Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

York University

$31,939.00

Multiscale methods for failure analysis of concrete materials

FY 2023-24

Summary
Multiscale analysis links material response at small scales such as individual fibers, grains of sand, or aggregates, referred to as the mesoscale or material scale, to the response of structural components. The goal is constitutive models at the larger structural scale that take into account the material microstructure. Advantages of such models are that they require fewer assumptions and a limited number of physical parameters of the individual components as compared to the many, often non-physical, i.e., non-measurable, parameters required to describe the many modes of failure of composite materials. Such models give the engineer flexibility to experiment computationally with different compositions, load conditions, and configurations so as to understand the effect of constituent properties on material behavior. The proposed research aims to develop machine learning-based multiscale modeling and simulation strategies with validated predictive capability to characterize the evolution of damage up to the occurrence of failure of structural components under a variety of loading conditions focusing on ultra high performance concrete (UHPC) as a paradigm of emerging significance. At the material scale, failure of UHPC is characterized by distributed matrix damage in the form of microcracks and the breaking, debonding, and kinking of fibers. Individual models for each of these effects will be provided such as an interface bonding model between matrix and fiber and a cohesive interface model for matrix microcracks. The latter will be developed building and expanding on the proposer's previous work related to energy-based cohesive fracture. The material scale resolution will be carried in damage homogenized form to the structural scale using data-driven machine learning, which is the technical innovative core of the proposed research. The computational tools necessary for this approach, aside from extensive fine scale computational experiments, include clustering algorithms, projection algorithms, and a method to deduce a state transition mapping such as regression with a machine learning model. The research will also be used to understand the localization of damage in lightly reinforced members and the implications this may have on strength, mode of failure, and the size effect in fracture related models of structural members.  The well-designed energy-based interplay between fracture and damage mechanics will allow for capturing the limit of a localising damage zone to a dominant crack in one seamless computation, which is an active research problem in computational mechanics. Integral to the research is the training of highly qualified personnel (HQP). Computer-aided design for civil engineering is a multibillion-dollar industry, but none of the major players are headquartered in Canada. The proposed research incorporates undergraduates and grad students and will develop Canadian expertise in this crucial area.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-05086
Name nom
Papoulia, Katerina
Department département
Civil Engineering
Organization id
38
Institution établissement
York University
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
31939
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1509
Committee name en
Civil, Industrial and Systems Engineering
Committee name fr
Génie civil, industriel et des systèmes
Area of application code
601
Area of application group en
Construction, urban and rural planning
Area of application group fr
Construction, urbanisme et aménagement rural
Area of application en
Structural engineering
Area of application fr
Génie des charpentes
Research subject code
2800
Research subject group en
Artificial intelligence
Research subject group fr
Artificial intelligence (computer vision, use2603)
Research subject en
Intelligence artificielle
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Application title
Multiscale methods for failure analysis of concrete materials
Keyword
cohesive fracture; computational mechanics; cortical bone; damage mechanics; data science; fiber reinforced composites; high performance concrete; machine learning; multiscale methods; polymer composites
Application summary
Multiscale analysis links material response at small scales such as individual fibers, grains of sand, or aggregates, referred to as the mesoscale or material scale, to the response of structural components. The goal is constitutive models at the larger structural scale that take into account the material microstructure. Advantages of such models are that they require fewer assumptions and a limited number of physical parameters of the individual components as compared to the many, often non-physical, i.e., non-measurable, parameters required to describe the many modes of failure of composite materials. Such models give the engineer flexibility to experiment computationally with different compositions, load conditions, and configurations so as to understand the effect of constituent properties on material behavior. The proposed research aims to develop machine learning-based multiscale modeling and simulation strategies with validated predictive capability to characterize the evolution of damage up to the occurrence of failure of structural components under a variety of loading conditions focusing on ultra high performance concrete (UHPC) as a paradigm of emerging significance. At the material scale, failure of UHPC is characterized by distributed matrix damage in the form of microcracks and the breaking, debonding, and kinking of fibers. Individual models for each of these effects will be provided such as an interface bonding model between matrix and fiber and a cohesive interface model for matrix microcracks. The latter will be developed building and expanding on the proposer's previous work related to energy-based cohesive fracture. The material scale resolution will be carried in damage homogenized form to the structural scale using data-driven machine learning, which is the technical innovative core of the proposed research. The computational tools necessary for this approach, aside from extensive fine scale computational experiments, include clustering algorithms, projection algorithms, and a method to deduce a state transition mapping such as regression with a machine learning model. The research will also be used to understand the localization of damage in lightly reinforced members and the implications this may have on strength, mode of failure, and the size effect in fracture related models of structural members.  The well-designed energy-based interplay between fracture and damage mechanics will allow for capturing the limit of a localising damage zone to a dominant crack in one seamless computation, which is an active research problem in computational mechanics. Integral to the research is the training of highly qualified personnel (HQP). Computer-aided design for civil engineering is a multibillion-dollar industry, but none of the major players are headquartered in Canada. The proposed research incorporates undergraduates and grad students and will develop Canadian expertise in this crucial area.