Discipline: Technology and Engineering
Subcategory: Civil/Mechanical/Manufacturing Engineering
Session: 3
Room: Senate
Bryant Rodriugez - Florida Agricultural & Mechanical University
Co-Author(s): Sean Psulkowski, Florida State University, Tallahassee; Dr. Tarik Dickens, Florida Agricultural & Mechanical University
This work describes the use of additive processing, design, and intelligent manufacturing, facilitated by collaborative robotics, to understand the linkages between process-structure-property relationships further. As the greater research community has gravitated towards knowledge-based design, in-situ monitoring provides real-time information concerning the processing of polymeric and semi-reinforced composite structures. Moreover, the necessity for multifunctionality and structural health monitoring capabilities has been underscored since the advent of third-generation composite structures. The following method outlines a closed-loop integration of both strategies in which iterative machine learning governs the placement and process parameters of novel mechanoluminescence smart material embedded in a polymeric structure. Thermomechanical simulation and in-situ sensing establish a priori knowledge to which to train a neural network for the placement, orientation, and printing conditions for smart sensor embedment to maximize sensing efficiency and part performance. The devised experiment utilizes compact tension standards to study the effect of printed weld lines across these generated polymer combinations, targeting the orthotropic properties inherent to additive assemblages. Experimental results i) expand the material library of smart material within additive processing, detailing fabrication criteria and characterization to standardize future applications, and ii) enrich the knowledge pool of digital design discovery via iterative machine learning. Therefore, the capabilities of cyber-physical systems for multifunctional composite fabrication are showcased and augmented through machine learning-driven design generation and executed with collaborative robotic agents to capture holistic process data. Future research involves utilizing the synergy of these ideas to evolve significantly more consistent additive manufacturing quality control informed by dynamic print defect detection and subsequent compensation.
Funder Acknowledgement(s): My research is funded by an NSF CREST HBCU-RISE grant awarded to Dr. Tarik Dickens.
Faculty Advisor: Tarik Dickens, dickens@eng.famu.fsu.edu
Role: My role in this research involved developing and implementing the collaborative robotic systems and several sensors utilized. In addition, I worked on the training, development, and implementation of the neural network used in the study and running experiments for validation and data acquisition.