Doctor of Engineering (DEng)
The Department of Engineering Science
As a new class of smart materials, shape memory polymer (SMP) is gaining great attention in both academia and industry. One challenge is that the chemical space is huge, while the human intelligence is limited, so that discovery of new SMPs becomes more and more difficult. In this dissertation, by adopting a series of machine learning (ML) methods, two frameworks are established for discovering new thermoset shape memory polymers (TSMPs). Specifically, one of them is performed by a combination of four methods, i.e., the most recently proposed linear notation BigSMILES, supplementing existing dataset by reasonable approximation, a mixed dimension (1D and 2D) input model and a type of dual-convolutional-neural-network model. The framework is able to predict the recovery stress and glass transition temperature for TSMP and screen 14 new TSMPs from a large chemical space. The other leverages transfer learning, variational autoencoder and weighted vector combination method, and the developed ML framework can design ultraviolet (UV) curable TSMPs with desired properties.
With new SMPs discovered by ML, as well as other new SMPs continuously developed in the labs, there is an urgent need to develop thermomechanical models so that the new SMPs can be used in structural design. Through the framework of solid mechanics, three different constitutive models are presented for classical one-way thermoset shape memory polymer (TSMP), two-way semi-crystalline SMP and enthalpy-driven four-chain SMP with large recovery stress, respectively. Among them, a new two-phase sphere model based on the physical growth process of the frozen phase from nuclei is proposed, which tends to bring more underly physical mechanism for the classical storage strain-based phase transition model. By introducing Gibbs energy and a transition of the molecule deformation mechanism, a enthalpy-driven thermomechanical model with new representative unit cell is developed, which could reasonably elucidates the large recovery stress for a new branch of TSMPs. Multiple mechanisms, involving phase transition law, damage evolution, and relaxation are introduced into the model for two-way semi-crystalline SMP, which is able to reveal the mechanisms of three different 2W-SMEs.
Yan, Cheng, "Machine Learning Assisted Discovery of Shape Memory Polymers and Their Thermomechanical Modeling" (2022). LSU Doctoral Dissertations. 5792.