Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


41848.pdf (8.02 MB)

ETD Abstract Container

Abstract Header

Fast Powder Bed Fusion Additive Manufacturing (PBFAM) Simulation and Optimization for Minimizing Part Distortions

Li, Lun
http://orcid.org/0000-0001-8481-8165
http://rave.ohiolink.edu/etdc/view?acc_num=ucin1659533599527459

Abstract Details

2022, PhD, University of Cincinnati, Engineering and Applied Science: Mechanical Engineering.
The Powder Bed Fusion Additive Manufacturing process has emerged as an important industrial process that is capable of manufacturing complex part features, such as hollow, lattice structures, and other unique design structures. In this process, part distortion caused by repeatedly heating and cooling is a major concern that needs to be addressed. Predicting distortion and optimizing the building parameters for distortion mitigation is a major focus of current AM research. In this work, novel algorithms based on inherent strain are presented to predict and mitigate part distortion, optimize hatch patterns, and support structures in order to decrease part GD&T errors. First, a neural network-based method is presented to predict inherent strain for any given hatch pattern that is adopted during the part build. The neural network was trained by the inherent strain calculated through thermo-mechanical simulation. The results show that the trained neural network can predict the inherent strain of any arbitrary hatch pattern within an acceptable error. Next, a novel Backward Interpolation (BI) model is presented for fast estimation of part distortion based on inherent strain and distortion factor. The as-built distortion before cutoff from the substrate is calculated based on distortion factors and the internal forces generated by inherent strain. Subsequently, the spring back distortion is calculated based on the interfacial reaction forces caused by the as-built distortion. The total part distortion is then finally calculated as the summation of as-built distortion and the spring back distortion. Experimental validations were conducted, and the predicted distortion results appeared to agree well with the distortion of the sample part and published data. The BI approach is then used for hatch pattern optimization in order to decrease part GD&T errors. A Genetic algorithm (GA) was formulated in conjunction with the BI approach to optimize hatch patterns to minimize flatness form error. The hatch angle of each layer was selected as the optimization variable, while the flatness error of the sample part was chosen as the fitness variable in the GA process. The results showed that increasing the number of islands in each layer appears beneficial for achieving lower flatness errors. Furthermore, the flatness errors of four standard benchmark hatch patterns were calculated to make a comparison with the optimization results. The comparison showed that the flatness error of the sample part with an optimized hatch pattern performs better than the benchmark hatch patterns. Finally, a Particle Swarm Optimization was developed to optimize support structures to decrease the part GD&T errors. The support structures are simplified as beam elements and connected to the solid part with multi-parameter coupling using ANSYS. The distortion is calculated by the macro layer method by applying a temperature load on the model with orthotropic expansion properties. The flatness and cylindricity errors of the part with the optimized support structures outperformed the results of three benchmarks and demonstrated the optimization model's effectiveness.
Sam Anand, Ph.D. (Committee Member)
Michael Alexander-Ramos, Ph.D. (Committee Member)
Kumar Vemaganti, Ph.D. (Committee Member)
Jing Shi, Ph.D. (Committee Member)
Manish Kumar, Ph.D. (Committee Member)
142 p.
Mechanical Engineering
Additive Manufacturing; Hatch Pattern; Inherent Strain; Distortion Prediction; GDT; Support Structure

Recommended Citations

Citations

  • Li, L. (2022). Fast Powder Bed Fusion Additive Manufacturing (PBFAM) Simulation and Optimization for Minimizing Part Distortions [Doctoral dissertation, University of Cincinnati]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1659533599527459

    APA Style (7th edition)

  • Li, Lun. Fast Powder Bed Fusion Additive Manufacturing (PBFAM) Simulation and Optimization for Minimizing Part Distortions. 2022. University of Cincinnati, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=ucin1659533599527459.

    MLA Style (8th edition)

  • Li, Lun. "Fast Powder Bed Fusion Additive Manufacturing (PBFAM) Simulation and Optimization for Minimizing Part Distortions." Doctoral dissertation, University of Cincinnati, 2022. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1659533599527459

    Chicago Manual of Style (17th edition)

ucin1659533599527459
132
© 2022, all rights reserved.
This open access ETD is published by University of Cincinnati and OhioLINK.