Search the Community
Showing results for tags 'weighted compliance'.
Found 3 results
Hello all, I've been trying to recreate the example problem seen here: https://connect.altair.com/CP/SA/training/self_paced/optimization/content/opt3/pdf/ex1a.pdf But with a couple of modifications: 1. I have written a script to run the model in batch mode, however, the script cannot handle RBEs at this point. I've substituted all RBEs with loads and constraints assigned directly to the surfaces, effectively creating tractions instead of RBEs. 2. The script meshes everything in tetra elements. Everything else should be the same (material definitions, boundary conditions, objective function, etc). I'm thinking that these two modifications shouldn't change the outcome very much, but maybe that's an incorrect assumption? I keep getting an error in Optistruct that the compliance is huge and optimization can't be performed. Any ideas as to why? The model is attached. Any help or guidance is greatly appreciated! vf0.3_min30_wcomp.mvw vf0.3_min30_wcomp_des.h3d vf0.3_min30_wcomp.hm vf0.3_min30_wcomp.fem vf0.3_min30_wcomp.log vf0.3_min30_wcomp_003.out
Hello, i am working with the optimization of the wheel disc. I have non design space weight as 7.14 kg and design space weight as 26 kg. I used mass constraint and had set the upperbound to 0.0035 (3.5 kg) and weighted compliance as my response. How ever when i used ISO with 0.3 the end weight of the design space came out as 4.7 kg. How does mass constraint works? can anyone please solve this concern. Moreover when i look the mass response graph it is showing almost the same value throughout the iteration. I have attched my output file, mass and compliance responce. Please explain me how mass constraint actually works? If mass value is same throughout my iterations, then how my stifness is reducing for both load cases? Thanking you Punith Guptha 2920774.out
Hello, I am using VWT to access to the surface pressures on a model. These are then to be exported and applied to the model within HyperMesh via linear interpolation. I recently came to realise that if I move the model in VWT then the nodal pressure coordinates that I can then export to a .txt file are wrong. They do not align with the nodal coordinates of the model within HyperMesh, but refer to the region of space that they occupy in VWT. Thus, you must translate and rotate the model in HyperMesh first. This is justifiable and works for just one loading scenario. The problem I am having is that I wish to do a weighted compliance topology optimisation with multiple loading scenarios that require the angle of incidence of the flow to change. Therefore, if I rotate the model, the pressure files nodal coordinates will change and I will not be able to use the same hypermesh file (as it will have to be rotated accordingly too). So, is there a way to change the direction of the flow or rotate the windtunnel within VWT - so that my model can stay constrained in its geometrical space but the loading scenario can change? Thank you in advance.