際際滷

際際滷Share a Scribd company logo

Additive_Manufacturing_Projects_Sushmit_Chowdhury

U
University of Cincinnati

This document describes two projects using artificial neural networks to help compensate for thermal deformations in additive manufacturing processes. The first project trains a neural network on deformation data to identify and implement geometric modifications to an STL file that compensate for thermal effects. Case studies showed this reduced geometric errors up to 64%. The second project optimizes part build orientation and uses a neural network to compensate for thermal deformations. Tools were developed to detect manufacturing concerns and optimize orientation to minimize these concerns.

1 of 3
Downloaded 16 times
DESIGN FOR ADDITIVE MANUFACTURING PROJECTS
ARTIFICIAL NEURAL NETWORK BASED GEOMETRICAL COMPENSATION FOR THERMAL DEFORMATIONS IN ADDITIVE MANUFACTURING PROCESSES MOTIVATION: Determine and implement geometrical modifications to be made to the part STL model to compensate for shrinkage and thermo-mechanical deformations undergone by Additive Manufactured parts. METHODOLOGY Input: Part deformation data obtained from thermo- mechanical simulation of AM process 1 ANN Training on surface deformation data 2 Application of fully trained ANN on part STL file 3 Output : Compensated STL file of the part 4 Publication: Chowdhury, S., and Anand, S., Artificial Neural Network based Geometric Compensation for Thermal Deformations in Additive Manufacturing Processes, Proceedings of the 11th ASME MSEC 2016, Blacksburg, VA. RESULTS Multi-layer feed forward neural network model is trained on part geometry deformation data. Trained network is able to successfully identify & implement required geometric modifications to part STL to compensate the thermal effects of Additive Manufacturing Processes Case studies show reduction in geometric errors of up to 64%. HIGHLIGHTS
PART BUILD ORIENTATION OPTIMIZATION AND GEOMETRIC COMPENSATIONS FOR ADDITIVE MANUFACTURING PROCESSES MOTIVATION: Optimize part build orientation and geometry for Additive Manufacturing (AM) processes using developed design tools and an artificial neural network based approach to compensate for thermal deformations. Publication: Chowdhury, S., Mhapsekar, K., and Anand, S., Part Build Orientation Optimization and Geometry Compensations for Additive Manufacturing Processes, 2016. (To be published) Developed individual design tools to detect potential part quality and manufacturing concerns for AM. Optimized part build orientation to minimize identified concerns. Implemented Neural Network based compensation for part geometry modifications to counteract thermal effects of AM processes such as shrinkage, deformation, etc. HIGHLIGHTS

More Related Content

Additive_Manufacturing_Projects_Sushmit_Chowdhury

  • 1. DESIGN FOR ADDITIVE MANUFACTURING PROJECTS
  • 2. ARTIFICIAL NEURAL NETWORK BASED GEOMETRICAL COMPENSATION FOR THERMAL DEFORMATIONS IN ADDITIVE MANUFACTURING PROCESSES MOTIVATION: Determine and implement geometrical modifications to be made to the part STL model to compensate for shrinkage and thermo-mechanical deformations undergone by Additive Manufactured parts. METHODOLOGY Input: Part deformation data obtained from thermo- mechanical simulation of AM process 1 ANN Training on surface deformation data 2 Application of fully trained ANN on part STL file 3 Output : Compensated STL file of the part 4 Publication: Chowdhury, S., and Anand, S., Artificial Neural Network based Geometric Compensation for Thermal Deformations in Additive Manufacturing Processes, Proceedings of the 11th ASME MSEC 2016, Blacksburg, VA. RESULTS Multi-layer feed forward neural network model is trained on part geometry deformation data. Trained network is able to successfully identify & implement required geometric modifications to part STL to compensate the thermal effects of Additive Manufacturing Processes Case studies show reduction in geometric errors of up to 64%. HIGHLIGHTS
  • 3. PART BUILD ORIENTATION OPTIMIZATION AND GEOMETRIC COMPENSATIONS FOR ADDITIVE MANUFACTURING PROCESSES MOTIVATION: Optimize part build orientation and geometry for Additive Manufacturing (AM) processes using developed design tools and an artificial neural network based approach to compensate for thermal deformations. Publication: Chowdhury, S., Mhapsekar, K., and Anand, S., Part Build Orientation Optimization and Geometry Compensations for Additive Manufacturing Processes, 2016. (To be published) Developed individual design tools to detect potential part quality and manufacturing concerns for AM. Optimized part build orientation to minimize identified concerns. Implemented Neural Network based compensation for part geometry modifications to counteract thermal effects of AM processes such as shrinkage, deformation, etc. HIGHLIGHTS