Characterization, Generation, and Computational Modeling of Asphalt Nanostructures for Predicting Their Mechanical Response

Asphalt binder is an intrinsically heterogenous material that consists of various chemical groups and structures. Asphalt heterogeneity is further exasperated in modified asphalt binders, which incorporate chemical and/or physical modifiers. Asphalt chemical composition is responsible for the formation of nanoscopic phases exhibiting various size distributions, dispersion, and material properties. Consequently, the variation in the nanoscopic phases within and among asphalt binders manifests itself in uncertainty in predicting binder mechanical properties and durability. Therefore, this dissertation presents a coupled approach for characterization, generation, and computational modeling of asphalt nanostructures in order to predict the mechanical response of asphalt binders. Furthermore, this approach helps to investigate the effect of variability in asphalt binders’ nanoscale properties on uncertainty in their mechanical response. To this end, nanoscale tests including the PeakForce Quantitative Nanomechanical Mapping (PFQNM) test, and the nanoscale Dynamic Mechanical Analysis (nDMA) test are carried out using the Atomic Force Microscope (AFM) to characterize the nanomechanical properties of asphalts and obtain spatial maps of their nanostructures. The efficacy of these methods is demonstrated in the characterization of unmodified and modified asphalt binder. Furthermore, this study complements the experimental measurements with computational methods to better understand of the effects of nanostructures on asphalt mechanical responses. For this purpose, random fields (RF) and generative adversarial networks (GANs) methods are utilized to generate computational replicates of asphalt nanostructure based on the experimental measurements, and use these replicates to account for mechanical response variability. Finally, computational finite element (FE) models are developed to simulate the generated nanostructures, predict the response of binders, and evaluate the effect of variability on asphalts’ mechanical response. This coupled experimental and computational approach promotes sustainability by bringing into the analysis the natural variability of asphalts without the need of conducting abundant resource-intensive laboratory tests. The developed approach enables the evaluation of hypothetical scenarios that would be impractical to investigate in a laboratory environment. Consequently, the outcome is the design of asphalt blends with tailored properties and enhanced performance. The findings of this work introduce a variability measure in the analysis of asphalt binders, and hence expand the body of knowledge toward probabilistic analysis and uncertainty quantification of asphalts’ mechanical behavior.

Name thesis:degree_name

Doctor of Philosophy

Level thesis:degree_level

Doctoral

Discipline thesis:degree_discipline

Civil Engineering

Grantor

Texas A&M University

Year dc:date.issued

2022

Author dc:creator

• Aljarrah, Mohammad Fuad Amin

Advisors dc:contributor.advisor

• Masad, Eyad
• Little, Dallas

Committee members dc:contributor.committeemember

• Epps Martin, Amy
• Sukhishvili, Svetlana

• asphalt
• nanoscale
• viscoelastic
• AFM
• random fields
• generative adversarial networks
• LDPE

Language dc:language.iso

en

Handle dc:identifier.uri

https://hdl.handle.net/1969.1/197986

OAI record oai:identifier

https://oaktrust.library.tamu.edu/server/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Aoaktrust.library.tamu.edu%3A1969.1%2F197986