clear in your wording
People need to know some details about your problem in order to understand what sort of answers would satisfy you, but nobody cares about how angry your boss or how bad your teacher or how tight your deadline is. Talk about the problem and only the problem. If you don't speak English well, you should probably post in your native language as well as providing a Google Translation of your question.
3. Attach minimal versions of all the relevant files
If you have a GH/GHX file you have a question about, attach it to the post. Don't expect that people will recreate a file based on a screen-shot because that's a lot of pointless work. It's also a good idea to remove everything non-essential from a GH file. You can use the 'Internalise Data' menu option to cut everything to the left of a parameter:
If you're importing curves or Breps or meshes from Rhino, you can also internalise them so you won't have to post a 3DM file as well as a GH file. If you do attach large files, consider zipping them first. Do not use RAR, Ning doesn't handle it.
It is especially a good idea to post files that don't require any non-standard components if at all possible. Not everyone has Kangaroo or Hoopsnake or Geco installed so if your file relies on those components, it might not open correctly elsewhere.
4. Include a detailed image of the GH file if it makes sense
If your question is about a specific (group of) components, consider adding a screenshot of the file in the text of the post. You can use the Ctrl+Shift+Q feature in Grasshopper to quickly create nice screenshots with focus rectangles such as this:
5. Include links to online resources if possible
If you have a question about Schwarz Minimal surfaces, please link to a website which talks about these.
6. Create new topics rather than continuing old ones
It's usually better to start a fresh question, even if there's already a discussion that kinda sorta tangentially touches upon the same issue. Please link to that discussion, but start anew.
7. This is not a 'do my work for me' group
Many of us like to help, but it's good to see effort on our part being matched by effort on your part. Questions in the form of 'I need to do X but cannot be bothered to try and learn the software' will (and should) go unanswered.
7b. Similarly, questions in the form of 'How do I quickly recreate this facade that took a team of skilled professionals four months to figure out?' have a very low success rate.
--
David Rutten
Lead Grasshopper Development
Robert McNeel & Associates…
Added by David Rutten at 12:58pm on October 1, 2013
: ----------------------------------------------------------------------------------------------
1)
Hi Clemens I've analysed a plate structure using Karamba and wanted to do a convergence analysis on results computed as a function of the number of elements.
Now, when strictly looking at the result magnitudes of internal energy (IE) and maximum displacement (w_max), it's acceptable, that their relative deviations are very small. But I cannot explain the tendencies of their graphs. From what I know, FEM should always compute underestimated results when compared to analytical solutions. So I don't understand why both the IE and w_max seem to be decreasing for an increasing number of elements.
But my main concern is the behaviour of the peak moment, it seems to be simply hill climbing untill suddenly a singularity kicks in. I initially wanted to use the peak moment as a fitness value for optimisation, but with this behaviour, I don't think that would make sense. I've attached my GH file as well.
It would be much appreciated if you could enlighten me on these subjects. Cheers Daniel Andersen
2)
Hi Daniel,
I could not run your definition because I have not all the plug-ins installed that you use.
You are basically right that the displacement should increase with a finer mesh. However the result of the shell analysis also depends on the shape of the triangles (well formed vs. very distorted). In order to test this, I think it would be interesting to use a very simple example (e.g. rectangular plate with one column) where you can easily control mesh generation. Would you like to start a discussion on this in the karamba group at http://www.grasshopper3d.com/group/karamba?
It is not a good idea to use the bending moment at a singularity for optimization because the result will be heavily mesh dependent. Also real columns do have a certain diameter and modeling them as point supports introduces an error.
Best,
Clemens
3)
oh, and by the way!
Here's some relevant literature on handling peak moments: https://books.google.dk/books?id=-5TvNxnVMmgC&pg=PA219&lpg=PA219&dq=blaauwendraad+plates+and+fem&source=bl&ots=SdDcwnrSA1&sig=6HulPmKNIhqKx4_rGxitteMC4CU&hl=da&sa=X&ved=0CDEQ6AEwA2oVChMIg66k0LPaxgIVgY1yCh1KPAeY#v=onepage&q=chapter%2014&f=false (Blaauwendraad, J., 2010. Plates and FEM : Surprises and Pitfalls, see Chapter 14) It would be great if a feature dealing with peak moments could be incorporated in Karamba. In my work, I ended up exporting my models to Robot in order to verify the moment values. Best, Daniel
4)
Hi Daniel,
thank you for your reply and the link to Blaauwendraads excellent book!
At some point I hope to include material nonlinearity in Karamba which will help in dealing with stress singularities.
If you want you could open a discussion with a title like 'moment peaks in shells at point-supports'. Then we could copy and paste the text of our conversation into it.
Best,
Clemens
----------------------------------------------------------------------------------------------…
ight be able to provide more insight). Whenever you run a new simulation in Radiance, it is not always necessary to re-write all of the initial simulation files from scratch. These initial simulation files include both a .rad geometry file as well as a separate .pts file that contains the test point locations. If all that you are changing in a given parametric run is the locations of the test points (like your case), it is not necessary to re-write (or reinterpret) the entire .rad geometry file. My guess is that there is some type of check for this built into either code Mostapha wrote or radiance functions that Mostapha is calling. As such, it seems that the rad geometry file is not being re-written (or re-interpreted by radiance) completely when all that you change is the test points and this actually seems to be saving you an extra 10 seconds each time that you run the component without changing the materials or the building geometry. Other times (like when you plug in custom radParameters), it seems that it re-writes (or re-interprets) the .rad geometry file from scratch since this file is probably affected by customized rad parameters.
So far, if this explanation is holding, it seems like there would be no concern on your end but I also recognize that the difference between these long and short simulations is giving you radiation results that are ever so slightly different from each other (by my estimates, they differ by about 0.2%). Compared to the other types of assumptions that the radiance model is making, though, these are mere rounding errors that probably originate from the number of decimal places in the vertices of the rad geometry file. Rather than worrying about whether your simulations are giving you the right rounding errors to give you matching results, I would encourage you to instead contemplate how much your radiance results are matching reality given all of the assumptions that you are making about the climate (with the epw file for a "typical" year) and with the number of light bounces in the radiance simulation. To give you an example, I ran your model with a higher quality of simulation type (3 ambient bounces) and this gives you results that differ by 1.1% from the original simulation that you were running with only 2 ambient bounces (this is practically an order of magnitude larger than 0.2%).
To address your unease I will say that, for a long time, I also felt uneasy any time that I encountered something that seemed unpredictable in software that I was using. Once I started coding my own stuff, though, I realized quickly that unpredictable behavior is an unavoidable aspect of all software. There is always a tradeoff between accurate results and the time it takes to get them, which produces a multitude of possible ways to arrive at a solution. Add into this complex situation the fact that you might have an almost infinite number of possible inputs to a given set of code.
Because of the unpredictable multitude of cases, there is no application that is completely free from limitations and assumptions. In this light, what ends up being more important than the actual calculation method used is the social infrastructure that is in place to help understand what is being run under the hood, hence why both Radiance and Honeybee are open source and why we try to build a robust community of support through forums like this one!
-Chris…
lly it should not make much of a difference - random number generation is not affected, mutation also is not. crossover is a bit more tricky, I use Simulated Binary Crossover (SBX-20) which was introduced already in 1194:
Deb K., Agrawal R. B.: Simulated Binary Crossover for Continuous Search Space, inIITK/ME/SMD-94027, Convenor, Technical Reports, Indian Institue of Technology, Kanpur, India,November 1994
Abst ract. The success of binary-coded gene t ic algorithms (GA s) inproblems having discrete sear ch sp ace largely depends on the codingused to represent the prob lem variables and on the crossover ope ratorthat propagates buildin g blocks from pare nt strings to childrenst rings . In solving optimization problems having continuous searchspace, binary-co ded GAs discr et ize the search space by using a codingof the problem var iables in binary st rings. However , t he coding of realvaluedvari ables in finit e-length st rings causes a number of difficulties:inability to achieve arbit rary pr ecision in the obtained solution , fixedmapping of problem var iab les, inh eren t Hamming cliff problem associatedwit h binary coding, and processing of Holland 's schemata incont inuous search space. Although a number of real-coded GAs aredevelop ed to solve optimization problems having a cont inuous searchspace, the search powers of these crossover operators are not adequate .In t his paper , t he search power of a crossover operator is defined int erms of the probability of creating an arbitrary child solut ion froma given pair of parent solutions . Motivated by t he success of binarycodedGAs in discret e search space problems , we develop a real-codedcrossover (which we call the simulated binar y crossover , or SBX) operatorwhose search power is similar to that of the single-point crossoverused in binary-coded GAs . Simulation results on a number of realvaluedt est problems of varying difficulty and dimensionality suggestt hat the real-cod ed GAs with t he SBX operator ar e ab le to perform asgood or bet t er than binary-cod ed GAs wit h t he single-po int crossover.SBX is found to be particularly useful in problems having mult ip le optimalsolutions with a narrow global basin an d in prob lems where thelower and upper bo unds of the global optimum are not known a priori.Further , a simulation on a two-var iable blocked function showsthat the real-coded GA with SBX work s as suggested by Goldberg
and in most cases t he performance of real-coded GA with SBX is similarto that of binary GAs with a single-point crossover. Based onth ese encouraging results, this paper suggests a number of extensionsto the present study.
7. ConclusionsIn this paper, a real-coded crossover operator has been develop ed bas ed ont he search characte rist ics of a single-point crossover used in binary -codedGAs. In ord er to define the search power of a crossover operator, a spreadfactor has been introduced as the ratio of the absolute differences of thechildren points to that of the parent points. Thereaft er , the probabilityof creat ing a child point for two given parent points has been derived forthe single-point crossover. Motivat ed by the success of binary-coded GAsin problems wit h discrete sear ch space, a simul ated bin ary crossover (SBX)operator has been develop ed to solve problems having cont inuous searchspace. The SBX operator has search power similar to that of the single-po intcrossover.On a number of t est fun ctions, including De Jong's five te st fun ct ions, ithas been found that real-coded GAs with the SBX operator can overcome anumb er of difficult ies inherent with binary-coded GAs in solving cont inuoussearch space problems-Hamming cliff problem, arbitrary pr ecision problem,and fixed mapped coding problem. In the comparison of real-coded GAs wit ha SBX operator and binary-coded GAs with a single-point crossover ope rat or ,it has been observed that the performance of the former is better than thelatt er on continuous functions and the performance of the former is similarto the lat ter in solving discret e and difficult functions. In comparison withanother real-coded crossover operator (i.e. , BLX-0 .5) suggested elsewhere ,SBX performs better in difficult test functions. It has also been observedthat SBX is particularly useful in problems where the bounds of the optimum
point is not known a priori and wher e there are multi ple optima, of whichone is global.Real-coded GAs wit h t he SBX op erator have also been tried in solvinga two-variab le blocked function (the concept of blocked fun ctions was introducedin [10]). Blocked fun ct ions are difficult for real-coded GAs , becauselocal optimal points block t he progress of search to continue towards t heglobal optimal point . The simulat ion results on t he two-var iable blockedfunction have shown that in most occasions , the sea rch proceeds the way aspr edicted in [10]. Most importantly, it has been observed that the real-codedGAs wit h SBX work similar to that of t he binary-coded GAs wit h single-pointcrossover in overcoming t he barrier of the local peaks and converging to t heglobal bas in. However , it is premature to conclude whether real-coded GAswit h SBX op erator can overcome t he local barriers in higher-dimensionalblocked fun ct ions.These results are encour aging and suggest avenues for further research.Because the SBX ope rat or uses a probability distribut ion for choosing a childpo int , the real-coded GAs wit h SBX are one st ep ahead of the binary-codedGAs in te rms of ach ieving a convergence proof for GAs. With a direct probabilist ic relationship between children and parent points used in t his paper,cues from t he clas sical stochast ic optimization methods can be borrowed toachieve a convergence proof of GAs , or a much closer tie between the classicaloptimization methods and GAs is on t he horizon.
In short, according to the authors my SBX operator using real gene values is as good as older ones specially designed for discrete searches, and better in continuous searches. SBX as far as i know meanwhile is a standard general crossover operator.
But:
- there might be better ones out there i just havent seen yet. please tell me.
- besides tournament selection and mutation, crossover is just one part of the breeding pipeline. also there is the elite management for MOEA which is AT LEAST as important as the breeding itself.
- depending on the problem, there are almost always better specific ways of how to code the mutation and the crossover operators. but octopus is meant to keep it general for the moment - maybe there's a way for an interface to code those things yourself..!?
2) elite size = SPEA-2 archive size, yes. the rate depends on your convergence behaviour i would say. i usually start off with at least half the size of the population, but mostly the same size (as it is hard-coded in the new version, i just realize) is big enough.
4) the non-dominated front is always put into the archive first. if the archive size is exceeded, the least important individual (the significant strategy in SPEA-2) are truncated one by one until the size is reached. if it is smaller, the fittest dominated individuals are put into the elite. the latter happens in the beginning of the run, when the front wasn't discovered well yet.
3) yes it is. this is a custom implementation i figured out myself. however i'm close to have the HypE algorithm working in the new version, which natively has got the possibility to articulate perference relations on sets of solutions.
…
o express my gratitude. I've been experimenting with your definitions (and still am), but let me extend my question.
Actually what I'm trying to achieve, is to recreate another project by Andrew Kudless, the spore lamp (I mentioned the Chrysalis at the beginning just because of the animation, which wasn't included in the Spore Lamp presentation).
Basically the spore lamp seems to me to be something like a preliminary study to the Chrysalis III project (I think it's a similar approach).
Andrew stated on his site that he used kangaroo for this project, so the Spore Lamp consists in my opinion either of a relaxed voronoi 3d diagram (b-rep, b-rep intersection) on a sphere which then has been planarized, or more likely it is a sort of relaxed facet dome.
The trick is to:
1. obtain a nicely-balanced voronoish diagram (or facet dome cells)
2. keep each cell/polyline planar (or force them with kangaroo to be planar) in order to move scale and loft them later on.
Here is what I have by now. (files: matsys spore lamp attempt)
That's the closest appearance that I got so far (simple move scale and loft of facet dome cells with the amount of transformations being proportional to the power of the initial cell area: bigger cell = bigger opening etc.) - with no relaxation of the diagram. But it's obviously not the same thing as the matsys design.
Here are some of my attempts of facet dome relaxation, but well, it certainly still not the right approach, and most importantly I don't know how to keep or force the cells to be planar after the relaxation.
1. pulling vertices to a sphere - no anchor points. That obviously doesn't make sense at all, but the relaxation without anchor points gives at the beginning a pattern that is closer to what I am looking for. (files: relaxation 01)
2. pulling vertices to a sphere - two faces of the initial facet dome anchored (files: relaxation 02)
3. pulling vertices to the initial geometry (facet dome) no anchor points (files: relaxation 03)
The cell pattern of the lamp kinda looks like this:
you can find it here: http://www.grasshopper3d.com/forum/topics/kangaroo-0-095-released?g...
Done with Plankton (of course without the "gradient increase" appearance), but in fact not, I took a look at Daniel Parker's Plankton example files, and it's not quite the same thing. Also the cells aren't planar...
The last problem is that during the relaxation attempts that I did, the biggest initial cells became enormous, and it's not like that in the elegant project by Andrew Kudless, that I'd like to achieve.
So to sum up:
Goal no 1: Obtain an elegant voronoi /facet dome cell pattern on a sphere (or an ellipsoid surface, whatever).
Goal no 2: Keep the cells planar in order to be able to loft them later and obtain those pyramidal forms, and assemble easily
Any ideas? Or maybe there's a completely different approach to that?…