Neural network-based build time estimation for additive manufacturing

Oh, Y., Sharp, M., Sprock, T., & Kwon, S. (2021). Neural network-based build time estimation for additive manufacturing: a performance comparison. Journal of Computational Design and Engineering, 8(5), 1243-1256, https://doi.org/10.1093/jcde/qwab044

(Phase 1) Input generation

Requirements

• BLENDER 2.82 & PYTHON 3.7.4 (embeded script), https://www.blender.org/download/releases/2-82/
• 3D model files (We used Baumann's dataset, https://doi.org/10.3390/data3010005)
• Voxelization.exe, https://drive.google.com/file/d/15MQk9bI3UP7zcChl9Rj6KzI77ybdFGPQ/view?usp=sharing
  • https://github.com/soonjokwon/VOX4AM
• The CURA engine & a config file for a 3D printer

How to run

• The code of the following files should be sequentially run in the Python script embedded in Blender 2.82
  • (Step 1) PartNormalization.py
  • (Step 2) PartGeneration.py
  • (Step 3) BuildTimeCaluculation(CURA).py
  • (Step 4) STLtoOBJ.py
  • (Step 5) Voxelization.py

(Phase 2) Build time estimation based on neural networks

Requirements

• Tensorflow 2.2.0; Python 3.6.9; Keras 2.3.0
• A dataset for metadata (a CSV file)
  • This is generated in Phase 1
• A voxelization dataset (a H5 file)
  • This is generated in Phase 1

How to run

• The code runs on Google Colab, https://colab.research.google.com/drive/1xMy7s4hVNjw2u3koOJSOAFGp-53vvcfC?usp=sharing
  • ANN with RFs
  • CNN with voxels
  • ANN with voxels