Structural optimization of 3D printed designs with spatially-varying material properties

Thesis (Master's)--University of Washington, 2018Design with conventional, homogeneous materials has historically been limited to finding ideal geometry to fit a given engineering purpose. These designs are driven by necessary geometric discontinuities which cause high strain energy gradients when subjected to mechanical loads, and thus are more likely to fail in these regions. However, new advancements in 3D printing enable manufacturing a solid part with varying material properties; this research seeks to establish techniques for finding optimal designs that use this new technology for the greatest structural benefit. A sequential quadratic programming optimization algorithm was used to find an optimal distribution of Young's modulus that minimize strain energy gradients, as calculated using finite element analysis. This design method has been applied to the case of a thin plate with a circular hole, and has been proven to successfully reduce strain energy gradients and therefore stress concentrations. The resulting optimal design has been 3D printed using applicable technology and the computational model has been validated with experiments. Proposed investigations for future research includes studying the effect of heterogeneous material properties on failure and reliability, and improving applicability to physical systems. Enabling design engineers to customize material properties around geometric discontinuities will provide greater flexibility in reducing stress concentrations without modifying geometry or adding additional mass