Jump to content

Dylan Stelzer

Members
  • Content Count

    28
  • Joined

  • Last visited

About Dylan Stelzer

  • Rank
    Beginner

Profile Information

  • Gender
    Male
  • Are you University user?
    Yes

Recent Profile Visitors

1400 profile views
  1. Rahul R, I can see that you're manually creating the shape perturbations this the attached example. I forgot to mention this in my original post but I'm wondering if there is a way to translate this to a free shape optimization? Is it possible to go through the same method you've shown here? I'm going to do some personal investigating of the matter but as a "slight" solution at the moment, I am using a volume constraint with an upper and lower bound in such a way to arrive at a final shape that is within <2.5% of the original volume. I know it is not an exact solution if volume were constant but I'm using this as an estimation. Thoughts? Dylan Stelzer
  2. As stated by Rahul R on the same post but in a different thread: "Hi, In order to keep the same mass or volume please refer attach video. Regards Rahul R" mcon.swf
  3. Hi, I am trying to run a model which simulates that of a cantilever beam with circular cuts designed into it at regular intervals. I have a force load applying in compression along the top edge of the beam and would like to perform a shape optimization of the circular cuts while maintaining constant volume of the model under a minimizing compliance (maximizing stiffness) objective. I've tried it a few different ways but can't get the model to converge after more than two iterations (the optimization isn't progressing very far most likely due to the strict volume constraints. So what I would like to know is if it is possible to perform a shape optimization while maintaining constant volume of the model. I understand that this may be a difficult problem for the solver as I'm asking it to make changes while perhaps not giving it room to work with. If an upper and lower bound constraint of the original volume is too tight of a constraint, how loose do I have to make the constraint to allow for a reasonable shape optimization to be performed while maintaining (mostly) a constant total volume? Thanks for your input! I've attached the .HM file as well if you would prefer to take a closer look at the problem setup and optimization parameters. Dylan Stelzer Compressive_Load_(Uniform)_Optimizaton_Setup_1.hm
  4. Hi, I am trying to run a model which simulates that of a cantilever beam with circular cuts designed into it at regular intervals. I have a force load applying in compression along the top edge of the beam and would like to perform a shape optimization of the circular cuts while maintaining constant volume of the model under a minimizing compliance (maximizing stiffness) objective. I've tried it a few different ways but can't get the model to converge after more than two iterations (the optimization isn't progressing very far most likely due to the strict volume constraints. So what I would like to know is if it is possible to perform a shape optimization while maintaining constant volume of the model. I understand that this may be a difficult problem for the solver as I'm asking it to make changes while perhaps not giving it room to work with. If an upper and lower bound constraint of the original volume is too tight of a constraint, how loose do I have to make the constraint to allow for a reasonable shape optimization to be performed while maintaining (mostly) a constant total volume? Thanks for your input! I've attached the .HM file as well if you would prefer to take a closer look at the problem setup and optimization parameters. Dylan Stelzer Compressive_Load_(Uniform)_Optimizaton_Setup_1.hm
  5. Tinh, Thanks for responding! You're right, that is a very clear error indeed! That works and the remeshing can now start but I am running into another error within the remesh input/output files. Apparently a problem with the subcases. Do you have any input on the attached error? Thanks for all of your help. Dylan Stelzer
  6. Hi, I'm running a 3D model and am trying to perform a shape optimization on 5 similar structures on a beam. It is a beam with 5 circular cuts made throughout it with a specific loading applied to it. I ran the shape optimization once and it ran for 12 iterations and then gave the error that the mesh became too distorted and was unable to proceed with the optimization. I looked online and found that most people are using the "remesh" option in the opti-control setting to combat this type of termination error. I have applied the remesh option in opti-control and ran the optimization again but now I am receiving the error shown in the image below which I can't seem to find any information for on the forums. Would anybody be able to weigh in on this? Thanks ahead of time! Dylan Stelzer
  7. Hi, I have another post about this same topic but I wanted to create a new one in order to keep the problem concise and detailed as possible. My goal is that I want to be able to create a constant force load over a curved surface (specifically, the upper half of a semi-circle) and have every force load be normal to said surface while utilizing 2-D shell elements. Currently, this is how I am doing it. 1. The geometry (and mesh) is the upper half of a semi-circle with a radius of ~4cm (average element mesh size is 0.05cm). *I have been previously doing this model with only horizontal and vertical forces on the arced surface (never combined) E.g. Horizontal forces in the +y direction on the left third of the arc, vertical forces in the -z direction on the top third of the arc, and another set of horizontal forces in the -y direction on the right third of the arc. 2. My boss asked me to combine horizontal and vertical forces in such a way that I will have a (let's just use an easy number) constant 100N force applied to every node on the arc in a direction normal to the surface of said arc (this is hopefully supposed to simulate some sort of compression-like force load). 3. Selecting all of the nodes, I know there are 252 of them on the arc in total. E.g. This makes sense because the length of the arc is equal to s = pi*radius = 12.5664cm and Total # of Nodes = s/.05cm = 252 nodes (approximately) 4. Combining horizontal and vertical forces at each node just requires two simple equations and a little bit of Matlab code *thetaSEP = 180deg / 252 = 0.7143deg (approximately) and hypotenuse = 100N Iteration 1 (Node 1): cos(thetaSEP) = horizontal/hypotenuse = 100N -sin(thetaSEP) = vertical/hypotenuse = 0N Iteration 2 (Node 2): cos(thetaSEP) = horizontal/hypotenuse = 99.9922N -sin(thetaSEP) = vertical/hypotenuse = -1.2466N etc. etc. etc. *Horizontal equation changes to -cos AFTER the 126 iteration 5. After doing this (and hopefully my math is correct...), I have two vectors, H = [1 x 252] and V = [1 x 252] which have been transferred into Excel (I just like Excel a little better than Matlab) which represents the horizontal and vertical forces that I need to combine at each node to create my constant 100N force load along the arc. ***6. Here, I did my first model by hand entering all 252 force vectors, one at a time. ***This is where I'm hoping someone can give me help on how to automate the process from this point into applying these loads across the nodes I have selected in HyperMesh with correct values (which leads to angle as well). If anyone has any help or details (or more resources for that matter) that I can try to make this process faster, that would be much appreciated. User: Q.NGUYEN-DAI has already given me some great advice but to be completely honest, I've never used Tcl in my life nor have I really programmed that much. Even if someone has a more efficient idea of doing this, could you please explain where I'm making this harder than it needs to be (I understand the "one by one" by hand part is the serious problem)? Thanks ahead of time and sorry for the long post. Attached is an image describing the problem a bit more if you need some visual clues as to further explain my situation. Dylan Stelzer
  8. I tried your method and did some research but still haven't had any luck for a few days now. I think I'll make a more detailed post and see if I can get any more specific information on where I'm running into problems. The tough part is that I've never even dealt with Tcl scripts before so I'm wondering in foreign land here. Thanks for your help though! Dylan Stelzer
  9. I'll give it my best shot. Hopefully I can figure it out because doing it one by one is driving me mad! Thanks again. Dylan Stelzer
  10. I'll google "how to" on the TCL script part, that doesn't sound too hard but where do I make the jump from reading Excel data to assigning them...I know there are only a few ways of assigning (specifically for me, I need it to be a force): components, vector, equation, linear interpolation, etc. etc. Thanks for replying though!
  11. Hi, I had a question concerning the integration of an array from Excel (or for that matter, any type of data where it is a string of numbers) into a boundary condition loading in HyperMesh. I have two arrays of length 126; is there any way to apply them to each of my 126 selected nodes in my HyperMesh model (it is composed of 2-D shell elements)? I've attached a picture of a section of the arc if that helps. I am trying to divide up a normal load over an arc into horizontal and vertical forces efficiently and quickly. If anyone has any suggestions, that would be greatly appreciated. Thanks! Dylan Stelzer
  12. Depending on your model, your optimization parameters, and your objective? Yes, 3 iterations could lead to an "optimal result". If you're not satisfied with the answer, try changing your constraints around or re-working the boundary conditions for your problem. That would be my "noob" opinion
  13. Hi Prakash, Thanks for answering! What do you mean by "update" the elements for conduction? I'll attempt my own interpretation on this but could you explain any further? I believe you mean reapplying the conduction interface to the elements found on this geometry surface but I want to be clear. Thanks. Dylan Stelzer
  14. Hi, Would anyone have any answers as to why the OSSmooth version of my model is creating errors for itself with respect to the conduction interface? I had an original version of HyperWorks 13.0 and never had a problem with OSSmooth and when I had to uninstall and download it again, I started getting errors whenever I tried to use OSSmooth on my models. I've attached some pictures below to help give some context to the problem as well as the error but if anyone has any suggestions or needs more background to the problem, I'm more than willing to attach the entire model file or fill you in on the specifics (properties, boundary conditions, etc. etc.). Any help would be really appreciated! Thanks. Dylan Stelzer
  15. Thanks for replying! I'll try doing an isolated test on doing what you suggested (switching from a nodal force to an elemental pressure) and see if that gives the type of results I'm looking for. I never thought of using pressure instead of force but yes, I presume the answers would be very similar if done correctly. Dylan Stelzer
×
×
  • Create New...