Date of Award
2026
Degree Type
Dissertation
Degree Name
Doctor of Philosophy in Civil and Environmental Engineering
Department
Civil and Environmental Engineering
First Advisor
Sumanta Das
Abstract
Computational modeling and machine learning offer powerful pathways for designing polymer-based protective systems subjected to dynamic loading, particularly for mitigating the early shock-dominated response generated by near-field underwater explosions (UNDEX). As naval, offshore, and submerged infrastructure systems continue to grow in strategic importance, there is an increasing need for lightweight, damage-tolerant protective solutions capable of reducing transmitted pressure, deformation, and energy transfer under extreme impulsive environments. Since large-scale experimental testing under such conditions are costly and limited, validated numerical frameworks provide an efficient means to evaluate polymeric coatings, architected metastructures, and data-driven predictive tools across broad design spaces. The first study examines fusion bonding between acrylonitrile-butadiene-styrene (ABS) and thermoplastic polyurethane (TPU) through a multiscale framework combining reactive molecular dynamics and finite element analysis. Interfacial adhesion behavior is predicted across length scales and compared with tensile experiments on additively manufactured multi-material specimens. The numerical results show good agreement with experiments while clarifying the influence of interfacial fracture properties and manufacturing-induced porosity on the macroscopic bond response. The focus then shifts to the protection of composite structures subjected to near-field UNDEX using polymer-based systems. Spatially distributed polyurea coatings applied to composite cylinders are first evaluated under constant mass constraints. The results show that coating placement, thickness, orientation, and stand-off distance strongly influence pressure attenuation, deformation reduction, and energy absorption. This demonstrates that protective efficacy depends not only on coating mass, but also on how the material is distributed around the structure. Building on this concept, architected auxetic protection systems are investigated using polyurea reentrant lattice encasements surrounding composite cylinders. A validated coupled Eulerian-Lagrangian (CEL) fluid-structure interaction framework shows that reentrant auxetic geometries can substantially reduce composite displacement, strain, and energy transfer by redirecting blast-induced loading through controlled geometric collapse. Parametric studies further demonstrate that arc angle, layer count, slant geometry, thickness, charge size, and stand-off distance strongly govern the protective response. The analysis is subsequently extended to mass-equivalent auxetic topologies including reentrant, double arrowhead, missing-rib, anti-tetra missing-rib, and chiral-based architectures. Despite identical external envelopes and total mass, the results reveal pronounced topology-dependent differences in collapse mechanisms, energy partitioning, and load transfer to the protected composite. Architectures governed by progressive hinging and controlled ligament collapse exhibit superior mitigation compared with more rotation-dominated systems, highlighting topology itself as a critical design variable in underwater blast protection. To accelerate exploration of this high-dimensional design space, high-fidelity UNDEX simulations are integrated with interpretable machine learning. A structured dataset of 576 simulations is generated from the validated reentrant auxetic framework, and multiple regression approaches are evaluated to predict maximum displacement, kinetic energy transfer, and internal energy accumulation during the early-time loading window. Among the models considered, neural networks provide the strongest predictive capability, while SHapley Additive exPlanations (SHAP) identify explosive weight and stand-off distance as the dominant variables, followed by shell thickness and auxetic geometry. Finally, polymer coating behavior is revisited through a comparative study of polyurea- and polyisocyanate-oxazolone (POZD)-coated composite plates subjected to near-field UNDEX. Both coatings improve plate survivability relative to the uncoated case, but POZD consistently produces lower deformation and more favorable energy redistribution, indicating that higher stiffness and improved adhesion characteristics can further enhance underwater blast mitigation performance. Collectively, these studies establish a unified computational and data-driven framework for the design of polymer-based protective systems under extreme dynamic loading. By linking multiscale material modeling, high-fidelity fluid-structure interaction simulations, architected polymer geometries, and interpretable machine learning, this body of work advances efficient design strategies for next-generation protective systems intended for marine, naval, and offshore composite structures.
Recommended Citation
Villada, Jonathan Tate, "COMPUTATIONAL MODELING AND MACHINE LEARNING FOR THE DESIGN OF POLYMER-BASED PROTECTIVE SYSTEMS UNDER DYNAMIC LOADING" (2026). Open Access Dissertations. Paper 4559.
https://digitalcommons.uri.edu/oa_diss/4559
Comments
The contents of this document are Controlled Unclassified Information.