- training and validation data is overlapping
- seperate some additional test data apart from training and validation data?
- make geometry/mesh/analysis of deformation in Simcenter 3D
- close every file in Simcenter 3D
- Heeds (Anaylsis Portal: Simcenter 3D_2306)
- Process -> Simcenter 3D
- Input sim file (named Hachathon on desktop)
- Select tags and variables
- press auto tag -> CAD expressions (lattice_d_cell, lattice_d_rod, scalingfactor_x_y, lattice_num, and all)
- select min/max intervals
- Output (response)
- select blue icon and everything
- Run
- study->methods->design set-> add item -> put numbers in the variables) and experiment with it
- go to RUN-tab and run
variables (CAD expressions) | computer 1 | computer 2 | computer 3 |
---|---|---|---|
lattice_d_cell [1.9,2.75] | 2.375,2.75 | 1.9 | 1.9,2.375,2.75 |
lattice_d_rod [0.2,1.2] | step 0.05 | step 0.05 | step 0.05 |
scaling_factor_YZ [1,6] | 3 | 3 | 6 |
lattice_num [1,4] | 2 | 2 | 2 |
load | 100N | 100N | 100N |
the lattice_d_rod has the highest impact on porosity
-
Young modulus
The first a few lines are CAD expressions/parameters, then the displacement vector
$U$ . We need to calculate the Young modulus/effective stiffness$E$ from the mean value of$U$ , i.e.$\Delta U$ $$E=\frac{\sigma}{\epsilon}=\frac{FU_0}{A\Delta U}$$ where$\sigma = \frac{F}{A}$ is the stress, F is the fixed load on area$A =$ h_total_z$\times$ w_total_y (depending on scaling factor), and$U_0$ is d_total_x -
Porosity
For presentation, we need to calculate porousity
$P$ $$1-P = \frac{V_{mesh}}{V_{total}}$$ where$V_{mesh} = V_{lattice} + 2V_{plates}$ ,$V_{total} = V_{cube} + 2V_{plates}$ $$V_{lattice} = 8\times\frac{\sqrt{3}d_{cell}}{2}\times\pi\times(\frac{d_{rod}}{2})^{2}\times ScalingFactor^{2} \times LatticeNum$$ $$V_{cube} = d_{cell}^{3}\times ScalingFactor^{2} \times LatticeNum$$ $$V_{plates} = 1.5 \times d_{cell}^{2}\times ScalingFactor^{2} \times LatticeNum$$ }