Comprehensive Exam Presentation - Md Asif Rahman

Md Asif Rahman – Data Science

Zoom link

Title: Understanding mesoscopic chemo-mechanical distress and mitigation mechanisms of concrete subject to ASR

Abstract:

Concrete is one of the promising building materials and frequently used in roads and highways. Concrete is beneficial over other building materials due to its durability, long service life and great economy. However, concrete goes under degradation over time because of the detrimental chemical attacks on concrete structure and severe environmental conditions. These are causing billions of dollars in damages every year in the United States.
Alkali-silica reaction (ASR) is a frequently occurred undesirable phenomena for concrete damage worldwide. ASR starts with the chemical contact between reactive aggregates and cement paste in concrete. Portland cement is a widely used construction material. However, silica from reactive aggregates reacts with alkali hydroxide of Portland cement to form alkali silica gel in a solid form. This gel further absorbs water from the surrounding and swells which induce increasing internal pressure in the cementitious matrix of concrete structure. This pressure can break the coherence between the components of concrete structure and develop crack in the concrete domain with the progression of time. This concrete ASR problem can be influenced by many factors: surrounding environment, concrete properties, mechanical loadings, aggregate/mineral types, use of supplementary materials etc. High temperature and humidity lead chemical reaction to the right direction and thus, initiate the ASR gel production. Moreover, extreme weather events in recent years have increased the complexity of ASR kinetics to some extent. Thus, the research effort documented in this work focused on the development of a multiphysics model to capture the ASR mechanisms under these factors, and precisely predict ASR-induced expansion with the application of machine learning. The overarching goal of this project is to develop an artificial neural network (ANN) model aimed at incorporating uncertainty quantification and simulating damage propagation of concrete meso-structure considering complex chemo-physio-mechanical phenomena and predict service life. Research goal is to develop a trained and tested ANN model using a comprehensive database collecting from different sources including lab tests, computational model and literature. Complex phenomena under research study includes Alkali Silica Reaction (ASR) in the concrete structure subjected to mechanical loading.

Committee:

Yang Lu, Ph.D.
Tim Andersen, Ph.D.
Mojtaba Sadegh, Ph.D.
Kevin Roche, Ph.D.
Michal Kopera, Ph.D.