Particle Spring Material Simulation
Project Completed under Robert Stuart-Smith, Penn Design - Summer 2018
Java, Processing
Autonomous Manufacturing Lab
Penn Design Robotics
This complex geometry script was run through the simulation prior to attempting to print it in order to visualize loop lengths and geometry structure. The simulation successfully handled the complex details of this script, including particles that vertically stacked at an angle, sharp turns, integration of layers, and a high final particle count.
Penn Design’s Autonomous Manufacturing Lab explores the usage of robotics in design methodologies. The lab aims to create autonomous and semi-autonomous generative design processes by employing programming, real-time robotics, and computer vision technologies. Currently, the lab is developing a ceramic 3D printing process that involves an industrial robot with a clay extrusion end-effector tool. I worked on developing a physics-based material simulation of the clay extrusions. The simulation aimed to truthfully convey the degree of unpredictability due to material dynamics.
Link to the larger project on the AML’s site.
OVerview
My primary role in this project was to write a particle-spring physics-based material simulation in order to predict the final geometry of the prints. With the assistance of the material simulation, waypoints and speeds in the script can be adjusted before running the script with the robots. The simulation is written in Java and launches a Processing application window.
Consulted: 3D Verlet Physics
Development
The development of this simulation can be broken into a few discrete steps:
This diagram shows the simulation’s mesh export data imported into Rhino and the mesh after a few basic manipulations in modeling program. These models have been constructed using a 6-sided low polygon mesh.
Creation of Physics Environment: Believable particle-spring model that responds to gravity, spring forces, collisions
Incorporation of Material Properties: Adjustable stacking and sticking properties to emulate clay
Construction of Low Polygon Mesh: N-sided polygon mesh that coats the particle-spring model to mimic the physical appearance of clay
Introduction of Dynamic Flowrate: Allows simulation to either print constant distance waypoints or dynamic plotting based on speed
In this leaning structure geometry the particle-spring model mimics the sticking properties of clay. After a certain height, the particles stack unevenly due to the height and angle difference of the layers. Visible disruptions in the simulation geometry can signify a need to readjust waypoints or make structural changes to the final print geometry.
In test prints, the disruptions began around the predicted layers, however appeared on the side of the print instead of the back of the print.
Note that the print was not run to finish whereas the simulation was. This means that the simulation images contain more layers than the photographs of the simulation tests.
Project was funded by Penn Undergraduate Research Mentorship (PURM) Grant by the Center for Undergraduate Research & Fellowship.
Presented project at Penn's 2018 Fall Research Expo hosted by the Center for Undergraduate Research & Fellowships.
Project Abstract on CURF’s website.